deﬁnition of the supraclavicular and infraclavicular...

Chika N. Madu, BSDouglas J. Quint, MDDaniel P. Normolle, PhDRobin B. Marsh, CMDEdwin Y. Wang, MDLori J. Pierce, MD

Index terms:Breast neoplasms, 00.32Breast neoplasms, therapeutic

radiology, 00.125Lymphatic system, CT, 997.12912,

997.92Lymphatic system, therapeutic

radiology, 997.33, 997.92Treatment planning

Published online before print10.1148/radiol.2212010247

Radiology 2001; 221:333–339

Abbreviations:IFV 5 infraclavicularIMRT 5 intensity-modulated

radiation therapySCV 5 supraclavicular

1 From the Department of RadiationOncology (C.N.M., D.P.N., R.B.M., L.J.P.),Department of Radiology (D.J.Q., E.Y.W.)University of Michigan School of Medi-cine, 1500 E Medical Center Dr, AnnArbor, MI 48109. Received January 2,2001; revision requested February 17;revision received May 7; accepted May10. Address correspondence to L.J.P.(e-mail: [email protected]).© RSNA, 2001

Author contributions:Guarantor of integrity of entire study,C.N.M., L.J.P., D.J.Q.; study conceptsand design, all authors; literature re-search, C.N.M., L.J.P.; clinical studies,L.J.P.; data acquisition, C.N.M., D.J.Q.,L.J.P., R.B.M., E.Y.W.; data analysis/in-terpretation, all authors; statisticalanalysis, D.P.N.; manuscript prepara-tion, definition of intellectual content,editing, revision/review, and final ver-sion approval, all authors.

Definition of theSupraclavicular andInfraclavicular Nodes:Implications forThree-dimensional CT-basedConformal RadiationTherapy1

PURPOSE: To delineate with computed tomography (CT) the anatomic regionscontaining the supraclavicular (SCV) and infraclavicular (IFV) nodal groups, to definethe course of the brachial plexus, to estimate the actual radiation dose received bythese regions in a series of patients treated in the traditional manner, and tocompare these doses to those received with an optimized dosimetric technique.

MATERIALS AND METHODS: Twenty patients underwent contrast material–en-hanced CT for the purpose of radiation therapy planning. CT scans were used tostudy the location of the SCV and IFV nodal regions by using outlining of readilyidentifiable anatomic structures that define the nodal groups. The brachial plexuswas also outlined by using similar methods. Radiation therapy doses to the SCV andIFV were then estimated by using traditional dose calculations and optimizedplanning. A repeated measures analysis of covariance was used to compare the SCVand IFV depths and to compare the doses achieved with the traditional andoptimized methods.

RESULTS: Coverage by the 90% isodose surface was significantly decreased withtraditional planning versus conformal planning as the depth to the SCV nodesincreased (P , .001). Significantly decreased coverage by using the 90% isodosesurface was demonstrated for traditional planning versus conformal planning withincreasing IFV depth (P 5 .015). A linear correlation was found between brachialplexus depth and SCV depth up to 7 cm.

CONCLUSION: Conformal optimized planning provided improved dosimetric cov-erage compared with standard techniques.

Increasing emphasis is being placed on the importance of regional therapy in the man-agement of early-stage breast cancer. Recent randomized trials have shown a significantbenefit in survival by the addition of postmastectomy radiation therapy in patients withpositive axillary nodes (1–3). Comprehensive radiation therapy fields used in these trialshave included not only the chest wall but also the supraclavicular (SCV), infraclavicular(IFV) (axillary apex), internal mammary, and axillary lymph nodes. Although the exactcontribution of regional treatment to the survival benefit is uncertain, it is clear from thesestudies that treatment to regional sites is of therapeutic benefit.

Management of the axilla in the treatment of breast cancer is primarily surgical, withaxillary recurrence in approximately 1%–2% of patients following a standard level I, IIdissection (4–6). However, as the number of involved axillary nodes identified in a levelI, II dissection increases, the risk of involvement of level III, also known as the apex of the

Radiation Oncology

333

axilla or the IFV region, and the SCVnodes increases as well (7,8). Radiationtherapy has traditionally been used totreat patients at high risk for microscopicresidual disease of the SCV and IFV re-gions, resulting in rates of regional failureas first failure of only 0.0%–1.5% (4,5).

Historically, the SCV nodes and the ax-illary apex have been treated with radia-tion therapy by using a single anteriorfield, as shown in Figure 1, with full doseprescribed to a point in the SCV fossa 3cm deep to the surface of the skin. Thisfield has generally extended superiorly tothe thyrocricoid membrane, inferiorly tothe inferior aspect of the clavicular head,medially to the lateral aspect of the ster-nocleidomastoid muscle, and laterally tothe humeral head. Although the nodes inthese regions considered at risk for micro-scopic disease are present within these fieldborders, tissues not at risk, including shoul-der musculature, are also included, whichincreases the potential for radiation-associ-ated complications (9). Uniform dosing toa depth of 3 cm in all patients for an arbi-trarily defined nodal volume is also of con-cern, given the range in body habitus,weights, and depth of subcutaneous adi-pose tissue among individuals.

Therefore, it is desirable to define thenodal regions at risk for spread of breastcancer such that full-dose radiation ther-apy could be delivered conformally tothese target volumes. It is also importantto estimate the dose actually received byradiation-sensitive structures, such as thebrachial plexus, by using traditional do-simetry techniques and to compare thesevalues with the doses received by usingradiation therapy optimized to treat tar-get nodal volumes. Thus, the goals of thisstudy were to delineate on computed to-mographic (CT) scans the anatomic re-gions containing the SCV and ICV nodalgroups, to define the course of the bra-chial plexus, to estimate the actual radi-ation dose received by these regions in aseries of patients treated in the tradi-tional manner, and to compare thesedoses with those received by using anoptimized dosimetric technique.

MATERIALS AND METHODS

Contrast material–enhanced CT scansobtained in 20 patients consecutivelytreated with radiation therapy at the Uni-versity of Michigan School of Medicine,Ann Arbor, were used to study the anat-omy of the SCV fossa and axillary apex.Each patient had been diagnosed withstage II breast cancer with positive axil-

lary nodes and had undergone a modi-fied radical mastectomy with a level I, IInodal dissection. Three patients had bi-lateral disease; thus, 23 SCV regions wereavailable for evaluation. Of the 23 eligi-ble regions, two could not be used due topoor opacification of the blood vessels inthe SCV fossa. Following completion ofchemotherapy, customized cradles (Alpha;KGF, Chesterfield, Mich) (1) for each pa-tient were fabricated for immobilizationduring radiation therapy planning andtreatment with the patient’s ipsilateral (orbilateral) arm or arms above the head.

CT was then performed for three-di-mensional treatment planning of thechest wall, supraclavicular fossa, and su-perior internal mammary nodes per rou-tine protocol for treatment of node-pos-itive disease. Scanning was performedwith a variety of third-generation CTscanners with a field of view selected toinclude the entire patient. Patient CT sec-tions were contiguously obtained at 3- or5-mm intervals from the mid-neck to thediaphragm during bolus administrationof nonionic intravenous contrast mate-rial. For this study, the CT images in-cluded an area that extended from theneck to the inferior aspect of the clavic-ular head. With the aid of a senior radi-ologist (D.J.Q.), the following structureswere carefully outlined on each CT sec-tion by using the three-dimensional sys-tem: the anterior scalene muscle, the bra-chial plexus, the carotid sheath, thesternocleidomastoid muscle, the pectora-

lis major and minor muscles, and thesubclavian artery.

Anatomically, the SCV fossa is subdi-vided into two compartments: the lesserSCV fossa, which is the depression be-tween the two heads of the sternocleido-mastoid muscle, and the greater SCVfossa, which is at the base of the posteriortriangle of the neck and is also referred toas the omoclavicular triangle (10). Theentire SCV fossa, defined as the compos-ite of these two compartments, was thenoutlined on each CT section by usingreproducible anatomic boundaries. Medi-ally, the SCV fossa was considered to ex-tend to the lateral edge of the trachea,excluding the thyroid gland and thyroidcartilage superiorly. Anteriorly, the SCVfossa was bounded by the deep surface ofthe sternocleidomastoid muscle and thedeep cervical fascia. The posterolateralborder of the fossa was considered to beat the anterior and medial borders of theanterior scalene muscle, while the pos-teromedial border extended medially tothe carotid artery and internal jugularvein. At the inferior aspect of the SCVfossa, the posterior border was defined bythe subclavian artery.

The IFV fossa was also outlined onsuccessive CT sections by using easilydelineated anatomic structures. The su-perior border was defined as the mostsuperior aspect of the pectoralis minormuscle. The inferior border was definedat the level of the insertion of the clav-icle into the manubrium. The IFV fossa

Figure 1. Standard anterior radiation therapy field used to treat thesupraclavicular nodes and the apex of the axilla. Yellow lines encom-pass the field with a spinal cord block.

334 z Radiology z November 2001 Madu et al

extended laterally to the medial borderof the pectoralis minor muscle and me-dially to the lateral edge of the clavicle.The anterior extent of the fossa wasconsidered to be the deep surface of thepectoralis major muscle and the poste-rior extent was defined by the subclavi-an-axillary artery.

The course of the brachial plexus wasalso outlined on each CT section. Thebrachial plexus usually arises from theroots of the fifth cervical (C5) throughthe first thoracic (T1) nerves. At approx-imately the T1 vertebral level, the bra-chial plexus courses between the anteriorand middle scalene muscles. In the infe-rior portion of the supraclavicular fossa,the brachial plexus closely follows thepath of the subclavian artery. Specifi-cally, the neural bundles are located justanterior to the subclavian artery, andthen course immediately posterior andparallel to the subclavian artery. The ter-minal branches of the plexus then sur-round the subclavian-axillary artery.Thus, even though the brachial plexuscan be difficult to identify on transverseCT images, the anterior and middlescalene muscles and subclavian-axillaryarteries can be used to identify the ex-pected course of the plexus (11).

Bilateral neck dissections were carriedout on two cadavers to confirm the loca-tion of the SCV lymph node chains. Spe-cifically, by using blunt dissection, a ver-tical incision was made along the anteriormidline extending from just below themandible to the inferior aspect of themanubrium. Another incision was madefrom the most lateral aspect of the neck

just beneath the angle of the mandibletoward the acromion. Both cuts werethen connected superiorly by a horizon-tal incision running just below the man-dible. The skin, subcutaneous tissue, andthe platysma muscle were reflected downto the level of the manubrium. The deepcervical fascia investing the sternocleido-mastoid muscle was removed, and thesternocleidomastoid muscle was severedfrom its nuchal attachment to expose theSCV fossa. The omohyoid muscle wasalso detached from its hyoid attachmentand reflected laterally. By using a saw, themiddle one-third of the clavicle was re-moved, and the pectoralis major muscleattached to that portion of the claviclewas reflected toward the chest wall togain easy access to the SCV fossa. Thefossa was essentially a space filled withadipose tissue containing several lymphnodes embedded at various depths. Theinternal jugular and carotid vessels werenoted as the medial extent of this fossa.The fatty tissue in the SCV fossa waslifted and revealed the scalene musclesand the subclavian artery on the floor ofthe fossa.

Radiation therapy doses to the SCVand IFV regions were estimated by usingboth a traditional and a conformal opti-mized technique. For the traditionalmethod, a single anterior treatment fieldwas set at 97 cm source-to-skin distanceby using an asymmetric inferior field bor-der such that the lower jaw was placed atthe central axis, at the base of the head ofthe clavicle. The gantry was angled 10°away from the spinal cord, and a focusedblock was placed to maintain an off-cord

approach. The dose to the SCV field wascalculated to a point located 1.0–1.5 cmsuperior to the clavicular head approxi-mately 3 cm lateral to the block edge, at adepth of 3 cm. Doses received at the 90%isodose surface for the SCV and IFV vol-umes were then calculated by using theUM Plan planning system (12).

For the conformal optimized plans, thegantry angle was modified in 13 of 21cases (15° angle, six patients; 20° angle,six patients; and 25° angle, one patient)to cover the target SCV and IFV volumes.The depth of normalization was deter-mined such that the 90% isodose surfacecovered the optimized SCV and IFV vol-umes. Six-megavolt photons and lung in-homogeneity corrections were used forboth calculation techniques. The depthto the deepest aspect of the target volumewas measured with the gantry angle set at10° in all patients.

A repeated measures analysis of covari-ance was used to compare the IFV andSCV depths and doses between the tradi-tional and optimized methods. Thismethod is analogous to a paired-compar-ison t test in that it is used to explicitlyestimate the within-patient correlationbetween plans. These models were fit byusing software (PROCMIXED in SAS LINUX,version 8.2; SAS Institute, Cary, NC).

The minimum (min) and maximum(max) depths from the skin to the bra-chial plexus nerves were measured oneach scan. The mean brachial plexus (BP)depth, calculated as (BP min 1 BP max)/2,was then correlated to the depth of theSCV volume.

Figure 2. (a) Anatomy of the supraclavicular fossa taken from a cadaveric dissection and (b) the accompanying line drawing of the structures. Thesternocleidomastoid muscle has been removed to display underlying structures. 1 5 omohyoid muscle, 2 5 sternocleidomastoid muscle (sternalhead), 3 5 sternocleidomastoid muscle (clavicular head), 4 5 subclavian vein, 5 5 subclavian artery, 6 5 brachial plexus, 7 5 pectoralis majormuscle (reflected), 8 5 clavicle, 9 5 trapezius muscle, 10 5 supraclavicular lymph nodes.

Volume 221 z Number 2 Definition of Supraclavicular and Infraclavicular Nodes: Radiation Therapy Planning z 335

RESULTS

Figure 2a shows the left SCV fossa from acadaver, with the sternocleidomastoidmuscle transected to expose underlyingstructures. It is accompanied by a dia-gram illustrating the pertinent anatomi-cal structures (Fig 2b). Figure 3 shows thecorresponding structures on successivetransverse CT scans from one patienttreated with postmastectomy radiationtherapy. The SCV fossa is medial at thesuperior aspect of this region (Fig 3, a–e)and then extends laterally as the nodalregions approach the clavicle (Fig 3, f–j).The superior aspect of the IFV region is

noted in Figure 3, e, located at the supe-rior and medial extent of the pectoralisminor muscle. The medial extent of thisregion approaches midline inferiorly as itfollows the course of the clavicle (Fig 3, k).

Figure 4 shows a composite view of theSCV and IFV volumes superimposed on astandard anterior SCV radiation therapyfield. As compared with the SCV fieldshown in Figure 1, restricting radiationtherapy dose to defined nodal volumes atrisk spares tissues at the lateral and supe-rior aspects of a standard SCV field.

Of the 21 scanned regions suitable forstudy, the median maximum depth ofthe SCV nodes was 5.0 cm (range, 3.9–

8.3 cm). In the optimized conformalplans, 100% of the SCV volume was cov-ered by the 90% isodose surface. Thiscompares with a median of 93.3% of theSCV volume (range, 98.2%–44.7%) cov-ered by the 90% isodose surface in thetraditional plans. As shown in Figure 5,however, as the depth to the SVC nodesincreases, the percentage of the SCV vol-ume encompassed within the 90% isod-ose surface significantly decreases forcases of therapy planned by using tradi-tional planning versus the conformal op-timized plan (P , .001).

The maximum median depth of theIFV nodes as measured in the 20 CT scans

Figure 3. Transverse CT sections at 3-mm intervals (window width, 500 HU; window level, 50 HU) with outlines of the SCV and IFV fossae andthe course of the brachial plexus, with the muscles and vessels that radiographically define these volumes and nerves. CT section in a is at the levelof the thyrocricoid membrane, which clinically corresponds to the superior level of the SCV fossa. Successive 3–5-mm interval sections are presentedto the inferior level of the IFV fossa, which corresponds clinically to the base of the clavicular head (Fig 3 continues).


was 5.6 cm (range, 3.3–7.3 cm). One hun-dred percent of the IFV volume was en-compassed by the 90% isodose surface inthe optimized plan; this compares with amedian of 87.7% (range, 100.0%–47.8%)of the IFV volume covered by the 90%isodose surface in the traditional plan.Figure 6 shows the comparison of the IFVdepth with percentage of the IFV volumewithin the 90% isodose surface and dem-onstrates a significant difference betweenthe regression lines for the conformal op-timized and traditional plans (P 5 .015),which suggests that a decreased radiation

dose would be delivered to nodes at adepth of 4 cm and greater.

Because the proximal and distal courseof the brachial plexus nerves and thepatient body habitus varied, the distancefrom the skin to the plexus varied accord-ing to position of the nerves and accord-ing to each patient. The median minimumdepth to the plexus, per measurementsfrom CT scans, was 2.3 cm (range, 0.5–4.0cm) and median maximum depth to theplexus was 6.5 cm (range, 3.8–9.1 cm). TheSCV depth was correlated (r 5 0.77, P ,.01) with the mean brachial plexus, repre-

sented by BP, depth calculated with (BPmin 1 BP max)/2, as shown in Figure 7.The relationship appeared to be linearwhen SCV depth was less than 7 cm; therewere not enough subjects with mean SCVdepth greater than 7 cm to determine withcertainty if the mean brachial plexus depthstopped increasing when the SCV depthexceeded 7 cm. The doses to the brachialplexus volume for the techniques of thetwo plans were then compared by usingthe calculated dose distributions. Eighty-one percent (range, 49.0%–95.4%) of thebrachial plexus was encompassed within

Figure 3 (continued). Transverse CT sections at 3-mm intervals (window width, 500 HU; window level, 50 HU) with outlines of the SCV and IFVfossae and the course of the brachial plexus, with the muscles and vessels that radiographically define these volumes and nerves. CT section in ais at the level of the thyrocricoid membrane, which clinically corresponds to the superior level of the SCV fossa. Successive 3–5-mm interval sectionsare presented to the inferior level of the IFV fossa, which corresponds clinically to the base of the clavicular head.


the 90% isodose surface in the optimizedplans versus 42.2% (range, 2.4%–78.8%) inthe traditional plans.

DISCUSSION

By using information obtained from ca-daver dissection and CT-based imaging,we have identified anatomic structuresthat localize the regions containing theSCV and IFV lymph nodes. We have alsoshown the range in depths of these nodesin a cohort of patients. In identifyingthese structures, we have superimposedthe dose calculations traditionally deliv-ered to these regions and have estimatedthe doses actually received by thesenodal groups. We have shown that forboth the SCV and IFV nodes, prescriptionof dose to a uniform depth of 3 cm sig-nificantly underdoses these regions asthe nodal depths exceed 3 cm when com-pared with a conformal optimized tech-nique. Thus, body habitus and patientseparation (thickness) can significantlyalter dose delivered to these nodal groupsunless the actual location of these nodescan be identified and the plans can beoptimized appropriately.

As shown in the comparison of Figures1 and 4, more precise definition of thesenodal regions can minimize unnecessarytreatment to uninvolved tissues. Regionalradiation therapy can significantly in-crease shoulder movement dysfunctionas shown in the functional analysis ofshoulder motion following the Danishtrials in which women were randomlyassigned to receive postmastectomy radi-ation therapy (9). Thus, delivery of radi-ation therapy to only those tissues thatrequire treatment could potentially resultin decreased rates of radiation therapy–related shoulder morbidity. It is also pos-sible that the occurrences of arm edemacould be decreased by using radiationtherapy fields tailored to specific individ-ualized anatomic regions. The lateral ex-tent of the supraclavicular field has beenimplicated in the risk of radiation-associ-ated lymphedema, with extension of thelateral border of the SCV field associatedwith increasing risk of arm edema (13).As shown in Figure 4, additional sparingof lateral tissues is possible with carefuldelineation of the SCV and IFV fossae. Asboth limited arm mobility and lymph-edema can result in permanent arm dys-function in breast cancer survivors, mea-sures should be taken which could improvelong-term functional outcome.

The pathway of the brachial plexusthrough the SCV region has been definedby using structures that are readily iden-

tifiable on CT scans, such as the anteriorand middle scalene muscles and the sub-clavian artery. We have shown the vari-ation in the depth of the brachial plexusfrom its proximal to its distal extent be-tween patients. Estimations of dose re-ceived to the brachial plexus for tradi-tional versus optimized planning revealedthat higher doses are received to theplexus by using the optimized plans. Al-though not well documented, clinicalstudies of peripheral nerve tolerance sug-gest a 5% occurrence of radiation-in-duced brachial plexus injury at doses of60 Gy in 2-Gy fractions (14). This dosethreshold is consistent with the 0%–1%rate of brachial plexopathy observed inpatients treated with breast and regionalirradiation for breast-conserving therapy,where doses of 46–50 Gy have been de-livered to a traditional SCV field to adepth of 3 cm (15,16). Thus, despite thedelivery of a higher dose of radiation tothe brachial plexus volume by using theconformal optimized technique, a higherrate of complications would not be ex-pected with the dose restricted to 50 Gy.Knowledge of the course of the brachialplexus with respect to the SCV nodes shouldbe helpful, however, in estimating the riskof radiation-induced brachial plexus injuryin patients with bulky SCV nodes requir-ing high-dose radiation therapy.

In recent years, there has been substan-tial interest and work throughout the ra-diation oncology community involvingthe development of conformal therapy(17,18). Conformal radiation therapy re-fers to the delivery of radiation which

conforms in three dimensions to theshape of the defined target or targetswhile at the same time minimizes dose tonormal tissue. Various optimization algo-rithms or inverse planning techniqueshave been clinically applied to improvedose conformation. The most promisingoptimization technique currently in useis called intensity-modulated radiationtherapy (IMRT) in which the usual reli-ance on uniform-intensity radiation fields

Figure 4. Registration of the SCV nodal volume (in blue) and theinfraclavicular nodal volume (in red) on the standard clinical anteriorSCV radiation therapy field.

Figure 5. Line graph shows comparison ofthe depth to the SCV nodes and the volumewithin the 90% surface for the traditional andconformal optimized plans.


is replaced by a variable-intensity patternthat is determined with the aid of a com-puterized optimization algorithm (19).The combination of IMRT delivery withinverse planning optimization is expectedto achieve better dosimetric results thannormal plans. It is hopeful that IMRT willresult in either the improvement of localdisease control due to improved coverageof the target or reduced normal tissuedose while achieving the same tumorcoverage. To achieve these goals, a de-fined clinical target must be determined

that allows not only restriction of thetreatment field to the designated targetsbut also allows three-dimensional deliv-ery of therapy. In the case of breast can-cer treatment, the simple traditional par-adigm of treatment delivery in whichseparate fields are used to treat the breastor chest wall and the SCV or IFV regionscould be replaced by a three-dimensionaltreatment approach to a more carefullydefined target volume that includes thebreast or chest wall and specific nodalvolumes. Thus, the tissues in the thoraxand neck or shoulder region that do notrequire treatment could be spared. Whilethis is currently an area of intense re-search, defining nodal volumes, as donehere, is the first step to achieving dosedelivery optimization.

There may be limitations in the inter-pretation of these results. In patients atrisk for microscopic nodal disease, the SCVnodes cannot be visualized on CT scans.Thus, surrogate structures are used toidentify the tissue volumes containing thenodes. For that reason, after thoroughlyreviewing the regional anatomy in ana-tomic textbooks, we dissected the neckregions bilaterally in two cadavers andconfirmed the location of these nodes, asshown in Figure 2a. This series contains alimited number of patients, which limitsthe power of these observations. Despitethe small number of patients, however,this series contains a spectrum of bodyhabitus which allowed us to demonstratesignificant dose differences between tra-ditional and optimized radiation therapyplans with respect to tissue depth. There-fore, our quantitative results should beapplicable to larger data sets.

In summary, the three-dimensionaltissue volumes containing the SCV andIFV lymph nodes were defined on CTscans by using readily identifiable ana-tomic landmarks. Optimized radiationtherapy planning provided improvedcoverage of these target structures com-pared with traditional planning, withgreater sparing of uninvolved tissues. Thedevelopment of this systematic approachin defining the structures at risk and con-forming therapy to these target structuresis a necessary step toward the next levelof dose optimization in the use of radia-tion therapy for carcinoma of the breast.

References1. Overgaard M, Hansen PS, Overgaard J, et al.

Postoperative radiotherapy in high-risk pre-menopausal women with breast cancer whoreceive adjuvant chemotherapy: Danish BreastCancer Cooperative Group 82b Trial. N EnglJ Med 1997; 337:949–955.

2. Overgaard M, Jensen MB, Overgaard J, et al.

Postoperative radiotherapy in high-risk post-menopausal breast-cancer patients given ad-juvant tamoxifen: Danish Breast Cancer Co-operative Group DBCG 82c randomised trial.Lancet 1999; 353:1641–1648.

3. Ragaz J, Jackson SM, Le N, et al. Adjuvantradiotherapy and chemotherapy in node-positive premenopausal women with breastcancer. N Engl J Med 1997; 337:956–962.

4. Halverson KJ, Taylor ME, Perez CA, et al.Regional nodal management and patterns offailure following conservative surgery and ra-diation therapy for stage I and II breast cancer.Int J Radiat Oncol Biol Phys 1993; 26:593–599.

5. Recht A, Pierce SM, Abner A, et al. Regionalnodal failure after conservative surgery andradiotherapy for early-stage breast carci-noma. J Clin Oncol 1991; 9:988–996.

6. Vicini FA, Horwitz EM, Lacerna MD, et al.The role of regional nodal irradiation in themanagement of patients with early-stagebreast cancer treated with breast-conservingtherapy. Int J Radiat Oncol Biol Phys 1997;39:1069–1076.

7. Dahl IE. Recherches sur les metastases mi-croscopiques des cancers du sein dans les gan-glions lymphatiques parasternaux et suscla-vicularies. Mem Acad Clin 1952; 78:651–654.

8. Smith JA, Gamez-Araujo JJ, Gallager HS,White EC, McBride CM. Carcinoma of thebreast: analysis of total lymph node involve-ment versus level of metastasis. Cancer 1977;39:527–532.

9. Hojris I, Andersen J, Overgaard M, OvergaardJ. Late treatment-related morbidity in breastcancer patients randomized to postmastec-tomy radiotherapy and systemic treatmentversus systemic treatment alone. Acta Oncol2000; 39:355–372.

10. Hollinshead N. The head and neck. In: Anat-omy for surgeons. 3rd ed. Philadelphia, Pa:Harper & Row, 1982; 444–445.

11. Gebarski KS, Glazer GM, Gebarski SS. Bra-chial plexus: anatomic, radiologic, and patho-logic correlation using computed tomography.J Comput Assist Tomogr 1982; 6:1058–1063.

12. Fraass BA, McShan DL. 3-D treatment plan-ning: overview of a clinical planning system.In: Bruinvis IAD, ed. The use of computers inradiation therapy. Amsterdam, the Nether-lands: Elsevier Science, 1987; 273–276.

13. Pierce LJ, Strawderman MH, Douglas KR, Li-chter AS. Conservative surgery and radio-therapy for early-stage breast cancer using alung density correction: the University ofMichigan experience. Int J Radiat Oncol BiolPhys 1997; 39:921–928.

14. Emami B, Lyman J, Brown A, et al. Toleranceof normal tissue to therapeutic irradiation.Int J Radiat Oncol Biol Phys 1991; 21:109–122.

15. Fowble BL, Solin LJ, Schultz DJ, GoodmanRL. Ten year results of conservative surgeryand irradiation for stage I and II breast can-cer. Int J Radiat Oncol Biol Phys 1991; 21:269–277.

16. Pierce SM, Recht A, Lingos TI, et al. Long-term radiation complications following con-servative surgery (CS) and radiation therapy(RT) in patients with early stage breast can-cer. Int J Radiat Oncol Biol Phys 1992; 23:915–923.

17. McShan DL, Kessler ML, Fraass BA. Advancedinteractive planning techniques for confor-mal therapy: high level beam descriptionsand volumetric mapping techniques. Int JRadiat Oncol Biol Phys 1995; 33:1061–1072.

18. Mohan R, Wang X, Jackson A. Optimizationof 3-D conformal radiation treatment plans.Front Radiat Ther Oncol 1996; 29:86–103.

19. Verhey LJ. 3-D conformal therapy usingbeam intensity modulation. Front RadiatTher Oncol 1996; 29:139–155.

Figure 6. Line graph shows comparison ofthe depth of the IFV nodes and the volumewithin the 90% surface for the traditional andconformal optimized plans.

Figure 7. Line graph shows correlation of themean brachial plexus depth, defined as theaverage of the minimum and maximumdepths to the brachial plexus, and the depth ofthe SCV nodes from the 21 scanned regions.


deﬁnition of the supraclavicular and infraclavicular...

Documents