1

Ina Jürgenliemk-Schulz

MRI guided radiotherapy for cervical cancer UMC Utrecht clinical experience

11.06.2014 Advanced Radiotherapy Techniques

Cervical cancer

• 650 – 700 new patients per year in The Netherland

• About 300 advanced stages

• 20-25 patients treated in UMCU

• 2rd most frequent malignancy in women worldwide

• 1st in India (200 000 new patients per year)

Traditional radiotherapy for advanced cervical cancer

Combination of EBRT and Brachytherapy

• 45 - 50 Gy external beam radiotherapy to the pelvic/PAO region – Conventional or conformal EBRT

• About 25 - 30 Gy brachytherapy to the central part of the pelvis – Point A based treatment planning

• Optional – External boost in case of parametrial invasion or affected lymph nodes

• Overall treatment time about seven weeks

Results of traditional RT not satisfying Logsdon, Int J Radiat Oncol Biol Phys. 1999

• Loco-regional control for locally advanced cervical cancer 53 %

• 5-year disease-free and overall survival less than 45 %

• Grade 3 - 5 toxicity about 17 %

WHO alert in 1999

• Improvement by combining radiotherapy with chemotherapy

• 5 trials published, about 12 % benefit in 5-year DSS

• 2001 Meta-analysis; Green et al. Lancet

• 16 % benefit 5-year DSS

• increased grade 3 and 4 hemato and gastro-intestinal toxicity

Outcome related to periods

RT (± HT) ≤ 1999 RT ± CT 1999-2004

Number of patients 95 54

Median FU in months 49 46

Years of treatment 1983 – 1990 1999-2004

FIGO stages IB1 - IVa IB1 - IVa

Planning aim EBRT Gy 46 45

Planning aim BT (LDR/PDR) Gy 24 35

Total planning aim dose Gy 70 80

Local control % at 3 years 62 81

NED % at 3 years 52 53

Overall survival % at 3 years 47 54

UMCU data Conclusions

Concomitant chemo might not be enough

Can we improve the radiotherapy technique?

Might MRI guidance help?

Aim

increase loco-regional control

increase survival rates

decrease toxicity

2

Challenges

• Combination of macroscopic tumor (primary and lymph node metastases) and a large elective area (uterus, parametria,

vagina, lympatics)

• Internal motion of primary tumor and organs at risk

• Tumor regression during treatment

• Large difference between tolerance dose to organs at risk (small bowel) and required dose to GTV

OAR

OAR

OAR Small bowel

Cervical tumor

MRI guided UMCU Developments from 2004 on

Imaging • From X-ray to MRI guidance

Radiotherapy technique • EBRT: From conventional irradiation to IMRT and IGRT

• Brachy: From point A to GEC ESTRO recommendations

From 2D to 4D adaptive treatment approach !

Impact of MRI on EBRT

EBRT imaging modalities

Clinical investigation at time of diagnosis

Exophytic tumor in the whole cervix

Infiltration of the middle part of the left parametrium

No infiltration into the vagina

Cystoscopy negative

No hydronephrosis

Rectum not involved

FIGO IIB

EBRT Imaging modalities for nodes

Match PET-MRI

Imaging to define tumor spread CT-PET

EBRT Imaging modalities for cervical tumor

Cervical tumor as depicted on CT or MRI

• FIGO IIB, cT2b

Contrast enhanced CT T2 weighted MRI

3

EBRT Imaging modalities

Cervical tumor as depicted on CT or MRI • FIGO IB2, cT4a uterine corpus, rectum and bladder wall

Advantage of MRI to define primary GTV

MRI versus clinical investigation and CT Overall staging accuracy of MRI is superior to clinical examination or CT with respect to

• parametrial infiltration

• infiltration into the uterus

• infiltration into bladder and rectum Boss et al. 2000, Ozarlak et al. 2003, Bipat et al. 2003

Sagittal Axial

GTV = intermediate or high signal on T2 weighted MR images

Diagnostic imaging for cervix

Staging of uterine cervical cancer with MRI:

Consensus guidelines of the European Society of Urogenital Radiology Eur Radiol 2011,21,1102

Morphologic T2 weighted images

a) Hyperintense lesion invading cervical stroma

b) Low signal intensity excluding parametrial invasion

c) Enlarged lymphnode in para-aortic space

c

Functional imaging

DWI images ADC map (B 800) Review: Imaging cervical cancer: recent advances and future directions.

Curr Opin Oncol. 2011 Sep;23(5):519-25.

“Particularly, functional MRI techniques have improved accuracy of disease staging …….”

“Once standardized and validated, the techniques should enable individualized patient treatment and optimization of outcome.”

Treatment volume definition

X-rays CT MRI

Impact of imaging modality on primary tumor depiction

- PTV - CTV - GTV

Contouring guidelines

Consensus contouring guidelines for 3D treatment planning available for CT and MRI approaches (Taylor 2005, Small 2008, Lim 2011)

• Bare chance on more conformity for target delineation and delineation of organs at risk

• Some discrepancies still exist

• At at this moment not clinically proved

Delineation of the targets and OAR

Consensus guidelines Lim et al 2011

4

Consensus guidelines “Small” versus “Taylor” Small, Red Journal, 2008, 71, 428 Taylor, Red Journal, 2005, 63, 1604

Impact of MRI on EBRT treatment techniques

Conventional

Based on bony anatomy

AP/PA fields

3 - 4 treatment fields

Image based

3DC-RT or IMRT

3DC-RT : 3 - 4 treatment fields

IMRT: 5 or more fields

MRI based treatment planning

3DC-RT

4-field box

box

IMRT

7 beam IMRT

IMRT

VMAT with SIB to pathological nodes

MRI based treatment planning

0

100

200

300

400

500

600

700

bowel 30 Gy bowel 42.75 Gy

vo

lum

e in c

c

conventional

conformal

imrt

0

20

40

60

80

100

120

bladder 30 Gy bladder 42.75 Gy

conventional

conformal

imrt

0

20

40

60

80

100

rectum 30 Gy rectum 42.72 Gy

conventional

conformal

imrt

vd Bunt 2006

IMRT allows organ sparing !

Improved Organ sparing

Chance on • Reduced morbidity • Improved QoL

But !!

Impact on treatment margins Patient related:

• Motion and deformation of tumor and OAR • Tumor regression

Externally:

• Contouring uncertainties • Position verification method • Set-up uncertainties

PTV

Courtesy Taran Paulsen-Hellebust

Higher conformity gives higher chance on geographical miss

5

Inter-fraction motion of primary GTV

GTV CTV PTV

• 5 consecutive MRI’s during EBRT • Impact of changes in bladder and

bowel filling on position changes of uterus

• Not only one organ is responsible

Low impact

Low impact

High impact of bladder and bowel

High impact of bladder Van de Bunt et al 2008

Inter-fraction motion of primary GTV

• Extreme difference in bladder filling • Extreme difference in position of the uterus • 6 – 40 mm in different publications

CTV PTV

Pre-treatment MRI MRI in first week of EBRT

Intra-and inter-fraction motion

week1 week4

• 4 consecutive sagittal images taken within 30 minutes of MR imaging

• Impact of patient’s movement en pelvic rotation!

• Impact of bowel filling

Kerkhof et al 2009

week1 week4 Pre-treatment week2

• Impact of changes in bladder and bowel filling, pelvic rotation

• Impact of volume changes of uterus

Kerkhof 2008

Intra-and inter fraction motion

Inter-fraction motion of node GTV

• Affected nodes change their position (< 8 mm) • Position shifts less extensive than for primary GTV

Maximal margins (computer)

0

1

2

3

4

5

6

7

8

9

med lat ant post sup inf

marg

in (

ml)

• 48 nodes, 15 patients • Position shifts in 6 directions

Schippers, submitted

Position shifts of primary tumor and nodes

6

Motion depicted on repetitive MRI

• Organ motion and deformation contribute essentially to the treatment margin concept !

• Order of magnitude between position shifts of primary tumor and affected nodes is different !

MRI to monitor primary tumor regression

• Primary tumors individually shrink during EBRT

• Impact on elective CTV and PTV is low

Before treatment After 30 Gy EBRT

0

100

200

300

400

500

600

700

800

vo

lum

e i

n c

c

GTV CTV PTV

pre-treatment intra-treatment

GTV 49.2%

CTV 13.4%

PTV 9.7%

Average decrease in volume after 30 Gy

MRI to monitor nodal tumor regression

At diagnosis

Enlarged (15 mm) Irregular shape Inhomogeneous

4th week of EBRT

Partial response in week 4 EBRT total dose = 55 Gy (45 + 10)

Node > 10 mm on MRI, PET positive, boost given

Treatment margins

PTV

bowel

rectum

GTVprim

GTV nodes

GTV

CTV

Recommended CTV-PTV margins for elective EBRT

• For primary CTV 15 – 20 mm • For nodal regions 10 mm • Position verification on bony landmarks, off-line protocol • Under investigation for soft tissue contrast, on-line protocols, replanning

Impact of treatment margins on coverage

Planning aim dose:

50 Gy

Comparison:

Four Field Box Large margin IMRT 10-20 mm Small margin IMRT 5 mm For GTV en CTV not that much difference between planning aim and delivered dose

Pelvic Radiotherapy for Cancer of the Cervix: Is What You Plan Actually What You Deliver?

Lim et al 2009

Ongoing

Image based margin concept for highly conformal adaptive EBRT

• Currently subject of investigation (repetitive CT/MRI)

• Promising developments • Image based adaptive “plan of the day” strategies

• Several treatment plans based on repetitive CT with variable bladder filling (Rotterdam group)

• Margin choice depends on position verification method

• Bony anatomy, soft tissue, internal markers • on-line, off-line

7

Ongoing research

Adaptive RT: variable bladder filling scans • Separate movers/non-movers: large/smaller margins →

direct clinical gain

• CBCT based introduction and monitoring

• Repeat MR

30 mm 5 mm

Position verification

Intergration of Cone-beam CT and accelerator

Intergration of MRI and accelerator

EPID • Bony anatomy • Gold markers

Possible gain of an MRI accelerator

Online MRI guidance Reduction CTV-PTV margin to 4 mm Clear reduction of in-field amount of OAR

Kerkhof 2008

MRI in postoperative situation

15 mm 0 mm Superior

28 mm 11.5 mm Posterior

13 mm 4 mm Anterior

10 mm 3 mm Right lateral

20 mm 5 mm Left lateral

max. in 95% median Margins

Endometrium en cervical cancer after hysterectomy 5 consecutive MRI’s during EBRT

5 CTV’s with one encompassing PTV

Toet et al 2007

Impact of MRI on Brachytherapy

Regions of interest for Brachytherapy

• macroscopic tumor after EBRT

• suspected microscopic tumor extension

• pre-treatment tumor extension

• most exposed volume of rectum, sigmoid and bladder

8

From ICRU rules to GEC ESTRO recommendations

According to Manchester System/ICRU 38 • X-ray guided (2D)

• Point A

• ICRU rectum point

• ICRU bladder point

• Point B

• 60 Gy volume

• TRAK

According to GEC ESTRO and ABS

recommendations • MRI/CT guided adaptive

• GTV (macroscopic tumor)

• HR-CTV (GTV + suspected microscopic tumor)

• IR-CTV (pre-treatment extension)

• 2cc rectum

• 2cc sigmoid

• 2cc bladder

• TRAK

Pötter et al., Radiotherapy and Oncology, 2006, 78, 67

ICRU 38, 1985

X-ray guided brachytherapy

Standard type Fletcher application for LDR, PDR or HDR treatments

Standard pear shaped isodose distributions

5.1 cm

4.9 cm

3.1 cm

Iinitial

aEllipsoid formula applied: V=width*thickness*height*0.5

At BT

4.7 cm

2.9 cm

3.6 cm

aV= 24.8 cm3

aV= 38. cm3

MRI guided brachytherapy MRI guided delineation

MRI with applicator in situ Regions of interest delineated

sagittal

axial

coronal

GTV

HR-CTV

IR-CTV

bladder

rectum

bowel

X-ray guided versus MRI guided intracavitary brachy planning

Ke1 hr-ctv bladder rectum sigmoid bowel

in cGy (PD 60 cGy) D90 D2 D2 D2 D2

conventional plan 48.4 60.0 43.1 51.2 49.8

optimized plan 53.1 62.0 49.3 53.6 43.5

Improved dose volume parameters !

Aimed tumor dose EBRT+ BT > 84 Gy (60cGy per PDR pulse)

OAR constraints D2cc • < 90 to Bladder (63 cGy pp)

• < 75 to rectum, sigmoid, bowel (53 cGy pp)

Locations with inadequate coverage by intracavitary applications

A1 A2

Inverse optimization

Extra needles

25% of the

cases

sigmoi

d bowel

s

10% of the

cases

90% of the

cases

Results of planning study with 20 applications

75% of the

cases

9

Adaptation of a the applicator

Applicator produced by Nucletron

150

250

Start clinical introduction July ‘07

Two BT applications First one without needles

Second with needles, if necessary

• Pre-planning for 2nd application based on dose distribution and DVH analysis of first application:

• Needles necessary?

• Where? How deep?

From 3D to 4D

Example

2 needles left side

treatment 3 hr-ctv bladder rectum sigmoid bowel

in cGy (PD 60 cGy) D90 D2 D2 D2 D2

conventional plan 38,3 72,9 39,0 27,6 38,5

optimized plan IC 35,0 64,3 32,3 26,9 38,4

optimized plan IC/IS 3ndl 68,7 56,2 48,5 36,6 32,6

Improved coverage by combining intracavitary and interstitial approach

Optimized IC

Optimized IC/IS 3 needles used

Aimed tumor dose EBRT+ BT > 84 Gy (60cGy per PDR pulse)

OAR constraints D2cc • < 90 to Bladder (63 cGy pp)

• < 75 to rectum, sigmoid, bowel (53 cGy pp)

Effect of IC/IS applicator on HR-CTV dose during learning curve

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

D90

hr-

ctv

in G

yab

10

patient nr

optimized IC

optimized IC/IS

Effect of IC/IS applicator

Optimized plan IC/IS – Optimized plan IC

(in Gyab10)

N=15 avg SD

D90 5.3 3.3

D100 3.9 2.6

EQD2 in Gyab ‘optimized’ plan

average SD

HR-CTV 88 7

Bladder 83 6

Rectum 67 6

Sigmoid 61 5

Bowel 64 9

Dosimetric gain for one application Total dose EBRT + BT

10

D90 HR-CTV as function of Volume HR-CTV

50

55

60

65

70

75

80

85

90

95

100

105

110

115

120

0 20 40 60 80 100 120 140

volume HR-CTV in cc

D9

0 H

R-C

TV

Gyab

filmplan

filmpl scaled

Log. (filmplan)

Log. (filmpl scaled)

50

55

60

65

70

75

80

85

90

95

100

105

110

115

120

0 20 40 60 80 100 120 140

volume HR-CTV in cc

D9

0 H

r-C

TV

Gyab

optimized

IC/IS

Log. (optimized)

MRI guided optimization and additional interstitial component allow to cover larger HR-CTV volumes adequately, has dosimetric benefit

standard optimised

Uncertainties OAR motions and volume changes during HDR

• Repetitive images around HDR fraction • One for treatment planning, one directly before and one after irradiation

• Protocol for bowel preparation

• Protocol to controlled bladder filling

• Differences in bladder filling

• Differences in rectum filling: Gas! • After adaptation comparable DVH parameters

MRI for HDR –treatment planning

MRI pre-irradiation about 1.5 hours later

MRI after insertion of rectal probe

MRI after HDR about 20 min later

Work in progress

Repeated and interventional imaging easier to realize when MR machine is part of brachytheater

• Application, imaging and irradiation in the same room

• Offers possibility to adapt when necessary

• Is less time consuming

• Offers possibility to work with new tools

Courtesy R.Moerland R.Schokker

Difference PDR / HDR

Christel Nomden et al

Treatment periods in UMCU

MRI guided EBRT and BT from 2006 on

46 Gy EBRT 24 Gy LDR-BT 14 Gy Boost (P/N)

Traditional treatment prior 1999

Eventually hyperthermia, IC BT, parametrial boost if indicated

45 Gy EBRT 35 Gy PDR-BT

Traditional treatment 1999-2004

Combined with cisplatinum, ICBT, parametrial/nodal boost if indicated

45 Gy EBRT 40Gy PDR-IGBT 10-14 Gy Boost (N)

MRI guided treatment 2006-2011

Combined with cisplatinum, IC or IC-IS BT, nodal boost if indicated

45 Gy EBRT 40Gy HDR-IGABT 10-14 Gy Boost (N)

MRI treatment 2012-currently

Combined with cisplatinum, IC or IC-IS BT, nodal boost if indicated

14 Gy Boost (P/N)

Retro-_

Impact of MRI guidance on outcome

UMCU RT (± HT) ≤ 1999

RT ± CT 1999-2004

RT ± CT 2006-2009

(Retro-EMBRACE)

RT ± CT 2008-

(EMBRACE)

RT ± CT 2015-

(EMBRACE II)

Number of patients

95 54 46 68

Median FU in months

49 46 41

Years of treatment

1983 – 1990 1999-2003 2006-2008

FIGO stages IB1 - IVa IB1 - IVa IB1 - IVa

Planning aim dose Gy EQD2

> 70 > 80 > 84

Local control % 62 81 93

PFS % (3 years) 52 53 71

OS % (3 years) 47 54 65

Late ≥ G3 morbidity %

ns 20

(RTOG)

9.5

(CTCAE 3.0)

11

Ongoing and future UMCU effort

IGART for primary and post-op treatments • Treatment margin reduction

• SIB techniques (vulva, vagina, pelvic and PAO nodes)

• MRI based daily position verification

• Deformable registration tools

• Adaptive re-planning strategies

Affected nodes/Nodal recurrence • Detection of small nodes possible on anatomical MRI

• Impact of PET-CT, contrast enhanced or functional MRI

• More precise dose calculation from EBRT and BT

• Better understanding of dose response relation

All Node positive Node negative

n=46 n=20 n=26

Three years outcome % % % P

Overall survival 65 50 77 0.032

Cancer specific survival 76 60 89 0.022

Local control 93 94 92 0.799

Pelvic control 84 78 88 0.370

Distant metastasis free survival 77 68 85 0.143

Progression free survival 71 53 85 0.013

Nodal recurrences

13 % UMCU MRI guided treatment 2006-2011

Patterns of regional recurrence after standard chemo-radiotherapy

Beadle et al 2010, MD Anderson

> 1800 cervical cancer patients, definitive radiotherapy • Pelvic control about 82 % • Local control about 86 %

Inside the field = dosimetric problem Outside and marginal = geographical miss

Node example on MRI

At diagnosis

Irregular appearance

and irregular border

4th week of EBRT

Nearly gone

in week 4

EBRT dose 45 Gy

6 weeks after treatment

Nearly no rest

6 weeks after

treatment

12 months after treatment

Nodal failure

Node 8 mm, PET negative, no boost

Summary

• MRI guided treatment for cervical cancer

– Supports better defining the disease at time of diagnosis

– Supports monitoring motion and deformation during treatment

– Supports adaptive treatment approaches

– Supports development of advanced EBRT and BT techniques

– Results in improved local control and survival

– Will hopefully help to improve nodal control

– Will hopefully help to reduce morbidity and improve QoL (also important for intensifying chemo)

UMCU Team

Judith Roesink Robbert Tersteeg Christel Nomden Astrid de Leeuw Petra Kroon -van Loon Rien Moerland Hans de Boer Marielle Phillipens Nico van den Berg Many others Radiologists Gynecologists Medical Oncologists Pathologists