towards mri-guided radiotherapy in early-stage … mri-guided radiotherapy in early-stage breast...

Mariska D

. den Hartogh

Towards M

RI-guided radiotherapy in early-stage breast cancer patients

in early-stage

TowardsMRI-guided radiotherapybreast cancer patients

Mariska D. den Hartogh

Towards MRI-guided radiotherapy in early-stage breast cancer patients

Mariska D. den Hartogh

Towards MRI-guided radiotherapy in early-stage breast cancer patientsPhD thesis, Utrecht University – with a summary in Dutch ©M.D. den Hartogh, Utrecht 2014

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means without permission in writing from the author. The copyright of the articles that have been published or have been accepted for pub-lication has been transferred to the respective journals.

ISBN: 978-90-393-6259-4Lay-out: Roy Sanders Printing: Ridderprint, Ridderkerk

Deze uitgave is tot stand gekomen met financiële steun vanElekta B.V., UtrechtVC en de Röntgenstichting Utrecht.

Towards MRI-guided radiotherapy in early-stage breast cancer patients

“Naar MRI-geleide radiotherapie voor patiënten met vroeg-stadium mammacarcinoom”

(met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan,

ingevolge het besluit van het college voor promoties in het openbaar te verdedigen

op donderdag 18 december 2014 des middags te 16.15 uur

door

Mariska Deborah den Hartoghgeboren op 16 februari 1986

te Nijkerk

Promotoren: Prof. dr. M. van Vulpen Prof. dr. M.A.A.J. van den Bosch

Co-promotoren: Dr. H.J.G.D. van den Bongard Dr. B. van Asselen

Thesis outline

Chapter 1

Introduction 9

Chapter 2

Excised and irradiated volumes in relation to the tumor size in breast-conserving therapy.

25

Chapter 3

MR-guided breast radiotherapy: Feasibility and magnetic-field impact on skin dose.

45

Chapter 4

Post-lumpectomy CT-guided tumor bed delineation for breast boost and partial breast irradiation: Can additional pre- and postoperative imaging reduce interobserver variability?

65

Chapter 5

Full-thickness closure in breast-conserving surgery: The impact on radiotherapy target definition for boost and partial breast irradiation. A multimodality image evaluation.

81

Chapter 6

Preoperative MRI and CT imaging for target volume delineation in breast- conserving therapy: Investigating the potential for a neoadjuvant irradiation.

95

Chapter 7

MRI-guided single fraction ablative radiotherapy for early-stage breast cancer: A brachytherapy versus VMAT planning study.

113

Chapter 8

Summary and general discussion 127

Chapter 9Nederlandse samenvattingList of publicationsCurriculum VitaeDankwoord

143151155159

Chapter 1Introduction

10

Chapter 1

Introduction

The purpose of radiotherapy is to kill malignant cells while sparing the surrounding organs at risk as much as possible. Within daily radiotherapy practice, a precarious balance must therefore be achieved between tumor control and treatment toxicity. Since normal cells have a larger repair capacity than tumor cells, the total dose that is required to kill a tumor is usually divided over multiple treatment fractions. Ho-wever, dose fractionation alone is not enough to avoid treatment toxicity. Accurate definition of the radiotherapy target volume is also essential. For target definition, the differentiation between the target tissue and the surrounding organs at risk is key. As a result, the ability to visualize and define the radiotherapy target largely determines the treatment accuracy and toxicity. Other main factors influencing tre-atment accuracy are daily set-up errors and inter- and intra-fraction motion.Radiotherapy target definition is most commonly performed using computed to-mography (CT) imaging. Due to the low soft-tissue contrast of CT, other imaging modalities are often used to aid target definition. Magnetic resonance imaging (MRI) with its variety of sequences, provides superior soft-tissue contrast and can reflect different tissue characteristics. The potential of MRI-guided radiotherapy tar-get definition has already been shown in other tumor sites as cervix, prostate and rectal cancer [1-5]. Another attractive advantage of MRI is the possibility to acquire images with a high temporal resolution. Dynamic visualization of the radiotherapy target during treatment would allow for direct motion compensation, which could increase the treatment accuracy by allowing smaller margins. Direct motion com-pensation may thus lead to a reduced dose to organs at risk and/or an increased dose to the tumor. Considering the aforementioned advantages of MRI, the department of Radiation Oncology at the University Medical Center Utrecht is developing a radiotherapy ac-celerator with fully integrated diagnostic MRI functionality in cooperation with Elek-ta AB and Philips (Figure 1). In this MRI-linac design, a 1.5 T cylindrical MRI scanner (Philips, Best, The Netherlands) is integrated with an 8 MV radiotherapy accelerator (Elekta AB, Stockholm, Sweden), which is mounted on a ring-shaped gantry [6, 7].The MRI-linac has the potential to improve radiotherapy accuracy for various tumor sites including cervical, prostate and rectal cancer. Among women worldwide, bre-ast cancer remains the most common malignancy. Since the introduction of breast cancer screening, most patients are diagnosed with breast cancer in an early stage and treated by breast-conserving therapy consisting of breast-conserving surgery followed by whole breast irradiation. The potential of MRI-guided radiotherapy has already been shown for other tumor sites. In order to explore whether early-stage breast cancer patients could benefit from future treatment by MRI-linac, the main focus of this thesis is to study the potential of MRI-guided radiotherapy in early-sta-ge breast cancer patients. Several aspects of MRI-guidance are described, such as

11

Introduction

potential magnetic field effects, MRI-guided target definition and the potential for an MRI-guided ablative radiotherapy approach.

Breast cancer

Breast cancer is the most common malignancy among woman worldwide 1. In-cidence rates are the highest in Western European countries. In the Netherlands, women are at a 13% risk of being diagnosed at some point during their life 2. With 14,000 patients newly diagnosed with invasive breast cancer in the Netherlands in 2012, the incidence is still rising since the introduction of breast cancer screening in the early nineties (Figure 2) [8]. As our population ages and diagnostic imaging mo-dalities are improving, we are likely to see a further increase in incidence [9]. In the majority of women diagnosed with breast cancer, the tumor is detected at an ear-ly-stage [10-12]. Therefore, most patients can be treated locally with breast-conser-ving therapy with or without adjuvant systemic therapies. Breast-conserving the-rapy, which is a combination of lumpectomy followed by radiotherapy, has shown survival equivalence with mastectomy, while preserving the affected breast [13]. However, breast-conserving treatment has some limitations. Due to its fractionated character, the radiotherapy treatment duration of several weeks makes it costly and time consuming [14]. Furthermore, the treatment volumes as excised by the surge-on or irradiated by the radiation oncologist can be large. As described by several studies, these large treatment volumes can lead to increased treatment toxicity and impaired cosmetic results [15-18].

Figure 1 A Schematic design of the MRI-linac concept [6]. The 1.5T MRI is shown in blue (1), the 8 MV accelerator (2) is located in a ring around the MRI. The split gradient coil (3) is shown in yellow and the superconducting coils (4) in orange. The light blue ring around the MRI indicated the low magnetic field torus (5) in the fringe field. B Prototype of the MRI-linac.

12

Chapter 1

Breast Radiotherapy

Whole breast and accelerated partial breast irradiation

Radiotherapy following breast-conserving surgery is an integral part of breast-con-serving treatment [13]. Traditionally, the whole breast is irradiated, with an additio-nal boost dose to the tumor bed in patients with a higher risk of local recurrence [19]. The conventional whole breast irradiation took 5 to 7 weeks of daily treatment, depending on disease characteristics and the applied fractionation scheme. This long treatment duration resulted in costly treatment and inconvenience for the pa-tient. The UK and Canadian hypofractionation studies in whole breast irradiation showed that the use of hypofractionated schedules of 3 to 5 weeks was equally ef-fective in terms of local control, survival, toxicity and cosmetic results [20-23]. Sever-al years ago, these hypofractionated schedules have therefore also been adopted in the Netherlands, and has resulted in decreased overall treatment time (i.e. 16 to 23 radiation treatments). Despite the shorter periods of irradiation treatment, the still prolonged treatment duration causes some patients to opt for mastectomy or to undergo suboptimal treatment by omitting radiotherapy after lumpectomy, or to choose endocrine tre-atment instead of any surgery [24-26]. Because of these downsides, further treat-ment acceleration up to 5 fractions (of 5.2 Gy or 5.4 Gy) in 1 week, is currently under investigation in the multicenter FAST-forward trial [27]. Another treatment option currently intensively studied as an alternative to whole breast irradiation in early-stage breast cancer patients is accelerated partial breast irradiation (APBI). ABPI is based on the principle that since most breast cancer recur-rences occur in the region of the original tumor bed, irradiation of the entire breast

Figure 2 Incidence of invasive breast cancer in the Netherlands 1989-2012 [8].

13

Introduction

may potentially be omitted in carefully selected patients with low risk of local recur-rence. [28][29, 30]. Low risk disease criteria are patient age over 50 or 60 years, fa-vorable histological subtype (i.e. invasive ductal carcinoma), unifocal tumor below 2 or 3 cm, negative resection margins, absence of lymphovascular invasion, estro-gen status positive, no extensive DCIS around the tumor, and negative lymph node status. By reducing treatment volumes using APBI, higher radiation doses may be delivered over the course of a shorter treatment period. Such larger radiation doses could potentially reduce overall treatment time resulting in decreased inconvenien-ce for the patients and health care costs [14]. Moreover, the dose to the surrounding organs at risk including heart and lungs can be reduced.

Several different APBI techniques have been developed since its introduction, in-cluding interstitial and balloon-based brachytherapy, radioactive breast seed brac-hytherapy and intraoperative radiotherapy [31-34]. Furthermore, APBI can also be delivered by standard external beam radiotherapy (EBRT) [35]. Each technique has its own strengths and weaknesses. For EBRT, its non-invasive character, the wide availability of linear accelerators, and the extensive experience with this technique, make it arguably one of the most accessible treatment options for APBI. This thesis will mainly focus on EBRT.

Breast tumor bed definition

The objective of both boost radiotherapy and APBI is to adequately irradiate mi-croscopic disease in the tumor bed and its surrounding rim of tissue in order to prevent disease from recurring locally. In daily clinical practice, tumor bed definition is performed on a radiotherapy planning CT scan, which is acquired several weeks after lumpectomy. On this CT scan, the radiation oncologist contours the tumor bed based on titanium clips inserted by the surgeon, architectural distortions, sero-ma formation, scar position, preoperative imaging and clinical breast examination (Figure 3). Subsequently, this contour is expanded to a clinical target volume (CTV) to account for subclinical microscopic disease in the rim of tissue surrounding the tumor bed [36]. To ensure adequate coverage of the CTV during radiotherapy, ano-ther margin expansion is applied to account for geometric uncertainties like setup variation and organ motion. This final volume is called the planning target volume (PTV). In addition, organs at risk of radiation exposure in the radiation field are also contoured on the CT. Based on both the PTV and organs at risk, a treatment plan is developed which aims to deliver a designated dose to the tumor bed region, while keeping the dose to the surrounding organs at risk as low as possible.

14

Chapter 1

Figure 3 CT-guided target definition of the postoperative tumor bed (green), clinical tar-get volume (yellow) and planning target volu-me (blue).

Since the tumor bed is delineated after breast-conserving surgery, when the tu-mor is not in situ anymore, the accuracy of post-lumpectomy tumor bed delineati-on cannot be validated by pathology studies. As an alternative, consensus among observers is often used to assess the precision of target volume delineation. The degree of consensus is referred to as the interobserver variability (IOV). Several different measures to quantify IOV are described in the literature. The conformity index (CI), which is the volume of agreement among observers divided by the to-tal encompassing volume, is often used. This measure is usually accompanied by the distance between the centers of mass (dCOM) among different target volumes. Multiple contouring studies have shown that tumor bed delineation on CT can be greatly inconsistent, which is reflected by the high IOV among radiation oncologists (Figure 4) [37-42]. In order to prevent geometrical miss of the target, accurate defi-nition of the tumor bed is of great importance, especially in APBI, in which only the tumor bed is irradiated and whole breast radiotherapy is omitted. Furthermore, as mentioned previously, target volumes can be large due to postoperative seroma and hematoma formation. In addition, target volumes can be variable in size due to the observers’ uncertainty about what to delineate, since there is no longer a tumor in situ. Irradiation of these large target volumes can lead to extended subcutaneous fibrosis and decreased cosmetic results [15-18]. Moreover, large treatment volumes can cause low-risk patients aiming for APBI to be ineligible for this treatment due to the inability to meet the dose-volume constraints of APBI [43].

Figure 4 Variation in tumor bed definition among radiation oncologists as represented by different colors.

15

Introduction

Definition of the tumor bed area is generally considered the weakest link in breast radiotherapy. Therefore, boost radiotherapy or APBI could be optimized by incre-asing the accuracy of tumor bed delineation. Such optimization could influence the future implementation of APBI, which is currently investigated in clinical trials into daily clinical practice. Furthermore, reducing treatment volumes would reduce radi-ation-induced toxicity and potentially allow for further treatment acceleration. This would increase patient eligibility for APBI and, thereby, patient convenience and treatment accessibility.

MRI-guided breast radiotherapy

When attempting to improve delineation accuracy on the standard radiotherapy planning CT scan, one of the main aims is to obtain a better differentiation between target tissue and non-target tissue. In the case of tumor bed delineation this would imply the differentiation between tumor bed and surrounding breast tissue. The use of breast MRI instead of CT could offer a promising alternative due to the supe-rior soft-tissue contrast of MRI. MRI is already applied in breast cancer diagnostics (Figure 5A). In diagnostic breast MRI, sequences can be varied to obtain a multitude of contrasts reflecting different tissue characteristics. Sequences that are common-ly used in diagnostic breast MRI are T1-weighted sequences with and without fat suppression, T2-weighted sequences, and dynamic contrast-enhanced sequences. In the radiotherapy setting, the combination of these sequences, together with the T2-weighted sequence with fat suppression, could offer superior differentiation be-tween, for instance, fatty breast tissue, blood, seroma and fibroglandular tissue. One of the aims of this thesis is, therefore, to consider the potential value of postopera-tive breast MRI for tumor bed definition in radiation therapy. However, diagnostic breast MRI is performed in prone position, while most radiotherapy departments treat breast cancer patients in supine position and therefore acquire the standard planning CT in supine treatment position (Figure 5B). Consequently, the CT imaging could not match the standard diagnostic breast MRI, which is performed in prone position. Moreover, a patient in radiotherapy supine position with the arms above the head would not fit into a standard MRI scanner due to the limited bore size. In order to solve these issues, we developed an MRI protocol in radiotherapy supine position with the arms in abduction and the hands above the head using a wide bore MRI scanner (Figure 5C and D).In the diagnostic setting, the dynamic contrast-enhanced series are highly valued, since they can reflect characteristic parameters of the tumor. Furthermore, diagnos-tic MRI has a high sensitivity for tumor detection and a good correlation with the microscopic tumor size [44, 45]. For these reasons, a preoperative target definiti-on in radiation treatment could be beneficial compared to standard postoperative

16

Chapter 1

target definition. Moreover, preoperative contrast-enhanced imaging could allow for superior differentiation between tumor and surrounding tissues compared to the postoperative soft-tissue differences between a tumor bed and its surrounding tissues. This could make target definition for radiotherapy more accurate. Further-more, a preoperative tumor volume would be smaller compared to a postoperative tumor bed, due to the absence of seroma formation. This could result in a reduced radiation dose delivery on the ipsi- and contralateral breast and other organs at risk. In the literature, the potential of a preoperative radiation treatment has been descri-bed by several groups [43, 46, 47]. These studies all describe a CT-guided approach. Due to the aforementioned advantages of dynamic contrast enhanced MRI, this thesis will consider the potential for a preoperative MRI-guided target definition. The smaller treatment volumes in combination with the possibilities for more con-sistent target definition in the preoperative setting could enable further treatment acceleration. This treatment acceleration could potentially be extended into an abla-tive radiotherapy approach, as proposed by Horton et al [48]. Providing a single treat-ment dose to the tumor and its surroundings could minimize the patient’s treatment burden. In a future MRI-linac setting treatment could be even further improved, sin-ce the MRI-linac offers high-contrast real time image guidance, which would allow for online set-up correction and motion compensation during irradiation. However, in the future setting of the MRI-linac, a static magnetic field is always present during treatment. Therefore, it is essential to first investigate the induced effects of the magnetic field itself on the dose distribution. Secondary electrons, which deliver the actual dose to the tissue, moving in a magnetic field, are acted

Figure 5 A Prone position, diagnostic dynamic contrast-enhanced MRI, T1-weighted with fat sup-pression B Supine radiotherapy position, postoperative planning CT scan C Newly developed preoperative dynamic contrast-enhanced MRI in supine radiotherapy position, T1-weighted with fat suppression (DIXON). D Newly developed postoperative MRI in supine radiotherapy position, T1-weighted with fat suppression (DIXON).

17

Introduction

upon by the Lorentz force [49]. In MRI-linac breast irradiation, this electron return effect (ERE) could result in a dose increase in the skin, which could influence skin toxicity and cosmetic results. Since the ERE depends on the beam inclinations at the skin surface, which differ between whole breast irradiation and APBI, the induced effects of the magnetic field are expected to differ between these treatments. Con-sequently, the feasibility of treating patients by MRI-linac is likely to depend on the treatment technique.

Aim and outline of this thesis

In order to explore whether early-stage breast cancer patients could benefit from future treatment by MRI-linac, the main focus of this thesis is to study the potential of MRI-guided radiotherapy in early-stage breast cancer patients. Several aspects of MRI-guidance are described, such as potential magnetic field effects, precision of MRI-guided target definition, its influence on treatment volumes, and the potential for an MRI-guided ablative treatment approach.Treatment volumes are associated with treatment toxicity and cosmetic results. In breast-conserving therapy, the size of the lumpectomy specimen and the irradiated tumor bed region can be large due to uncertainties in surgical localization and radi-otherapy target definition. The relations between these treatment volumes and the microscopic tumor size are evaluated in Chapter 2. Future treatment by MRI-linac could provide high contrast image guidance with the possibility to adapt the radiotherapy plan during irradiation. However, in the presence of a magnetic field, skin dose could be increased due to the ERE. This could potentially influence toxicity and cosmetic results in breast cancer patients. In Chapter 3, the effect of the magnetic field on the skin dose, for both whole bre-ast irradiation and APBI, is simulated by generating IMRT plans for 0 T and 1.5 T MRI using specially developed MRI-linac treatment planning software. Diagnostic MRI is performed in prone position, while most radiotherapy institutions perform treatment planning and irradiation with the patient in supine, radiothera-py position. Therefore, an MRI protocol had to be developed in RT supine position in a wide bore MRI scanner. This is described in Chapter 4.Our results described in Chapter 3 show that, in the presence of a magnetic field, skin dose would be significantly increased in whole breast irradiation compared to an APBI setting. Therefore, the value of MRI for definition of the tumor bed is further investigated in the subsequent chapters. Chapter 4 describes the value of MRI in addition to standard CT-guided tumor definition. In this study, tumor bed definition is performed by radiation oncologists from the UMC Utrecht, The Nether-lands. Chapter 5 describes tumor bed definition on MRI-only. In this chapter, tumor bed definition is performed by radiation oncologists from the Sunnybrook Health

18

Chapter 1

Science Center, Toronto, Canada. In this study, the influence of full-thickness closure on tumor bed definition is also investigated. To investigate whether target definiti-on using MRI is superior to the current standard, MRI-guided tumor bed definition is compared to standard CT-guided tumor bed definition in both chapters. Chapter 4 describes the value of MRI in addition to standard CT-guided tumor bed definition. In addition, Chapter 5 describes tumor bed definition on MRI-only. MRI in combination with administration of a contrast agent would provide detail on tu-mor morphology, enhancement and kinetics. This could potentially enable more consistent target definition combined with small treatment volumes. Therefore, preoperative target definition is investigated in Chapter 6. To identify whether dy-namic contrast enhanced MRI is the best imaging modality for preoperative target volume delineation, preoperative delineation is performed on both contrast-en-hanced CT and MRI in radiotherapy supine position. The advantages of small preoperative target volumes and more consistent target definition on preoperative MRI could allow further treatment acceleration up to a single ablative dose. This single fraction radiotherapy would increase patient con-venience and eligibility. To investigate whether a single ablative dose to the tumor and its surrounding tumor bed is feasible, a planning study is performed. This plan-ning study, in which both brachytherapy and IMRT plans were acquired and compa-red, is described in Chapter 7.Chapter 8 gives a summary of the most important results and discusses the poten-tial for MRI-guided radiotherapy as part of breast-conserving therapy.

19

Introduction

References

1. Haie-Meder C, Potter R, Van Limbergen E et. al. Recommendations from Gynaecological (GYN) GEC-

ESTRO Working Group (I): concepts and terms in 3D image based 3D treatment planning in cervix

cancer brachytherapy with emphasis on MRI assessment of GTV and CTV. Radiother Oncol 2005:

74:235-245.

2. Lim K, Small W,Jr, Portelance L et. al. Consensus guidelines for delineation of clinical target volume

for intensity-modulated pelvic radiotherapy for the definitive treatment of cervix cancer. Int J Radiat

Oncol Biol Phys 2011: 79:348-355.

3. Rasch C, Steenbakkers R, van Herk M. Target definition in prostate, head, and neck. Semin Radiat

Oncol 2005: 15:136-145.

4. O’Neill BD, Salerno G, Thomas K, Tait DM, Brown G. MR vs CT imaging: low rectal cancer tumour

delineation for three-dimensional conformal radiotherapy. Br J Radiol 2009: 82:509-513.

5. Seierstad T, Hole KH, Saelen E, Ree AH, Flatmark K, Malinen E. MR-guided simultaneous integrated

boost in preoperative radiotherapy of locally advanced rectal cancer following neoadjuvant che-

motherapy. Radiother Oncol 2009: 93:279-284.

6. Raaymakers BW, Lagendijk JJ, Overweg J et. al. Integrating a 1.5 T MRI scanner with a 6 MV accelera-

tor: proof of concept. Phys Med Biol 2009: 54:N229-37.

7. Lagendijk JJ, Raaymakers BW, Raaijmakers AJ et. al. MRI/linac integration. Radiother Oncol 2008:

86:25-29.

8. . Breast cancer in the Netherlands: Incidence, prevalence, survival, mortality and risk. Comprehensive

cancer center the Netherlands. www.cijfersoverkanker.nl.

9. Bluekens AM, Holland R, Karssemeijer N, Broeders MJ, den Heeten GJ. Comparison of digital scree-

ning mammography and screen-film mammography in the early detection of clinically relevant

cancers: a multicenter study. Radiology 2012: 265:707-714.

10. Verbeek AL, Broeders MJ, Otto SJ et. al. Effects of the population screening into breast cancer. Ned

Tijdschr Geneeskd 2013: 157:A5218.

11. Miller AB, To T, Baines CJ, Wall C. Canadian National Breast Screening Study-2: 13-year results of a

randomized trial in women aged 50-59 years. J Natl Cancer Inst 2000: 92:1490-1499.

12. Autier P, Hery C, Haukka J, Boniol M, Byrnes G. Advanced breast cancer and breast cancer mortality

in randomized controlled trials on mammography screening. J Clin Oncol 2009: 27:5919-5923.

13. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Darby S, McGale P et. al. Effect of radi-

otherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death:

meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet 2011:

378:1707-1716.

14. Suh WW, Pierce LJ, Vicini FA, Hayman JA. A cost comparison analysis of partial versus whole-breast

irradiation after breast-conserving surgery for early-stage breast cancer. Int J Radiat Oncol Biol Phys

2005: 62:790-796.

15. Mukesh MB, Barnett G, Cumming J et. al. Association of breast tumour bed seroma with post-opera-

tive complications and late normal tissue toxicity: results from the Cambridge Breast IMRT trial. Eur

J Surg Oncol 2012: 38:918-924.

20

Chapter 1

16. Collette S, Collette L, Budiharto T et. al. Predictors of the risk of fibrosis at 10 years after breast con-

serving therapy for early breast cancer: a study based on the EORTC Trial 22881-10882 ‘boost versus

no boost’. Eur J Cancer 2008: 44:2587-2599.

17. Vrieling C, Collette L, Fourquet A et. al. The influence of patient, tumor and treatment factors on

the cosmetic results after breast-conserving therapy in the EORTC ‘boost vs. no boost’ trial. EORTC

Radiotherapy and Breast Cancer Cooperative Groups. Radiother Oncol 2000: 55:219-232.

18. Borger JH, Kemperman H, Smitt HS, Hart A, van Dongen J, Lebesque J, Bartelink H. Dose and volume

effects on fibrosis after breast conservation therapy. Int J Radiat Oncol Biol Phys 1994: 30:1073-1081.

19. Bartelink H, Horiot JC, Poortmans PM et. al. Impact of a higher radiation dose on local control and

survival in breast-conserving therapy of early breast cancer: 10-year results of the randomized

boost versus no boost EORTC 22881-10882 trial. J Clin Oncol 2007: 25:3259-3265.

20. Whelan TJ, Pignol JP, Levine MN et. al. Long-term results of hypofractionated radiation therapy for

breast cancer. N Engl J Med 2010: 362:513-520.

21. Hopwood P, Haviland JS, Sumo G, Mills J, Bliss JM, Yarnold JR, START Trial Management Group. Com-

parison of patient-reported breast, arm, and shoulder symptoms and body image after radiothera-

py for early breast cancer: 5-year follow-up in the randomised Standardisation of Breast Radiothera-

py (START) trials. Lancet Oncol 2010: 11:231-240.

22. START Trialists’ Group, Bentzen SM, Agrawal RK et. al. The UK Standardisation of Breast Radiotherapy

(START) Trial A of radiotherapy hypofractionation for treatment of early breast cancer: a randomised

trial. Lancet Oncol 2008: 9:331-341.

23. START Trialists’ Group, Bentzen SM, Agrawal RK et. al. The UK Standardisation of Breast Radiotherapy

(START) Trial B of radiotherapy hypofractionation for treatment of early breast cancer: a randomised

trial. Lancet 2008: 371:1098-1107.

24. van de Water W, Bastiaannet E, Dekkers OM et. al. Adherence to treatment guidelines and survival

in patients with early-stage breast cancer by age at diagnosis. Br J Surg 2012: 99:813-820.

25. Field TS, Bosco JL, Prout MN et. al. Age, comorbidity, and breast cancer severity: impact on receipt of

definitive local therapy and rate of recurrence among older women with early-stage breast cancer.

J Am Coll Surg 2011: 213:757-765.

26. Morgan J, Wyld L, Collins Karen A, Reed Malcolm W. Surgery versus primary endocrine therapy for

operable primary breast cancer in elderly women (70 years plus). Cochrane Database of Systematic

Reviews 2014.

27. The FAST Trialists’ Group. Fast Forward: A randomised clinical trial testing a 1 week course of curative

whole breast radiotherapy against a standard 3 week schedule in terms of local cancer control and

late adverse effects in women with early breast cancer. Study information provided by: UK Clinical

Research Network Study Portfolio public.ukcrn.org.uk. ISRTCTN 19906132.

28. Mannino M, Yarnold J. Accelerated partial breast irradiation trials: diversity in rationale and design.

Radiother Oncol 2009: 91:16-22.

29. Polgar C, Van Limbergen E, Potter R et. al. Patient selection for accelerated partial-breast irradiation

(APBI) after breast-conserving surgery: recommendations of the Groupe Europeen de Curiethera-

pie-European Society for Therapeutic Radiology and Oncology (GEC-ESTRO) breast cancer working

group based on clinical evidence (2009). Radiother Oncol 2010: 94:264-273.

21

Introduction

30. Smith BD, Arthur DW, Buchholz TA et. al. Accelerated partial breast irradiation consensus statement

from the American Society for Radiation Oncology (ASTRO). Int J Radiat Oncol Biol Phys 2009:

74:987-1001.

31. Njeh CF, Saunders MW, Langton CM. Accelerated Partial Breast Irradiation (APBI): A review of availa-

ble techniques. Radiat Oncol 2010: 5:90.

32. Hepel JT, Wazer DE. A comparison of brachytherapy techniques for partial breast irradiation. Brachy-

therapy 2012: 11:163-175.

33. Pignol JP, Rakovitch E, Keller BM, Sankreacha R, Chartier C. Tolerance and acceptance results of a

palladium-103 permanent breast seed implant Phase I/II study. Int J Radiat Oncol Biol Phys 2009:

73:1482-1488.

34. Klepczyk LC, Keene KS, De Los Santos JF. Accelerated partial breast irradiation for early-stage breast

cancer: controversies and current indications for use. Curr Treat Options Oncol 2013: 14:51-65.

35. Njeh CF, Saunders MW, Langton CM. Accelerated partial breast irradiation using external beam con-

formal radiation therapy: a review. Crit Rev Oncol Hematol 2012: 81:1-20.

36. ICRU Report. No. 50 1993.

37. van Mourik AM, Elkhuizen PH, Minkema D, Duppen JC, Dutch Young Boost Study Group, van

Vliet-Vroegindeweij C. Multiinstitutional study on target volume delineation variation in breast radi-

otherapy in the presence of guidelines. Radiother Oncol 2010: 94:286-291.

38. Hurkmans C, Admiraal M, van der Sangen M, Dijkmans I. Significance of breast boost volume chan-

ges during radiotherapy in relation to current clinical interobserver variations. Radiother Oncol

2009: 90:60-65.

39. Coles CE, Wilson CB, Cumming J et. al. Titanium clip placement to allow accurate tumour bed lo-

calisation following breast conserving surgery: audit on behalf of the IMPORT Trial Management

Group. Eur J Surg Oncol 2009: 35:578-582.

40. Petersen RP, Truong PT, Kader HA et. al. Target volume delineation for partial breast radiotherapy

planning: clinical characteristics associated with low interobserver concordance. Int J Radiat Oncol

Biol Phys 2007: 69:41-48.

41. Landis DM, Luo W, Song J et. al. Variability among breast radiation oncologists in delineation of the

postsurgical lumpectomy cavity. Int J Radiat Oncol Biol Phys 2007: 67:1299-1308.

42. Struikmans H, Warlam-Rodenhuis C, Stam T, Stapper G, Tersteeg RJ, Bol GH, Raaijmakers CP. Interob-

server variability of clinical target volume delineation of glandular breast tissue and of boost volu-

me in tangential breast irradiation. Radiother Oncol 2005: 76:293-299.

43. Nichols EM, Feigenberg SJ, Marter K et. al. Preoperative Radiation Therapy Significantly Increases

Patient Eligibility for Accelerated Partial Breast Irradiation Using 3D-conformal Radiotherapy. Am J

Clin Oncol 2012: 36:232-238.

44. Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH. Meta-analysis of MR imaging

in the diagnosis of breast lesions. Radiology 2008: 246:116-124.

45. Schmitz AC, van den Bosch MA, Loo CE et. al. Precise correlation between MRI and histopathology -

exploring treatment margins for MRI-guided localized breast cancer therapy. Radiother Oncol 2010:

97:225-232.

22

Chapter 1

46. Palta M, Yoo S, Adamson JD, Prosnitz LR, Horton JK. Preoperative single fraction partial breast radio-

therapy for early-stage breast cancer. Int J Radiat Oncol Biol Phys 2012: 82:37-42.

47. van der Leij F, Elkhuizen PH, Janssen TM et. al. Target volume delineation in external beam partial

breast irradiation: Less inter-observer variation with preoperative- compared to postoperative deli-

neation. Radiother Oncol 2014: 110:467-470.

48. Horton JK, Blitzblau RC, Yoo S et. al. Preoperative Single-Fraction Partial Breast Radiation Therapy:

A Novel Phase 1 Dose-Escalation Protocol and Exploration of Breast Cancer Radiation Response.

International Journal of Radiation Oncology*Biology*Physics 2013: 87:S229.

49. Raaijmakers AJ, Raaymakers BW, Lagendijk JJ. Integrating a MRI scanner with a 6MV radiotherapy ac-

celerator: dose increase at tissue-air interfaces in a lateral magnetic field due to returning electrons.

Phys Med Biol 2005: 50:1363-1676.

Chapter 2Excised and irradiated volumes

in relation to the tumor size in breast-conserving therapy

M.D. den Hartogh, B. van Asselen, E.M. Monninkhof, M.A.A.J. van den Bosch, M. van Vulpen, P.J. van Diest, K.G.A. Gilhuijs,

A.J. Witkamp, L. van de Bunt, W.P.T.M. Mali, H.J.G.D. van den Bongard

Based on : Breast Cancer Res Treat 2011: 129:857-865

26

Chapter 2

Abstract

Purpose

In early stage breast cancer and DCIS patients, breast-conserving therapy is today’s standard of care. The purpose of this study was to evaluate the relation between the microscopic tumor diameter (mTD), the excised specimen (ES) volume, and the ir-radiated postoperative complex (POC) volume, in patients treated with breast-con-serving therapy.

Methods

In 186 patients with pTis-2N0 breast cancer, the mTDs, ES and POC volumes (as deli-neated on the radiotherapy planning CT scan), were retrospectively determined. Li-near regression analysis was performed to study the association between the mTD, and the ES and POC volumes. The explained variance (r2) was calculated to establish the proportion of variation in the outcome variable that could be explained by the determinant (p≤0.05). Moreover, the influence of tumor characteristics, age, surgi-cal procedures, and breast size was studied.

Results

Median mTD was 1.2 cm (range 0.1-3.6 cm), median ES volume was 60 cm3 (range 6-230 cm3) and median POC volume was 15 cm3 (range 0.5-374 cm3). The POC was not clearly visible on the majority of the CT scans, based on a median assigned ca-vity visualization score of 3 (range 1-5). The explained variance (r2) for the mTD on the ES volume was low (r2=0.08, p<0.001). A slightly stronger association was obser-ved in palpable tumors (r2=0.23, p<0.001) and invasive lobular carcinomas (r2=0.39, p=0.01). Furthermore, weak associations were observed between POC volume and mTD (r2=0.04, p=0.01), and POC and ES volume (r2=0.23, p<0.001). A weak associati-on was observed between breast volume and ES volume (r2=0.27, p<0.001).

Conclusion

Both the excised and the irradiated POC volumes did not show a clinically relevant association with the mTD in women with early-stage breast cancer treated with bre-ast-conserving therapy. Future studies should focus on improvement of surgical lo-calization, development of image-guided, minimally invasive operation techniques, and more accurate image-guided target volume delineation in radiotherapy.

27

Treatment volumes in breast-conserving therapy

Introduction

The introduction of screening programs and the improvement in breast imaging techniques, have led to an increased detection of early-stage breast cancer and ductal carcinoma in situ (DCIS) [1, 2]. In these patients, breast-conserving therapy is today’s standard of care, since it has demonstrated survival equivalence to mas-tectomy in many DCIS and early breast cancer clinical trials [3-6]. Breast-conserving therapy consists of breast-conserving surgery and radiotherapy.In breast-conserving surgery, it is important to obtain tumor free resection margins, since margin status is an important prognostic factor for local recurrence after bre-ast- conserving therapy [7-9]. Moreover, a margin of breast tissue surrounding the tumor has to be treated, due to the potential presence of additional microscopic tumor foci [10-13]. This margin is established by both the surgeon and the radi-ation oncologist. The surgeon excises the tumor including a 1-2 cm macroscopic margin (Figure 1). After breast- conserving surgery, another 1-2 cm margin has to be treated by the radiation oncologist. Therefore, the excision cavity walls have to be identified on a planning CT scan and delineated by the radiation oncologist. This area will be referred to as the postoperative complex (POC). After delineation, the POC is expanded with another 1-2 cm margin, resulting in an area called the Clinical

Figure 1 Schematic, theoretical margin assessment (A). Schematic, post-operative margin assess-ment without full-thickness closure (B) and with full-thickness closure (C). T tumor, ES excised specimen margin of 1-2 cm around the tumor, CTV clinical target volume

28

Chapter 2

Target Volume (CTV). Besides irradiation of the whole breast, this CTV receives a higher radiotherapy dose, known as the radiotherapy boost. This additional boost has demonstrated to further reduce local recurrence rates in BCT with nearly 50% in patients ≤70 years [14, 15]. Exact localization of the POC is essential, since local recurrences are mainly found in this area [16]. It would be expected that both the excised specimen (ES) volume and the POC volume correlate with the microscopic tumor size. One other study, performed by Christiaens et al., showed lack of correlation between the microscopic tumor dia-meter and the ES size [17]. Weak associations between irradiated target volumes, and the mTD and ES volumes, have been previously described by Al Uwini et al. [18]. Other investigators observed no relation between mTD and POC volume [19]. Un-derstanding the relation between the original tumor size and the treated volumes is important, including the influence of tumor and patient characteristics, surgical techniques and breast size. It can show us if surgical techniques and/or radiation therapy procedures can be further improved. Therefore, the aim of this study was to determine more in detail, the relation be-tween the microscopic tumor diameter, the excised specimen volume, and the volume of the irradiated postoperative complex, in patients undergoing bre-ast-conserving therapy. Moreover, we studied the influence of patient and tumor characteristics, surgical techniques and breast size on these associations.

Methods

In this retrospective study, the medical files of female patients with pT1-2N0M0 breast cancer and/or DCIS treated with BCS, were analyzed. Patients were referred to our institute between July 2008 and December 2009 for standard postoperative local radiation therapy of the breast. Patients with multicentric breast cancer were excluded, as well as patients who were treated with neo-adjuvant chemotherapy, or who underwent an axillary lymph node dissection.

Pathology and surgery

In all patients, the pathology reports were studied to obtain the tumor characte-ristics, e.g. histology (invasive ductal carcinoma, invasive lobular carcinoma, DCIS), excised specimen dimensions (mm), microscopic tumor dimensions (mm) and tumor-free margins (yes/no). Margins were considered as ‘not free’, when margins were reported to be involved. Since in most cases only the maximum microscopic tumor dimension was reported by the pathologist, the microscopic tumor diameter (mTD), rather than volume was considered.

29


Surgical reports were assessed, to obtain information about excision and closure techniques. We differentiated between resection from the ventral fascia of Scarpa to the pectoral fascia (yes/no), and the execution of full-thickness closure (yes/no). More detailed information about the excision technique, e.g. whether lumpectomy, wide local excision, sector resection or quadrantectomy was performed, was not available. Furthermore, information about the placement of titanium clips (yes/no), the number of clips, and re-excision (yes/no) was extracted from the medical files. Obscurities or uncertainties in the operation report were clarified by inquiring the corresponding surgeon.In the pathology reports, the maximum length (mm), width (mm) and height (mm) of the ES was described. Since the weight of the ES was only described in 39 pa-tients, we estimated the total excised tissue volume, by the assumption that the specimen was ellipsoidally-shaped. The volume of an ellipsoid can be calculated by: Volume = ⅙ π length * width * height. This model was verified in the subgroup whe-re the specimen weight was known. No-intercept linear regression analysis showed a 0.96 correlation of the calculated ES volume with the specimen weight, with a slope of 0.84 (CI 0.76-0.92). Subsequently, ES volumes were multiplied by 0.84.If a re-excision was performed, the same procedure was executed and the volumes were summed up to calculate the total excised volume.

Postoperative complex and radiotherapy

The patients were referred for standard postoperative whole breast irradiation (50 Gy) including a POC boost (16 Gy). If resection margins were not free, and no re-ex-cision was performed, a high POC boost was given (20 Gy). Before the start of radio-therapy, a standard supine radiotherapy-planning CT scan, obtained at a 3 mm slice thickness and a minimal in-plane resolution of 1 x 1 mm, was performed in each patient. A radio-opaque wire was placed around the palpable edge of breast tissue and on the scar, in order to define the fibroglandular tissue of the whole breast and the incision location, respectively. An in-house developed software tool (Volumetool®) was used for delineation of the POC on each CT-slice [20]. The POC volume was calcu-lated by rasterizing the delineation on the planning CT and multiplying the number of rastered voxels by the voxel volume. The POC was delineated based on the combina-tion of pre-operative radiological imaging, any physical examination, postoperative seroma, architectural distortion, position of surgical clips and the scar localization (Figure 2). In order to quantify the POC visibility on the radiotherapy-planning CT scans, a cavity visualization score (CVS) according to Smitt et al., was retrospectively assigned to each POC volume [21]. The CVS ranged from not distinguishable (CVS 1) to clearly defined (CVS 5) (Table 1).

30

Chapter 2

Figure 2 Transversal section of a postope-rative radiotherapy-planning CT scan, showing the radiotherapy target volumes. Surgical clip in situ. POC postoperative complex, CTV clinical target volume

Table 1 Cavity Visualisation Score (CVS)

CVS Cavity description n %

CVS 1 Cavity not visualized 5 3

CVS 2 Heterogeneous cavity with indistinct margins 68 36

CVS 3 Heterogeneous cavity with some distinct margins 46 25

CVS 4 Mildy heterogeneous cavity with mostly dinstict margins 32 17

CVS 5 Homogeneous cavity with clearly defined margins 35 19

Delineation of the POC was performed with a fixed window level (20 Hounsfield Units (HU)) and width (500 HU). In this retrospective study, delineation was perfor-med by 12 radiation oncologists and 10 supervised radiation oncologist in training. If substantial seroma was observed on the first planning CT, a second CT was per-formed approximately 1-2 weeks before the start of the radiation boost treatment. Breast size was expected to have a possible influence on the ES volume. It was hypo-thesized that ES volume would be increased in patients with larger breasts. Informa-tion about the pre-operative breast size could not retrospectively be extracted. We assumed the contralateral breast volume to resemble the pre-operative ipsilateral breast volume. Therefore, we used the planning CT scan and accompanying soft-ware tool to delineate the contralateral breast and determine the volume (Volume-tool®) [20].

Statistical Analysis

Statistical Package for Social Sciences (version 16.0, SPSS, Chicago IL) was used to analyze the data. Scatter plots were established to visualize the association between the mTD, and the ES and POC volumes. We performed linear regression analyses to determine the associations between these variables. The following associations were studied:

31


1. mTD as determinant and ES volume as outcome2. mTD as determinant and POC volume as outcome 3. ES volume as determinant and POC volume as outcome

The explained variance (r2) was calculated to establish the proportion of variation in the outcome variable that could be explained by the determinant. The r2 was consi-dered as statistically significant when p ≤ 0.05. Furthermore, beta coefficients were determined. Since ES and POC volumes are volumetric variables which showed skewed distributions, values were transformed by calculating the cubic root (3√). Transformed values of the ES and POC volumes were used for regression analyses. In subgroup analyses, we studied whether tumor characteristics (i.e. histology, pal-pable vs. not palpable, localization), age or re-excision modified the associations. Furthermore, we studied the influence of surgical resection and closure techniques and breast size on the association between mTD and the ES volume. The influence of localization technique on the association was not studied. The majority of the non-palpable tumors were localized by wire localization and the other subgroups would become too small for statistical significance.

Results

In total, 208 patients were potentially eligible and charts were reviewed. In 7 pa-tients, all with invasive ductal carcinoma, the size of the ES was not reported. In 15 patients, all with ductal carcinoma in situ, the mTD was not mentioned. Conse-quently, these 22 patients were excluded. The characteristics of the resulting 186 patients included for analysis are presented in Table 2. Most patients were diagno-sed with an invasive ductal carcinoma with or without DCIS (86%), predominantly pathological stage T1 (80%). The median mTD was 1.2 cm (range 0.1-3.6 cm). The mTD and treated volumes are shown in Table 3.

Surgery

The median ES volume was 60 cm3 (range 6-230 cm3). Pre-operatively, lesions were mostly localized by palpation or by wire guided localization in non-palpable lesi-ons (Table 2). Excision to the pectoral fascia was performed in 153 patients. In 171 patients, titanium clips were placed intra-operatively (median number of 4 clips, range 1-10). Full-thickness closure of the excision cavity was performed in most pa-tients (n=140). In 17 patients, information about the closure procedure could not be extracted. In 24 patients, resection margins were microscopically involved. In 12 of these patients, a re-excision was performed, 12 patients received a high radio-therapy boost dose.

32

Chapter 2

Table 2 Patient and treatment characteristics

N=186 %

Age (years)Median

24-6051

Tumor stagepTispT1apT1bpT1cpT2

917409327

59

215015

Histologyductal carcinomaductal carcinoma and DCISlobular carcinomaDCISother

93671592

5036851

Affected breastleftright

9096

4852

Tumor locationupper lateral quadrantlower lateral quadrantinner quadrantscentral

104174421

569

2411

Palpableyesno

9987

5347

Surgical localization techniquepalpationwire-guided localizationROLL*ultrasoundnot described

82756

167

4540375

Excision to pectoral fasciayesno

15333

8218

Titanium clips placedyesno

17115

928

Full-thickness closureyesnonot described

1402917

75169

Resection margins involvedRe-excisionHigh radiotherapy boost

241212

135050

33


Radiotherapy

The median time between surgery and the first planning CT was 24 days. The me-dian POC volume was 15 cm3 (range 0.5-374 cm3). The median CTV was 78 cm3 (ran-ge 13-658 cm3). A second radiotherapy-planning CT was performed in 67 patients, due to the presence of substantial seroma on the first planning CT. This resulted in a median decrease in POC volume from 42 to 21 cm3 (range 2-374 cm3 compared to 1.4-245 cm3 on the second planning CT). The median time between the 2 CT scans was 38 days, corresponding with a median POC volume reduction of 0.6 cm3 per day. Mean CVS on the first CT scan (in patients who underwent a second CT scan) was 3.94, compared to 2.67 in the rest of the study population, due to the presence of more seroma in the first group of patients. Mean CVS on the second CT scan was 3.50.

Regression analyses

The associations between the mTD, and the ES and POC volumes are visualized in Figure 3. The coefficient of correlation (r) and explained variance (r2), are depicted in the upper corner of each figure. No significant associations were found between the mTD and the ES and POC volumes. The explained variance for the mTD on the ES volume is 0.08, reflecting that only 8% of the variation in ES volumes could be ex-plained by the mTD. Only 4% of the variation in POC volume could be explained by the mTD. A slightly stronger association was found between ES and POC volumes, with an explained variance of 0.23. These associations did not alter in the subgroup where a second CT scan was performed. Due to the weak associations observed, the beta coefficients were not indicative and therefore, not shown. The influence of several tumor characteristics (i.e. palpable vs. non-palpable tumor, re-excision, histology, location, excision technique, and patient age) on the associa-tion between the mTD and ES volume was studied.

Table 3 Microscopic tumor size and treated volumes

Median Range

Microscopic tumor diameter (cm) 1.2 0.1-3.6

Excised specimen volumes:

Excised specimen volume (cm3) 60 6-230

Patients who underwent a re-excision 93 43-230

Patients who did not underwent a re-excision 55 6-206

Irradiated volumes:Delineated postoperative complex volume (cm3) 15 0.5-374Clinical target volume (cm3) 77 13-658

34

Chapter 2

Figure 3 Scatter plots, showing the association between mi-croscopic tumor diameter and excised specimen volume (A) microscopic tumor diameter and postoperative complex volume (B) and excised specimen volu-me and postoperative complex volume (C). r coefficient of corre-lation, r2 explained variance

35


The results of the subgroup analysis are presented in Table 4. The association bet-ween the mTD and ES volume was slightly strengthened in invasive lobular carci-nomas (r2 = 0.39), invasive ductal carcinomas without DCIS component (r2 = 0.11), palpable tumors (r2 = 0.23), and in patients ≤ 50 years of age (r2 = 0.18). No major effect modification of tumor location or breast size on the association between the mTD and the ES volume was found. Furthermore, the influence of the surgical closure technique on the association be-tween ES and POC volume was studied. Full-thickness closure of the breast tissue was performed in 140 patients and not performed in 29 patients (missing data in 17 patients). The explained variance of ES on the POC volume decreased from 0.38 to 0.20 (both p<0.001) when full-thickness closure was performed (data not shown). In Figure 4, the association between breast volume and the ES volume is shown, the T-stages, according to the TNM-classification, are depicted in different symbols [22]. An explained variance of 0.27 was observed. No association between breast volume and ES volume according to T-stage was observed.

Table 4 Association between microscopic tumor diameter (mTD) and excised specimen (ES) volumes among subgroups

n r2 p-valuePalpable

Yes 99 0.23 <0.001No 87 0.02 0.26

Re-excisionYes 16 0.02 0.66No 170 0.11 <0.001

HistologyInvasive ductal 93 0.11 0.001Invasive ductal + DCIS 67 0.03 0.15Invasive lobular 15 0.39 0.01DCIS 9 0.01 0.85Other 2 1 -

LocationUpper lateral 104 0.09 0.02Lower lateral 17 0.12 0.17Inner 44 0.03 0.31Central 21 0.10 0.16

Age< 50 years 85 0.18 <0.00151-60 years 101 0.03 0.07

36

Chapter 2

Discussion

In this study, statistically significant, but weak, and therefore clinically not relevant associations were observed between the microscopic tumor diameter (mTD), the excised specimen (ES) volume, and the irradiated postoperative complex (POC) vo-lume in 186 DCIS and early stage breast cancer patients treated with breast-conser-ving therapy. ES volume was not strongly associated with mTD in our study, which is in accor-dance with previous findings by Christiaens et al. [17]. We observed that different amounts of breast tissue were excised for a similar tumor size. This implicates that surgeons could not excise a tumor with fixed macroscopic margins of healthy bre-ast tissue. ES volume showed a slightly better relation with breast volume (r2=0.27) than with tumor size (r2=0.08). This might be explained by the surgical technique being performed. Surgeons generally dissect the mammary gland entirely from the ventral fascia of Scarpa to the pectoral fascia, to guarantee radicality in the ventral and dorsal direction [23]. This could theoretically result in larger ES volumes in larger breast. Nevertheless, subgroup analysis did not show any differences in association between mTD and ES volume, whether the resection was performed to the pectoral

Figure 4 Scatter plot, showing the correlation between the breast volume in cm3 and the excised specimen volume according to TNM T-stage. * DCIS, o T1-stage, • T2-stage, r2 coefficient of determi-nation

37


fascia (n=153) or not (n=33). This can be caused by the small patient numbers in the other subgroup in which no excision to the pectoral fascia was performed. We observed a slightly stronger association between mTD and ES volume in pa-tients with palpable tumors. Presumably, small and non-palpable tumors are more difficult to localize, whereby fixed macroscopic margins are even more difficult to achieve. Accurate lesion localization has become more important due to the higher incidence of small and non-palpable lesions that is caused by the introduction of screening programs [1, 2]. Wire-localized breast biopsy (WLBB) is the current stan-dard method for the surgical excision of non-palpable breast tumors. WLBB has se-veral limitations, e.g. discrepancy between the entry site of the wire and the target lesion, or wire displacement, which can lead to insufficient tumor excision or the re-moval of excess breast tissue [24]. Consequently, further optimization of image-gui-ded localization techniques, which facilitate surgical removal of non-palpable bre-ast tumors, is required. Alternatives such as radio-guided occult lesion localization, intraoperative ultrasound and radioactive seed localization, are being studied [24, 25]. A slightly stronger association between the mTD and ES volume was found in pa-tients with isolated ductal or lobular carcinoma. A possible explanation could be the difficult visualization of the whole DCIS lesion on pre-operative imaging. Fu-rthermore, since DCIS is often non-palpable, it is often difficult for the surgeon to determine the precise excision location, and to keep a fixed macroscopic margin of healthy breast tissue. Delineated POC volumes showed very weak associations with both the mTD and the ES volumes. These findings are in line with findings by other investigators. [18]. Our subgroup analyses showed that the association between POC and ES volumes slightly further decreased when the surgeon performed full-thickness closure by approximating the excision cavity walls, which resulted in a reduction of the POC volume [26-28]. However, POC delineation has been shown not to be accurate on CT, and can therefore influence the observed associations with mTD and ES volume. The actual POC might be larger than delineated on CT, since CT cannot visualize all postoperative soft-tissue abnormalities [29]. Several studies have shown that deli-neation of the POC volume on a planning CT results in high interobserver variability despite guidelines [30-34]. This interobserver variability in target delineation can be clinically and dosimetrically significant [35]. Moreover, the surgical clips, which are inserted in the excision cavity intra-operatively in order to improve the POC visibi-lity, are often not inserted according to a standard protocol. In our study, we obser-ved that only 36% of the CT-scans showed a POC with most or all margins clearly de-fined (CVS 4 and 5, respectively), due to the presence of any seroma. Furthermore, it was shown that in the subgroup that underwent a second CT due to substantial seroma, a higher CVS score was assigned. Seroma causes pressure on the excision cavity walls, resulting in a better visibility of the POC margins. Since full-thickness

38

Chapter 2

closure, which can result in less seroma formation, becomes more widely practiced in order to improve cosmetic outcome, POC visibility on the radiotherapy-planning CT can further decrease [27, 28]. Our study has some limitations. Due to the retrospective data-analysis, we could not extract all the desired information, which could have led to a selective sub-group. Tumor volumes could not be calculated, but were substituted by the mTD. ES volume calculation was based on the assumption that excision specimens were ellipsoidally-shaped. This model was validated using the specimen weight for the known subjects. ES volumes and mTD in our study might be underestimated since the possible influence of fixating techniques on the mTD and specimen dimensions was omitted [36]. Information about the surgical excision technique, e.g. whether lumpectomy, wide local excision, sector resection or quadrantectomy was perfor-med, could not be extracted. Consequently, the determined association mTD and ES volume could not be corrected for surgical excision technique. In this population, CTV boost margins were not compensated for the pathological tumor-free resec-tion margin width. An asymmetric CTV boost that corrects for asymmetry of the surgical excision has the potential to reduce boost volume [19, 37]. Moreover, since many surgeons and radiation oncologists were involved in the treatment of these patients, interobserver variation may have played a substantial role. However, this is a representation of routine daily clinical practice.The results of this study show that disproportionate volumes are being excised and irradiated in relation to the mTD. The exact localization of both the tumor during bre-ast-conserving surgery, and the POC in radiotherapy-treatment is important. Local recurrences are mainly found in this area, and large excision and radiation volumes can both decrease cosmetic outcome [14, 17, 38-42]. Further optimization could be performed at different levels in breast-conserving therapy. First, lesion localization techniques could be optimized, as discussed previously. Secondly, intra-operative imaging, or image-guided minimally invasive treatment could offer interesting tre-atment alternatives in an accurately selected subgroup of patients. The latter might result in more proportional treatment volumes in small breast tumors. Radiofre-quency ablation, cryoablation, laser-induced thermal therapy, microwave ablation and high-intensity focused ultrasound have shown to be technically feasible in the breast [43]. However, further treat-and-resect studies are needed to compare the efficacy of these techniques with conventional breast-conserving surgery, before they can be implemented in clinical practice. Thirdly, new techniques for accurate POC delineation and subsequent optimal planning of radiotherapy should be ex-plored. Both for the current POC boost irradiation as for partial breast irradiation. The current CT-based delineation of the POC is not accurate enough, to precisely determine the POC volume, as previously discussed [30-35]. Several groups have studied MRI for postoperative imaging and radiotherapy planning [29, 33, 44-47].

39


MRI can have several advantages: 1) increased discrimination between fibroglan-dular tissue and fat in the affected breast, and 2) increased distinction between the POC and the surrounding breast tissue. The surgical cavity (with or without seroma) has a high signal intensity on T2-weighted images, while surrounding breast tissue has a low signal intensity on this sequence. Future studies are needed to investiga-te prospectively whether this may indeed lead to more accurate POC delineation. Another interesting approach is that more precise delineation of the POC could possibly be achieved by delineating and irradiating the breast tumor pre-opera-tively [48, 49]. If the breast cancer is in situ during irradiation, contrast-enhanced MRI could be used for accurate delineation of the invasive tumor within the breast. After intravenous administration of contrast, the breast cancer typically presents as an enhancing mass with high signal intensity on the T1-weighted images, whereas surrounding tissue has a low signal intensity due to fat suppression [50, 51]. In conclusion, in our study, both the excised and irradiated volumes did not show a clinically relevant association with the microscopic tumor diameter in women with early-stage breast cancer treated with breast-conserving therapy. This can be due to uncertainties in e.g. lesion localization, macroscopic margin assessment or target volume visibility in radiotherapy. The current breast-conserving therapy is suscepti-ble to imprecision. Although recurrence rates are low, accurate tumor excision and delineation of the postoperative complex is important, since most recurrences oc-cur within this area. Furthermore, cosmesis can decrease with increased excised and irradiated volumes. Future studies should focus on improving the different aspects of breast-conserving therapy: 1) Optimizing lesion localization for surgical excisi-on, 2) More proportionate excision volumes, i.e. image-guided minimally invasive treatments in small tumors and 3) More accurate (e.g. pre-operative), and further optimization of image-guided target volume delineation in radiotherapy.

40

Chapter 2

References

1. Fracheboud J, Otto SJ, van Dijck JA, Broeders MJ, Verbeek AL, de Koning HJ, National Evaluation

Team for Breast cancer screening (NETB). Decreased rates of advanced breast cancer due to mam-

mography screening in The Netherlands. Br J Cancer 2004: 91:861-867.

2. Nystrom L, Andersson I, Bjurstam N, Frisell J, Nordenskjold B, Rutqvist LE. Long-term effects of mam-

mography screening: updated overview of the Swedish randomised trials. Lancet 2002: 359:909-

919.

3. Clarke M, Collins R, Darby S et. al. Effects of radiotherapy and of differences in the extent of surgery

for early breast cancer on local recurrence and 15-year survival: an overview of the randomised

trials. Lancet 2005: 366:2087-2106.

4. van Dongen JA, Voogd AC, Fentiman IS et. al. Long-term results of a randomized trial comparing

breast-conserving therapy with mastectomy: European Organization for Research and Treatment

of Cancer 10801 trial. J Natl Cancer Inst 2000: 92:1143-1150.

5. Cuzick J, Sestak I, Pinder SE et. al. Effect of tamoxifen and radiotherapy in women with locally ex-

cised ductal carcinoma in situ: long-term results from the UK/ANZ DCIS trial. Lancet Oncol 2011:

12:21-29.

6. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Correa C, McGale P et. al. Overview of

the randomized trials of radiotherapy in ductal carcinoma in situ of the breast. J Natl Cancer Inst

Monogr 2010: 2010:162-177.

7. Dunne C, Burke JP, Morrow M, Kell MR. Effect of margin status on local recurrence after breast con-

servation and radiation therapy for ductal carcinoma in situ. J Clin Oncol 2009: 27:1615-1620.

8. Meijnen P, Gilhuijs KG, Rutgers EJ. The effect of margins on the clinical management of ductal carci-

noma in situ of the breast. J Surg Oncol 2008: 98:579-584.

9. Menes TS, Tartter PI, Bleiweiss I, Godbold JH, Estabrook A, Smith SR. The consequence of multiple

re-excisions to obtain clear lumpectomy margins in breast cancer patients. Ann Surg Oncol 2005:

12:881-885.

10. Holland R, Veling SH, Mravunac M, Hendriks JH. Histologic multifocality of Tis, T1-2 breast carcino-

mas. Implications for clinical trials of breast-conserving surgery. Cancer 1985: 56:979-990.

11. Borger JH. The impact of surgical and pathological findings on radiotherapy of early breast cancer.


12. Vaidya JS, Vyas JJ, Chinoy RF, Merchant N, Sharma OP, Mittra I. Multicentricity of breast cancer: who-

le-organ analysis and clinical implications. Br J Cancer 1996: 74:820-824.

13. Gump FE, Shikora S, Habif DV, Kister S, Logerfo P, Estabrook A. The extent and distribution of cancer

in breasts with palpable primary tumors. Ann Surg 1986: 204:384-390.




15. Romestaing P, Lehingue Y, Carrie C et. al. Role of a 10-Gy boost in the conservative treatment of early

breast cancer: results of a randomized clinical trial in Lyon, France. J Clin Oncol 1997: 15:963-968.

41


16. Mannino M, Yarnold J. Accelerated partial breast irradiation trials: diversity in rationale and design.


17. Christiaens MR, Cataliotti L, Fentiman I et. al. Comparison of the surgical procedures for breast con-

serving treatment of early breast cancer in seven EORTC centres. Eur J Cancer 1996: 32A:1866-1875.

18. Al Uwini S, Antonini N, Poortmans PM et. al. The influence of the use of CT-planning on the irradia-

ted boost volume in breast conserving treatment. Radiother Oncol 2009: 93:87-93.

19. Hanbeukers B, Borger J, van den Ende P et. al. Customized computed tomography-based boost

volumes in breast-conserving therapy: use of three-dimensional histologic information for clinical

target volume margins. Int J Radiat Oncol Biol Phys 2009: 75:757-763.

20. Bol GH, Kotte AN, van der Heide UA, Lagendijk JJ. Simultaneous multi-modality ROI delineation in

clinical practice. Comput Methods Programs Biomed 2009: 96:133-140.

21. Smitt MC, Birdwell RL, Goffinet DR. Breast electron boost planning: comparison of CT and US. Radi-

ology 2001: 219:203-206.

22. Union for International Cancer Control (UICC). TNM Classification of Malignant Tumors, 7th edition

2009.

23. Aspegren K, Holmberg L, Adami HO. Standardization of the surgical technique in breast-conserving

treatment of mammary cancer. Br J Surg 1988: 75:807-810.

24. Jakub JW, Gray RJ, Degnim AC, Boughey JC, Gardner M, Cox CE. Current status of radioactive seed

for localization of non palpable breast lesions. Am J Surg 2010: 199:522-528.

25. Rovera F, Frattini F, Marelli M et. al. Radio-guided occult lesion localization versus wire-guided loca-

lization in non-palpable breast lesions. Int J Surg 2008: 6 Suppl 1:S101-3.

26. Holmberg L, Zaren E, Adami HO, Bergstrom R, Burns T. The patient’s appraisal of the cosmetic result

of segmental mastectomy in benign and malignant breast disease. Ann Surg 1988: 207:189-194.

27. Paterson ML, Nathanson SD, Havstad S. Hematomas following excisional breast biopsies for invasi-

ve breast carcinoma: the influence of deep suture approximation of breast parenchyma. Am Surg

1994: 60:845-848.

28. Indelicato D, Grobmyer SR, Newlin H, Morris CG, Haigh LS, Copeland EM,3rd, Mendenhall NP. Asso-

ciation between operative closure type and acute infection, local recurrence, and disease surveil-

lance in patients undergoing breast conserving therapy for early-stage breast cancer. Surgery 2007:

141:645-653.

29. Whipp EC, Halliwell M. Magnetic resonance imaging appearances in the postoperative breast: the

clinical target volume-tumor and its relationship to the chest wall. Int J Radiat Oncol Biol Phys 2008:

72:49-57.






42

Chapter 2




33. Kirby AM, Evans PM, Nerurkar AY et. al. How does knowledge of three-dimensional excision margins

following breast conservation surgery impact upon clinical target volume definition for partial-bre-

ast radiotherapy?. Radiother Oncol 2010: 94:292-299.



2009: 90:60-65.

35. Li XA, Tai A, Arthur DW et. al. Variability of target and normal structure delineation for breast can-

cer radiotherapy: an RTOG Multi-Institutional and Multiobserver Study. Int J Radiat Oncol Biol Phys

2009: 73:944-951.

36. Graham RA, Homer MJ, Katz J, Rothschild J, Safaii H, Supran S. The pancake phenomenon contributes

to the inaccuracy of margin assessment in patients with breast cancer. Am J Surg 2002: 184:89-93.

37. Stroom J, Schlief A, Alderliesten T, Peterse H, Bartelink H, Gilhuijs K. Using histopathology breast

cancer data to reduce clinical target volume margins at radiotherapy. Int J Radiat Oncol Biol Phys

2009: 74:898-905.

38. Van Limbergen E, Rijnders A, van der Schueren E, Lerut T, Christiaens R. Cosmetic evaluation of bre-

ast conserving treatment for mammary cancer. 2. A quantitative analysis of the influence of radiati-

on dose, fractionation schedules and surgical treatment techniques on cosmetic results. Radiother

Oncol 1989: 16:253-267.



40. Curran D, van Dongen JP, Aaronson NK, Kiebert G, Fentiman IS, Mignolet F, Bartelink H. Quality

of life of early-stage breast cancer patients treated with radical mastectomy or breast-conserving

procedures: results of EORTC Trial 10801. The European Organization for Research and Treatment of

Cancer (EORTC), Breast Cancer Co-operative Group (BCCG). Eur J Cancer 1998: 34:307-314.







43. Zhao Z, Wu F. Minimally-invasive thermal ablation of early-stage breast cancer: a systemic review.

Eur J Surg Oncol 2010: 36:1149-1155.

44. Ahn KH, Hargreaves BA, Alley MT, Horst KC, Luxton G, Daniel BL, Hristov D. MRI guidance for acce-

lerated partial breast irradiation in prone position: imaging protocol design and evaluation. Int J

Radiat Oncol Biol Phys 2009: 75:285-293.

45. Sabine B, Giovanna D, Peter P, Clara J, Bert P, John K. Open low-field magnetic resonance (MR) versus

CT scanner (CT) imaging in breast radiotherapy treatment planning. Int J Radiat Oncol Biol Phys

2005: 63:S232-S233.

43


46. Lee G, Fyles A, Tran P et. al. Magnetic resonance imaging assessment for clinical target volume

delineation of the postoperative cavity in breast radiotherapy. Int J Radiat Oncol Biol Phys 2010:

78:S211.

47. Giezen M, Kouwenhoven E, Scholten AN et. al. MRI- versus CT-based volume delineation of lumpec-

tomy cavity in supine position in breast-conserving therapy: an exploratory study. Int J Radiat Oncol

Biol Phys 2012: 82:1332-1340.

48. Nichols EM, Dhople AA, Mohiuddin MM, Flannery TW, Yu CX, Regine WF. Comparative analysis of the

post-lumpectomy target volume versus the use of pre-lumpectomy tumor volume for early-stage

breast cancer: implications for the future. Int J Radiat Oncol Biol Phys 2010: 77:197-202.

49. Palta M, Yoo S, Adamson JD, Prosnitz LR, Horton JK. Preoperative Single Fraction Partial Breast Radi-

otherapy for Early-Stage Breast Cancer. Int J Radiat Oncol Biol Phys 2010.

50. Lee CH, Dershaw DD, Kopans D et. al. Breast cancer screening with imaging: recommendations

from the Society of Breast Imaging and the ACR on the use of mammography, breast MRI, breast

ultrasound, and other technologies for the detection of clinically occult breast cancer. J Am Coll

Radiol 2010: 7:18-27.

51. Sardanelli F, Boetes C, Borisch B et. al. Magnetic resonance imaging of the breast: recommendations

from the EUSOMA working group. Eur J Cancer 2010: 46:1296-1316.

Chapter 2

Chapter 3MR-guided breast radiotherapy:

feasibility and magnetic-field impact on skin dose

T.C.F. van Heijst, M.D. den Hartogh, J.J.W. Lagendijk, H.J.G.D. van den Bongard, B. van Asselen

Based on: Phys Med Biol 2013: 58:5917-5930

46

Chapter 3

Abstract

Purpose

The UMC Utrecht MRI/linac (MRL) design provides image guidance with high soft-tissue contrast, directly during radiotherapy (RT). Breast cancer patients are a potential group to benefit from better guidance in the MRL. However, due to the electron return effect, the skin dose can be increased in presence of a magnetic field. Since large skin areas are generally involved in breast RT, the purpose of this study is to investigate the effects on the skin dose, for whole-breast irradiation (WBI) and accelerated partial-breast irradiation (APBI).

Methods

In ten patients with early-stage breast cancer, targets and organs at risk (OARs) were delineated on postoperative CT scans co-registered with MRI. The OARs included the skin, comprising the first 5 mm of ipsilateral-breast tissue, plus extensions. Three intensity-modulated RT (IMRT) techniques were considered (2x WBI, 1x APBI). Indi-vidual beam geometries were used for all patients. Specially developed MRL treat-ment-planning software was used.

Results and Conclusion

Acceptable plans were generated for 0 T, 0.35 T, and 1.5 T, using a class solution. The skin dose was augmented in WBI in presence of a magnetic field, which is a potential drawback, whereas in APBI the induced effects were negligible. This opens possibilities for developing MR-guided partial-breast treatments in the MRL.

47

Magnetic-field impact on breast skin dose

Introduction

Magnetic resonance imaging (MRI) guidance for radiotherapy (RT) has the potential of fast, high soft-tissue contrast visualization of tumors and organs at risk (OARs), directly on the treatment table. At the UMC Utrecht a hybrid MRI/linac (MRL) is constructed which integrates a 1.5 T closed-bore MRI scanner (Philips, Best, The Netherlands) with a linear accelerator (Elekta AB, Stockholm, Sweden), mounted on a ring-shaped gantry [1]. Several other groups are currently working on MR-guided RT in cancer treatment as well [2, 3]. MRI combined with RT offers possibilities for introducing high-precision MR-guided treatments for breast cancer patients.The standard of care for early-stage breast cancer and ductal carcinoma in situ (DCIS) patients is breast-conserving therapy (BCT) which consists of breast-conser-ving surgery (BCS), followed by whole-breast irradiation (WBI). The dose is generally delivered by intensity-modulated RT (IMRT) using two tangential beams. Further-more, in low-risk breast cancer patients, accelerated partial-breast irradiation (APBI) studies using IMRT are ongoing [4-8].In both WBI and APBI, image guidance is currently performed using portal imaging or cone-beam CT [9, 10]. Direct visualization of the target volumes is especially im-portant when considering boosts to the target volume, which both modalities do not allow with high tissue-contrast. MR guidance is potentially useful for on-line high-contrast visualization of the tumor bed in postoperative RT, or for tumor de-tection in a preoperative setting [11-14]. However, apart from developing novel tar-geting techniques, it is of great importance to investigate the induced effects of the magnetic field itself on the dose distribution. In the MRL a static magnetic field is always present during treatment. Charged parti-cles moving in a magnetic field are acted upon by the Lorentz force, perpendicularly to their velocity direction. As a consequence, secondary electrons emanating from the skin into air can be bent back, resulting in a dose increase at the surface (Figure 1). This electron return effect (ERE) is clearly observed at boundaries between layers with large density differences, and can induce a significant increase of skin dose, as shown by Raaijmakers et al [15]. The magnetic field also results in a shorter build-up distance which may play a role in a higher skin dose.The ERE depends on the inclinations of the beams to the skin surface, with oblique angles inducing the highest increase and opposing beams compensating for the effect at perpendicular interfaces [15, 16, 16]. In WBI treatments, oblique beam/sur-face inclinations are ubiquitous and the target volume is superficial and relatively large. This results in a large irradiated area of skin. Treatments by APBI, however, are performed with several fields which do not necessarily all have oblique orientati-ons. Additionally, APBI fields are generally smaller since they target smaller volumes and, moreover, the target volumes are not necessarily superficial. Hence, a smaller

48

Chapter 3

region of skin generally receives high dose, so that induced effects of the magnetic field are expected to be less prominent in APBI relative to WBI.Current cosmetic results for WBI and APBI are good to excellent [7, 8, 17, 18]. Further increase in skin dose could lead to a higher rate and severity of negative side effects of RT. Complications of the skin should be minimized in the treatment of breast can-cer; therefore the main objective of this study is to investigate the physical effects of the magnetic field on the skin dose for WBI and for APBI.

Methods and materials

Patients

Ten BCT patients treated at the Radiotherapy Department of the UMC Utrecht in 2011 or 2012 were enrolled in our treatment-planning study. The median age of the women was 63 years (range: 39 – 72 years). One patient had breast cancer on both sides and for this study both tumors were considered separately. In total, 6 of the tumors were right-sided and 5 were located in the left breast. The majority of the excised tumors were stage T1c (7), while the rest were stage T1b (2) or stage T2 (2). No patients had tumor-positive lymph nodes. The median volume of the ipsilateral breast was 1003 cc (range: 705 – 2373 cc).

Figure 1 Illustration of the ERE, for left-breast WBI by means of two tangential fields. The edges of the photon beams are depicted by the blue lines. Trajectories of secondary electrons, crossing the skin-air boundary on either exit side of the irradiated breast, are represented by the arrows. The ERE may result in a higher skin dose when comparing the situation of no magnetic field (left) to that of a non-zero magnetic field directed into the plane (right).

49


Radiotherapy techniques

The magnetic-field-induced effects were studied for both WBI and APBI. In WBI a dose of 42.56 Gy was prescribed to the breast delivered in 16 fractions of 2.66 Gy. The first WBI technique considered was the classical tangential two-field set-up using IMRT (WBI-2). The beam angles in WBI-2 were determined such that the two beam edges inside each patient aligned, which was done for each patient indivi-dually. In order to investigate the influence of the magnetic field on multiple beam directions, also a seven-field IMRT technique was used for WBI (WBI-7) in which the beam orientations were chosen individually for each patient. For APBI, target pres-cription was 10 fractions of 3.85 Gy, to a total dose of 38.5 Gy, according to the 2008 IRMA study protocol of the European Organization for Research and Treatment of Cancer Radiation Oncology Group (EORTC ROG). A seven-field IMRT technique with individually determined beam angels was chosen to for the APBI approach (APBI-7).

Volumes of interest

All patients underwent a planning CT with 3 mm slice thickness, at three weeks (median: 21 days, range: 14 – 50 days) after surgery, in supine RT position. The pal-pable breast tissue was indicated with a copper wire. For WBI, the clinical target volume for the whole breast (CTVWBI) was delineated according to the Radiotherapy Oncology Group (RTOG) Breast Cancer Contouring Atlas. A 5 mm margin was ap-plied to the CTVWBI to generate the planning target volume (PTVWBI), excluding the first 5 mm under the surface (Figure 2). The median PTVWBI was 741 cc (range: 517 – 2028 cc). Delineations of the target volumes in APBI were performed according to the 2008 IRMA study protocol of EORTC ROG. The gross tumor volume (GTVAPBI) was contoured using preoperative diagnostic imaging, preoperative MRI in treatment position, surgical clips, and possible postoperative seroma. A 15 mm margin in all directions was applied to the GTVAPBI to generate the clinical target volume (CTVAPBI), while excluding the chest wall and skin. The PTVAPBI was delineated as the CTVAPBI with a 5 mm margin, while excluding the skin (Figure 2). The median PTVAPBI was 148 cc (range: 88 – 248 cc). For both WBI and APBI, the delineated OARs were the heart, lungs, contralateral breast, body (comprising all unspecified tissue) and the skin (Fi-gure 2). The skin considered for the analysis of the impact of the ERE was defined to be the first 5 mm under the surface of the ipsilateral breast. It was further extended up to 35 mm in anterior, posterior, medial, and lateral directions relative to the PT-VWBI in order to include skin tissue expected to be irradiated in WBI-2 (Figure 2). An extension of 15 mm was made in the caudal and cranial directions to include skin tissue receiving scattered dose. The median volume of the skin was 318 cc (range: 256 – 474 cc).

50

Chapter 3

Figure 2 Transversal slice of a patient CT scan with VOIs delineated: PTVWBI (red contour), GTVAPBI (purple), PTVAPBI (bright green), ipsilateral lung (yellow), contralateral lung (dark green), heart (sky blue), unspecified tissue (dark blue), contralateral breast (brown), and skin (orange, filled here with light grey). An extension made to the skin is visible here in the medial and lateral directions relative to the PTVWBI, as indicated by the white oval-shaped markings. The location of the isocentre is mar-

IMRT planning

Specially developed MRL treatment planning software (MRLTP) was used to gene-rate IMRT plans while taking into account the presence of a magnetic field. MRLTP is a combination of GPUMCD – which is a Graphics Processing Unit (GPU)-oriented Monte Carlo dose calculation algorithm as described by Hissoiny et al and the Cen-tral Processing Unit (CPU)-based Fast Inverse Dose Optimization (FIDO) as described by Goldman et al [19]. A more detailed description of this system can be found in Bol et al [20, 21]. FIDO has the distinct advantage that the optimization process is executed within a short time frame. The MRL’s characteristics are incorporated in the MRLTP. The 6 MV photon beam can rotate in full 360 degrees around the patient with maximum field sizes of 24 cm in caudocranial direction and 56 cm in anteri-or-posterior direction. As a consequence, the isocentre is fixed, at 14 cm above the treatment table. The isocentre-source distance is 142.7 cm. With MRLTP, plans can be generated at any magnetic-field strength required. IMRT plans were generated for 0 T and 1.5 T, which is the field strength in the MRL. Also, plans were calculated at the intermediate value of 0.35 T. For each IMRT technique (WBI-2, WBI-7, APBI-7) a class solution of optimization objectives was developed at 0 T, which was then applied for all patients. These sets of objectives were unaltered when applied at

51


higher field strengths. The plans were based on the fully optimized fluence maps with a 5 mm grid resolution and the acquired dose distributions consisted of 2x2x2 mm3 voxels.

Comparison of plans

For the comparison of the plans, dose volume histograms (DVHs) for all VOIs and cor-responding dose parameters were calculated. For the PTV the D95% - the percentage of the PTV receiving at least 95% of the prescribed dose – to describe PTV coverage, and the D107%, describing the amount of overdose, were calculated. The starting point of the plan comparisons was a similar D95% and D107%. Several parameters were compared for the OARs. The mean lung dose (MLD) over two lungs was computed, which was to be a maximum of 7 Gy. For the ipsilateral lung, V10Gy – the fraction of the volume receiving 10 Gy or more –, V20Gy and V30Gy were derived for the WBI plans while for APBI the V5Gy, V10Gy and V20Gy were found. Similar parameters were calculated for the heart. Moreover, the D2cc – maximum dose en-compassing 2 cc of the structure – was calculated for the heart as an indication of the maximum dose. The range 30-40 Gy is considered to be relevant for the skin dose, so the mean dose and parameters V35Gy and D2cc were calculated for the skin. For the V35Gy, box-whisker plots are derived to determine how the values at different magnetic-field strengths are distributed. Also, the D2cc of the unspecified tissue (body) and the mean dose of the contralateral breast were acquired, both of which were to be kept low. The statistical significance of differences between all dosimetric parameters was tested by performing a paired student’s t-test. A value of p < 0.05 was considered to represent a statistically significant result. Additionally, dose-difference maps for all plans were derived, which depict for the same patient the dose differences per voxel relative to the situation at 0 T, i.e. dose va-lues at 0.35 T versus 0 T and at 1.5 T versus 0 T. This allows further analysis of the spatial distribution of any dose differences caused by the magnetic field.

Results

Optimization process

IMRT plans were successfully generated using MRLTP. The time required to calculate the Monte Carlo beamlets varied with size of the PTV, the number of beam angles and the magnetic-field strength (i.e. number of electron steps). For this, a computer with 4 GB RAM, a 2.27 GHz CPU, and a GeForce GX 580 GPU was used. Beamlet cal-culations for a WBI-2 plan took 15 minutes at 0 T which went up to 25 minutes at 1.5 T, while a typical WBI-7 set-up took 40 minutes for 0 T and up to an hour for a 1.5 T plan. Calculating beamlets for APBI-7 typically took 20 minutes for a 0 T plan, up to 25 minutes for a plan at 1.5 T.

52

Chapter 3

For each of the three RT techniques, a class solution of objectives was employed for the optimization of all plans. Once a suitable template of objective parameters was found for one patient, an automated script was activated to apply it for optimizing the plans of all other patients. Some individual plans required additional optimizati-on. With the use of a 32 GB RAM, 3.40 GHz, eight-core CPU, the optimization time of FIDO for a WBI-2 plan or an APBI-7 plan was typically in the order of a few seconds, while a WBI-7 plan took only half a minute to optimize.

Induced effects for WBI-2

At the three different magnetic-field strengths 0 T, 0.35 T and 1.5 T, similar target coverage is achieved. Furthermore, no target overdose is observed (Table 1).

Table 1 Dose parameters calculated for WBI-2 (columns 2-4) and WBI-7 (columns 5-7), at 0 T, 0.35 T, and 1.5 T, respectively, for all breast tumors (n = 11). Prescription was 16 x 2.66 Gy = 42.56 Gy. Values are denoted as: ‘mean (standard deviation)’.

WBI-2 WBI-70 T 0.35 T 1.5 T 0 T 0.35 T 1.5 T

PTVD95% (%) 96.5 (0.5) 96.7 (0.7) 96.7 (0.6) 95.9 (0.6) 96.1 (0.7) 96.1 (0.7)D107% (%) 0.0 (0.0) 0.0 (0.1) 0.0 (0.1) 0.1 (0.3) 0.1 (0.3) 0.1 (0.2)

Ips. lungMLD (Gy) 2.9 (0.9) 3.0 (0.9) 3.0 (0.9) 5.2 (1.1) 5.8 (1.1) 5.9 (1.1)V10Gy (%) 16.7 (4.9) 16.2 (4.8) 16.2 (4.7) 36.0 (7.7) 35.5 (7.6) 36.6 (7.8)V20Gy (%) 13.6 (4.5) 13.5 (4.5) 13.6 (4.4) 17.3 (7.3) 17.2 (7.0) 17.4 (6.9)V30Gy (%) 7.1 (3.7) 8.1 (3.8) 9.3 (3.8) 3.5 (3.1) 3.7 (3.2) 3.9 (3.1)

HeartD2cc (Gy) 20.5 (18.6) 20.2 (18.5) 19.9 (18.5) 28.3 (9.2) 28.3 (9.1) 28.2 (9.0)V25Gy (%) 1.5 (2.0) 1.4 (2.0) 1.4 (2.0) 3.4 (4.3) 3.1 (3.9) 3.2 (4.1)

Unspec. tissueD2cc (Gy) 46.0 (2.7) 46.7 (2.7) 47.2 (2.8) 45.0 (2.3) 45.0 (2.4) 45.2 (2.4)

Contr. breast Mean (Gy) 0.2 (0.3) 0.2 (0.3) 0.2 (0.3) 3.3 (1.7) 3.1 (1.6) 3.2 (1.7)

SkinD2cc (Gy) 43.3 (0.6) 44.6 (1.0) 45.6 (1.1) 43.3 (0.3) 44.4 (1.4) 45.6 (2.9)Mean (Gy) 29.5 (1.4) 32.3 (1.6) 33.2 (1.7) 27.9 (1.4) 30.2 (1.7) 29.8 (1.3)

D95% fraction of the volume receiving at least 95% of the prescribed dose, or 40.43 Gy; D107% fracti-on of the volume receiving at least 107% of the prescribed dose, or 45.54 Gy; Ips. lung ipsilateral lung; MLD mean lung dose (over both lungs); VXGy fraction of volume receiving at least X Gy; D2cc dose received by 2 cc of the structure; Unspec tissue Unspecified tissue; Contr. Breast contralateral breast.

53


Relative to the situation at 0 T, the mean skin dose in WBI-2 is raised by 9.5% (p < 0.01) and 12.5% (p < 0.01) at 0.35 T and 1.5 T, respectively (Table 1). This dose incre-ase is clearly visible in a typical DVH and dose-difference maps as shown in Figure 3(a) and Figure 4 (a,b), respectively. It can also be observed in Figure 4 (a,b) that the increase in skin dose is especially manifested at the most superficial voxels of the skin. The most significant change in dose distribution in the skin is found in the range of around 30-40 Gy (Figure 3(a)). Further analysis shows that the ERE causes the skin volume receiving more than 35 Gy (V35Gy) at both non-zero magnetic-field strengths to increase considerably. This can be observed in the box-whisker plot in Figure 5(a), where the whiskers of the boxes of V35Gy at both 0.35 T and 1.5 T do not overlap with those at 0 T, indicating significant differences (p < 0.01). Also, the maxi-mum skin dose (D2cc) is slightly raised, with an increase of 2.3 Gy at 1.5 T compared to no magnetic field (Table 1).

Figure 3 DVHs at different field strengths, 0 T (full), 0.35 T (dash-dotted), and 1.5 T (dashed), acqui-red from the plans for one typical patient. For each separate RT technique, (a) WBI-2; (b) WBI-7; (c) APBI-7, the PTV and OARs, including the skin, are shown. See legend for color specifications. Here, “ips.”= ipsilateral, “contr.” = contralateral, “unspec. tissue” = all unspecified tissue. For WBI, the skin at different magnetic-field strengths is also indicated by arrows with corresponding text boxes. Pres-cribed dose for WBI and APBI was 42.56 Gy and 38.5 Gy, respectively.

54

Chapter 3

The mean lung dose (MLD) in WBI-2 is low (~3.0 Gy) and unaffected by the magnetic field (Table 1). This is also true for the V10Gy of the ipsilateral lung. However, it receives more high dose at increasing field strength as indicated by the V30Gy going up by 2.2% (p < 0.01) at 1.5 T relative to 0 T. This high-dose volume can also be observed in the dose-difference maps of Figure 4 (a,b). The heart dose is unaffected by the presence of the magnetic field in WBI-2. Both the V25Gy and D2cc of the heart remain constant. This is also true for the contralateral-breast mean dose, which remains very low. The unspecified tissue receives a higher dose locally as indicated by the D2cc which is on average 1.2 Gy more (p = 0.02) at 1.5 T than at 0 T (Table 1).

Figure 4 Maps of dose differences (in Gy) per voxel relative to the situation of no magnetic field. Examples of transversal slices are depicted in each consecutive row for each RT technique on three different patients, while the different magnetic-field strengths are arranged per column, i.e. (a) WBI-2 at 0.35 T vs. 0 T; (b) WBI-2 at 1.5 T vs. 0 T; (c) WBI-7 at 0.35 T vs. 0 T; (d) WBI-7 at 1.5 T vs. 0 T; (e) APBI-7 at 0.35 T vs. 0 T; (f ) APBI-7 at 1.5 T vs. 0 T. Voxel size is 2x2x2 mm3. Differences range from -5 Gy (dark blue) to +10 Gy (dark red).

55


Induced effects for WBI-7

Similar target coverage is achieved for all patients in WBI-7 at all field strengths and, again, no target overdoses are observed (Table 1).Although the mean skin dose in WBI-7 is lower than in WBI-2, it is raised by the ERE. This augmentation, relative to the 0 T situation, is on average 8.2% (p < 0.01) and 6.8% (p < 0.01) at 0.35 T and 1.5 T, respectively (Table 1). The fact that this increase is less severe than in WBI-2 can be seen in the typical DVH example in Figure 3(b) and the dose-difference maps of Figure 4 (c,d). The principle change in skin dose is in the range of around 30-40 Gy, which was also observed for WBI-2. Indeed, V35Gy is increased considerably (p < 0.01) at 0.35 T and at 1.5 T. This can also be seen in the box-whisker plot in Figure 5(b), where the interquartile ranges (height of the boxes) of V35Gy, at both 0.35 T and at 1.5 T, do not overlap with those at 0 T, indicating that there is a significant difference between the medians. It nevertheless shows a smaller increase compared to the tangential technique. The D2cc of the skin is similar to that in WBI-2 and is in presence of a magnetic field raised by a similar amount (Table 1).The MLD in WBI-7 is higher than in WBI-2 and is increased with respectively 0.6 Gy and 0.7 Gy at 0.35 T and 1.5 T (Table 1). The volumes of the ipsilateral lung receiving at least 10 Gy and 20 Gy, respectively, are larger than in WBI-2 and remain almost constant. However, the V30Gy is smaller relative to WBI-2, and is only enlarged from on average 3.5% (p = 0.17) at 0 T to 3.9% (p = 0.11) at 1.5 T. Due to the beam orien-tation in WBI-7, the V25Gy and D2cc for the heart and mean dose of the contralateral breast are higher than for WBI-2, but they remain constant with increasing mag-netic-field strengths. Also, the maximum dose indicated by D2cc of the unspecified tissue is increased with on average only 0.2 Gy more (p = 0.02) at 1.5 T than at 0 T (Table 1). This effect is smaller relative to WBI-2 (Figure 4 (c,d)).

Induced effects for APBI-7

For the APBI-7 plans, similar dose coverage is achieved for all magnetic-field values. Furthermore, no overdose is observed for the three situations (Table 2). The impact of the ERE on the skin dose is less prominent in APBI than in WBI. The mean skin dose in APBI-7 is low when compared to the WBI techniques and is only slightly raised when a magnetic field is present (Table 2). This can also be observed in the typical DVH in Figure 3(c) and is also reflected in the dose-difference maps of Figure 4 (e,f ). Since the skin dose is already observed to be low, the V35Gy for the skin in APBI-7 is low and does not increase considerably (Figure 5(c)). This can be clearly seen by the boxes of V35Gy at 0.35 T and at 1.5 T both largely overlapping with the box at 0 T (p = 0.13 and p < 0.01, respectively). Although the increase at 1.5 T is sig-nificant, it does not reflect a clinically relevant difference since the absolute change is very small. The maximum skin dose (D2cc) is lower when compared to WBI and remains unchanged at non-zero magnetic-field strengths (Table 2).

56

Chapter 3

Table 2 Dose parameters calculated for APBI-7 at 0 T, 0.35 T, and 1.5 T, respectively, for all breast tumors (n = 11). Prescription was 10 x 3.85 Gy = 38.5 Gy. Values are denoted as: ‘mean (standard deviation)’.

APBI-70 T 0.35 T 1.5 T

PTV D95% (%) 97.0 (0.5) 97.0 (0.5) 97.0 (0.7)D107% (%) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)

Ipsilateral lungMLD (Gy) 2.1 (1.0) 2.0 (1.0) 1.8 (0.8)V5Gy (%) 25.1 (12.9) 23.1 (12.6) 20.3 (9.6)V10Gy (%) 10.3 (6.1) 9.8 (7.1) 8.0 (4.5)V20Gy (%) 2.3 (1.5) 2.3 (1.6) 1.9 (1.3)

HeartD2cc (Gy) 6.9 (2.5) 6.2 (2.0) 5.8 (2.5)V5Gy (%) 8.0 (13.3) 6.2 (10.2) 6.0 (10.9)V10Gy (%) 0.4 (1.2) 0.0 (0.0) 0.0 (0.0)

Unspecified tissueD2cc (Gy) 37.9 (0.5) 38.2 (0.6) 38.4 (0.7)

Contralateral breast Mean (Gy) 1.1 (0.5) 1.0 (0.4) 1.4 (0.6)

SkinD2cc (Gy) 35.5 (4.7) 35.2 (4.9) 35.6 (4.4)Mean (Gy) 5.2 (2.1) 5.6 (2.4) 5.8 (2.4)

D95% fraction of the volume receiving at least 95% of the prescribed dose, or 36.6 Gy; D107% fraction of the volume receiving at least 107% of the prescribed dose, or 41.2 Gy; MLD mean lung dose (over both lungs); VXGy fraction of volume receiving at least X Gy; D2cc dose received by 2 cc of the structure

The MLD in APBI-7 is low (~2 Gy) and remains almost constant for non-zero field strengths. Furthermore, there are no observed significant differences (p > 0.05) in the presence of a magnetic field in the V5Gy and V10Gy of the ipsilateral lung, while also the V20Gy remains unchanged. The V5Gy, V10Gy, and D2cc of the heart in APBI-7 remain constant relative to 0 T at both 0.35 T and 1.5 T (p > 0.05 for all parameter differen-ces). The heart V25Gy stays zero in APBI-7 (not given in Table 2). Relative to WBI, the mean dose of the contralateral breast remains low, whereas the unspecified tissue D2cc seems to be raised slightly. The maximum raise is 0.5 Gy at 1.5 T compared to 0 T, however, this is statistically insignificant (p = 0.17). Small increases are visible in the dose-difference maps in Figure 4 (e,f ).

57


Discussion

The results of this treatment-planning study show effects on the dose distribution when treating breast-cancer patients with a 6 MV photon beam in presence of a magnetic field. The induced dosimetric effects on the skin are observed at 0.35 T and 1.5 T. A significant increase in skin dose is observed when conventional treat-ment by tangential-field WBI is performed. This observation could impair the cli-nical acceptability of the plans in the MRL. Therefore, a seven-field WBI technique was introduced, in an attempt to reduce the skin dose, as predicted in a previous film-based phantom study by Almberg et al [22](2011), although the other OARs would receive more dose than with the tangential technique. However, an increase in skin dose is observed there as well. For the APBI technique, the impact on the skin dose is small in comparison with WBI. Doses to the OARs in WBI are comparable to those from other studies in the 0 T cases, except for the V20Gy of the ipsilateral lung in WBI-7, which is relatively high [23-26]. Differences in the dose distribution between the WBI-2 and WBI-7 techni-

Figure 5 Box-whisker plots of the skin dose parameter V35Gy, as obtained from all plans (11 breast tumors). The box whiskers are re-presented for (a) WBI-2; (b) WBI-7; (c) APBI-7. They are evaluated for field strengths 0 T (red), 0.35 T (orange), and 1.5 T (blue). The line in each box represents the median; the lower and upper quartiles determine the height of each box – or interquartile ran-ge; the minimum and maximum values are represented by the whiskers. The p-values acquired from the student’s paired t-test are depicted to indicate significances rela-tive to 0 T.

58

Chapter 3

ques are mainly in the low-dose areas of the OARs as a consequence of the different beam orientations, and in the high-dose region of the ipsilateral lung. Although the MLD is higher in seven-field WBI, the high-dose volume is smaller compared to the tangential set-up, since in the tangential plans the dose is less conformal near the lung. In the seven-field set-up the mean skin dose is lower relative to the tangenti-al-field technique. However, a magnetic-field-induced effect was still observed, thus the plans were clinically not better than for the tangential technique.In APBI, the observed effects on the skin dose are found to be negligible in presen-ce of a magnetic field. Moreover, the absolute dose values on the other OARs are low relative to WBI and comparable to values found in other studies at 0 T, while it should be kept in mind that the fractionation in APBI is different from WBI [26, 27]. Since the OAR doses are not affected adversely under influence of the magnetic field, it can be concluded that the performance of APBI treatments in the MRL is not impaired by the ERE. Current cosmetic outcome for treatments by WBI after BCS is good to excellent [17, 18, 28, 29]. For APBI, at least in the studies by Lewin et al and Shaikh et al, the current cosmetic outcome was shown to be good to excellent [7, 8]. With regard to the skin dose, the presence of the magnetic field is clearly disadvantageous for WBI. Further increases to the skin dose could result in a higher complication rate and severity of complications [30]. The actual clinical implications of the observed raised skin dose – especially in the regions receiving 30-40 Gy or more – have to be investiga-ted further, since precise data and calculations on the biological effects of high skin dose are unavailable. However, negative side effects induced by high skin dose are observed, implying a potential drawback when considering the performance of WBI treatments in the MRL [31, 32]. A fluence-based optimization process is used in MRLTP to optimize the dose distri-bution, which means that the sequencing step for actual dose delivery is not inclu-ded. Thus, any quality degradation of the plan caused by a sequencer is omitted. The aim of our study, however, was to investigate the induced physical influences on the dose caused by the magnetic field. Although the absolute dose values may differ slightly when including a sequencer, the physical effects on the skin dose in general are already reflected in the dose plans. The highest dose increases in the skin were observed in the most superficial layer of the skin (Figure 4). In our analysis a voxel size of 2x2x2 mm3 was used. Further investigation of the magnetic-field-induced effects on the dose distribution with a higher-resolution dose grid was performed by Oborn et al [33]. In specific phantom set-ups, layers of tiny voxels (in the order of 10 mm per voxel) were constructed and analyzed for a 6 MV photon spectrum and a range of magnetic-field strengths (0 T – 3.0 T) using Monte Carlo simulations. It was shown that at non-zero magnetic fields the first 70 mm under the surface receive a larger increase in skin dose than

59


what was to be expected based on lower-resolution calculations. As a consequence, there is a possibility that even in APBI treatments, a dose peak may appear in the top layers of the skin. This should be properly measured and monitored extensive-ly before the clinical implementation of APBI treatments in the MRL. Moreover, it was shown in Monte Carlo dose calculations by Oborn et al that the contribution of contaminant electrons to both the entry dose and entry-surface dose can cause an increase at 0 T [34]. However, this effect is again limited to about the first 70 mm of the tissue. Contaminant electrons are not incorporated in the MRLTP. The current study was based on static-image treatment-planning, which means that effects due to intrafractional movements are not accounted for. However, Fra-zier et al showed that dose plans for WBI are relatively insensitive to the effects of breast motion during normal breathing [35]. Furthermore, in a study by George et al it was concluded that were no significant effects observed in the IMRT delivery when respiratory motion is compensated for in breast RT [36]. Additionally, the MRL is primarily designed for tracking the target volume and performing dynamic plan-ning, thus compensating for all motion.The added value of MRI for breast RT in the MRL will especially be the visualization of the tumor bed in high contrast to the surrounding anatomy. For WBI alone, the breast boundaries can also be visualized with other modalities such as kV or MV ra-diographs. However, the purpose of the research was to investigate the dosimetric effects of the magnetic field on RT treatments, including the standard technique of WBI, which involves relatively large field sizes. This can be of special relevance when WBI is to be performed in the MRL in combination with, for example, the delivery of a (stereotactic) boost. In future studies, more novel treatments in the MRL will be investigated as well. A single modality which can perform both standard and more sophisticated RT techniques at once could lead to a very high efficiency in the treatment of the patient. The use of MRI allows for very precise RT performances in the MRL, from which bre-ast-cancer patients could potentially benefit. However, from the previous discussion we find that the direct application in the MRL of currently standard treatments such as WBI is not straightforward. In both tangential and seven-field WBI, the skin-dose increase caused by the ERE is probable to cause negative side effects, which implies a potential drawback for those treatments. On the other hand, in the APBI set-up considered, negative side effects are much more unlikely, which makes APBI in the MRL feasible. APBI treatments could fully benefit from the capabilities of MRI, such as the high soft-tissue contrast, enabling direct, on-line position verification of the target area itself. This opens new possibilities for developing novel treatment tech-niques for breast cancer, aiming directly at the primary tumor location [37]. Therefo-re, our future research will focus on MR-guided partial-breast treatments.

60

Chapter 3

In conclusion, the treatment of patients in the MRL using a conventional tangential WBI set-up or a seven-field WBI technique, at 0.35 T or 1.5 T, implies an increase in skin dose of the ipsilateral breast. In APBI treatments the skin dose and other OAR doses are relatively low and only minimally affected by the magnetic field. This opens new possibilities for developing MR-guided partial-breast treatment techni-ques in the MRL.

61


References



2. Fallone BG, Murray B, Rathee S et. al. First MR images obtained during megavoltage photon irradia-

tion from a prototype integrated linac-MR system. Med Phys 2009: 36:2084-2088.

3. Dempsey J, Dionne B, Fitzsimmons J et. al. A real-time MRI guided external beam radiotherapy

delivery system . Med Phys 2006:2254.

4. Oliver M, Chen J, Wong E, Van Dyk J, Perera F. A treatment planning study comparing whole breast

radiation therapy against conformal, IMRT and tomotherapy for accelerated partial breast irradiati-

on. Radiother Oncol 2007: 82:317-323.

5. Njeh CF, Saunders MW, Langton CM. Accelerated Partial Breast Irradiation (APBI): A review of availa-

ble techniques. Radiat Oncol 2010: 5:90.

6. Livi L, Buonamici FB, Simontacchi G et. al. Accelerated partial breast irradiation with IMRT: new

technical approach and interim analysis of acute toxicity in a phase III randomized clinical trial. Int J

Radiat Oncol Biol Phys 2010: 77:509-515.

7. Lewin AA, Derhagopian R, Saigal K et. al. Accelerated partial breast irradiation is safe and effective

using intensity-modulated radiation therapy in selected early-stage breast cancer. Int J Radiat On-

col Biol Phys 2012: 82:2104-2110.

8. Shaikh AY., LaCombe MA., Du H.’, Raghaven VT., Nanda RK., Bloomer WD. Accelerated partial breast

irradiation using once-daily fractionation: analysis of 312 cases with four years median follow-up

2012: 7:17.

9. Fein DA, McGee KP, Schultheiss TE, Fowble BL, Hanks GE. Intra- and interfractional reproducibility

of tangential breast fields: a prospective on-line portal imaging study. Int J Radiat Oncol Biol Phys

1996: 34:733-740.

10. Jaffray DA, Siewerdsen JH, Wong JW, Martinez AA. Flat-panel cone-beam computed tomography

for image-guided radiation therapy. Int J Radiat Oncol Biol Phys 2002: 53:1337-1349.

11. Sabine B, Giovanna D, Peter P, Clara J, Bert P, John K. Open low-field magnetic resonance (MR) versus

CT scanner (CT) imaging in breast radiotherapy treatment planning. Int J Radiat Oncol Biol Phys

2005: 63:S232-S233.



72:49-57.

13. Kirby AM, Yarnold JR, Evans PM, Morgan VA, Schmidt MA, Scurr ED, desouza NM. Tumor bed deline-

ation for partial breast and breast boost radiotherapy planned in the prone position: what does MRI

add to X-ray CT localization of titanium clips placed in the excision cavity wall?. Int J Radiat Oncol

Biol Phys 2009: 74:1276-1282.



Biol Phys 2012: 82:1332-1340.

62

Chapter 3



Phys Med Biol 2005: 50:1363-1676.

16. Raaijmakers AJ, Raaymakers BW, van der Meer S, Lagendijk JJ. Integrating a MRI scanner with a 6 MV

radiotherapy accelerator: impact of the surface orientation on the entrance and exit dose due to

the transverse magnetic field. Phys Med Biol 2007: 52:929-939.

17. Taylor ME, Perez CA, Halverson KJ et. al. Factors influencing cosmetic results after conservation the-

rapy for breast cancer. Int J Radiat Oncol Biol Phys 1995: 31:753-764.

18. Munshi A, Kakkar S, Bhutani R, Jalali R, Budrukkar A, Dinshaw KA. Factors influencing cosmetic out-

come in breast conservation. Clin Oncol (R Coll Radiol) 2009: 21:285-293.

19. Goldman SP, Thurnbull D, Johnson C, Chen JZ, Battista JJ. Real-time fast inverse dose optimizati-

on for image guided adaptive radiation therapy enhancements to fast inverse dose optimization

(FIDO). J Appl Phys 2009: 105:102008.

20. Bol GH, Hissoiny S, Lagendijk JJ, Raaymakers BW. Fast online Monte Carlo-based IMRT planning for

the MRI linear accelerator. Phys Med Biol 2012: 57:1375-1385.

21. Hissoiny S, Ozell B, Bouchard H, Despres P. GPUMCD: A new GPU-oriented Monte Carlo dose calcu-

lation platform. Med Phys 2011: 38:754-764.

22. Almberg SS, Lindmo T, Frengen J. Superficial doses in breast cancer radiotherapy using conventio-

nal and IMRT techniques: a film-based phantom study. Radiother Oncol 2011: 100:259-264.

23. Hong L, Hunt M, Chui C et. al. Intensity-modulated tangential beam irradiation of the intact breast.

Int J Radiat Oncol Biol Phys 1999: 44:1155-1164.

24. Popescu CC, Olivotto I, Patenaude V, Wai E, Beckham WA. Inverse-planned, dynamic, multi-beam,

intensity-modulated radiation therapy (IMRT): a promising technique when target volume is the

left breast and internal mammary lymph nodes. Med Dosim 2006: 31:283-291.

25. Johansen S, Cozzi L, Olsen DR. A planning comparison of dose patterns in organs at risk and predic-

ted risk for radiation induced malignancy in the contralateral breast following radiation therapy of

primary breast using conventional, IMRT and volumetric modulated arc treatment techniques. Acta

Oncol 2009: 48:495-503.

26. Moran JM, Ben-David MA, Marsh RB, Balter JM, Griffith KA, Hayman JA, Pierce LJ. Accelerated partial

breast irradiation: what is dosimetric effect of advanced technology approaches?. Int J Radiat Oncol

Biol Phys 2009: 75:294-301.

27. Moon SH, Shin KH, Kim TH et. al. Dosimetric comparison of four different external beam partial

breast irradiation techniques: three-dimensional conformal radiotherapy, intensity-modulated ra-

diotherapy, helical tomotherapy, and proton beam therapy. Radiother Oncol 2009: 90:66-73.




29. Immink JM, Putter H, Bartelink H et. al. Long-term cosmetic changes after breast-conserving treat-

ment of patients with stage I-II breast cancer and included in the EORTC ‘boost versus no boost’ trial.

Ann Oncol 2012: 23:2591-2598.

63




J Surg Oncol 2012: 38:918-924.

31. Hopewell JW. The skin: its structure and response to ionizing radiation. Int J Radiat Biol 1990: 57:751-

773.

32. Archambeau JO, Pezner R, Wasserman T. Pathophysiology of irradiated skin and breast. Int J Radiat

Oncol Biol Phys 1995: 31:1171-1185.

33. Oborn BM, Metcalfe PE, Butson MJ, Rosenfeld AB. Monte Carlo characterization of skin doses in 6 MV

transverse field MRI-linac systems: effect of field size, surface orientation, magnetic field strength,

and exit bolus. Med Phys 2010: 37:5208-5217.

34. Oborn BM, Metcalfe PE, Butson MJ, Rosenfeld AB. High resolution entry and exit Monte Carlo dose

calculations from a linear accelerator 6 MV beam under the influence of transverse magnetic fields.

Med Phys 2009: 36:3549-3559.

35. Frazier RC, Vicini FA, Sharpe MB et. al. Impact of breathing motion on whole breast radiotherapy: a

dosimetric analysis using active breathing control. Int J Radiat Oncol Biol Phys 2004: 58:1041-1047.

36. George R, Keall PJ, Kini VR et. al. Quantifying the effect of intrafraction motion during breast IMRT

planning and dose delivery. Med Phys 2003: 30:552-562.

37. Schmitz AC, Pengel KE, Loo CE et. al. Pre-treatment imaging and pathology characteristics of invasi-

ve breast cancers of limited extent: potential relevance for MRI-guided localized therapy. Radiother

Oncol 2012: 104:11-18.

Chapter 4Post-lumpectomy CT-guided tumor

bed delineation for breast boost and partial breast irradiation:

Can additional pre- and postoperative imaging reduce

interobserver variability?

M.D. den Hartogh, M.E.P. Philippens, I.E. van Dam, C.E. Kleynen, J.H.A. Tersteeg, A.N.T.J. Kotte, M. van Vulpen, B. van Asselen, H.J.G.D. van den Bongard

Submitted

Abstract

Purpose

For breast boost radiotherapy or accelerated partial breast irradiation, the tumor bed is delineation by the radiation oncologist on a planning CT scan. The purpose of this study was to investigate whether the interobserver variability of the tumor bed delineation could be reduced by providing the radiation oncologist with additional MRI or CT imaging.

Methods

Fourteen cT1-2 patients underwent a standard planning CT in supine treatment position after lumpectomy and additional pre- and postoperative imaging in the same position. Post-lumpectomy tumor beds were delineated by 4 breast radiation oncologists on standard postoperative CT and on CT registered to an additional imaging modality. Additional imaging modalities were: postoperative MRI, preope-rative contrast-enhanced (CE) CT and preoperative CE-MRI. A cavity visualization score (CVS) was assigned to each standard postoperative CT by each observer. Con-formity index (CI), volume, and distance between centers of mass (dCOM) of the tumor bed delineations were calculated.

Results

On CT, median CI was 0.57, volume 22 cm3 and dCOM 5.1 mm. Addition of postoperative MRI increased the median tumor bed volume significantly to 28 cm3 (p<0.001), while the CI (p=0.176) and dCOM (p=0.110) were not influenced. Addi-tion of preoperative CT or MRI increased the tumor bed volume to 26 and 25 cm3 (both p<0.001). The CI increased to 0.58 and 0.59 (both p<0.001), and dCOM decre-ased to 4.7 and 4.6 mm (p=0.004 and p=0.001). In patients with CVS ≤ 3, median CI was 0.40 on CT, which significantly increased by all additional imaging modalities up to 0.52, accompanied by a median volume increase up to 6 cm3.

Conclusion

The addition of postoperative MRI, preoperative CE-CT or preoperative CE-MRI did not result in a considerable reduction of the interobserver variability in postopera-tive CT-guided tumor bed delineation, while target volumes slightly increased. The value of additional imaging may be dependent on CVS.

66

Chapter 4

Introduction

In early-stage breast cancer, radiotherapy following breast-conserving surgery is an integral part of breast-conserving treatment [1]. Traditionally, the whole breast is irradiated, with an additional boost dose to the tumor bed (TB) in patients with a higher risk of local recurrence [2]. Even though the geometric precision of modern radiation dose delivery is high, target delineation uncertainties are often large [3]. Especially delineation of the TB area, which is performed on the postoperative ra-diotherapy planning CT scan, is highly variable among radiation oncologists [4-8]. This could potentially lead to treatment inaccuracies, which are of particular con-cern with the increasing use of Accelerated Partial Breast Irradiation (APBI), in which only the breast tissue surrounding the TB is irradiated. Furthermore, postoperative seroma formation can lead to large TB volumes, which are associated with an incre-ased risk of subcutaneous fibrosis and worse cosmetic results [9-12]. There are several potential options to improve TB visualization for standard CT-gui-ded TB delineation. MRI might have additional value, since it has an excellent soft-tissue contrast compared to CT. The use of different MRI sequences makes it possible to differentiate between fibroglandular tissue, fluid and fat, and it can show the heterogeneity and irregularity of seromas [13]. Another opportunity to improve CT-guided TB delineation might be to increase the observer’s knowledge of the original tumor location. In the current situation, the preoperative diagnostic mammogram or MRI are some of the features used by the radiation oncologist to reconstruct the original tumor location on the planning CT scan with. This diagnos-tic imaging is not acquired in the supine radiotherapy treatment position which makes it difficult to interpret the original tumor position on the supine planning CT. Preoperative imaging in treatment position might improve the observer’s knowled-ge about the original tumor location and thereby potentially reduce the interob-server variability (IOV). The addition of a preoperative contrast-enhanced (CE) CT in supine radiotherapy treatment position to the standard postoperative planning CT was investigated by Boersma et al. [14]. They reported a minor reduction of the IOV. However, we showed in a recent delineation study on preoperative breast tumor delineation, that CE-MRI in radiotherapy supine position was superior to CE-CT for tumor detection and the visualization of tumor irregularities and spiculations [15]. Therefore, in this study, CE-CT, which is most commonly available in radiotherapy institutes, and CE-MRI were both investigated as an additional imaging modality to improve post-lumpectomy CT-guided tumor bed definition.The purpose of this study was to investigate whether the IOV of standard CT-gui-ded post-lumpectomy tumor bed delineation could be reduced by using additional postoperative MRI, preoperative CE-CT or preoperative CE-MRI in supine radiothe-rapy treatment position.

67

Additional imaging for postoperative tumor bed definition

Materials and methods

Patients and selectionThe study population included NTR3198 study patients, who received both CT and MRI before and after lumpectomy as part of the study, which was approved by our institutional review committee and registered in the International Clinical Trials Re-gistry [15]. Patients eligible for inclusion had a clinical T1-T2, N0 staged adenocarci-noma of the breast and were scheduled for breast-conserving therapy. Patients with lobular carcinoma, a history of ipsilateral breast surgery, contra-indications for 1.5 Tesla MRI, iodine allergy, and patients who received neoadjuvant treatment were not eligible.

Image acquisitionPatients underwent preoperative CE-CT and CE-MRI before lumpectomy, and a stan-dard planning CT directly followed by an MRI at a median of 21 days (range 14-50) after lumpectomy. All imaging was performed in the supine treatment position. For both CT and MRI, patients were placed with the arms in abduction and hands above the head at 10° inclination and with the use of a knee support (C-Qual and Thorawedge for CT and MRI respectively, CIVCO medical solutions, Reeuwijk, The Netherlands). The tumor or surgical scar was marked on the skin with a CT/MRI com-patible wire. For MRI, a wide bore (70 cm) MRI scanner (Ingenia 1.5T, Philips Medical Systems, Best, The Netherlands) and anterior receive coil were used. To prevent breast deformation by the anterior receive coil, a polymethyl methacrylate (PMMA) support was desig-ned, which is adjustable to patient habitus and breast size. The following 3D high resolution MRI images were used in this study: T1 weighted (T1w) fast field echo (FFE) with and without fat suppression (Dixon), T2 weighted (T2w) turbo spin echo (TSE) with fat suppression (SPAIR) and preoperative dynamic series of CE T1w Dixon images. MRI sequence parameters are provided in Table 1. The total acquisition times of the pre- and postoperative MRI protocols were 21 and 14 minutes, respectively.

Target volume delineationAll images were transferred to our in-house developed delineation software [16]. The postoperative MRI was registered to the postoperative planning CT by rigid mutu-al information registration on a box around the tumor, using the T1w images with fat-suppression. The preoperative CE-CT and CE-MRI were registered to the planning CT by automatic registration on the chest wall. Four experienced breast radiation oncologists independently delineated the TB, using written delineation instructions. These instructions were formulated in a consensus meeting with all observers, su-pervised by an experienced breast radiologist. The consensus meeting was repeated once, to answer questions regarding MRI and to discuss ambiguities in the delineati-on instructions. Delineation took place as follows: Firstly, the observer delineated the

68

Chapter 4

TB on standard postoperative planning CT and assigned a Cavity Visualization Score (CVS) which ranges from 1 (no cavity visible), 2 (heterogeneous cavity with indistinct margins), 3 (heterogeneous cavity with some distinct margins), 4 (mildly heteroge-neous cavity with mostly distinct margins), to 5 (homogeneous cavity with clearly identified margins) [17]. Secondly, the CT-based delineation was duplicated and ad-justed according to findings at the co-registered 1) postoperative MRI, 2) preopera-tive CT and 3) preoperative MRI. The original CT-guided delineation was duplicated to prevent influencing our data by intraobserver variations and to solely study the influence of additional imaging on this standard delineation method.

Data analysis

The conformity index (CI) and distance between the centers of mass (dCOM) of the TB contours were calculated for all possible observer pairs. The CI per observer pair was calculated by using the following formula: CI = . Consequently, CI = 1 implies a perfect agreement among observers, while CI = 0 means there is no overlap. For dCOM, a value of 0 means that two delineations are centered at the same position. Median values and accompanying ranges were used to describe the data since not all variables were normally distributed. A Wilcoxon signed-rank test was performed to compare paired variables using IBM SPSS Statistics 20 (Chicago, IL, USA) with a significance level of α=0.05. To visualize the influence of additional imaging on the CI, the change in CI per observer pair was plotted against the original CI of that ob-server pair on CT by using GraphPad Prism 6 (La Jolla, CA, USA). Furthermore, these outcomes were categorized by CVS ≤ 3 and CVS ≥ 4.

69


Table 1 MRI sequence parameters

Postoperative Postoperative Preoperative dynamicT2 TSE SPAIR T1 DIXON FFE T1 DIXON FFE

Orientation Transverse Transverse TransverseAcquisition mode 3D 3D 3DFOV(mm) 250x450x200 250x448x200 250x448x200Matrix size 200x357x167 252x447x182 208x388x167Acquired voxel size (mm) 1.25x1.25x2.40 0.99x1.00x2.20 1.20x1.21x2.40Reconstructed voxel size (mm) 0.78x0.78x1.2 0.95x0.93x1.10 1.16x1.18x1.20TR/TE (ms/ms) 2000/172 7.1/1.71 6.1/1.87Flip angle 90 12 10Refocusing angle 120 n.a. n.a.Turbo factor 74 117 84NSA 1 2 1Fat suppression SPAIR Dixon DixonAcquisition time (minutes) 5:40 7:51 4:13

FOV field of view, TR repetition time, TE echo time, TSE turbo spin echo, NSA number of averages, FFE fast field echo, SPAIR spectral adiabatic inversion recovery, n.a. not applicable

Results

Fourteen patients were prospectively included in this study (Table 2). Most patients underwent full-thickness closure of the excision cavity, which consists of suturing the deep and superficial layers of the cavity’s breast tissue. A standard postoperati-ve CT, registered to postoperative MRI, preoperative CE-CT and preoperative CE-MRI in one patient is shown in Figure 1. The different features of the TB as visualized by different MRI sequences are shown in Figure 2. Fat suppressed T1w (Figure 2b) and T2w (Figure 2d) images enable distinction between fibroglandular tissue and sero-ma. Surgical clips can be visualized by the T1w not fat-suppressed images (Figure 2c)

Table 2 Patient characteristics

Characteristic Value

Age (years) median

48-7061

Microscopic tumor diameter (mm) median

6-2912

Histology ductal carcinoma ductulolobular carcinoma tubular carcinoma

1031

Side left right

77

Days between surgery and postoperative imaging median

Surgical technique Open cavity Full-thickness closure

Number of clips placed median

Cavity visualization scores (mean)1 − no cavity visible2 − heterogeneous cavity, indistinct margins3 − heterogeneous cavity, some distinct margins4 − mildly heterogeneous cavity, mostly distinct

margins5 − homogenous cavity, clearly identified margins

14-5021

212

4-65

31525

1

70

Chapter 4

Figure 1 Sixty-four year old patient with pT1cN0(sn) ductal carcinoma of the right breast, cavity visualization score was 2 on CT. CT-guided tumor bed delineation of 4 observers, shown in different colors, on the same transversal slice on A) postoperative CT, B) postoperative T2w MRI, C) preoperative CE-CT and D) preoperative CE-MRI.

Figure 2 Different features of the tumor bed as shown on (A) Postoperative planning CT,(B) Postoperative T1w MRI with fat suppression, (C) Postopera-tive T1w MRI, (D) Postoperative T2w MRI with fat suppression. Arrows: Blue fibroglandular tissue, Red seroma, Orange surgical clip, Green area with intermediate signal intensity.

71


Interobserver variability and volumesTB delineation by the 4 observers resulted in wide ranges in volume, CI and dCOM on standard postoperative planning CT (Table 3), which did not improve with any of the additional imaging methods. The lower limit of 0.00 of the range in CI was the result of an absolute non-agreement among observers, with no overlap of TB delineations. This non-agreement occurred in 1 patient, with a centrally located tumor bed, with a CVS of 1 (no cavity visible) assigned unanimously by al observers. One observer con-toured a region different from the other 3 observers. Excluding this outlier from the analysis did not influence the outcomes. This observer did not cause outliers in any of the other patients. Data analysis showed there was no observer who structurally deviated from the other observers concerning volume, CI and dCOM. Addition of a postoperative MRI to the standard postoperative planning CT did not influence the CI (p=0.176) or dCOM (p=0.110) (Table 3). However, the TB volumes in-creased significantly (p<0.001) with a median increase of 6 cm3. Addition of a preoperative CT or MRI significantly increased the CI (both p<0.001) and dCOM (p=0.004 and p=0.001) (Table 3). A significant absolute volume increase, 4 cm3 and 3 cm3, was observed after addition of preoperative CE-CT and CE-MRI, respectively (both p<0.001). In Figure 3, the change in CI per observer pair after addition of an imaging method on the Y-axis is plotted against the original CI of that observer pair on standard postope-rative CT on the X-axis.

72

Chapter 4

Table 3 Volume, conformity index and dCOM of the tumor bed delineations.median range p-value#

Volume (cm3)CT 22 4 - 934CT + postoperative MRI 28 3 - 964 <0.001CT + preoperative CT 26 6 - 933 <0.001CT + preoperative MRI 25 7 - 933 <0.001

Conformity indexCT 0.57 0.00 - 0.90CT + postoperative MRI 0.61 0.00 - 0.89 0.176CT + preoperative CT 0.58 0.00 - 0.90 <0.001CT + preoperative MRI 0.59 0.00 - 0.89 <0.001

dCOM (mm)CT 5.11 0 - 53CT + postoperative MRI 3.72 0 - 52 0.110CT + preoperative CT 4.69 1 - 42 0.004CT + preoperative MRI 4.56 0 - 48 0.001

#p-value of the median of differences between the additional imaging modality and standard postoperative CT

Figure 3 Conformity index (CI) per observer pair on standard postope-rative CT (x-axis), plotted against the change in CI (y-axis) after addition of (A) postoperative MRI (B) preopera-tive CT (C) preoperative MRI ▼ = CVS ≤ 3; ∙ = CVS ≥ 4

These outcomes were categorized by CVS ≤ 3 and CVS ≥ 4. In general, CI was higher in patients with a high CVS score, as represented by the circles is Figure 3. In these patients, conformity did not increase with the use of additional imaging modali-ties and even significantly decreased after addition of postoperative MRI (p= 0.016). Contrarily, in patients with a low CI and a more heterogeneous TB (CVS ≤ 3), an in-crease in CI was observed from a median 0.40 on CT up to 0.52 on CT with additional preoperative MRI. In this same subgroup, volumes increased from 17 cm3 on CT up to 23 cm3 on CT with additional postoperative MRI.

73


Discussion

To our knowledge, this was the first study investigating the value of both additional pre- and postoperative CT and MRI with all imaging acquired in radiotherapy supi-ne position by using wide bore CT and MRI scanners. The addition of postoperative MRI did not improve the IOV of standard CT-guided TB delineation. This was against our expectations, since the use of different MRI sequences enables differentiation between different soft tissues (Figure 2). In the fat-suppressed images, MRI is unique for its clear contrast between seroma and fibroglandular tissue. However, the interpretation of the different MRI sequences in combination with the available patient information seems to be observer-de-pendent, despite the training and written delineation instructions. We found that observers rather expand their target volume than reduce their original CT-guided delineation based on the information provided by additional imaging. For example, in Figure 1B the area of seroma and architectural distortion on T2-weighted MRI was included. Furthermore, when observers are provided with preoperative imaging, they only expand the original CT-based delineation in the direction of the origi-nal tumor and do not adjust, for instance, the medial borders. They seem willing to expand their contour based on additional information, but not seem to reduce it when an area might not be part of the tumor bed. In that case, they do seem to favor their interpretation of the standard planning CT which they are most familiar with. This finding can also explain the volume increase that is observed after provi-ding additional imaging.The heterogeneity in CI, as showed by Figure 3, indicates that the subgroup of pa-tients with CVS ≤ 3 potentially benefits. However, the clinical relevance of this fin-ding is debatable since the median CI after addition of preoperative MRI for this patient subgroup was still only 0.52. Moreover, these findings should be interpreted with caution, since delineated volumes also showed a median increase up to 6 cm3. Though, it might be interesting to focus on the subgroup of patients with CVS ≤ 3 in future studies in a larger patient cohort, especially since a higher incidence of these CVS scores could be expected with the increasing use of full-thickness closure after lumpectomy. In line with our study results, Kirby et al. reported increased TB volumes delineated on CT-MRI datasets [18]. In our current study, we investigated whether additional MRI, registered to the standard postoperative planning CT, could reduce the IOV of tumor bed delineation. Tumor bed delineation on MRI-only was previously repor-ted [19]. In that study the conformity among observers was even lower on MRI-only, which is also in line with results reported by Giezen en al [20]. However, in a study by Jolicoeur et al. IOV improved and volumes were smaller on MRI-only, compared to CT [21]. This contradiction might be caused by MRI quality and the definition of

74

Chapter 4

the TB. In the study by Jolicoeur et al., the TB was defined as an architectural change on primarily T2 weighted MRI, while in the other studies the TB was reconstructed according to architectural changes, original tumor location on preoperative diag-nostic imaging, physical examination and the placement of surgical clips. Further-more, Jolicoeur et al. primarily used T2 weighted sequences, while in our study, multiple sequences were used. Moreover, the use of different operation techniques might have influenced the difference in study outcomes. Most patients in our stu-dy underwent full-thickness closure, while Jolicoeur et al. excluded patients who underwent oncoplastic techniques, which probably included full-thickness closure as well. After suturing the cavity walls, seroma might follow the suturing lines, the shapes of which might be more subject to interpretational differences compared to clearly defined cavity walls in superficially closed cavities (Figure 2, red and green arrows).Both Giezen et al. and Jolicoeur et al. investigated TB delineation on MRI separate-ly and compared it with CT-guided delineation. In our study, we investigated the additional value of MRI registered to standard postoperative CT, since this would be the application in clinical practice. In this setting, we found no added value of postoperative MRI for the general postoperative patient population. However statistically significant, no clinically relevant influence of either additional preoperative MRI or CT imaging on the CI of postoperative CT-guided TB delineati-on was found in our study. Our findings are in line with Boersma et al., who found no increase in CI after addition of a preoperative CE-CT [14]. Study results could be influenced by structural observer outliers, observer training and observer knowledge of MRI interpretation. Since there is no gold standard to validate the ‘correct’ imaging modality for TB delineation, consensus among obser-vers is used as an alternative method. At inspection of the data, no observer was structurally deviating in volume, CI or dCOM. Even in the presence of guidelines, training and an adequate number of clips, considerable variation still exists [22]. A possible limitation of this study is the small number of patients. However, with both increases and decreases in IOV and volumes, a larger cohort will probably not change the average difference considerably. For the overall patient population, additional imaging will not improve consistency in TB delineation. The question arises as to whether it is possible to further improve the IOV of postoperative CT-guided TB delineation. Several groups have proposed irradiation of the high-risk breast tissue surrounding the tumor before lumpectomy, when the tumor is still in situ [15, 23-25]. This preoperative approach would result in a much lower IOV compared to the current standard postoperative treatment [15, 25]. Moreover, preoperative image guided target volume definition could be valida-ted by pathology studies as a gold standard, which might improve confidence in an accurate treatment of the high-risk area [26].

75


In conclusion, the addition of postoperative MRI, preoperative CE-CT or preoperati-ve CE-MRI did not result in a considerable reduction of the interobserver variability of postoperative CT-guided tumor bed delineation, while target volumes slightly increased. The influence of additional imaging may be dependent on CVS.

76

Chapter 4

References




378:1707-1716.




3. Malinen E, Muren LP. Image guided therapy - do we get the picture?. Acta Oncol 2014: 53:3-5.









2009: 90:60-65.






9. den Hartogh MD, van Asselen B, Monninkhof EM et. al. Excised and irradiated volumes in relation to

the tumor size in breast-conserving therapy. Breast Cancer Res Treat 2011: 129:857-865.









J Surg Oncol 2012: 38:918-924.



72:49-57.

14. Boersma LJ, Janssen T, Elkhuizen PH et. al. Reducing interobserver variation of boost-CTV deline-

ation in breast conserving radiation therapy using a pre-operative CT and delineation guidelines.


77


15. den Hartogh MD, Philippens ME, van Dam IE et. al. MRI and CT imaging for preoperative target

volume delineation in breast-conserving therapy. Radiat Oncol 2014: 9:63.




ology 2001: 219:203-206.




Biol Phys 2009: 74:1276-1282.

19. den Hartogh MD, van den Bongard HJ, Davidson MT et. al. Full-Thickness Closure in Breast-Conser-

ving Surgery: The Impact on Radiotherapy Target Definition for Boost and Partial Breast Irradiation.

A Multimodality Image Evaluation. Ann Surg Oncol 2014.



Biol Phys 2012: 82:1332-1340.

21. Jolicoeur M, Racine ML, Trop I, Hathout L, Nguyen D, Derashodian T, David S. Localization of the

surgical bed using supine magnetic resonance and computed tomography scan fusion for planifi-

cation of breast interstitial brachytherapy. Radiother Oncol 2011: 100:480-484.

22. Kirby AN, Jena R, Harris EJ, Evans PM, Crowley C, Gregory DL, Coles CE. Tumour bed delineation

for partial breast/breast boost radiotherapy: what is the optimal number of implanted markers?.






Clin Oncol 2012: 36:232-238.






97:225-232.

78

Chapter 4

Chapter 5Full-thickness closure in

breast-conserving surgery:The impact on radiotherapy target

definition for boost and partial breast irradiation

A multimodality image evaluation

M.D. den Hartogh, H.J.G.D. van den Bongard, M.T.M. Davidson, A.N.T.J Kotte, H.M. Verkooijen, M.E.P. Philippens, M. van Vulpen, B. van Asselen, J.P. Pignol

Based on: Ann Surg Oncol 2014: 21: 3774-3779

82

Chapter 5

Abstract

Purpose

During breast conserving surgery (BCS), surgeons increasingly perform full-thick-ness closure (FTC) to prevent seroma formation. This could potentially impair preci-sion of target definition for boost and accelerated partial breast irradiation (APBI). The purpose of this study was to investigate the precision of target volume defi-nition following BCS with FTC among radiation oncologist, using various imaging modalities.

Methods

Twenty clinical T1-2N0 patients, scheduled for BCS involving clip placement and FTC, were included in the study. Seven experienced breast radiation oncologists contoured the tumor bed on CT, MRI and fused CT-MRI datasets. A total of 121 indi-vidual image sets and 360 observer pairs per image modality were analyzed. A pair-wise conformity among the generated contours of the observers and the distance between their centers of mass (dCOM) were calculated.

Results

On CT, median conformity was 44% (interquartile range (IQR) 28-58%) and median dCOM was 5.5 mm (IQR 3-9 mm). None of the outcome measures improved when MRI or fused CT-MRI were used. In 2 patients, superficial closure was performed instead of FTC. In these 14 image sets and 42 observer pairs, median conformity increased to 70%.

Conclusions

Localization of the radiotherapy target after FTC is imprecise, on both CT and MRI. This could potentially lead to a geographical miss in patients at increased risk of local recurrence receiving a radiation boost, or for those receiving APBI. These fin-dings highlight the importance for breast surgeons to clearly demarcate the tumor bed when performing FTC.

83

Full-thickness closure and multi-modality tumor bed definition

Introduction

With the introduction of breast cancer screening programs, the majority of wo-men with breast cancer are currently diagnosed at an early stage [1, 2]. In these patients, breast-conserving surgery (BCS) followed by whole breast radiotherapy is the standard of care and leads to excellent local control rates while preserving the affected breast [3, 4]. The addition of a radiation boost to the tumor bed further reduces the risk of breast recurrences [5]. During BCS, surgeons increasingly per-form full-thickness closure (FTC) of the excision cavity, instead of a circumferential tumorectomy followed by superficial closure (SC). In FTC, both deep and superficial layers of breast tissue are repositioned and approximated after an excision which is generally performed down to the pectoral fascia. The advantages of FTC include a better hemostatic control, lower risk of post-surgical infections, less fibrosis and increased sensitivity of follow-up mammography in detecting local recurrence [6-8]. The downside of FTC is that it may hinder the visualization of the tumor bed on CT-images, which is an essential step in planning the radiation boost or partial bre-ast irradiation. While SC usually leaves a seroma-filled cavity behind which is easily identifiable on planning CT, seroma is often absent after FTC or follows indistinct surgical boundaries which are not always reflecting the original tumor location. Consequently, the surgical closure technique may impair precision of tumor bed definition, which drives the targeting of radiotherapy doses, and ultimately deter-mines the treatment quality and efficacy. The purpose of this study was to investigate the precision of tumor bed definition following BCS with FTC among radiation oncologist, by using various imaging mo-dalities.

Methods and Materials

Study populationThe study population included NTR3198 study patients, who received both CT and MRI after BCS as part of the study, which was approved by our institutional review committee and registered in the Netherlands Trials Register (NTR) and the Interna-tional Clinical Trials Registry Platform. Eligible patients included women who under-went ipsilateral BCS for a clinically T1-2, N0 breast cancer referred to the University Medical Center Utrecht (The Netherlands). Patients with neoadjuvant treatment or contra-indications for imaging with 1.5 Tesla MRI were excluded. All patients had surgical clips placed in the surgical cavity to guide radiotherapy target definition and underwent FTC, which is routine practice in the Netherlands. Seven experien-ced radiation oncologists from the Sunnybrook, (Odette Cancer Center, Toronto, Canada), routinely using CT-imaging for tumor bed contouring and familiar with breast MRI were asked to participate in this study.

84

Chapter 5

Imaging

In standard clinical practice, the radiotherapy target definition is performed by con-touring the tumor bed on a postoperative planning CT scan. Since MRI has been recently shown to improve contouring precision in patients having undergone SC, we also investigated whether MRI could aid the target definition after FTC [9]. Patients underwent standard planning CT and MRI at a median of 3 weeks after lumpectomy (range 14 – 50 days). CT and MRI were performed in the same supine treatment position with the hands above the head. Patients were positioned on an MRI compatible 10° wedge board (Thorawedge, CIVCO medical solutions, Reeu-wijk, The Netherlands) with knee support. The surgical scar was marked on the skin with a CT/MRI compatible wire. CT images were obtained at 3 mm slice thickness and a minimal in-plane resolution of 1 x 1 mm2 (Brilliance, Philips Medical Systems, Best, The Netherlands), For MRI, a polymethyl methacrylate (PMMA) coil support was used to prevent breast deformation by the anterior receive coil. 3D high-resolu-tion MR images were acquired using a wide bore MRI scanner (Ingenia 1.5T, Philips Medical Systems, Best, The Netherlands) using the following three sequences with a reconstructed voxel size of ≤ 1x1x1 mm3. T1 weighted (T1w) fast field echo (FFE) i) with and ii) without fat suppression (Dixon) and iii) T2 weighted (T2w) turbo spin echo (TSE) with fat suppression (SPAIR).

Radiotherapy target definition

CT and MRI image sets were transferred to the Pinnacle treatment planning soft-ware version 9.2 at the Canadian study site (Philips Medical Systems Inc, Cleveland, Ohio, USA). The tumor bed was first contoured on the standard postoperative plan-ning CT for each patient by seven experienced breast radiation oncologists from the Canadian study site. To investigate whether tumor bed definition could be im-proved by MRI, the observers also contoured the tumor bed on MRI images, as well as on fused CT-MRI datasets independently, with at least a 2-week interval between modalities. For fused CT-MRI datasets, images were co-registered by rigid mutual information registration on a box around the tumor bed area. Observers were blinded for each other’s contours and also to their own contours at the other imaging modality. For this purpose, images were anonymized and presen-ted in a random sequence. The radiation oncologists were provided with each pa-tient’s preoperative information (mammography, ultrasound) and surgical report, including the scar position and location of surgical clips. Observers were provided with written contouring guidelines formulated in a consensus meeting and each observer was individually trained on tumor bed appearances on the different MRI sequences.

85


Outcome parameters

The following outcome measures of interobserver variability were captured:

1. The conformity Index (Cx), corresponding to the ratio between the volume of agreement of the defined target volumes divided by the encompassing volu-me, for each observer pair (Figure 1).

2. The distance between observers’ target volumes (dCOM), which was quantified by the distance between the centers of mass of the tumor bed contours of each observer pair.

3. Volume of the defined targets 4. Cavity Visualization Score (CVS). For each set of images, the visualization of the

tumor bed area was scored by each observer assigning a CVS [10]. This is a 5-point-scale ranging from 1 (no cavity visible), 2 (heterogeneous cavity with indistinct margins), 3 (heterogeneous cavity with some distinct margins), 4 (mildly heterogeneous cavity with mostly distinct margins), to 5 (homogene-ous cavity with clearly identified margins).

Figure 1 Conformity index as calculated by the volume of agreement divided by the encompassing volume per observer pair.

Figure 1 Conformity index as calculated by the volume of agreement divided by the encompassing volume per observer pair.

86

Chapter 5

Statistical analysisTo compare target definition between CT and MRI/CTMRI-fusion with a confidence level of 95% and a power of 80%, a 5% in- or decrease in conformity (defined as a 0.05 change in Cx, with a sigma of 0.3) would require the inclusion of 283 observer pairs. The inclusion of 20 patients and 7 observers would account for 140 individual image sets and 420 observer pairs per imaging dataset (CT, MRI and CTMRI-fusion).Median values and accompanying interquartile ranges (IQR) were calculated, as Sha-piro-Wilk normality tests showed variables were not normally distributed. Wilcoxon tests were performed to compare Cx, dCOM and volume between image modalities. Only complete data sets, i.e. CT, MRI and CTMRI data available per observer, were included in the analysis. The 95% confidence interval of the median difference (95% CI) and p-values were reported.IBM SPSS Statistics 20 (Chicago, IL, USA) was used, with a significance level of α=0.05. Box-and-whisker-plots were created by using GraphPad Prism 6 (La Jolla, CA, USA).

Results

PatientsCharacteristics of all 20 patients are shown in Table 1. In 18 patients, the standard FTC procedure was performed. In 2 patients, the surgeon performed a SC technique, leaving a rounded seroma cavity. A SC technique is sometimes preferred in patients with low-density breast tissue with a major fatty composition, since this determines the ability to perform breast undermining and reshaping without complications [11]. Data of all 20 patients were analyzed according to an intention-to-treat principle. Data on 2 SC patients were also reported separately, to enable comparison with pre-vious studies on target definition in SC patients. Ten contours, divided over 9 pa-tients, were inevaluable due to either the observers’ impossibility of defining a target, or contouring a target in the opposite breast in patients previously treated for con-tralateral breast carcinoma. These cases were excluded from analysis. Only observer pairs with all three imaging modalities available were included in the analysis. In to-tal, 361 complete observer pairs per imaging modality were included in the analysis.

Outcome parameters1. Cx - For the total cohort, median conformity of the radiotherapy target on stan-

dard planning CT between the 7 observers was low at 44% and ranges were wide (IQR 28-58%), suggesting large discrepancies between observers (Figure 2A). A Cx of 0.02 reflects that in one patient, two observers contoured targets with only a 2% overlap. On MRI, conformity statistically significantly decreased to 32% (Cx=0.32, IQR 0.16-0.52, 95% CI (-0.1 to -0.04), p<0.001). When using CT-MRI-fusion, conformity decreased to 39% (Cx=0.39, IQR 0.21-0.55, 95% CI (-0.08 to -0.03), p<0.001).

87


Table 1 Patient characteristicsCharacteristic ValueAge (years) 39-72 median 61Microscopic tumor diameter (mm) 6-29 median 12Histology 14 ductal 3 ductulolobular 1 lobular 2 tubularSide left 11 right 9Days between surgery and postoperative imaging 14-50 median 21Surgical technique full-thickness closure 18 open cavity 2Number of clips 4-6 median 5

2. dCOM - On standard CT, targets were centered a median of 6 mm apart from each other (IQR 3-9 mm) (Figure 2B). MRI decreased the median dCOM to 5 mm (IQR 3-11 mm, 95% CI (-0.28 to 0.74), p=0.03)). No significant difference was ob-served for fused CT-MRI relative to the standard CT (median dCOM was 5 mm, IQR 3-12 mm, 95% CI (-0.33 to 0.54), p=0.07).

3. Volume - On standard CT the median volume of the targets was 23 cc (IQR 15-43 cc) (Figure 2C). One patient developed a large postoperative seroma (mean volu-me of 835 cc), and was excluded from the graph to allow appropriate scaling of the x-axis. Targets on MRI had a median volume of 20 cc (IQR 9-39 cc, 95% CI (-4.7 to -0.50), p=0.03) and fused CT-MRI had a median volume of 21 cc (IQR 12-32 cc, 95% CI (-4.4 to -0.7), p<0.001).

4. CVS - The median CVS was 3 (heterogeneous cavity with some distinct margins), ranging from 1 to 5, on all imaging modalities.

Superficial closureThe subgroup who received SC instead of FTC, consisted of 14 contours and 42 ob-server pairs and was also analyzed separately. Median conformity this SC subgroup was 70% (Cx 0.70, IQR 0.63-0.74), median dCOM was 1.5 mm (IQR 1.0-2.3 mm) and median volume was 52 cc (IQR 17-63 cc). Median CVS in SC patients was scored as 5 (homogeneous cavity with clearly identified margins).

88

Chapter 5

Figure 3 Tumor bed delineations by 7 radiation oncologists (different colors) after A) Full-thickness closure of the excision cavity in breast-conserving surgery, and B) Superficial closure.

Figure 2 Box-and-whisker plots of variation among observers according to imaging technique.

89


Discussion

This study evaluated the consequences of FTC, a surgical closure technique which is increasingly performed by breast surgeons, on radiotherapy target definition. In addition, this study evaluated whether MRI or CTMRI-fusion could improve target definition compared to standard planning CT in this patient population. A large variability in target definition was found between the different radiation oncologists on all imaging datasets used in this study. The interobserver agree-ment was poor on CT images, with a median conformity of 44%. Surprisingly, the agreement did not improve but significantly degraded when targets were defin-ed on MRI or fused CT-MRI datasets, with an even lower conformity of 32% and 39%, respectively (both p<0.001). This suggests that the inclusion of MRI datasets added some degree of confusion rather than increasing interobserver agreement. Those findings are at odds with previously reported results by Jolicoeur et al., which showed a good agreement between observers, with a median conformity on CT of 66%, and an improvement up to 96% with MRI [10]. However, in that study most pa-tients underwent a SC technique as surgical method, which was comparable to the 2 patients that underwent SC in our own series with a median conformity of 70%. These results suggest that it is the type of surgery and closure technique rather than the quality of imaging or the contouring skill of the radiation oncologist that is the cause of the degraded radiotherapy target definition found in our study. Finally, in our study, the median CVS score of 3 was low. This is suggesting that in FTC, the ab-sence of a distinct seroma cavity resulted in the increased variability of radiotherapy target definition. When comparing the Cx among studies, it has to be considered that Cx is volume dependent. The smaller the volume being studied, the more the CI is influenced by small interobserver differences. The current shift towards the use of FTC is unlikely to change as it presents several clinical benefits [6-8, 12]. However, there may be a selected group of patients where those advantages should be balanced against the increased risk of local recurren-ce due to a poor target definition. The first is the group of patients at high-risk of ipsilateral breast relapse who are eligible to receive a tumor bed boost after BCS. This is especially relevant for young women, typically less than 50 years old, with aggressive tumors or patients with positive surgical margins [5, 13]. A second group will be patients opting for accelerated partial breast irradiation (APBI). In APBI, only the tumor bed area is irradiated, such that any error in the target definition puts the patient at risk of a geographical miss, and hence at risk of cancer relapse. In clinical practice, the tumor bed delineation is always expanded with 1-2 cm to account for any microscopic disease and to compensate for treatment setup errors. This at least compensates for the geographical inaccuracies in some extend. However, while tre-ating a smaller volume may put the patient at risk of local recurrence, it should also

90

Chapter 5

be stressed that irradiating a larger volume than necessary may also put the patient at a higher risk of long-term side effects, including poorer cosmetic outcomes, fi-brosis and pain [8, 14-16]. This is of particular importance in APBI. Recently publis-hed results of the Canadian RAPID trial, in which patients are randomized between whole breast irradiation and APBI, showed increased rates of adverse cosmesis in patients treated with APBI at a median follow-up of 3 years [17]. The authors sug-gest that treatment volumes might have been too large in proportion to the breast volume.Are there alternative possibilities or strategies to improve the accuracy of radiothe-rapy target definition in FTC patients? Marking the tumor bed with surgical clips, would be the most logical solution in this situation, as tissues are being mobilized and repositioned during FTC. However, recent studies about the use of boost ra-diotherapy in patients treated with oncoplastic surgical techniques reported that only a minority of studies effectively reported marking of the tumor bed during sur-gery [18, 19]. This, while Pezner et al. showed that 73% of patients who underwent oncoplastic surgery had the final tumor bed extended beyond the original index quadrant [20]. In our study, however, the precision in target definition was low de-spite the presence of an adequate number of clips [21]. Though, in the absence of clips, precision would probably be even lower [21, 22]. One limitation of our study was that clips were not placed according to a predefined protocol. Therefore, we recommend for FTC patients that clip placement should be performed according to a predefined protocol, known by both surgeons and radiation oncologist, and that this placement being detailed in the surgical report [23]. According to Kirby et al., five implanted markers (one deep and four radial) are likely to be adequate for the purposes of target definition for APBI and boost radiotherapy using tangential fields [21]. Another potential option to improve radiotherapy treatment accuracy for patients in whom FTC or oncoplastic procedures are scheduled, would be to change the timing of radiotherapy treatment delivery to a moment when the surgical cavity or the tumor location is better defined. Such options can include the delivery of preoperative radiotherapy, or intraoperative or intracavitary partial breast radiothe-rapy [24-28]. While the delivery of intraoperative radiotherapy has its own logisti-cal challenges, like the need for radiotherapy facilities and quality assurance in the operating room, preliminary results seems to suggest that it represents an efficient option [29].In conclusion, localization of the radiotherapy target after FTC is imprecise, on both CT and MRI. This could potentially lead to a geographical miss in patients at incre-ased risk of local recurrence receiving a radiation boost, or for those receiving APBI. These findings highlight the importance for breast surgeons to clearly demarcate the tumor bed when performing FTC.

91


References

1. Nystrom L, Andersson I, Bjurstam N, Frisell J, Nordenskjold B, Rutqvist LE. Long-term effects of mam-

mography screening: updated overview of the Swedish randomised trials. Lancet 2002: 359:909-919.

2. Fracheboud J, Otto SJ, van Dijck JA, Broeders MJ, Verbeek AL, de Koning HJ, National Evaluation

Team for Breast cancer screening (NETB). Decreased rates of advanced breast cancer due to mam-

mography screening in The Netherlands. Br J Cancer 2004: 91:861-867.




378:1707-1716.

4. Whelan TJ, Pignol JP, Levine MN et. al. Long-term results of hypofractionated radiation therapy for

breast cancer. N Engl J Med 2010: 362:513-520.




6. Indelicato D, Grobmyer SR, Newlin H, Morris CG, Haigh LS, Copeland EM,3rd, Mendenhall NP. Asso-

ciation between operative closure type and acute infection, local recurrence, and disease surveil-

lance in patients undergoing breast conserving therapy for early-stage breast cancer. Surgery 2007:

141:645-653.

7. Paterson ML, Nathanson SD, Havstad S. Hematomas following excisional breast biopsies for invasi-

ve breast carcinoma: the influence of deep suture approximation of breast parenchyma. Am Surg

1994: 60:845-848.



J Surg Oncol 2012: 38:918-924.





ology 2001: 219:203-206.

11. Clough KB, Kaufman GJ, Nos C, Buccimazza I, Sarfati IM. Improving breast cancer surgery: a classifica-

tion and quadrant per quadrant atlas for oncoplastic surgery. Ann Surg Oncol 2010: 17:1375-1391.

12. Shaikh T, Narra V, Goyal S et. al. Lumpectomy closure technique does not affect dosimetry in pa-

tients undergoing external-beam-based accelerated partial breast irradiation. Ann Surg Oncol

2013: 20:1323-1328.

13. Poortmans PM, Collette L, Horiot JC et. al. Impact of the boost dose of 10 Gy versus 26 Gy in patients

with early stage breast cancer after a microscopically incomplete lumpectomy: 10-year results of

the randomised EORTC boost trial. Radiother Oncol 2009: 90:80-85.

92

Chapter 5









17. Olivotto IA, Whelan TJ, Parpia S et. al. Interim Cosmetic and Toxicity Results From RAPID: A Randomi-

zed Trial of Accelerated Partial Breast Irradiation Using Three-Dimensional Conformal External Beam

Radiation Therapy. J Clin Oncol 2013: 31:4038-4045.

18. Schaverien MV, Stallard S, Dodwell D, Doughty JC. Use of boost radiotherapy in oncoplastic bre-

ast-conserving surgery - A systematic review. Eur J Surg Oncol 2013: 39:1179-1185.

19. Azu M, Goyal S, Patel U, Haffty B, Kearney T. Has placement of surgical clips in the lumpectomy bed

fallen out of favor?. Ann Surg Oncol 2011: 18:1529-1532.

20. Pezner RD, Tan MC, Clancy SL, Chen YJ, Joseph T, Vora NL. Radiation therapy for breast cancer pa-

tients who undergo oncoplastic surgery: localization of the tumor bed for the local boost. Am J Clin

Oncol 2013: 36:535-539.




22. Shaikh T, Chen T, Khan A et. al. Improvement in interobserver accuracy in delineation of the lumpec-

tomy cavity using fiducial markers. Int J Radiat Oncol Biol Phys 2010: 78:1127-1134.






Clin Oncol 2012: 36:232-238.






27. Barry M, Ho A, Morrow M. The evolving role of partial breast irradiation in early-stage breast cancer.

Ann Surg Oncol 2013: 20:2534-2540.


volume delineation in breast-conserving therapy. Radiat Oncol 2014: 9:63.

29. Vaidya JS, Wenz F, Bulsara M et. al. Risk-adapted targeted intraoperative radiotherapy versus who-

le-breast radiotherapy for breast cancer: 5-year results for local control and overall survival from the

TARGIT-A randomised trial. Lancet 2014: 383:603-613.

Chapter 6Preoperative MRI and CT imaging for

target volume delineation in breast-conserving therapy:

Investigating the potential for a neoadjuvant irradiation

M.D. den Hartogh, M.E.P. Philippens, I.E. van Dam, C.E. Kleynen, J.H.A. Tersteeg, R.M. Pijnappel, A.N.T.J. Kotte,

H.M. Verkooijen, M.A.A.J. van den Bosch, M. van Vulpen, B. van Asselen, H.J.G.D. van den Bongard

Based on: Radiat Oncol 2014: 9:63

96

Chapter 6

Abstract

Purpose

Accurate tumor bed delineation after breast-conserving surgery is important. Ho-wever, consistency among observers on standard postoperative radiotherapy plan-ning CT is low and volumes can be large due to seroma formation. A preoperative delineation of the tumor might be more consistent. Therefore, the purpose of this study was to determine the consistency of preoperative target volume delineation on CT and MRI for breast-conserving radiotherapy.

Methods

Tumors were delineated on preoperative contrast-enhanced (CE) CT and newly de-veloped 3D CE-MR images, by four breast radiation oncologists. Clinical target vo-lumes (CTVs) were created by addition of a 1.5 cm margin around the tumor, exclu-ding skin and chest wall. Consistency in target volume delineation was expressed by the interobserver variability. Therefore, the conformity index (CI), center of mass distance (dCOM) and volumes were calculated. Tumor characteristics on CT and MRI were scored by an experienced breast radiologist.

Results

Preoperative tumor delineation resulted in a high interobserver agreement with a high median CI for the CTV, for both CT (0.80) and MRI (0.84). The tumor was missed on CT in 2/14 patients (14%). Leaving these 2 patients out of the analysis, CI was higher on MRI compared to CT for the GTV (p<0.001) while not for the CTV (CT (0.82) versus MRI (0.84), p=0.123). The dCOM did not differ between CT and MRI. The median CTV was 48 cm3 (range 28-137 cm3) on CT and 59 cm3 (range 30-153 cm3) on MRI (p<0.001). Tumor shapes and margins were rated as more irregular and spiculated on CE-MRI.

Conclusions

This study showed that preoperative target volume delineation resulted in small target volumes with a high consistency among observers. MRI appeared to be ne-cessary for tumor detection and the visualization of irregularities and spiculations. Regarding the tumor delineation itself, no clinically relevant differences in interob-server variability were observed. These results will be used to study the potential for future MRI-guided and neoadjuvant radiotherapy.

97

Preoperative MRI and CT for target volume definition

Background

The standard treatment of early-stage breast cancer is lumpectomy, or wide local excision, followed by whole breast irradiation with an additional boost dose to the tumor bed (TB) in patients with a higher risk of local recurrence [1, 2]. Since most local recurrences occur in or nearby the TB, several accelerated partial breast irradi-ation (APBI) studies are ongoing in early-stage breast cancer patients. APBI targets the breast tissue immediately surrounding the TB. The advantages of APBI are a shorter overall treatment time and a potential dose reduction in the normal tissues (i.e. breast, heart and lung) compared to whole breast irradiation [3]. Accurate TB delineation on the radiotherapy planning CT scan after lumpectomy is important for both TB boost irradiation and APBI. However, in radiotherapy practice, there is no gold standard to validate the accuracy of our target volume delineation after lumpectomy. As an alternative, consensus among observers is often used to assess the precision of our target volume delineation. The degree of consensus is generally called the interobserver variability (IOV), and quantified by a conformity index (CI) which is the volume of agreement among observers divided by the total encom-passing volume. The current CT guided delineation after lumpectomy is prone to a high IOV. Several studies showed a low conformity index (CI) and a large distance between the centers of mass (dCOM) among observers [4-13]. Besides the high IOV in the current postoperative radiotherapy setting, there is also the concern of large postoperative treatment volumes due to seroma and hemato-ma formation. Irradiation of these disproportionally large target volumes can lead to extended subcutaneous fibrosis, poor cosmetic results and even missing the tar-get [14-17]. Furthermore, these large volumes can cause low-risk patients aiming for APBI to be ineligible for this treatment due to the inability to meet the dose-vo-lume constraints [18, 19] The poor consistency in target volume definition and large volumes after lumpec-tomy might be avoided by irradiating the tumor preoperatively. Since the tumor is still in situ without any seroma formation, this would probably lead to a high de-lineation precision and small treatment volumes. Several groups are studying the potential for neoadjuvant irradiation in early stage breast cancer patients [18, 20, 21]. In these studies, IOV and normal tissue dose were reduced, which shows that neoadjuvant irradiation could result in more precise target volume definition and localization and smaller volumes [20-22]. Furthermore, Bondiau et al. reported the feasibility of a neoadjuvant stereotactic body irradiation in combination with ne-oadjuvant chemotherapy in locally advanced breast cancer patients [23].Alternatively, preoperative imaging in radiotherapy supine position might also have potential value to improve the standard post-lumpectomy TB delineation, since it provides additional information about the original tumor location [11].

98

Chapter 6

To correctly delineate the tumor, imaging quality is of great importance. Since it is unknown what the optimal imaging modality for preoperative target volume deli-neation is, delineation was studied on both contrast-enhanced (CE) CT and MRI. In daily clinical practice, CT is the standard imaging modality for target volume deli-neation in breast cancer patients. However, MRI has a superior soft tissue contrast which can be explored with different sequences to show endogenous contrast or the distribution of an administered contrast agent. This enables differentiation be-tween the tumor and benign lesions like post-biopsy hematomas or cysts. Further-more, MRI has a high sensitivity for detection of invasive breast cancer and a good correlation with histopathology findings [24, 25]. However, standard diagnostic MRI is performed in prone position, while patients in most departments are irradiated in supine position. Acquiring images in supine radiotherapy position is generally li-mited by narrow bore sizes of standard MRI scanners. Therefore a new MRI protocol was designed in a wide bore MRI scanner. The purpose of this study was to quantify the consistency of preoperative target vo-lume delineation for breast-conserving radiotherapy. To identify the best imaging modality for preoperative target volume delineation, preoperative delineation was performed on both CE CT and a newly developed 3D CE-MRI in supine radiotherapy position.

Methods

Patients and selectionThe study was approved by our institutional review board and registered in the In-ternational Clinical Trials Registry Platform (NTR3198). Fourteen early-staged bre-ast cancer patients, scheduled for lumpectomy at the University Medical Center Utrecht or St. Antonius hospital, were included in this study. All patients gave writ-ten informed consent. Patients eligible for inclusion had a clinical T1-T2, N0 staged adenocarcinoma of the breast and were scheduled for lumpectomy and sentinel node procedure. Patients with lobular carcinoma, a history of ipsilateral breast sur-gery, contra-indications for 1.5 Tesla MRI, iodine allergy, and patients who received neoadjuvant treatment were not eligible. In case of additional suspected findings on study MRI or CT imaging, patients were referred to their physician for additional diagnostic work-up.

Patient positioning and image acquisitionPatients underwent both CT and MRI in radiotherapy supine position prior to sur-gery. On CT, they were positioned with arms in abduction and hands above the head at 10° inclination and with the use of a knee support (C-Qual, CIVCO medical solutions, Reeuwijk, The Netherlands). If palpable, the tumor was marked on the skin with a CT/MRI compatible wire. CE-CT images were obtained at 3 mm slice

99


thickness and a minimal in-plane resolution of 1 x 1 mm2 (Brilliance, Philips Medical Systems, Best, The Netherlands), with a delay time of 120s after intravenous con-trast agent injection (Ultravist, 80 ml, 3ml/s) [11]. Delay time was shortened to 80s after the 6th patient according to Kuroki-Suzuki et al. in attempt to improve tumor enhancement [26].For MRI, patients were positioned on an MRI compatible 10° wedge board (Thora-wedge, CIVCO medical solutions, Reeuwijk, The Netherlands). To acquire MR ima-ges, an anterior receive coil was used. To prevent breast deformation by the anterior receive coil, a polymethyl methacrylate (PMMA) support was designed, which is ad-justable to patient habitus and breast size. MRI patient setup is shown in Figure 1. The bore of a standard MRI scanner is too narrow to acquire images in this position. Therefore, we used a wide bore (70 cm) MRI scanner (Ingenia 1.5T, Philips Medical Systems, Best, The Netherlands). The following 3D high resolution images were ac-quired: T1 weighted (T1w) fast field echo (FFE) ± fat suppression (Dixon), T2 weigh-ted (T2w) turbo spin echo (TSE) + fat suppression, and a dynamic series of con-trast-enhanced T1w images ± fat suppression after contrast agent administration.

Figure 1 MRI patient setup in radiotherapy supine position.

For T1w Dixon FFE MRI, acquired 3D resolution was 0.99 x 1.05 x 2.19 mm3 recon-structed to 0.95 x 0.95 x 1.1 mm3 using overcontiguous slices and for T2w TSE MRI,

100

Chapter 6

the voxels measured 0.78 x 0.78 x 1.2 mm3 acquired with a resolution of 1.25 x 1.32 x 2.41 mm3. In the dynamic T1w series, the first 3D image was acquired before and 6 images after intravenous contrast injection (Gadobutrol (Gadovist, Bayer), 0.1 mmol/kg, 1 ml/s), at 60s intervals with an acquired resolution of 1.20 x 1.21 x 2.41 mm3 reconstructed to 1.16 x 1.16 x 1.2 mm3 using overcontiguous slices. The total acquisition time of this protocol was 21 minutes. Small displacements between se-quences during image acquisition caused by patient motion were corrected for by using a rigid mutual information registration on a box around the tumor. No breast deformation by the anterior receive coil was observed.To quantify differences in tumor visualization on CT and MRI, the shape (1–round, 2–oval, 3–lobular, 4–irregular) and margin (1–smooth, 2–irregular, or 3–spiculated) of the tumor were rated by an experienced breast radiologist [27].

Target volume delineationFour experienced breast radiation oncologists independently delineated the GTV on both CT and MRI data, with at least a 4-week interval between delineation ses-sions, using an in-house developed software tool (Volumetool) [28]. Written deli-neation instructions were formulated in a consensus meeting with all observers, supervised by an experienced breast radiologist. MRI delineations were performed on preoperative 3D CE T1w images with an individually prescribed fixed window and level as determined by an experienced breast radiologist. Observers were al-lowed to consult other sequences, which were registered to the CE-MRI series to differentiate between structures, i.e. tumor (gadolinium uptake causes a high signal on CE T1w images), post-biopsy hematoma (blood causes a high signal on both CE and non-CE T1w images), and cysts (fluid causes a high signal on T2w images). Clini-cal target volumes (CTVs) were created by adding a 1.5 cm margin around the GTV, restricted by the chest wall and a 5 mm margin beneath the skin surface. Delinea-tion of a preoperative GTV different from the tumor location as confirmed during histopathological examination of the lumpectomy specimen (gold standard) was considered as ‘misdelineation’.

Data analysisThe conformity index (CI) and distance between the centers of mass (dCOM) for both the GTV and CTV contours as delineated by the 4 observers were calculated for all possible observer pairs. The CI per observer pair was calculated by using the following formula: CI = . Consequently, CI = 1 implies a perfect agreement among observers, while CI = 0 means there is no overlap. For dCOM, a value of 0 means that two delineations are centered at the same position. Median values and accompanying ranges were used to describe the data since not all variables were normally distributed. A Wilcoxon signed-rank test was performed to compare paired variables using IBM SPSS Statistics 20 (Chicago, IL, USA) with a significance level of α=0.05.

101


Results

Patients

Patient and tumor characteristics are shown in Table 1. Median age was 61 years (range 48-70). Median clinical tumor diameter (as measured on diagnostic ultra-sound/MRI) was 15 mm (range 7-30 mm), and median microscopic tumor diameter (as measured by histopathological examination) was 12 mm (range 6-29 mm). On CE-MRI, tumor margins were scored more spiculated compared to CE-CT (Table 1, Figure 2). Tumor shape was mainly scored as an irregular mass on CE-MRI and as a lobular mass on CE-CT.

Table 1 Patient and tumor characteristics Characteristic ValueAge (years) 48-70

median 61Side

left 7right 7

Histologyductal carcinoma 8ductulolobular carcinoma 5tubular carcinoma 1

Clinical tumor diameter (mm) 7-30median 15

Microscopic tumor diameter (mm) 6-29median 12

Tumour visualization score on CT (median)margin 2shape 3

Tumour visualization score on MRI (median)margin 3shape 4

Tumour visualization scores: Margin: 1 – smooth, 2 – irregular, or 3 – spiculatedShape: 1 – round, 2 – oval, 3 – lobular, 4 – irregular

Figure 2 Small peripheral branches in the transversal plane. (a) CE-MRI and (b) CE-CT.

102

Chapter 6

Interobserver variability and volumes

All observers delineated the GTV of all 14 patients by following the delineation in-structions. In Figure 3a and 3b, GTV delineations of the 4 observers are shown on both preoperative CE-CT and CE-MRI in one patient. To illustrate the comparison with the current standard CT delineations after lumpectomy, postoperative deline-ations of this patient are shown in Figure 3c as a clinical example. Preoperative tumor delineation resulted in a high median CI of the CTV, for both CT (0.80) and MRI (0.84)). However, the tumor was missed on CT in 2/14 patients (14%).

Figure 3 3D GTV delineations of 4 different observers in the transversal and sagittal plane in one patient on (A) preoperative CE-MRI, (B) preoperative CE-CT, and (C) clinical postoperative CT.

103


This resulted in wide ranges in CI on CT (range 0.00-0.93 for the CTV) compared to MRI (range 0.47-0.93). The first patient in which misdelineation occurred was a patient with multiple macrocalcifications in the breast, as seen on mammography. On CE-MRI all 4 observers contoured the tumor. On CE-CT a benign lesion was con-toured by 3 observers, resulting in a CI ranging from 0.00 to 0.52. The second patient had a tumor centrally located in the breast. On CE-MRI all observers contoured the tumor, while on CE-CT one observer contoured dense fibroglandular tissue resul-ting in a range in CI of 0.00-0.59. Outcomes of the analysis including misdelineati-ons are provided as Additional file 1. To only focus on differences in contouring the actual tumor and not on tumor de-tection, the 2 misdelineations were excluded from further IOV and volume analysis. Results of this analysis are shown in Table 2. The CI for the GTV was significantly higher on MRI (p<0.001) compared to CT. No difference in CI for the CTV was found (p=0.123). Delineated volumes were significantly larger on MRI for both the GTV and CTV (both p<0.001). There was no difference in dCOM between CT and MRI for both the GTV and CTV.

Table 2 Parameters of interobserver variability (misdelineations excluded from analysis) CT MRI

Median Range Median Range p-valueMean volume (cm3)

GTV 2.1 0.3 – 21.3 2.7 0.4 – 19.4 <0.001CTV 48.1 27.7 – 137.3 59.0 30.4 – 153.1 <0.001

Conformity indexGTV 0.56 0.11 – 0.83 0.61 0.37 – 0.78 <0.001CTV 0.82 0.39 – 0.93 0.84 0.47 – 0.93 0.123

Mean dCOM (mm) GTV 1.1 0.3 – 12.8 1.2 0.3 – 3.6 0.245CTV 1.4 0.3 – 13.6 1.8 0.1 – 4.8 0.836

CI Conformity Index, GTV gross tumor volume, CTV clinical target volume, dCOM center of mass distance.

Discussion

To our knowledge, this is the first study in which the feasibility of 3D CE-MRI of pa-tients in radiotherapy supine position using a wide bore MRI scanner has been de-monstrated. Different sequences of high resolution 3D CE and non-CE images were acquired with isotropic voxel sizes ≤ 1.2 mm. In the present study, target volume delineation before lumpectomy resulted in a high agreement and small treatment volumes among observers compared to stan-dard postoperative TB delineation as reported in literature (Table 3).

104

Chapter 6

Tabl

e 3

Stud

ies

rega

rdin

g th

e in

tero

bser

ver v

aria

bilit

y in

tum

ourb

ed, G

TV, C

TV a

nd P

TV d

elin

eatio

n af

ter b

reas

t-co

nser

ving

sur

gery

.

Num

ber o

f pa

tien

tsN

umbe

r of

obs

erve

rsSt

udie

d ta

rget

vo

lum

eVo

lum

e (c

c) ¶

Met

hod

Out

com

e ¶

a) C

TSt

ruik

man

s 20

05 4

185

TBM

ean

20 (6

.4-7

5.8)

CIpa

irsM

ean

0.56

(0.3

9-0.

74)

Pete

rsen

200

7530

3TB

Mea

n 48

.7 (1

0.3-

189.

3)CI

com

mon

dCO

MM

ean

0.61

(0.2

7-0.

84)

0.5-

1.1* (S

D 0

. 5-1

.8)

Hur

kman

s 20

096

104

TB40

CIco

mm

onM

ean

0.31

(ran

ge 0

.11

– 0.

52)

Cole

s 20

097

122

TB29

.0 (S

D 2

8.5)

18.7

(8-1

16)

dCO

MM

edia

n 3

(ran

ge 1

-9)

Jolic

oeur

201

1866

3TB

13.7

CIco

mm

onM

ean

0.66

Gie

zen

2012

915

4TB

27 (S

D 2

5)CI

gen

dCO

M0.

52 (S

D 0

.21)

4 (S

D 3

)

Van

Mou

rik 2

01010

813

CTV

n.a.

CIpa

irs

dCO

M0.

61# (r

ange

0.3

5 –

0.79

)7.

25 (2

.73

– 26

.87)

Boer

sma

2012

1126

5C

TV41

.9 (S

D 3

4)CI

pairs

dCO

M0.

36 (S

D 0

.21)

11 (S

D 8

)

Van

den

Ass

em 2

01212

193

CTV

37.6

CIpa

irs

dCO

MM

ean

0.59

Mea

n 5.

5

Land

is 2

00713

294

PTV

202

(65-

492)

CIpa

irs

dCO

M0.

76 (0

.52-

0.92

)2.

5 (0

. 7-1

1.5)

105


b) M

RIJo

licoe

ur 2

0118

663

TB10

.1CI

com

mon

Mea

n 0.

96

Gie

zen

2012

915

4TB

27 (S

D 2

6)CI

gen

dCO

M0.

32 (0

.25)

11(S

D 1

0)

c) C

T +a

dditi

onal

pr

eop

CT

Boer

sma

2012

1126

5C

TV36

(SD

31)

CIpa

irs

dCO

M0.

36 (S

D 0

.19)

10 (S

D 7

)

Van

den

Ass

em 2

01212

193

CTV

34.7

CIpa

irs

dCO

MM

ean

0.68

Mea

n 2.

9

d) P

reop

CT

Boer

sma

2012

11

+ vd

Lei

j 201

218

(sam

e da

tase

t)

265

GTV

CTV

0.99

37.5

CIpa

irs

dCO

MCI

pairs

dCO

M

0.45

(SD

0.2

2)4.

3 (S

D 7

.6)

Mea

n 0.

77

Mea

n 4.

4

This

stu

dy (

excl

udin

g 2

mis

delin

eatio

ns)

124

GTV

CTV

2.1

48.1

CIpa

irs

dCO

MCI

pairs

0.56

(0.1

1 –

0.83

)1.

1 (0

.3 –

12.

8)0.

82 (0

.39

– 0.

93)

1.4

(0.3

– 1

3.6)

e) P

reop

MRI

This

stu

dy (

excl

udin

g 2

mis

delin

eatio

ns)

124

GTV

CTV

2.7

59.0

CIpa

irs

dCO

MCI

pairs

dCO

M

0.61

0.3

7 –

0.78

)1.

2 (0

.3 –

3.6

)0.

84 (0

.47

– 0.

93)

1.8

(0.1

– 4

.8)

* Ra

nge

of th

e m

ean

dist

ance

from

the

COM

in x

-y a

nd z

-dire

ctio

n an

d ra

nge

in S

D#

CI v

alue

s of

one

obs

erve

r with

all

twel

ve o

bser

vers

wer

e av

erag

ed¶

Volu

me,

CI a

nd d

COM

: refl

ecte

d in

med

ian

(ran

ge) o

r mea

n ±

SD, u

nles

s st

ated

diff

eren

tly.

TB T

umor

Bed

vol

ume,

GTV

Gro

ss T

umor

Vol

ume,

CTV

Clin

ical

Tar

get V

olum

e, P

TV P

lann

ing

Targ

et V

olum

eCI

Con

form

ity In

dex,

met

hod

of a

naly

sis

(com

mon

, pai

rs, g

ener

al) a

s de

scrib

ed b

y Va

n Ko

uwen

hove

n et

al.

27

dCO

M D

ista

nce

from

the

Cent

re o

f Mas

s (m

m)

106

Chapter 6

Since the optimal imaging modality for preoperative target volume delineation was unknown, delineation was studied on both CT and MRI. MRI appeared to be essen-tial for tumor detection. For tumor delineation itself, the CI of the GTV was signifi-cantly higher on MRI and ranges on CT were wider. However, median differences were small (0.05) and may not be considered clinically relevant. For the CTV, no significant difference was found, since interobserver differences are blurred when expanding structures while uniformly excluding the skin and chest wall. However, more tumor spiculations and irregularities were observed on MRI due to its high spatial resolution (Table 1, Figure 2). This did not appear to result in a decreased GTV conformity on MRI compared to CT. The more irregular and spiculated tumor visualization on CE-MRI might have caused the significantly larger target volumes on MRI. Thin branches in the cranio-caudal or medio-lateral direction caused a relatively large volume expansion when applying a CTV margin. Even though large volumes can lead to increased toxicity and worse cosmesis, these effects do not outweigh the chances of not including peripheral tumor branches in the target volume, especially in APBI. However, despite the high level of consensus among observers, we acknowledge that no definitive statements can be made about the accuracy of the delineations, with the lack of pathologic validation of these branches as the gold standard. A pathology study must validate whether these branches are actual tumoral extensions or rather fibrotic strands or interstitial reactions, before standard inclusion of these branches in the preopera-tive GTV. With the implementation of high resolution imaging, the strict boundary between the GTV and the CTV with its microscopic spread might be fading. The appropriate preoperative CTV margin on MRI is therefore subject to debate and will also be further refined according to the future information about the appearance of local recurrences in the breast in the APBI studies [22]. In the last decade, several other attempts have been made to improve the current postoperative target volume delineation (Table 3b and 3c). Delineation on postope-rative MRI resulted in contradicting results [8, 9]. Jolicoeur et al. found an improved IOV and smaller volumes, while Giezen et al. found similar volumes with a decreased IOV. In two other studies, IOV was assessed on postoperative CT while preoperative CE-CT images in the same treatment position were provided [11, 12]. This resulted in an improved IOV in one of these studies. Preoperative delineation was studied on CE-CT by Boersma et al., resulting in a low IOV, which was in line with our study findings (Table 3d) [11]. Our reported findings on preoperative MRI-guided delineation resulted in high and stable conformity among observers (Table 3e). Furthermore, our preoperatively deli-neated GTVs were considerably smaller compared to postoperative volumes repor-ted in literature (Table 3a). CTVs were larger, although preoperative volumes would have less outliers since there is no seroma formation. The larger CTVs in our study were caused by a uniform 1.5 cm volume expansion, while the postoperative results in Table 3a reflect ‘boost’ volumes, in which the microscopic resection margin is of-

107


ten subtracted from this margin. PTVs were not compared in this study, since PTV margins are institution-dependent due to the method of position verification being practiced. These PTV margins might even be changed or improved in a preopera-tive setting, due to less volume distortions. Overall, the high conformity index in combination with the small and stable volumes in this study, imply that a future neoadjuvant irradiation would be more accurate and lead to less toxicity.When comparing our results to published data in Table 3, we have to be aware of the different methods used in the other studies. For instance, the method of CI cal-culation, observer backgrounds and multi-centricity of a study can influence the observed results regarding IOV [29]. Interobserver studies often use small patient groups due to the high workload (Table 3). Furthermore, it has to be noted that the CI is volume dependent. The smaller the volume being studied, the more the CI is influenced by small interobserver differences. This especially accounts for our small preoperative GTVs, but also emphasizes that when comparing different studies, the studied volume (i.e. GTV, TB, CTV or PTV) must be taken into account. Can we, from the results of this study, conclude that MRI superior to CT for preope-rative tumor delineation? In this study, MRI was essential for tumor detection. Ho-wever, alternatives for tumor detection can be considered, e.g. optimizing CT pa-rameters like contrast-enhancement of the tumor, or clearly marking the tumor by fiducials. This might be easier to implement, less time consuming and less expen-sive. When using preoperative imaging for a preoperative irradiation or ablative in-terventional techniques, treating another area but the GTV would be unacceptable. Furthermore, more detail could be visualized by MRI, which could contribute to an accurate target definition. Therefore, in our future studies, CE-MRI in radiotherapy supine position will be used in addition to CT as CT is required for treatment plan-ning. In our institute, an MRI linear accelerator is being developed in collaboration with Philips Medical Systems (Best, The Netherlands) and Elekta (Stockholm, Swe-den) [30]. This system can provide online tumor tracking using MRI during radiothe-rapy, which makes it possible to adapt the plan to the actual tumor position. The results of our study show that a preoperative irradiation of breast tumors could be beneficial in terms of delineation consistency and treatment volumes. There would be more certainty that the correct target is delineated when the tumor is in situ. Moreover, target volumes would probably be more stable in the absence of seroma formation, and not being subject to seroma shrinkage [17, 31]. The advantages of preoperative CE-MRI for treatment planning will be further studied with respect to the dosimetric consequences [32]. CE-MRI in supine position could also be used for other purposes. For instance, it might provide additional information to improve the consistency in target volume definition in standard postoperative CT-guided delineation [11]. Furthermore, it might aid tumor localization for breast conserving surgery or interventional procedures [33].

108

Chapter 6

In conclusion, preoperative target volume delineation resulted in small treatment volumes with a high consistency among observers. MRI appeared to be necessary for tumor detection and the visualization of irregularities and spiculations. Regar-ding delineation of the tumor itself, no clinically relevant differences in interobser-ver variability among imaging modalities were observed. These results will be used to study the potential for a future MRI-guided and neoadjuvant radiotherapy.

Additional file 1. Parameters of interobserver variability including misdelineations on CT, which resulted in CIs of 0.00 and high dCOMs. In the manuscript, these outliers were excluded from analysis and shown in Table 2.

CT MRIMedian Range Median Range p-value

Mean volume (cm3)GTV 2.2 0.3 – 21.3 2.6 0.4 – 21.3 0.009CTV 48.8 27.7 – 168.4 59.0 30.4 – 153.1 <0.001

Conformity indexGTV 0.54 0.00 – 0.83 0.60 0.37 – 0.78 <0.001CTV 0.80 0.00 – 0.93 0.84 0.47 – 0.93 0.003

Mean dCOM (mm) GTV 1.3 0.3 – 77.9 1.2 0.2 – 3.6 0.004CTV 1.6 0.3 – 77.6 1.7 0.1 – 4.8 0.172

CI Conformity Index, GTV gross tumor volume, CTV clinical target volume, dCOM center of mass distance.

109


References




378:1707-1716.




3. Theberge V, Whelan T, Shaitelman SF, Vicini FA. Altered fractionation: rationale and justification for

whole and partial breast hypofractionated radiotherapy. Semin Radiat Oncol 2011: 21:55-65.






Biol Phys 2007: 69:41-48.



2009: 90:60-65.









Biol Phys 2012: 82:1332-1340.




11. Boersma LJ, Janssen T, Elkhuizen PH et. al. Reducing interobserver variation of boost-CTV deline-

ation in breast conserving radiation therapy using a pre-operative CT and delineation guidelines.


12. van den Assem MB, Visser J, Zonderland HM, van Tienhoven G, Crama KF, Bijker N. 459 Pre-operative

CT Scan in Breast Conserving Therapy for Determination of the Boost Volume for Radiotherapy. Eur

J Cancer 2012: 48, Supplement 1:S180-S181.



110

Chapter 6







16. Benda RK, Yasuda G, Sethi A, Gabram SG, Hinerman RW, Mendenhall NP. Breast boost: are we mis-

sing the target?. Cancer 2003: 97:905-909.

17. den Hartogh MD, van Asselen B, Monninkhof EM et. al. Excised and irradiated volumes in relation to

the tumor size in breast-conserving therapy. Breast Cancer Res Treat 2011: 129:857-865.



Clin Oncol 2012: 36:232-238.

19. Pignol JP, Rakovitch E, Keller BM, Sankreacha R, Chartier C. Tolerance and acceptance results of a

palladium-103 permanent breast seed implant Phase I/II study. Int J Radiat Oncol Biol Phys 2009:

73:1482-1488.

20. van der Leij F, Elkhuizen PHM, Janssen TM et. al. External beam partial breast irradiation: Difference

in pre- and postoperative target volume delineation. Radiotherapy and Oncology 2012: 103, Sup-

plement 1:S246.



22. Kim L, Shaitelman S. What is the target volume for preoperative accelerated partial breast irradiation

(APBI)? In regards to Nichols et al. (Int J Radiat Oncol Biol Phys 2010;77:197-202) and Palta et al. (Int J

Radiat Oncol Biol Phys 2010, in Press). Int J Radiat Oncol Biol Phys 2011: 80:314-5; author reply 315-6.

23. Bondiau PY, Courdi A, Bahadoran P et. al. Phase 1 clinical trial of stereotactic body radiation therapy

concomitant with neoadjuvant chemotherapy for breast cancer. Int J Radiat Oncol Biol Phys 2013:

85:1193-1199.



97:225-232.



26. Kuroki-Suzuki S, Kuroki Y, Ishikawa T, Takeo H, Moriyama N. Diagnosis of breast cancer with multi-

detector computed tomography: analysis of optimal delay time after contrast media injection. Clin

Imaging 2010: 34:14-19.

27. Ikeda DM, Hylton NM, Kinkel K et. al. Development, standardization, and testing of a lexicon for

reporting contrast-enhanced breast magnetic resonance imaging studies. J Magn Reson Imaging

2001: 13:889-895.



111


29. Kouwenhoven E, Giezen M, Struikmans H. Measuring the similarity of target volume delineations

independent of the number of observers. Phys Med Biol 2009: 54:2863-2873.

30. Lagendijk JJ, Raaymakers BW, Raaijmakers AJ et. al. MRI/linac integration. Radiother Oncol 2008:

86:25-29.

31. Tersteeg RJ, Roesink JM, Albregts M, Warlam-Rodenhuis CC, van Asselen B. Changes in excision

cavity volume: prediction of the reduction in absolute volume during breast irradiation. Int J Radiat

Oncol Biol Phys 2009: 74:1181-1185.

32. van Heijst TC, den Hartogh MD, Lagendijk JJ, van den Bongard HJ, van Asselen B. MR-guided breast

radiotherapy: feasibility and magnetic-field impact on skin dose. Phys Med Biol 2013: 58:5917-5930.

33. Alderliesten T, Loo C, Paape A, Muller S, Rutgers E, Peeters MJ, Gilhuijs K. On the feasibility of MRI-gui-

ded navigation to demarcate breast cancer for breast-conserving surgery. Med Phys 2010: 37:2617-

2626.

Chapter 6

Chapter 7MRI-guided single

fraction ablative radiotherapy for early-stage breast cancer:

A brachytherapy versus VMAT planning study

K.R. Charaghvandi, M.D. den Hartogh, A.M.L.N. van Ommen,J.H.W. de Vries, V. Scholten, M.A. Moerland, M.E.P. Philippens,

M. van Vulpen, B. van Asselen, BH.J.G.D. van den Bongard

Manuscript in preparation

114

Chapter 7

Abstract

Purpose

To evaluate the dosimetric potential for MRI-guided single fraction ablative radio-therapy for early-stage breast cancer by comparing volumetric modulated arc ther-apy (VMAT) with an interstitial multicatheter brachytherapy (IMB) approach.

Methods

The tumors of 20 patients with early-stage breast cancer were delineated on a pre-operative contrast-enhanced planning CT-scan, co-registered with a contrast en-hanced MRI, both in radiotherapy supine position. A 15 Gy dose was prescribed to the planned target volume of the clinical target volume (PTVCTV), and 20 Gy in-tegrated boost to the PTV of the gross tumor volume (PTVGTV). Treatment plans for IMB and VMAT were optimized for adequate target volume coverage and minimal organs at risk (OAR) dose.

Results

The median PTVGTV/CTV receiving at least 95% of the prescribed dose was ≥ 99% with both techniques. The median volume PTVCTV that received 95% of the prescribed PTVGTV dose was 80.1% and 5.6% with IMB and VMAT, respectively. OAR doses were comparable with both techniques.

Conclusion

MRI-guided single fraction radiotherapy with an integrated ablative boost to the GTV is dosimetrically feasible using VMAT. IMB is less suitable for clinical implemen-tation due to PTVCTV overdosage.

115

MRI-guided single fraction ablative radiotherapy

Introduction

Radiotherapy planning has evolved during the past decade due to image-guided treatment. At our department, MRI-guidance is applied for high-dose rate brac-hytherapy in cervical cancer patients due to MRI availability in the brachytherapy room. Furthermore, the MRI-linac, a hybrid system consisting of a 8 MV accelera-tor and an integrated 1.5 Tesla MRI scanner, is currently being investigated at the Utrecht University Medical Center (UMCU) for several tumor sites [1]. For the field of breast oncology, we focus on the use of MRI-guided radiotherapy developments for early-stage breast cancer. The current treatment standard in early-stage breast cancer is breast-conserving therapy (BCT), consisting of breast-conserving surgery (BCS) followed by whole breast irradiation (WBI) with or without a boost [2, 3]. De-spite its proven effectiveness [4], the protracted radiotherapy (RT) duration ranging from 3 to 7 weeks with a hypofractionated or conventional regimen, can provide a substantial treatment burden to selected patients. Post-operative accelerated partial breast irradiation (APBI) to the tumor bed offers a promising alternative to WBI in low-risk breast cancer patients due to its potential to reduce BCT duration. Furthermore, Palta et al have shown that APBI in preoperative setting can result in a treatment volume reduction that could accomplish further treatment acceleration [5]. When aiming for an alternative to BCT with minimal treatment burden, the role of RT could further be extended to a radiosurgical approach [6], comparable to a stereotactic treatment for certain lung or cerebral neoplasms. In case of a radiosur-gical approach, accurate tumor localization is critical. Since tumor size on MRI is highly correlated to microscopic tumor size, MRI-guidance is required in addition to planning CT-scan, in order to adequately identify the tumor extent [7]. The purpose of this study was to evaluate the dosimetric potential for one MRI-guided single fraction ablative RT for early-stage breast cancer. We conducted a planning study by comparing a volumetric modulated arc therapy (VMAT) versus interstitial multi-catheter brachytherapy (IMB) approach.

Materials and methods

Patient characteristics

This study included patients from the pre-existing NTR3198 study, which was ap-proved by our institutional review board and results have been previously reported [8]. Patients eligible for inclusion were diagnosed with early-stage breast cancer and were scheduled for breast-conserving surgery and whole breast irradiation. Baseli-ne characteristics of the 20 eligible patients are shown in Table 1. Patients under-went a contrast-enhanced (CE) CT and CE-MRI in supine RT treatment position, with the arms in abduction above the head. Details on patient positioning and preope-

116

Chapter 7

rative imaging parameters were previously reported [8]. Patients with tumors up to 30 mm were eligible for the current planning study, regardless of the distance from the tumor to the chest wall or skin.

Target definition and organs at risk

All delineations (Figure 1) were performed using delineation software developed at our department Volumetool® [9]. Gross tumor volumes (GTVs) and organs at risk (OAR) were delineated by an experienced breast radiation oncologist on CE-CT, co-registered with CE-MRI (Figure 1). The GTV was uniformly expanded by 2 cm to create a clinical target volume (CTV), thereby excluding the skin and chest wall. For VMAT, both GTV and CTV were uniformly expanded by 3 mm to obtain the planning target volumes (PTV) PTVGTV and PTVCTV. For IMB planning, the PTVGTV and PTVCTV were equal to the GTV and CTV, respectively. The skin was excluded from the PTV. The ipsilateral breast was contoured using the additional CT/MRI compatible de-marcation wire as placed by clinical examination. The lungs were automatically con-toured. The skin was defined as the area within the first 5 mm under the ipsilateral breast surface, extended with a uniform 3.5 cm margin from the breast borders. The chest wall was delineated as one structure thereby including the bony structures (i.e. ribs, sternum, scapula) and muscles (i.e. intercostal, pectoral and (a part of the) rotator-cuff). The heart contour commenced below the pulmonary trunk bifurcati-on and the lateral borders included the pericard [10].

Treatment plan acquisition

The preoperative planning CT-images and delineations were exported from Volu-metool® into the planning software. Two radiotherapy dose levels were concomi-tantly prescribed in one single fraction: 15 Gy to the PTVCTV and 20 Gy to the PTVGTV.

Table 1 Tumor baseline characteristics

Characteristic Value Interquartile rangeSide

left 9 (45%)right 11 (55%)

Location in breastlateral 13 (65%)medial 4 (20%)central 3 (15%)Median clinical diameter (mm) 14.0 8.5-16.0

Median distance (mm)to skin 10 6.3-19.5to chest wall 9 6.0-22.3

117


Figure 1 Planning target volumes & organs at risk delinea-

The 20 Gy single dose is equivalent to a 73.7 Gy dose in 2 Gy fractions (EQD2, α/β 4.7), which results in a 100% 5 year tumor control probability for cT1N0 tumors [11]. The single 15 Gy dose corresponds to an EQD2 of 44.1Gy (α/β 4.7), which is equiv-alent to the standard hypofractionated schedule at our institute consisting of 16 fractions of 2.66 Gy. For IMB treatment planning, predominantly double and triple-plane implants were simulated into the PTV. The implants were inserted alternating in a triangular con-figuration. IMB plans were generated using inverse planning with the OncentraB-rachy 4.3® software (Elekta Ltd). In order to achieve a more conformal coverage of the PTV, some catheters were subsequently displaced. The spacing between the needles was below 20 mm. The VMAT plans were created using two partial arcs (clockwise and counter wise) with a total angle of 210 to 240°. Starting at an angle of 180°, VMAT plans were generated using Monaco 3.2® software (Elekta Ltd).The plans were optimized for adequate target volume coverage whereas an as low as possible dose was aimed for the OARs. Adequate target volume coverage was defined as 99% or more of the planning target volume receiving at least 95% of the prescribed dose, thus PTVGTV receiving at least 19 Gy and PTVCTV receiving at least 14.3 Gy.

118

Chapter 7

Dose constraints in OARs were set to minimize the volume of normal tissue receiv-ing the prescription dose without compromising target volume coverage. Dose constraints for the lung and heart were converted from the QUANTEC recommen-dations to the corresponding single dose equivalent in 2 Gy fractions using an α/β of 3 Gy-1 [12]. The QUANTEC recommendation for both lungs, a mean lung dose < 7 Gy (physical dose) was converted to a mean lung dose < 3.6 Gy in a single dose. In this study, a constraint of the mean ipsilateral lung dose < 3.6 Gy was maintained. In concordance to the standard of care at our department, the heart constraint was tightened to V5Gy < 10% (physical dose), and was added to the V25Gy < 10% (physical dose) QUANTEC recommendation. This implied V2.8Gy < 10% of the heart, in a sin-gle dose delivery. The chest wall objective was extrapolated from studies on ste-reotactic lung RT and associated chest wall pain [13]. The chest wall objective was formulated as D20cc < 16.3 Gy. Regarding an acceptable skin dose, there is limited availability on single dose constraints. Since 15 Gy was prescribed to the CTV and no restrictions on tumor location towards skin were maintained, a relative skin ob-jective D1cc< 16 Gy (single dose) was formulated.

Plan evaluation

The plans were evaluated with respect to target volume coverage, high dose volu-mes, maximal dose and OAR dose. For the target volumes, high dose volume was defined as the percentage volume of the PTVCTV (with the exclusion of PTVGTV) re-ceiving at least 19 Gy. The percentage volume of the ipsilateral breast receiving 15 Gy outside the PTVCTV was also considered to be relevant for high dose volumes evaluation. In addition, the maximal doses for the PTVGTV/CTV were assessed. For the OARs, the mean ipsilateral lung dose, the V2.8Gy of the heart and the D1cc in the chest wall, skin and contralateral breast were evaluated. Moreover, the median number of required catheters in the IMB plans was calculated.

Data analysis

For each patient and each planning modality, median values on percentage PTV coverage, high dose volumes and OAR doses were evaluated. Non-parametric tes-ting by using a Wilcoxon signed-rank test was performed to compare the paired variables between the two treatment modalities by using IBM SPSS Statistics 20 (Chicago, IL, USA) with a significance level of α below 0.05.

Results

The median breast volume was 867.8 cc. A comparison of the baseline volumetric characteristics of the ipsilateral breast and PTVs for VMAT and IMB plans are pre-sented in Table 2. The median PTVCTV was 1.3 times larger in VMAT plans by compa-

119


rison to IMB plans. Table 3 illustrates the dosimetric parameters for the VMAT and IMB techniques. Figure 4 illustrates a representative dose-volume histogram for the VMAT and IMB plans in one patient.

Table 2 Baseline volumetric characteristics

Characteristic(n=20)

IMB VMAT

median range median rangeVolume PTVGTV (cc) 1.8 0.2-12.7 6.1 1.4 -24.3Volume PTVCTV (cc) 74.8 44.4-147.4 100.9 61.6-183.2Ratio PTVCTV to ipsilateral breast (%) 6.9 4.4-21.0 9.2 5.4-26.0

Table 3 Dosimetric parameters for IMB and VMAT treatment plans for a single dose ablative radio-therapy, with a 20 Gy dose prescription to GTV and a 15 Gy dose prescription to the CTV.

IMB VMATmedian range median range

PTV evaluationCoverage

V19Gy PTVGTV (%) 100 98.7-100 99.7 98.9-100V14.3Gy PTVCTV (%) 99.2 89.8-100 99.2 98.8-100Number of catheters 9 7-14

High dose volumesV19Gy PTVCTV without PTVGTV (%) 80.5 73.1-88.9 4.4 2.6-9.7V15Gy ipsilateral breast without PTVCTV (%) 2.0 0.4-9.0 0.5 0-1.1.1D1cc PTVGTV (Gy)* 60.0 29.7-64.0 21.8 20.7-23.7D1cc PTVCTV (Gy) 119.0 82.8-198.6 20.2 19.6-21.6

OAR evaluationD1cc contralateral breast (%) 0 0-0.8 0.9 0.2-3.2D1cc skin (Gy) 15.8 4.5-21.6 15.2 10.2-18.3D1cc chest wall (Gy) 14.6 4.1-21.4 15.5 6.1-19.7Mean ipsilateral lung dose (Gy) 1 0-1.7 1.3 0.4-2.5V2.8Gy heart (%) 0 0-1.3 0.3 0-5.9

VMAT volumetric modulated arc therapy, IMB interstitial multicatheter brachytherapy, V19 Gy percentage of certain volume receiving at least 95% of the prescribed dose (19 Gy), V14.3Gy percentage of certain volume receiving at least 95% of the prescribed dose (14.3 Gy) D1cc the maximum 1cc dose in a PTV or OAR, V2.8 Gy heart percentage volume of the heart receiving 2.8 Gy, GTV gross tumor volume delineated on registered MRI-CT, CTV GTV+2cm, thereby excluding skin and chest wall, PTV planning target volume, PTVGTV GTV+3 mm, thereby excluding skin PTVCTV CTV+3 mm, thereby excluding skin, OAR organs at risk, * patients with a PTVGTV below 1 cc in IMB plans (n=7) were excluded from the analysis

120

Chapter 7

Target volume coverage

The median PTVGTV and PTVCTV receiving at least 95% of the prescribed dose was ≥99% for both treatment modalities. Figure 2 illustrates representative isodoses and differences between an IMB versus VMAT treatment plan in one patient. Overall, in the 20 IMB plans a median number of 9 catheters (range 7-14) was required for adequate PTVCTV, and PTVGTV coverage. However, in 3 IMB plans, no adequate co-verage was achievable due to an extremely lateral tumor location (n=2) or a tumor adjacent to the chest wall (n=1), not adequately accessible for interstitial catheter placement. For these IMB plans, the maximum achievable PTVCTV coverage was 89.8 %, 96.4% and 96.7%, respectively.

Figure 3 Examples of extent of high dose volu-mes in IMB plan

Figure 2 Representative dose distribution IMB (A) versus VMAT (B)

121


High dose volumes

Figure 3 illustrates representative high dose volume distribution in an IMB treat-ment plan. The PTVCTV (with the exclusion of PTVGTV), receiving 95% of the prescribed PTVGTV dose (V19 Gy) differed substantially, i.e. 4.4% and 80.5% for the VMAT and IMB plans, respectively. The percentage volume of the ipsilateral breast receiving 15 Gy outside the PTVCTV was 0.5% and 2% in VMAT and IMB plans, respectively. The median PTVGTV was below 1cc in 7 IMB plans. Therefore, the median D1cc PTVGTV was evaluated in 13 patients, with 21.8 Gy and 60.0 Gy for VMAT and IMB plans, respec-tively. The median D0.1cc dose was 36.5 Gy for the IMB plans.

Figure 4 Representative dose volume histograms for a VMAT and IMB plan from one pa-

Dose to organs at risk

The median OAR dose values were compared. The heart dose was low and did not differ between the treatment modalities. The mean ipsilateral lung dose was comparable. The predefined chest wall objective D20cc < 16.3 Gy was achieved in all plans. The median D1cc to the skin was 15.2 Gy for the VMAT plans and 15.8 Gy for the IMB plans. The median D1cc to the contralateral breast was 0.9 Gy with VMAT and 0 Gy with IMB.

122

Chapter 7

Discussion

A comparative planning study on single fraction treatment was performed to eva-luate the dosimetric feasibility of MR-guided ablative radiotherapy for early-stage breast cancer. The VMAT treatment approach resulted in adequate target volume coverage for all cases. With IMB, adequate coverage was hindered in some cases with a lateral or adjacent to chest wall tumor location . In addition, Dmax and high dose volumes in both target volumes and ipsilateral breast were considerably lower with VMAT compared to IMB. The dose to the OAR was comparable between the two treatment approaches. To the best of our knowledge, no other comparative VMAT (or external beam RT) versus IMB planning studies have been performed to investigate a single fraction ablative RT approach. However, preoperative single fraction RT has previously been evaluated as feasible by Palta et al using a three-dimensional conformal radiation therapy (3D-CRT) planning technique. In 17 virtual plans, 15 Gy dose was delivered to T1 tumors [5]. The 15 Gy prescription dose was lower and the CTV 1.5 cm margin was less extensive compared to our planning study (20 Gy and 2 cm, respectively). In addition, Palta et al defined the skin as the first 3 mm of the breast surface whe-reas in our study 5 mm was employed. Furthermore, tumors were located at least 1 cm from the skin while our study had no restrictions regarding tumor location. Compared to our outcomes, the dosimetric results of Palta et al with respect to OAR dose were more favorable, in particular the skin dose (D1cc 9 Gy versus D1cc 15.2 Gy). Furthermore, Horton et al performed a phase I escalation protocol on radiosurgery IMRT based treatment. In this study breast cancer patients with T1 tumors were irradiated to a single dose of 15, 18 or 21 Gy followed by breast-conserving sur-gery at two weeks after RT [6]. During a median follow-up of 6.5 months, no dose limiting toxicity was observed, along with good or excellent cosmetic outcome in all cases of preoperative partial breast irradiation. These encouraging preoperative clinical results are not in accordance with the observations by Pinnarò et al who eva-luated single fraction partial breast 3D-CRT in post-operative setting. At a median follow-up of 3 years, a dose of 21 Gy significantly increased the treatment related toxicity and resulted in fair and poor cosmesis in 36% and 5% of the patients, res-pectively [14]. These conflicting toxicity and cosmetic results between the study by Horton et al en Pinnarò et al could be attributed to a substantial reduction in ipsila-teral breast tissue dose observed with preoperative versus post-operative APBI [5]. In the post-operative APBI study, grade 2 fibrosis or higher was associated with ≥ 66 cc of the ipsilateral breast receiving more than 21 Gy [14]. Contrary, in our study, no cases of V21Gy > 66 cc of the breast volume were observed in the VMAT plans. This is most likely due to our two-dose level RT approach including a lower, 15 Gy prescribed dose to the PTVCTV. A V21Gy > 66 cc was noted in 6 IMB plans (30.0% of the

123


plans). In addition, Pinnarò et al observed that impaired cosmesis was correlated with a mean ipsilateral breast dose above 9 Gy. In our study, no VMAT plans acquired a mean ipsilateral breast dose above 9 Gy whereas this was the case in 2 IMB plans (10.0%). In the current study, the PTVCTV of the IMB plans had a large volume and numerous catheters were required for adequate coverage due to the extensive 2 cm margin from the GTV. A median of 80.5% of the PTVCTV received the prescribed dose for the PTVGTV, thus PTVCTV overdosage was observed. This can be attributed to a more complex catheter configuration for the two dose level RT approach. In VMAT plans substantially less overdosage was observed, with a median of 4.4% for the PTVCTV. However, delivering a single dose with IMB implied a considerable treatment vo-lume reduction of 25.9 % by comparison with VMAT plans. Nevertheless, since the substantial PTVCTV overdosage associated with IMB can result in excessive toxicity, e.g. fibrosis or fat necrosis and decreased cosmetic results, we prefer the VMAT tech-nique for implementing the single dose ablative RT in the clinical practice. An important aspect on the clinical applicability of the single dose ablative RT is local tumor control. The importance of managing high and low-risk patient crite-ria is illustrated in the randomized controlled trials on intraoperative (IORT) single dose APBI versus WBI. The ELIOT trial evaluating post-lumpectomy single 21 Gy dose electron IORT showed an acceptable 5-year local recurrence (LR) of 1.9 % in good candidates but an 7.4% LR in possible candidates and 7.7% in patients with a contraindication for APBI according to Groupe Européen de Curie thérapie Eu-ropean Society for Therapeutic Radiology and Oncology (GEC-ESTRO) [15]. When considering the delivery of APBI in low-risk patients, an utterly precise imaging me-thod is crucial for targeting high-risk (breast or tumor) tissue only and minimizing the treatment volume in favor of toxicity and cosmesis. MRI-guidance in addition to mammography findings and histopathology features has the potential to accurate-ly identify breast cancers of limited extent [7] and is therefore suitable for the de-livery of the ablative treatment in clinical practice. Furthermore, in order to ensure accurate treatment delivery, the GTV was extensively expanded in our study, with 2 cm for CTV formation. A drawback of this study regards the reproducibility of our planning study observa-tions in the clinical practice, especially for IMB. The static CT-based catheter tracks in this IMB planning study tracks can differ from implants achievable in daily practice and, for example, adequate coverage could still be feasible for extremely laterally lo-cated tumors. Nevertheless, PTVCTV overdosage due to two dose level RT approach, remains the main limiting factor for IMB implementation of single fraction ablative RT. Furthermore, in this comparative study, differences in dose calculation algorit-hms between VMAT and IMB planning software are possible. In addition, we obser-ved minor differences in OAR dose with VMAT and IMB plans despite statistically

124

Chapter 7

significant differences evaluated with the Wilcoxon signed rank test. Therefore only the substantial dosimetric difference such as PTVCTV overdosage (80.5% with IMB versus 4.4% with VMAT) was considered clinically relevant. Finally, for the clinical implementation of a single fraction ablative RT, there is a lack of single dose OAR constraints. The linear-quadratic model is not applicable for frac-tion sizes as high as 15-20 Gy and the conversion to an acceptable constraint for OARs is therefore limited. In particular for the skin, the permitted dose could not be easily determined given the limited amount of studies. In one post-operative single fraction APBI study, late toxicity such as grade 1 or 2 telangiectasia was observed in 26.5% of the patients and was associated with a previous acute erythema episode [14]. Acute erythema correlated to the mean skin dose of 5.4 Gy. We therefore esti-mate that high doses to the skin are reasonably tolerated. However a comparison with the mean skin dose in our study is impracticable due to differences in skin volume definition. We observed low median skin doses of 2.7 Gy and 2.2 Gy for IMB and VMAT plans and therefore anticipate acceptable skin toxicity. Nevertheless, OAR toxicity has to be confirmed in further clinical studies. In conclusion, we found that MRI-guided single fraction radiotherapy with an inte-grated ablative boost to the GTV is dosimetrically feasible using VMAT. IMB is less suitable for clinical implementation due to PTVCTV overdosage. Within the develop-ment of the MRI-linear accelerator, we will initiate a feasibility study that will focus on MRI-guided single fraction ablative RT using VMAT in selected early-stage breast cancer patient with low-risk on local recurrence.

The work described in this chapter was supported by a research grant from Pink Ribbon.

125


References

1. Lagendijk JJ, Raaymakers BW, van Vulpen M. The Magnetic Resonance Imaging-Linac System. Se-

min Radiat Oncol 2014: 24:207-209.

2. Litiere S, Werutsky G, Fentiman IS et. al. Breast conserving therapy versus mastectomy for stage I-II

breast cancer: 20 year follow-up of the EORTC 10801 phase 3 randomised trial. Lancet Oncol 2012:

13:412-419.







378:1707-1716.






7. Schmitz AC, Pengel KE, Loo CE et. al. Pre-treatment imaging and pathology characteristics of invasi-

ve breast cancers of limited extent: potential relevance for MRI-guided localized therapy. Radiother

Oncol 2012: 104:11-18.


volume delineation in breast-conserving therapy. Radiat Oncol 2014: 9:63-717X-9-63.



10. Feng M, Moran JM, Koelling T et. al. Development and validation of a heart atlas to study cardiac ex-

posure to radiation following treatment for breast cancer. Int J Radiat Oncol Biol Phys 2011: 79:10-18.

11. Van Limbergen E, Van der Schueren E, Van den Bogaert W, Van Wing J. Local control of operable

breast cancer after radiotherapy alone. Eur J Cancer 1990: 26:674-679.

12. Marks LB, Yorke ED, Jackson A et. al. Use of normal tissue complication probability models in the

clinic. Int J Radiat Oncol Biol Phys 2010: 76:S10-9.

13. Creach KM, El Naqa I, Bradley JD et. al. Dosimetric predictors of chest wall pain after lung stereotac-

tic body radiotherapy. Radiother Oncol 2012: 104:23-27.

14. Pinnaro P, Arcangeli S, Giordano C et. al. Toxicity and cosmesis outcomes after single fraction partial

breast irradiation in early stage breast cancer. Radiat Oncol 2011: 6:155-717X-6-155.

15. Leonardi MC, Maisonneuve P, Mastropasqua MG et. al. Accelerated partial breast irradiation with

intraoperative electrons: using GEC-ESTRO recommendations as guidance for patient selection.


Chapter 8Summary and

general discussion

128

Chapter 8

Summary and general discussion

Treatment accuracy in radiotherapy can be defined as treating tumor cells with op-timal sparing of the organs at risk. In clinical practice this definition can be trans-lated into the objective of optimal tumor control and limited treatment-related toxicity. Radiotherapy target definition is most commonly performed on computed tomography (CT) imaging. However, magnetic resonance imaging (MRI) offers the advantages of high-contrast tissue characterization, response monitoring and mo-tion analysis. Combining these advantages of MRI during radiation treatment would considerably increase treatment accuracy by decreasing uncertainties concerning target position and shape. Therefore, the Utrecht MRI-linac design features an 8 MV linear accelerator (Elekta AB, Stockholm, Sweden) mounted on a circular, continuo-usly rotating gantry around a 1.5 T cylindrical MRI scanner with a split gradient coil (Philips, Best, The Netherlands).

The MRI-linac has the potential to improve radiation treatment accuracy for various tumor sites. Breast cancer is the most common malignancy among women world-wide. In the western world breast cancer is often detected at an early stage, and local treatment for most women consists of breast-conserving surgery followed by whole breast irradiation. However, in whole breast irradiation, treatment burden can be substantial due to the multiple treatment sessions. For this reason, accelera-ted partial breast irradiation (APBI) is extensively studied worldwide. In APBI, only the tumor bed is irradiated since most breast cancer recurrences occur in this area. Since whole breast irradiation is omitted in APBI, accurate tumor bed definition is essential in order to prevent disease from recurring locally. Furthermore, other fac-tors, such as large treatment volumes, can negatively influence treatment toxicity and cosmetic outcomes [1-4]. Considering the previously described advantages of MRI, an MRI-guided treatment could potentially improve breast-conserving treat-ment by improving target definition and thereby possibly reducing treatment vo-lumes.

Therefore, this thesis addresses the potential for an MRI-guided radiation treatment in early-stage breast-cancer patients.

Since treatment volumes are known to affect treatment-related toxicity and cos-metic results, treatment volumes in the current breast-conserving treatment were evaluated (Chapter 2). Subsequently, to establish whether a treatment by MRI-linac would be feasible and beneficial, several topics were explored. Firstly, the effect of the magnetic field on the skin dose for both whole breast and partial breast irradi-ation was quantified (Chapter 3). Secondly, an MRI protocol in radiotherapy supine

129


position was developed (Chapter 4). Thirdly, the value of MRI for tumor bed defini-tion was investigated in addition to standard planning CT (Chapter 4) or as a single modality (Chapter 5). Fourthly, pre-lumpectomy tumor delineation was studied on dynamic contrast enhanced CT and MRI (Chapter 6). Finally, the feasibility of an MRI-guided single ablative dose to the tumor was assessed (Chapter 7).

Treatment volumes

For boost radiotherapy and especially for APBI, an accurate definition of the tumor bed is important. In breast tumor bed definition, the balance between including the tissue at risk with microscopic spread and excluding normal tissue, e.g. fibroglandu-lar breast tissue, lung and heart, is delicate. While treating a smaller volume may put the patient at increased risk of local recurrence, irradiating a larger volume than ne-cessary may put the patient at risk of increased treatment-related toxicity, including fibrosis, pain and reduced cosmetic outcome [1-4]. This volume effect might be of particular importance in APBI in which a higher dose per fraction is delivered. Fu-rthermore, large volumes can cause low-risk patients aiming for APBI to be ineligible for this treatment due to the inability to meet the dose-volume constraints. Recent-ly published results of the Canadian RAPID trial, in which patients were randomized between whole breast irradiation (42.5 Gy in 16 or 50 Gy in 25 daily fractions) and external beam APBI (38.5 Gy in 10 fractions, twice daily), showed increased rates of adverse cosmesis in patients treated with APBI at a median follow-up of 3 years [5]. Results of various single-arm APBI studies show conflicting results [6-11]. These adverse cosmetic results might be related to the treated volumes. Excision volume and the volume of the irradiated tumor bed region both influence treatment-related toxicity and cosmetic results in patients treated with breast-con-serving therapy [1-4]. It would be expected that these treatment volumes would somehow be related to the size of the original tumor. Understanding this relation is important, since it can show us whether surgery or radiotherapy techniques should be further improved. Therefore, in Chapter 2, the relation between the microscopic tumor size and the volumes treated by surgery and boost radiotherapy was deter-mined in 186 early-stage breast cancer patients treated with breast-conserving the-rapy. Both surgical and radiotherapy treatment volumes did not relate to the tumor diameter. Different amounts of breast tissue were excised for similar tumor sizes. These results illustrate that it is not feasible to excise a tumor with fixed margins of healthy breast tissue. Moreover, it was shown that the relation between tumor size and excision volume was stronger related to tumor characteristics and surgi-cal techniques. A slightly stronger association between microscopic tumor size was observed in palpable tumors and purely ductal or lobular carcinomas. The associ-ation between the excision specimen and the radiotherapy target volume further

130

Chapter 8

decreased when the surgeon performed full-thickness closure of the excision cavity walls. Furthermore, it was found that the volume of the excision specimen was more strongly related to breast size than to tumor size, which in turn might be explained by the surgical technique. To guarantee treatment of potential microscopic tumor cells in the, for instance, 1.5 cm margin around the tumor, a CTV margin expansi-on of 1.5 cm is applied after definition of the tumor bed for radiation treatment planning. This margin is often reduced by the tumor-free resection margin width as measured at pathology investigation of the excision specimen. The lack of an association between the microscopic tumor size and the delineated tumor bed volume is not surprising. When delineating the radiotherapy target, the tumor is not in situ anymore and it was already found that the excised volume is not in pro-portion to the microscopic tumor size. Furthermore, the postoperative tumor bed is prone to seroma and hematoma formation, which can increase target volumes. Moreover, soft-tissue tumor bed appearances cannot be properly visualized on standard planning CT. Additionally, full-thickness closure of the excision cavity be-comes more widely practiced. Even though large seromas might be avoided by the use of full-thickness closure, radiotherapy target definition can be hindered due to less distinct tumor bed margins in the absence of seroma on the planning CT scan.

Therefore, two options to improve the disproportionate treatment volumes are de-scribed in this thesis. First, the value of MRI in definition of the postoperative the tumor bed was investigated. Therefore, the value of MRI in addition to standard CT-guided tumor bed definition was studied, as well as tumor bed definition on MRI-only. Second, the potential for a preoperative MRI-guided radiotherapy was described.

Breast radiotherapy in the presence of a magnetic field

A future breast radiation treatment using the MRI-linac could provide high contrast image guidance directly during radiotherapy. However, before further exploring the possibilities for an MRI-guided breast radiotherapy, the effect of the magnetic field on the dose distribution had to be determined. In the presence of a magnetic field, secondary electrons emanating from the skin into air will be bent back, an ef-fect which is known as the electron return effect (ERE) [12][13]. Since the ERE could result in a higher skin dose, skin toxicity and breast cosmesis could be negatively influenced in the case of treatment by MRI-linac. The expected effects of the ERE are dependent on the beam inclinations at the skin surface, which usually differ between whole breast irradiation and APBI. Other factors determining the possible induced effects of the ERE are the size of the radiation fields and the superficiali-ty of targets. Therefore, the induced effects of the magnetic field were expected

131


to differ between whole breast irradiation and APBI. For that reason, a treatment planning study was performed in Chapter 3 to show the induced effects on the dose distribution in the presence of a magnetic field for both types of breast irra-diation. For whole breast irradiation, standard tangential field treatment planning was performed, as well as a seven field technique in order to investigate whether the influence of the magnetic field decreased with the use of multiple beam direc-tions. A significant increase in skin dose was observed for whole breast irradiation in both cases. Since this increased skin dose can lead to increased skin toxicity and decreased cosmetic outcomes, this observation impairs the clinical acceptability of treating patients with whole breast irradiation by MRI-linac. The direct application of whole breast irradiation in the MRI-linac therefore seems unlikely. However, for APBI, the impact of the magnetic field on skin dose was non-significantly increased. Moreover, the absolute dose values of the organs at risk were low and comparable to reported values in the literature without the presence of a magnetic field [14, 15]. Consequently, in the subsequent chapters we focused on MRI-guided radiotherapy in APBI and not in whole breast irradiation.

Postoperative MRI-guided target definition

Post-lumpectomy CT-guided tumor bed delineation for boost radiotherapy or APBI is inconsistent among radiation oncologists [16-21]. These findings from the litera-ture were confirmed by the results of the CT-guided tumor bed delineation series as described in Chapter 4 and Chapter 5, with a median conformity on CT of only 57% and 44%, respectively. Even though the introduction of surgical clips as markers for the tumor bed has shown to decrease interobserver variation, the high variability among observers still causes target definition to be one of the largest uncertain-ties in the breast radiotherapy chain [18, 22]. A geographical miss would put the patient at increased risk of local recurrence, especially in APBI in which irradiation of the whole breast is omitted. Because there is no tumor to target, differentiation between target tissue and surrounding normal tissue on standard planning CT is problematic and cannot be validated by pathology studies. Since MRI has superior soft-tissue contrast compared to CT, the use of MRI in tumor bed definition was expected to improve differentiation between tumor bed and the surrounding tis-sues, which could lead to an improved treatment accuracy accompanied by smaller treatment volumes. Therefore, in this thesis, the value of MRI for breast tumor bed definition was further explored.

Before further investigating the value of MRI for tumor bed definition, an MRI pro-tocol and MRI patient setup in supine radiotherapy position had to be developed. Standard diagnostic breast MRI is performed with the patient in prone position,

132

Chapter 8

while most departments acquire planning CT imaging with the patient in supine radiotherapy position. A patient in radiotherapy position would not fit into a stan-dard MRI scanner. Furthermore, the use of an anterior receive coil in supine position would deform the breast. In Chapter 4, acquisition of breast MRI with the patient in supine radiotherapy position was shown to be feasible by using a wide bore MRI scanner (Ingenia 1.5 T, Philips, Best, The Netherlands). Perspex coil supports were developed in order to avoid breast deformation. The following 3D high resolution MRI images were acquired pre- and postoperatively: T1 weighted (T1w) fast field echo (FFE) with and without fat suppression (Dixon). T2 weighted (T2w) turbo spin echo (TSE) with fat suppression (SPAIR). Furthermore, a preoperative dynamic series of CE T1w Dixon images was acquired.

Subsequently, in Chapter 4, the value of MRI in addition to standard planning CT for tumor bed delineation was described. In this study, MRI was investigated as an imaging modality which provides information in addition to CT, rather than repla-cing CT.

The results presented in Chapter 4 show that the addition of both preoperative and postoperative MRI to standard CT-guided tumor bed delineation did not im-prove delineation consistency among the 4 observers from the University Medical Center Utrecht in 14 patients, while target volumes increased. Also, preoperative contrast-enhanced CT did not show to improve tumor bed definition. Observers seemed to rather expand their delineation than reducing it based on additional in-formation. This might be influenced by their experience with CT-guided tumor bed delineation, which makes them seem to favor their interpretation of the imaging modality with which they are most familiar. Observers were familiar with the use of diagnostic breast MRI in daily clinical practice. However, the postoperative breast shows different tissue-appearances on MRI compared to the diagnostic situation due the absence of a tumor and the presence of hematoma, seroma, surgical clips and architectural distortions caused by the surgical procedure. These postoperative changes should be carefully interpreted based on the different tissue characteristics as shown by the different sequences.

The increase in tumor bed volume in our study is in line with findings described in the literature by Kirby et al. [23]. According to our study methodology, observers duplicated their CT-based delineation and adjusted this delineation based on the additional information provided by the registered MRI. This methodology was cho-sen to avoid the influence of intraobserver variability. This methodology was diffe-rent in our subsequent study as described in Chapter 5. In this chapter, the value of MRI in addition to CT as well as the value of MRI instead of CT was studied. Seven

133


observers, from the Sunnybrook Health Science Center, delineated a new contour on CT registered to MRI and on MRI-only in 20 patients, at least 2 weeks after com-pletion of their former delineation session, and observers where blinded for their previous contour. In this setting, the consistency among observers on CTMRI-fusion or MRI-only was even found to be significantly lower compared to standard CT-gui-ded tumor bed delineation. Conformity decreased to 39% on CTMRI-fusion and 32% on MRI-only. This suggests that MRI added some degree of confusion rather than increasing the consistency among observers. This confusion could also be the result of limited observer experience with postoperative breast MRI. All observers were trained and instructed at the beginning of the study. However, more extensive training and perhaps also experience over time in interpreting the different tissue appearances on the various MRI sequences might improve study outcomes. In the current format, tumor bed contouring on MRI-only would decrease treatment ac-curacy. Therefore, postoperative tumor bed contouring on MRI-only should not be implemented in today’s clinical practice.

The decreased consistency in tumor bed definition on MRI-only found in this study is in line with findings by Giezen et al. [24]. However, in a study by Jolicoeur et al. MRI-guided tumor bed delineation was found to improve delineation consistency [25]. This contradiction might be caused by the difference in surgical closure tech-niques among studies. Most patients in our study underwent full-thickness closu-re, while Jolicoeur et al. excluded patients who underwent oncoplastic techniques, which probably included full-thickness closure as well. After suturing the cavity walls, seroma might follow the suturing lines, the shapes of which might be more subject to interpretive differences compared to clearly defined cavity walls in su-perficially closed cavities. This hypothesis can be strengthened by the higher con-formity of 66% on standard planning CT in the Jolicoeur study, which was already higher at baseline compared to our study. These findings suggest that it is the type of surgery that hinders tumor bed delineation to a large extent. Even though MRI offers an increased differentiation between the various soft-tissues surrounding the tumor bed, if tissues of the same origin, e.g. fibroglandular breast tissue, are sutu-red together, it would be difficult to visualize the postoperative tumor bed even with MRI. Furthermore, the seroma or edema following suturing lines might lead to differences in interpretation. Perhaps if delineation guidelines would be extremely detailed on what to include and what not to include in the tumor bed delineation, and perhaps if observers would be trained intensively, consistency would improve. However, who should formulate this guideline on what to include and what not? There still is no reference standard with which to validate the, since there is no lon-ger a tumor in situ.

134

Chapter 8

Summarizing, our results show that, in its current format, MRI has no additional value in standard CT-guided postoperative tumor bed delineation. Moreover, the implementation of MRI as a single modality for tumor bed definition could signifi-cantly reduce treatment accuracy. The key issue seems to be that there is no tumor to define as a target in the postoperative setting, and that target definition based on postoperative soft-tissue appearances is difficult, independent of the imaging modality being used. The use of MRI might be more advantageous in a preoperative treatment approach because of the presence of a tumor which can be visualized and characterized by the use of dynamic contrast enhanced sequences. Therefore, the following chapters concentrate on preoperative MRI-guided target definition.

Preoperative MRI-guided target definition

While in the postoperative radiotherapy setting an MRI-guided treatment did not show to offer substantial advantages regarding target definition and treatment vo-lumes, the use of MRI could be advantageous in a preoperative treatment approach with the tumor in situ. MRI has shown to have a high sensitivity for tumor detection [26]. Furthermore, in a study by Schmitz et al. the tumor diameter on MRI showed a good correlation with microscopic tumor size [27]. In the presence of a tumor, dynamic contrast-enhanced parameters can be used for target definition. For tu-mor delineation, contrast-enhancement could result in a superior differentiation between the tumor and surrounding tissues. The results described in Chapter 6 show that preoperative tumor delineation is consistent among 4 observers from the University Medical Center Utrecht, on both CT and MRI. Preoperative imaging of the tumor can be used in two different treatment approaches. The first is a preope-rative treatment approach, which will be described below. Furthermore, preopera-tive imaging could be used in addition to standard postoperative CT-guided tumor bed delineation by increasing the knowledge of the radiation oncologist about the original tumor location. However, the latter was not shown to increase delineation consistency, as presented in Chapter 4.

The high consistency in preoperative tumor delineation among observers presen-ted in Chapter 6, on both CT and MRI, is line with the results presented by Van der Leij et al. [28]. They investigated CT-guided preoperative tumor delineation. Furthermore, our results show that preoperative target volumes are small and less prone to outliers due to the absence of seroma formation. For a preoperative APBI treatment approach, these smaller target volumes could potentially decrease treat-ment toxicity. Several recent postoperative APBI studies reported adverse cosmetic outcomes [5-11]. These outcomes could be related to large volumes receiving high doses, which would be avoidable in the preoperative setting. Furthermore, preope-

135


rative APBI following the NSABP B-39 criteria dosimetrically showed to increase patient eligibility and decrease the dose to the normal tissues in a planning study performed by Nichols et al [29]. These advantages would ultimately influence the implementation of APBI in clinical practice.

Our results, as described in Chapter 6, show that preoperative MRI was superior to CT in terms of tumor detection. Furthermore, more tumor spiculations and irregu-larities were shown on MRI. As image quality further improves, the increased diffe-rentiation between tumor and surrounding tissues can further increase delineation accuracy and thereby decrease target volumes. However, our results described in Chapter 6 show that volumes can also increase. Since more tumor branches were delineated on MRI, MRI-defined volumes were larger compared to CT-defined volu-mes, especially after expansion of GTV to CTV. As techniques in MRI are continuo-usly improving, high resolution MRI can lead to the visualization of (partial) CTV, as compared to the standard planning CT that visualizes only the GTV. The improving visualization will ultimately start to influence our perspectives on the definition of the GTV and the microscopic spread as accounted for by the CTV according to the ICRU recommendations [30].

The superiority of MRI for tumor detection compared to CT is debatable, as descri-bed in Chapter 6. Alternatives improving tumor detection on CT could be consi-dered, like marking the tumor by fiducials or optimizing CT parameters. However, these options will not offer the other advantages of MRI like tumor characterization, segmentation, treatment response measurement, motion analysis and, ultimately, online treatment guidance.

ABLATIVE study

As previously described, a preoperative MRI-guided irradiation of the tumor would be highly consistent and treatment volumes would be small and stable. These advantages, in combination with the high contrast image guidance of the target directly during radiotherapy in a future MRI-linac setting, could allow further tre-atment acceleration. This treatment acceleration could potentially be further ex-tended into a radiosurgical ablative approach as is performed in lung and brain can-cer. In a recent conference abstract, Horton et al presented the results of a clinical phase I study designed to deliver a single ablative dose of 15, 18 or 21 Gy to tumors up to 1 centimeter in maximum diameter by intensity modulated radiotherapy (IMRT) followed by BCS two weeks later [31]. No dose-limiting toxicity was reported and cosmetic results were good to excellent.

136

Chapter 8

An ablative treatment approach would minimize treatment burden to the patient and treatment cost. Because of these benefits, the UMC Utrecht started the ABLATI-VE project. In this project, the delivery of an MRI-guided single ablative radiothera-py dose in early-stage, low risk breast cancer patients is investigated. Before starting a clinical trial, a planning study was commenced in 20 patients to test the dosime-tric feasibility of a single ablative dose to the tumor. This planning study, in which both interstitial multicatheter brachytherapy and external beam radiotherapy plans using volumetric modulated arc therapy (VMAT) were acquired and compared, is described in Chapter 7. A 15 Gy dose was prescribed to the PTV of the CTV (PTVCTV) with an integrated boost of 20 Gy to the PTV of the GTV (PTVGTV). Interstitial multi-catheter brachytherapy showed to be less suitable for this ablative approach. This was mainly due to the high number of catheters (median of 9 catheters) required for adequate coverage of the PTVGTV, while it also resulted in a common overdosage at the PTVCTV. The high number of catheters would decrease patient comfort during the procedure. Furthermore, the interstitial multicatheter approach would be less accessible compared to external beam radiotherapy, due to the required brachy-therapy facilities and level of expertise. The dose in organs at risk was comparable between both treatment approaches.

Since this study showed a single ablative dose, by using a VMAT approach to be dosimetrically feasible, the clinical applicability will be further investigated in a cli-nical trial. At the UMC Utrecht, we recently started the ABLATIVE [32]. In this study, the clinical feasibility of an MRI-guided, preoperative, single dose, ablative radiation treatment in low-risk patients will be investigated. The prescribed dose in this stu-dy is similar to the planning study described in Chapter 7. The primary outcome will be the pathologic complete response rate at six months after treatment, when breast-conserving surgery will be performed. Secondary objectives will be toxicity, quality of life and cosmetic result. Within this APBI study, careful patient selection is crucial. Only patients with low-risk disease characteristics (e.g. patients age ≥ 60 years, unifocal disease, tumor size ≤ 2 cm, non-lobular histological type, no lymph node metastases or indication for adjuvant systemic therapy) will be included in the study. The ultimate aim will be to treat the tumor precisely, while minimizing treatment burden to the patient with a good cosmetic outcome. If this single dose MRI-guided treatment demonstrates a complete pathologic response in combina-tion with limited toxicity and good breast cosmesis in a sufficient proportion of the patients, the necessity of omitting breast-conserving surgery could be studied in the future. However, if the treatment would not turn out to be ablative while toxicity profiles, cosmesis and disease free survival rates are good, this preoperative treat-ment by single dose could still be implemented in clinical practice as a replacement of the current 3-5 week treatment in selected patients.

137


Conclusions and future perspectives

The results of this thesis offer possible avenues for studying MRI-guided breast ra-diotherapy in a future MRI-linac setting. MRI-guided preoperative target definiti-on showed to be highly consistent and could be accompanied by small and stable treatment volumes. The prescription of a single ablative dose to the breast tumor and its surrounding CTV showed to be dosimetrically feasible by using a VMAT ap-proach. Furthermore, this thesis shows that the use of MRI does not improve the poor consistency of the current postoperative CT-guided tumor bed definition. Mo-reover, we showed that the use of postoperative MRI as a single imaging modality could significantly decrease consistency and treatment precision when implemen-ted in its current format.

The findings outlined in this thesis showed that it was feasible to start the ABLATIVE trial. In this trial, the feasibility of an MRI-guided single ablative dose of 20 Gy to the tumor and 15 Gy to the surrounding clinical target volume will be investigated in low-risk, early-stage breast cancer patients. This study could be a first step towards a more accessible and patient-friendly treatment for selected early-stage breast can-cer patients.

138

Chapter 8

References



J Surg Oncol 2012: 38:918-924.









5. Olivotto IA, Whelan TJ, Parpia S et. al. Interim Cosmetic and Toxicity Results From RAPID: A Randomi-

zed Trial of Accelerated Partial Breast Irradiation Using Three-Dimensional Conformal External Beam

Radiation Therapy. J Clin Oncol 2013: 31:4038-4045.

6. Lei RY, Leonard CE, Howell KT et. al. Four-year clinical update from a prospective trial of accelerated

partial breast intensity-modulated radiotherapy (APBIMRT). Breast Cancer Res Treat 2013: 140:119-133.

7. Leonard KL, Hepel JT, Hiatt JR, Dipetrillo TA, Price LL, Wazer DE. The effect of dose-volume parame-

ters and interfraction interval on cosmetic outcome and toxicity after 3-dimensional conformal

accelerated partial breast irradiation. Int J Radiat Oncol Biol Phys 2013: 85:623-629.

8. Formenti SC, Hsu H, Fenton-Kerimian M et. al. Prone accelerated partial breast irradiation after bre-

ast-conserving surgery: five-year results of 100 patients. Int J Radiat Oncol Biol Phys 2012: 84:606-611.

9. Polgar C, Fodor J, Major T, Sulyok Z, Kasler M. Breast-conserving therapy with partial or whole breast

irradiation: ten-year results of the Budapest randomized trial. Radiother Oncol 2013: 108:197-202.

10. Shah C, Badiyan S, Ben Wilkinson J et. al. Treatment efficacy with accelerated partial breast irradiati-

on (APBI): final analysis of the American Society of Breast Surgeons MammoSite((R)) breast brachy-

therapy registry trial. Ann Surg Oncol 2013: 20:3279-3285.

11. Liss AL, Ben-David MA, Jagsi R et. al. Decline of cosmetic outcomes following accelerated partial

breast irradiation using intensity modulated radiation therapy: results of a single-institution pros-

pective clinical trial. Int J Radiat Oncol Biol Phys 2014: 89:96-102.



Phys Med Biol 2005: 50:1363-1676.



14. Moran JM, Ben-David MA, Marsh RB, Balter JM, Griffith KA, Hayman JA, Pierce LJ. Accelerated partial

breast irradiation: what is dosimetric effect of advanced technology approaches?. Int J Radiat Oncol

Biol Phys 2009: 75:294-301.

139


15. Moon SH, Shin KH, Kim TH et. al. Dosimetric comparison of four different external beam partial

breast irradiation techniques: three-dimensional conformal radiotherapy, intensity-modulated ra-

diotherapy, helical tomotherapy, and proton beam therapy. Radiother Oncol 2009: 90:66-73.






2009: 90:60-65.






Biol Phys 2007: 69:41-48.












Biol Phys 2009: 74:1276-1282.



Biol Phys 2012: 82:1332-1340.








97:225-232.

140

Chapter 8






Clin Oncol 2012: 36:232-238.

30. ICRU Report. No. 50 1993.




32. Centrale Commissie Mensgebonden Onderzoek. Dossiernummer NL46017.041.13 www.toetsing-

online.nl .

Chapter 8

Chapter 9Nederlandse samenvatting

List of publicationsCurriculum Vitae

Dankwoord

144

Chapter 9

Het doel van radiotherapie is het vernietigen van kankercellen. Hierbij dienen de omliggende gezonde organen zoveel mogelijk gespaard te worden, om zo de bijwerkingen zoveel mogelijk te beperken. Dit evenwicht wordt bepaald door de nauwkeurigheid van de behandeling. Deze nauwkeurigheid wordt voor een groot deel bepaald door het zo precies mogelijk definiëren van het doelgebied dat be-straald dient te worden. Het definiëren van het doelvolume wordt doorgaans uitge-voerd door de radiotherapeut op een CT-scan. Deze CT-scan wordt voorafgaand aan de bestralingsbehandeling van de patiënt gemaakt, om zo een goed onderscheid te kunnen maken tussen het doelvolume en de omliggende weefsels. Aan de hand van de CT-scan wordt een bestralingsplan vervaardigd en wordt de patiënt meer-dere dagen tot weken bestraald. Het aantal keren dat een patiënt bestraald moet worden, wordt het aantal fracties genoemd. Een groot nadeel van CT is dat het con-trast tussen de verschillende weefsels niet optimaal is, waardoor de radiotherapeut het doelgebied niet altijd goed kan onderscheiden van de omliggende gezonde weefsels. MRI, daarentegen, is een beeldvormende techniek waarbij het contrast tussen de verschillende weefsels veel beter is. Daarnaast kan met MRI, zonder io-niserende stralingsbelasting voor de patiënt, de beweging van het doelgebied en de omliggende organen in beeld worden gebracht en de reactie van tumoren op de behandeling worden geanalyseerd. Als we de voordelen van MRI zouden kun-nen gebruiken tijdens de bestralingsbehandeling, zou dat de nauwkeurigheid van de behandeling kunnen verbeteren. Daarom heeft de afdeling Radiotherapie van het UMC Utrecht samen met Philips (Best, Nederland) en Elekta (Stockholm, Zwe-den) een bestralingsapparaat ontwikkeld met geïntegreerde MRI-functionaliteit, de zogenaamde MRI-versneller. Deze MRI-versneller bestaat uit een 8 MV lineaire versneller die in een ring rondom een 1.5 Tesla MRI scanner draait. De MRI-versnel-ler heeft de potentie om de radiotherapeutische behandeling te verbeteren voor verschillende tumorgebieden, waaronder baarmoederhalskanker, prostaatkanker en endeldarmkanker.

Borstkanker is de meest voorkomende vorm van kanker bij vrouwen wereldwijd. Omdat in de westerse wereld borstkanker vaak in een vroeg stadium wordt gedi-agnosticeerd, kan een groot deel van deze vrouwen behandeld worden middels een borstsparende operatie gevolgd door radiotherapie van de gehele borst. Deze vorm van radiotherapie kan echter als te belastend worden ervaren door patiën-ten, omdat ze meerdere weken lang, iedere dag behandeld moeten worden. Om deze reden wordt wereldwijd momenteel de zogenoemde ‘partiële borstbestraling’ onderzocht, waarbij alleen het gedeelte van de borst bestraald waar de tumor oor-spronkelijk heeft gezeten, het zogenoemde ‘tumorbed’. Door het kleinere bestra-lingsvolume bij de partiële borstbestraling kan de bestralingsdosis per fractie wor-den verhoogd en hoeft de patiënt minder bestralingsfracties te ondergaan. Omdat

145

Nederlandse samenvatting

bij deze partiële borstbestraling de bestraling van de rest van de borst achterwege wordt gelaten, is het nauwkeurig definiëren van het tumorbed essentieel ter voor-koming van het lokaal teruggroeien van tumorweefsel.

Naast het terugkomen van borstkanker ter plaatse van het oorspronkelijke tumor-bed, streeft men bij de borstsparende behandeling naar minimale bijwerkingen en een zo goed mogelijk cosmetisch resultaat. Aangezien grote doelvolumina de bijwerkingen en cosmetiek negatief beïnvloeden, is het belangrijk de bestralings-volumina zo klein mogelijk te houden. Gezien de bovengenoemde voordelen van MRI, zou een MRI-geleide bestraling de borstsparende behandeling van borstkan-kerpatiënten kunnen verbeteren. MRI zou onder andere het definiëren van het doelvolume, in dit geval het tumorbed, nauwkeuriger kunnen maken en daarmee de bestralingsvolumina verkleinen.

In dit proefschrift worden de mogelijkheden beschreven voor een MRI-geleide be-straling bij patiënten met vroeg-stadium borstkanker.

In Hoofdstuk 2 zijn de behandelvolumina binnen de huidige borstsparende be-handeling geëvalueerd. Binnen deze behandeling wordt continu naar een balans gestreefd: Het behandelen van een te klein doelgebied geeft een verhoogde kans op het lokaal recidiveren van de ziekte, terwijl een te groot doelvolume kan leiden tot meer bijwerkingen aan de omliggende organen en een slechter cosmetisch re-sultaat. Het volume van het borstweefsel dat is verwijderd door de chirurg en de grootte van het bestraalde tumorbed beïnvloeden beiden de toxiciteit en cosme-tiek van de borstsparende behandeling. Daarom werd de relatie bestudeerd tussen de grootte van de tumor en de geëxcideerde en bestraalde tumorbed-volumina bij 186 patiënten met vroeg-stadium borstkanker, behandeld middels borstsparende behandeling. Zowel de geëxcideerde volumina als de bestraalde volumina bleken niet gerelateerd te zijn aan de grootte van de oorspronkelijke tumor. Verschillen-de hoeveelheden borstweefsel werden verwijderd voor gelijke tumorgroottes. Het volume van het geëxcideerde borstweefsel bleek sterker gerelateerd te zijn aan de grootte van de borst dan aan de tumorgrootte. Daarnaast bleek dat de relatie tus-sen de tumorgrootte en het geëxcideerde volume beïnvloed werd door het wel of niet palpabel zijn van de tumor. De relatie tussen het geëxcideerde volume en het bestraalde tumorbed-volume bleek bovendien te worden beïnvloed door de chirurgische techniek die werd uitgeoefend.

Alvorens te onderzoeken of een bestralingsbehandeling met de MRI-versneller bij deze patiëntengroep voordelen zou bieden, werd in Hoofdstuk 3 onderzocht of de aanwezigheid van een magneetveld tijdens de borstbestraling invloed heeft op de

146

Chapter 9

verdeling van de bestralingsdosis in de huid. Middels een planningsstudie werden zowel de gehele borstbestraling als de partiële borstbestraling onderzocht. Deze twee behandelingen verschillen van elkaar wat betreft de hoeken van de bestra-lingsbundels ten opzichte van het huidoppervlak, de grootte van de bestralings-velden en de diepte van de doelvolumina. Voor de gehele borstbestraling is zowel een standaard schampveldentechniek als een zeven-bundel techniek bestudeerd om zo vast te stellen of de invloed van het magneetveld afnam bij het gebruik van meerdere bundelrichtingen. Een significante toename van de huiddosis werd waar-genomen voor beide technieken. Omdat de toegenomen huiddosis kan leiden tot meer huidtoxiciteit en slechtere cosmetische resultaten, zou de implementatie van de MRI-versneller voor een bestralingsbehandeling van de gehele borst tot toena-me van bijwerkingen kunnen leiden. Het toekomstig inzetten van de MRI-versneller voor de gehele borstbestraling is daarom niet aantrekkelijk. Voor een partiële borst-bestraling bleek de huiddosis echter niet significant verhoogd in de aanwezigheid van een magneetveld. Daarbij was de stralingsdosis in de omliggende risico-orga-nen laag en vergelijkbaar met waarden die in de literatuur staan beschreven in stu-dies zonder magneetveld. Zodoende is de focus in de volgende hoofdstukken ko-men te liggen op de partiële borstbestraling en niet op de gehele borstbestraling.Voor een partiële borstbestraling is het van belang het te bestralen tumorbed nauwkeurig te definiëren. In de literatuur is beschreven dat definitie van het tu-morbed op de standaard planning-CT, die wordt verricht voor het maken van het bestralingsplan, grote variabiliteit toont onder radiotherapeuten. Daarom werd in Hoofdstuk 4 en Hoofdstuk 5 onderzocht of MRI de nauwkeurigheid van de inteke-ning van het tumorbed kon verhogen. Alvorens dit te kunnen onderzoeken, werd een nieuw MRI scanprotocol ontwikkeld voor patiënten in de bestralingshouding. De standaard diagnostische MRI-mamma wordt namelijk uitgevoerd met de pati-ent in buikligging, terwijl de meeste radiotherapie-instituten de planning-CT in de bestralingshouding in rugligging maken omdat de bestraling in rugligging wordt uitgevoerd.

In Hoofdstuk 4 wordt het vervaardigen van MRI van de borsten bij de patiënt in de bestralingshouding in rugligging beschreven. Hierbij wordt een MRI-scanner met een extra grote opening gebruikt (Ingenia 1.5 T, Philips, Best, The Netherlands). Er werden steunen van Plexiglas ontwikkeld ter ondersteuning van de spoelen, om zo deformatie van de borst te voorkomen. De volgende 3D MRI beelden met hoge resolutie werden vervaardigd, zowel pre- als postoperatief: T1-gewogen (T1w) fast field echo (FFE) met en zonder vetsuppressie (Dixon) en T2-gewogen turbo spin echo (TSE) met vetsuppressie (SPAIR). Daarnaast werd een preoperatieve serie T1w beelden vervaardigd met het gebruik van gadolinium contrast. Vervolgens werd de waarde van MRI in aanvulling op de standaard planning-CT bestudeerd voor het intekenen van het tumorbed. Op standaard CT was de overeenstemming tus-

147

Nederlandse samenvatting

sen de 4 radiotherapeuten uit het UMC Utrecht mediaan slechts 57%. Deze waarde verbeterde niet na het toevoegen van zowel pre- als postoperatieve MRI, terwijl de volumina toenamen. In aanvulling op Hoofdstuk 4, waar de waarde van MRI als toevoeging op CT werd onderzocht, werd in Hoofdstuk 5 ook de waarde van MRI als enige beeldvormende modaliteit voor het intekenen van het tumorbed onderzocht. Zeven radiotherapeu-ten werkzaam in het Sunnybrook Health Science Center, Toronto, Canada, definieer-den een doelgebied op zowel CT, CTMRI-fusie en MRI-alleen bij 20 patiënten. Er zat minstens 2 weken tussen de inteken-sessies en intekenaars werden geblindeerd voor hun eerder gedefinieerde doelvolumina. In deze setting was de consisten-tie onder radiotherapeuten op CTMRI-fusie of MRI-alleen lager dan de standaard CT-geleide definitie van de doelvolumina. De overeenstemming nam af van 44% op CT tot 39% op CTMRI-fusie en 32% op MRI-alleen. Dit suggereert dat MRI een bepaalde mate van verwarring veroorzaakte, in plaats van het verduidelijken van het doelgebied en het verhogen van de overeenstemming. Evenals in hoofdstuk 2, werd ook in dit hoofdstuk de invloed van de chirurgische techniek bij de resultaten betrokken. Bij het merendeel van de patiënten zijn de wanden van de excisieholte door de chirurg geapproximeerd, wat de definitie van het tumorbed substantieel lijkt te beïnvloeden. In deze patiëntencategorie zal daarom de definitie van het tu-morbed op alleen MRI de nauwkeurigheid van de behandeling verkleinen. Om deze reden dient een puur MRI-geleide definitie van het doelvolume niet in zijn huidige vorm te worden geïmplementeerd in de huidige klinische praktijk. De bovenbeschreven resultaten tonen dat het postoperatieve tumorbed niet nauwkeuriger kan worden gedefinieerd met behulp van MRI. Dit zou in grote mate verklaard kunnen worden door het feit dat er geen tumor is om te definiëren. Dit maakt het intekenen van een doelvolume gebaseerd op weke-delenveranderingen in de borst gecompliceerd, ongeacht de beeldvormende techniek.

In Hoofdstuk 6 werd daarom de nauwkeurigheid van een preoperatieve definitie van het doelvolume onderzocht. In een preoperatieve setting is er een tumor in situ als doelvolume voor de bestraling, welke logischerwijs beter gedefinieerd kan worden op de beeldvorming. Mogelijk zou MRI bij de tumordefinitie meerwaarde kunnen bieden ten opzichte van CT, gezien de hoge sensitiviteit van MRI voor tu-mordetectie. Daarnaast is bekend uit de literatuur dat de tumordiameter op MRI goed correleert met de daadwerkelijke microscopische tumorgrootte zoals ge-meten bij pathologisch onderzoek. De resultaten in dit hoofdstuk laten zien dat preoperatieve tumordefinitie consistent is onder 4 radiotherapeuten werkzaam binnen het UMC Utrecht, op zowel CT als MRI. Daarnaast zijn de preoperatieve doel-volumina veel kleiner en minder gevoelig voor uitschieters dan we gezien hebben bij de standaard postoperatieve series. Dit is mede het gevolg van de afwezigheid

148

Chapter 9

van seroomvorming, zoals in de postoperatieve situatie het geval is. Deze kleine preoperatieve doelvolumina zouden in het geval van een partiële borstbestraling de bijwerkingen van de behandeling mogelijk kunnen verminderen, de cosmetiek kunnen verbeteren, verdere hypofractionering mogelijk maken en hiermee meer patiënten in aanmerking kunnen laten komen voor een borstsparende behande-ling. Daarnaast lieten de resultaten in dit hoofdstuk zien dat tumordetectie beter was op preoperatieve MRI dan op CT. Tevens toonde MRI meer sprieterige uitlopers. Deze uitlopers zorgden er bij expansie van de intekening naar het klinisch doelvo-lume (CTV) wel voor dat de MRI-gedefinieerde CTV’s significant groter werden dan de CT-gedefinieerde CTV’s.

Gezien de bevindingen in hoofdstuk 6 en het feit dat MRI mogelijkheden biedt voor tumorkarakterisatie, segmentatie, responsmetingen en bewegingsanalyse, werd de mogelijkheid van een preoperatieve MRI-geleide behandeling van borstkan-kerpatiënten verder onderzocht. Door de kleine preoperatieve behandelvolumina zou de bestralingsbehandeling mogelijk gehypofractioneerd kunnen worden (met een hogere fractiedosis in minder fracties) tot mogelijk zelfs een eenmalige abla-tieve dosis. Een ablatieve behandeling zou de belasting voor de patiënt kunnen verkleinen en mogelijk ook de kosten kunnen verlagen. Daarom is in Hoofdstuk 7 de dosimetrische haalbaarheid van een preoperatieve eenmalige MRI-geleide be-stralingsbehandeling onderzocht, met een hoge dosis op de borsttumor. In deze planningsstudie werd zowel een inwendige bestraling middels interstitiële mul-ticatheter brachytherapie als een uitwendige bestraling middels de zogenaamde volumetric modulated arc therapy (VMAT) onderzocht. Hierbij werd een dosis tot 20 Gy voorgeschreven op de primaire tumor en een dosis tot 15 Gy op de omlig-gende 2 cm aan borstweefsel met hierin mogelijke tumoruitbreiding. De resultaten van deze studie laten zien dat een eenmalige MRI-geleide ablatieve bestralingsbe-handeling middels VMAT dosimetrisch haalbaar is. De multicatheter brachytherapie techniek bleek dosimetrisch minder geschikt, vanwege een groot aantal naalden nodig voor het behalen van een adequate coverage van de tumor. Dit ging gepaard met een overdosering in de omliggende 2 cm aan borstweefsel. De dosis in de om-liggende organen was vergelijkbaar tussen de twee technieken. Om deze reden is een vervolgstudie gestart, waarbij de haalbaarheid van deze uitwendige bestra-lingsbehandeling in patiënten met vroeg-stadium borstkanker, wordt onderzocht.

In Hoofdstuk 8 worden de belangrijkste bevindingen samengevat en in het per-spectief van de literatuur, dagelijkse praktijk en toekomst geplaatst.

Chapter 9



Dankwoord

152

Chapter 9

Journal articles

den Hartogh MD, van Asselen B, Monninkhof EM et. al. Excised and irradiated volu-mes in relation to the tumor size in breast-conserving therapy. Breast Cancer Res Treat 2011: 129:857-865.

van Heijst TC, den Hartogh MD, Lagendijk JJ, van den Bongard HJ, van Asselen B. MR-guided breast radiotherapy: feasibility and magnetic-field impact on skin dose. Phys Med Biol 2013: 58:5917-5930.

den Hartogh MD, Philippens ME, van Dam IE et. al. MRI and CT imaging for preope-rative target volume delineation in breast-conserving therapy. Radiat Oncol 2014: 9:63.

den Hartogh MD, van den Bongard HJ, Davidson MT et. al. Full-Thickness Closure in Breast-Conserving Surgery: The Impact on Radiotherapy Target Definition for Boost and Partial Breast Irradiation. A Multimodality Image Evaluation. Ann Surg Oncol 2014: 21: 3774-3779

Popovic M, den Hartogh MD, Zhang L et. al. Review of international patterns of prac-tice for the treatment of painful bone metastases with palliative radiotherapy from 1993 to 2013. Radiother Oncol 2014: 111:11-17.

den Hartogh MD, Philippens MEP, van Dam IE, Kleynen CE, Tersteeg JHA, Kotte ANTJ, van Vulpen M, van Asselen B, van den Bongard HJGD. Post-lumpectomy CT-guided tumor bed delineation for breast boost and partial breast irradiation: Can additional pre- and postoperative imaging reduce interobserver variability? Submitted.

Charaghvandi KR, den Hartogh MD, van Ommen AMLN, de Vries JHW, Scholten V, Moerland MA, Philippens MEP, van Vulpen M, van Asselen B, van den Bongard HJGD. MRI-guided single fraction ablative radiotherapy for early-stage breast cancer: A brachytherapy versus VMAT planning study. 2014. Manuscript in preparation.

Conference proceedings

den Hartogh MD, van den Bongard HJGD, Davidson MTM, Kotte ANTJ, Philippens MEP, Van Vulpen M, Van Asselen B, Pignol JP. The use of full-thickness closure in bre-ast conserving surgery results in an imprecise target definition for boost or partial breast radiotherapy. European Breast Cancer Conference 2014, Glasgow, Scotland, UK

van Heijst TCF, Philippens MEP, Kleijnen JJE, den Hartogh MD, Lagendijk JJW, van den Bongard HJGD, van Asselen B. Quantification of intra-fraction motion of breast tumors using cine-MRI. AAPM 2014, Austin, USA

153

List of publications

Charaghvandi KR, den Hartogh MD, de Vries JHW, Scholten V, Moerland MA, Phi-lippens MEP, van Vulpen M, van Asselen B, van den Bongard HJGD. Single fraction ablative radiotherapy for early stage breast cancer: a brachytherapy versus IMRT planning study. ESTRO’s Annual Meeting 2014, Vienna, Austria

Den Hartogh MD, Philippens MEP, Van Dam IE, Kleynen CE, Tersteeg JHA, Kotte ANTJ, Van Vulpen M, Van Asselen B, Van den Bongard HJGD. The additional value of MR and CT imaging for target volume delineation in breast boost and partial breast irradiation. San Antonio Breast Cancer Symposium 2013, San Antonio, USA

Den Hartogh MD, Philippens MEP, Van Dam IE, Kleynen CE, Tersteeg JHA, Kotte ANTJ, Van Vulpen M, Van Asselen B, Van den Bongard HJGD. Small treatment volumes and a low interobserver variability in preoperative MRI-guided target volume delineati-on for accelerated partial breast irradiation. ASTRO’s Annual Meeting 2013, Atlanta, USA

Den Hartogh MD, Philippens MEP, Van Dam IE, Kleynen CE, Tersteeg JHA, Pijnappel RM, Fernandez Gallardo M, Kotte ANTJ, Verkooijen HM, Van den Bosch MAAJ, Van Vulpen M, Van Asselen B, Van den Bongard HJGD. High precision of MRI-guided tar-get volume delineation before breast-conserving surgery. ESTRO’s Annual Meeting 2013, Geneva, Switzerland

Van Heijst TCF, den Hartogh MD, Bol GH, Raaymakers BW, Lagendijk JJW, van den Bongard HJGD, van Asselen B. Magnetic field effects on the skin dose in MRI-guided breast RT. ESTRO’s Annual Meeting 2013, Geneva, Switzerland

Den Hartogh MD, Philippens MEP, Blanken N, Van den Bosch M.A.A.J., Van Vulpen M, Van Asselen B, Van den Bongard HJGD. Pre- and postoperative breast MRI in supine radiotherapy position. ESTRO’s Annual Meeting 2012, Barcelona, Spain

Doganos D, Den Hartogh MD, Philippens MEP, Lagendijk JJW, Van Vulpen M, Van den Bongard HJGD, Van Asselen B. Dose escalation in MRI guided preoperative accele-rated partial breast radiotherapy. ESTRO’s Annual Meeting 2012, Barcelona, Spain

Den Hartogh MD, Philippens MEP, Blanken N, Van Vulpen M, Van Asselen B, Van den Bongard HJGD. Feasibility study on preoperative MRI-guided irradiation in bre-ast-conserving therapy. European Breast Cancer Conference 2012, Vienna, Austria

Den Hartogh MD, Monninkhof EM, van Asselen B, van den Bosch MAAJ, van Vulpen M, van Diest PJ, Gilhuijs KGA, van de Bunt L, Mali WPTM, van den Bongard HJGD. Breast conserving therapy: Evaluation of the tumor diameter and treated volumes. ESTRO’s Annual Meeting 2011, London, UK

Chapter 9



Dankwoord

157

Curriculum vitae

Mariska den Hartogh was born on Februari 16, 1986 in Nijkerk, The Netherlands. In 2004 she graduated from secondary school at Farel College in Amersfoort. From 2004 to 2010, she studied medicine at Utrecht University. She was involved in the organization of the UMC Utrecht International Medical Summerschool for several years, until she started her internships. During her medical studies, she started con-ducting scientific research at the Department of Radiation Oncology at the Univer-sity Medical Center Utrecht which has led to the results presented in this thesis. From November 2010 to May 2014, she was fully dedicated to this project investi-gating the potential for MRI-guided radiotherapy for early-stage breast cancer pa-tients. The project was conducted under the supervision of Prof. M. van Vulpen, Prof. M.A.A.J. van den Bosch, Dr. H.J.G.D van den Bongard and Dr. B. van Asselen. In May 2013 she collaborated on an MRI-project under the supervision of Prof. J.P. Pignol at the Sunnybrook Health Science Center, Toronto, Canada. The results of this project are described in this thesis. In June 2014 she started her residency in radiation on-cology at the Department of Radiation Oncology at the University Medical Center Utrecht under the supervision of Dr. J.L. Noteboom.

Chapter 9



Dankwoord

160

Chapter 9

Tot slot wil ik iedereen bedanken die een bijdrage heeft geleverd aan het tot stand komen van dit proefschrift, waarbij in het bijzonder…

Allereerst de patiënten die enkele dagen na het stellen van de diagnose borstkan-ker belangeloos besloten deel te nemen aan het onderzoek.

Mijn promotoren, Prof. M. van Vulpen en Prof. M.A.A.J. van den Bosch. Beste Mar-co en Maurice, bedankt voor het enthousiasme waarmee jullie mij en de mensen om mij heen weten te motiveren en inspireren. Hierbij wil ik ook Prof. W.P.T.M. Mali noemen, die het mogelijk maakte dit onderzoek bij de radiologie van start te laten gaan. Professor J.J.W. Lagendijk, vanaf de zijlijn betrokken om zo nodig kritisch mee te denken.

Mijn co-promotoren Dr. H.J.G.D. van den Bongard en Dr. B. van Asselen. Desirée, ik heb veel bewondering voor je inhoudelijke kennis en doorzettingsvermogen. Be-dankt voor je grote betrokkenheid bij de studies en de mogelijkheden die je me binnen het project hebt geboden. Bram, bedankt voor de fysische input, het bij-brengen van de basisbegrippen binnen de radiotherapie fysica en je vermogen om zo nu en dan binnen de projecten ook even ‘uit te zoomen’.

Leden van de beoordelingscommissie, Prof. Dr. P.J. van Diest, Prof. Dr. P.H.M. Peeters, Prof. C.T.W. Moonen, Prof. Dr. R. van Hilligersberg en Prof. Dr. J.P. Pignol, bedankt dat jullie de tijd hebben genomen het manuscript te beoordelen. Jean-Philippe, thank you for your warm welcome in Toronto. I have enjoyed our collaboration and I look forward to continue learning from you in the future.

Mariëlle, jouw kennis en ideeën over de MRI sequenties hebben een onmisbare bijdrage geleverd aan dit proefschrift, waarvoor veel dank. Daarnaast natuurlijk bedankt voor het scannen van vrijwilligers in de avonduren en je bijdrage aan de manuscripten.

Iris van Dam, Karin Kleynen en Robbert Tersteeg, dank voor de enorme tijdsinves-tering die jullie hebben geleverd bij het intekenen van de vele CT’s en MRI’s. Iris, daarnaast extra dank voor de enorme rust die je als supervisor soms wist te bren-gen in de drukke tijden waar ik kliniek met het afronden van mijn proefschrift com-bineerde.

Mede-auteurs bedankt voor de plezierige samenwerking en jullie bijdrage. Alexis Kotte, bedankt voor alle analyses en het meedenken over de uitkomsten ervan. Len-ny Verkooijen, dank voor het kritisch meedenken over de studies en analyses.

Melanie Davidson, many thanks for all your expertise and support when acquiring the Sunnybrook data. Dr. Liu, Dr Chow, Dr. Paszat, Dr. Vesprini, Dr. Ackerman and Dr. Lee, thank you all very much for your time and help on the delineation project.

161

Dankwoord

Iedereen op de afdeling die zijdelings bij dit project betrokken is geweest, dank voor het meedenken met het onderzoek en alles wat daarbij komt kijken. Vincent, bedankt dat je me hebt leren plannen in Monaco. Radiotherapeuten, klinisch fysi-ci, onderzoekers, laboranten, onderzoekers, doktersassistenten, secretaresses, be-dankt voor de gezelligheid en de goede werksfeer op onze afdeling.

Laboranten van de MR9, dank voor het meedenken over de nieuwe scanposities en protocollen.

Trialbureau van de divisie Beeld met in het bijzonder Tineke, Marianne, Saskia en Shanta, bedankt voor het meedenken over METC en monitor gerelateerde zaken. Anneke, bedankt voor het inplannen van alle studie-MRI’s.

Afdeling multimedia van de divisie Beeld, dank voor maken van de vele figuren en congresposters.

Mammacare verpleegkundigen, NP’s, PA’s, mammachirurgen en -radiotherapeuten van het UMC Utrecht en St. Antonius Ziekenhuis, ontzettend bedankt voor jullie inspanningen rondom het includeren van studiepatiënten.

Hanne, Metha, Lisanne, Frederiek, Maarten, Joanne, Ramona, Danny, Peter, Max, Joris en Juliette. Soms zijn er zaken die alleen mede-onderzoekers begrijpen. Wat ben ik blij dat ik deel heb mogen uitmaken van een geweldige arts-onderzoekers-groep. Bedankt voor jullie enthousiasme en vooral voor jullie gezelligheid! Ramo-na, ik wens je veel succes met het ABLATIVE project! Tristan, wat was ik blij toen je de mamma-onderzoeksgroep kwam versterken. Dank voor de fysische inzichten en gezellige praatjes op de gang en succes met het afronden van je proefschrift. Mamma-onderzoekers binnen het UMC Utrecht, bedankt voor de gezelligheid op congressen en tijdens de vele uren mamma-MDO!

Mede-AIOS Irene, Martijn, Miranda, Hugo, Saskia, Paulien, Hanne, Christa, Laura, An, Debora en Carlijn. Dank voor jullie hulp en begrip tijdens het combineren van de kliniek met het afronden van mijn proefschrift. Ik hoop vanaf nu ook weer wat terug te kunnen doen!

Paranimfen bedankt dat jullie aan mijn zijde willen staan straks bij de verdediging. Lisanne, met onze CI zit het wel goed. Ik kijk enorm uit naar onze ‘over-5-jaar-la-chen- we-hierom-borrel’ die steeds dichterbij komt. Kris, we ontvingen op dezelfde dag onze artsen-bul en ons daaropvolgende harde werken werd zo nu en dan on-derbroken door een reisje en een wereldwonder op zijn tijd. Fijn dat je ook aan de andere kant van de wereld altijd bereikbaar was!

Lieve vrienden en vriendinnen, in onderzoeksland gaat het niet altijd over rozen. Dank voor alle luisterende oren, adviezen, afleiding en zeker ook het begrip als ik eens moest afhaken in tijden van drukte. Ik kijk uit naar alle gezellige avonden en weekenden vanaf december!

162

Chapter 9

Frits, Willemien, Frederieke en Nienke, wijzen komen uit het oosten. Dank voor jullie betrokkenheid, steun en Nienke natuurlijk voor de Engelse correcties.

Sabine, Fabian en Pascal, wat fijn dat ik weet dat jullie er altijd voor me zijn. Natuur-lijk altijd gepaard met de nodige dosis humor. Sabine, fijn dat we ‘het lief en leed dat promoveren heet’ konden delen!

Pap en mam, jullie hebben me alle mogelijkheden gegeven mezelf te ontwikkelen. Bedankt voor jullie oneindige vertrouwen in mijn kunnen en voor de deur die altijd open staat.

Lieve Roel, wat voelt het goed om thuis te komen. Dank voor je relativeringsvermo-gen en onvoorwaardelijke steun.

towards mri-guided radiotherapy in early-stage … mri-guided radiotherapy in early-stage breast...

Documents