Radiation Field Design and Patterns of Locoregional Recurrence Following Definitive Radiotherapy for Breast Cancer

Transcription

1 International Journal of Radiation Oncology biology physics Clinical Investigation: Breast Cancer Radiation Field Design and Patterns of Locoregional Recurrence Following Definitive Radiotherapy for Breast Cancer Susie A. Chen, MD,* David M. Schuster, MD, y Donna Mister, BS,* Tian Liu, PhD,* Karen Godette, MD,* and Mylin A. Torres, MD* *Department of Radiation Oncology, Winship Cancer Institute, Emory University, Atlanta, Georgia; and y Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Winship Cancer Institute, Emory University, Atlanta, Georgia Received Apr 12, 2012, and in revised form Jun 1, Accepted for publication Jun 6, 2012 Summary Radiotherapy promotes locoregional control and improves survival in breast cancer patients. To determine the relationship between locoregional recurrence and radiotherapy field design and dose, we fused radiotherapy planning computed tomography (CT) images and dose delivered to the initial primary breast cancer with positron emission tomography (PET)-CT images identifying subsequent locoregional recurrence. In our study, most locoregional recurrences occurred in areas not adequately covered by the prescribed radiation dose for the primary breast cancer. Purpose: Locoregional control is associated with breast cancer-specific and overall survival in select women with breast cancer. Although several patient, tumor, and treatment characteristics have been shown to contribute to locoregional recurrence (LRR), studies evaluating factors related to radiotherapy (XRT) technique have been limited. We investigated the relationship between LRR location and XRT fields and dose delivered to the primary breast cancer in women experiencing subsequent locoregional relapse. Methods and Materials: We identified 21 women who were previously treated definitively with surgery and XRT for breast cancer. All patients developed biopsy-result proven LRR and presented to Emory University Hospital between 2004 and 2010 for treatment. Computed tomography (CT) simulation scans with XRT dose files for the initial breast cancer were fused with 18 F-labeled fluorodeoxyglucose positron emission tomography (FDG PET)/CT images in DI- COM (Digital Imaging and Communications in Medicine) format identifying the LRR. Each LRR was categorized as in-field, defined as 95% of the LRR volume receiving 95% of the prescribed whole-breast dose; marginal, defined as LRR at the field edge and/or not receiving 95% of the prescribed dose to 95% of the volume; or out-of-field, that is, LRR intentionally not treated with the original XRT plan. Results: Of the 24 identified LRRs (3 patients experienced two LRRs), 3 were in-field, 9 were marginal, and 12 were out-of-field. Two of the 3 in-field LRRs were marginal misses of the additional boost XRT dose. Out-of-field LRRs consisted of six supraclavicular and six internal mammary nodal recurrences. Conclusions: Most LRRs in our study occurred in areas not fully covered by the prescribed XRT dose or were purposely excluded from the original XRT fields. Our data suggest that XRT technique, field design, and dose play a critical role in preventing LRR in women with breast cancer. Ó 2013 Elsevier Inc. Reprint requests to: Mylin A. Torres, MD, Department of Radiation Oncology, Emory University, 1365 Clifton Rd NE AT1307, Atlanta, GA. Tel: (404) ; Fax: (404) ; matorre@emory.edu This work was presented previously at the 2011 Breast Cancer Symposium, San Francisco, CA, Sept 8-10, Conflict of interest: none. Int J Radiation Oncol Biol Phys, Vol. 85, No. 2, pp. 309e314, /$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.

2 310 Chen et al. International Journal of Radiation Oncology Biology Physics Introduction Advances in systemic therapy have led to significantly lower rates of distant metastasis and improved overall survival (OS) in women with breast cancer. Due to the relatively lower risk of distant recurrence seen in patients treated with modern therapies, locoregional recurrence (LRR) has emerged as an important predictor of breast cancer-specific and OS even in the setting of early stage, node-negative disease (1). Various tumor, patient, and treatment factors are known to predict for LRR. Tumor characteristics associated with higher rates of LRR include high-grade histology results, angiolymphatic invasion, and triple-negative or HER2-positive receptor status (2, 3). Positive nodes and extranodal extension have also been predictive of LRR following mastectomy (4). Examples of patient characteristics contributing to LRR include young age and African American race (5). Treatment factors associated with higher rates of local control include surgical margins greater than 1 or 2 mm (3, 6) and systemic treatment with chemotherapy, trastuzumab, and endocrine therapy, when indicated (7). Several studies have also alluded to the importance of radiation therapy (XRT) technique in preventing LRR, but research thoroughly evaluating the influence of XRT field design and dose on LRR is limited. In fact, previous publications have focused more on total dose delivered (eg, boost dose) rather than actual field design (8). In the modern era, computed tomography (CT)-based planning has been used to individualize treatment and tailor XRT treatment fields to encompass the breast and draining lymphatics while minimizing dose to the heart, lungs, and spinal cord. Threedimensional planning has also enabled more accurate demarcation of the lumpectomy cavity at depth rather than relying solely on the surgical scar at the skin s surface (9). Furthermore, the Radiation Therapy Oncology Group (RTOG) breast contouring atlas has assisted radiation oncologists by establishing guidelines for contouring nodal volumes at risk and normal tissue structures that should be avoided during XRT (10). Collectively, these advances in technology have the potential to improve local control rates by optimizing dose coverage of the breast and draining lymph nodes. Moreover, while the importance of XRT technique in preventing LRR has previously been established in other malignancies such as cervical and head and neck cancers (11, 12), no study has systematically examined the importance of XRT field design with modern imaging modalities and treatment planning in preventing LRR in breast cancer patients. We examined a group of breast cancer patients who received definitive postoperative radiation therapy and subsequently developed LRR, treated at Emory University. We systematically evaluated the influence of XRT treatment technique, both field design and dose, on local control by superimposing the 18 F-labeled fluorodeoxyglucose positron emission tomography (FDG PET)/ CT-detected local recurrence on the original XRT treatment plan for the initial breast cancer diagnosis. Our goal was to determine whether XRT technique might play a role in preventing LRR. Methods and Materials After obtaining institutional review board approval, we reviewed the records of 791 breast cancer patients and identified 35 women with LRR treated at Emory University between 2004 and Of these 35 patients, 21 had CT simulation scans with dose files available from their initial treatment as well as [ 18 F]FDG PET/CT scans confirming subsequent LRR in DICOM (Digital Imaging and Communications in Medicine) format. Alignment of these images was needed to clearly depict the relationship between LRR location and planned XRT dose. Patient, tumor, and treatment characteristics were obtained from the medical records of each patient. Evaluation and treatment of the original breast cancer Patient median age at initial breast cancer diagnosis was 52 years (range, years). Twelve women were postmenopausal; 61% of patients were African American. Other tumor and treatment characteristics are listed in Table 1. All women were initially evaluated with mammograms and ultrasonograms to determine primary tumor location and axillary lymph node status. All patients received definitive surgery (lumpectomy [nz10] or mastectomy [nz11]) with lymph node assessment (sentinel lymph node biopsy or axillary lymph node dissection) and definitive XRT with or without systemic therapy, as determined by the treating medical oncologist. Among the 21 patients who Table 1 Tumor characteristics at initial diagnosis Characteristic No. (%) Tumor laterality Right 6 (29%) Left 15 (71%) T stage DCIS 2 (10%) T1 2 (10%) T2 7 (33%) T3 7 (33%) T4 3 (14%) N stage N0 8 (38%) N1 9 (43%) N2 3 (14%) N3 1 (5%) Tumor location Central 1 (5%) Upper inner 8 (38%) Lower inner 3 (14%) Upper outer 4 (29%) Lower outer 1 (5%) Not recorded 2 (9%) Lymphovascular space invasion Yes 4 (19%) No 17 (81%) Extracapsular extension Yes 1 (5%) No 21 (95%) Biologic subtype ER/PR þ, HER2 6 (29%) ER/PR/HER2 þ 2 (9%) ER/PR/HER 11 (52%) ER/PR, HER2 þ 2 (9%) Abbreviations: DCIS Z ductal carcinoma in situ; ER Z Estrogen Receptor; PR Z Progesterone Receptor.

3 Volume 85 Number Recurrence after definitive RT for breast cancer 311 developed LRR, 14, 4, and 3 patients received neoadjuvant, adjuvant, or no chemotherapy, respectively. Eight of nine hormone receptor-positive patients received endocrine therapy for a median of 1.7 years (range, years), although 1 patient refused endocrine therapy. After completing XRT treatment, all patients were subsequently followed with at least physical examinations and mammograms. [ 18 F]FDG PET scans or MRI scans were obtained when clinically indicated, and when a recurrence was identified, all patients underwent a restaging [ 18 F]FDG PET scan, according to our institutional practice. The median prescribed XRT dose was 50 Gy (range, Gy) to the breast or chest wall. The median boost dose was Gy (range, Gy). The most commonly prescribed XRT regimen was 45 Gy with a Gy boost delivered as a simultaneous-in-field boost (nz10 patients). Eleven patients received XRT with opposed tangents modulated with dynamic wedges. The remaining 10 patients were treated with intensity-modulated radiation therapy (IMRT) to the whole breast. Eleven patients received supraclavicular XRT. Internal mammary (IM) lymph nodes were deliberately excluded in all patients. Definition and classification of LRR LRR was defined as biopsy-result proven recurrent disease in the ipsilateral breast or chest wall or in the ipsilateral axillary, supraclavicular, infraclavicular, or IM lymph node basins. We intentionally limited the study to relatively recent period between 2004 and 2010 in an effort to identify true LRRs as opposed to second primaries. In addition, all LRRs fulfilled previously published criteria for LRRs (13). XRT plans (CT simulation and dose) used to treat the original breast cancer and the [ 18 F]FDG PET/CT images identifying the LRR were aligned using a B-spline deformable registration algorithm (VelocityAI; Velocity Medical Solutions, Atlanta, GA) to correct for posture and soft tissue changes. Recurrence volumes were defined as 30% maximum standardized uptake value (SUV), except for two subcentimeter volumes which were defined as 50% of maximum SUV due to partial volume effect (14). All LRR foci were identified by [ 18 F] FDG PET/CT and then confirmed by a nuclear radiologist with 10 years experience with PET/CT interpretation (DMS). The XRT dose delivered previously (in treating the original breast cancer) to the FDG-avid LRR volume was then calculated. LRRs were classified as in-field, marginal, or out-of-field (Figs. 1-3). In-field LRRs were defined as 95% of the LRR volume Fig. 2. A 53-year-old woman developed an isolated marginal LRR after receiving 50.4 Gy to the whole breast with a 10 Gy boost. Red Z LRR; yellow Z 50.4 Gy. receiving 95% of the prescribed whole-breast dose ( Gy) without accounting for the boost dose. Marginal recurrences were LRRs at the field edge and not receiving 95% of the prescribed dose to 95% of the volume. Out-of-field recurrences were LRR that were intentionally not treated with the original XRT plan (eg, supraclavicular or IM nodal recurrences). Statistical methods Statistical analysis was performed using SAS version 9.2 software (SAS Institute Inc, SAS Campus Drive, Cary, NC). Descriptive statistics and analyses were performed with Wilcoxon signed rank test for continuous variables and the Fisher exact test for categorical variables. Significance was assessed at a P value of less than.05. Results Patient characteristics Twenty-four consecutive LRRs were identified; 1 patient had two synchronous anatomically separate LRRs, and 2 other patients had two metachronous LRRs. There were 3 in-field recurrences and 9 marginal and 12 out-of-field recurrences, and 1 patient experienced both an in-field and an out-of-field recurrence synchronously. Median time to recurrence was 2.1 years (range, years). Twelve patients developed isolated LRRs as their site of first Fig. 1. A 59-year-old woman developed an in-field LRR after receiving 45 Gy to the whole breast with a Gy boost. Red Z LRR; blue Z 45 Gy; yellow Z 50 Gy; orange Z 60 Gy. Fig. 3. A 44-year-old woman developed an internal mammary lymph node recurrence after receiving 45 Gy to the whole breast with a 15 Gy boost to the cavity (not pictured). Red Z LRR; green Z 45 Gy; yellow Z 50 Gy.

4 312 Chen et al. International Journal of Radiation Oncology Biology Physics failure. Twelve of the LRRs occurred concurrently with distant disease, most commonly pulmonary metastasis. Nineteen LRRs were initially detected with imaging, most commonly on [ 18 F] FDG PET/CT, and the remaining 5 were detected with physical examination. In-field recurrences Among 11 patients prescribed 50 or 50.4 Gy to the whole breast, none developed in-field recurrences. However, among the 10 patients prescribed 45 Gy, 3 developed in-field recurrences within the 100% isodose line. Two in-field LRRs were marginal misses of the boost fields where the intended dose was 15 Gy, and the median dose delivered was only 8.5 Gy. Both patients (stage IIIA and IIIB, respectively) had high-grade estrogen receptor/progesterone receptor (ER/PR)-negative and HER2-positive tumors and postlumpectomy surgical margins greater than 2 mm. The third in-field LRR occurred in a patient with stage 0 disease, who had a highgrade ductal carcinoma in situ that was ER/PR-positive and HER2- negative and who developed an inflammatory in-field recurrence that was multicentric. Marginal recurrences There were 9 marginal LRRs. Among the patients receiving 50 and 50.4 Gy, 6 patients developed marginal LRRs with a median dose delivered to the recurrence volume of Gy (range, Gy). Among patients receiving 45 Gy, 3 patients developed marginal LRRs with a median dose of 39.6 Gy (range, Gy). Among the marginal LRRs, 2 occurred within breast parenchyma following lumpectomy, and 2 occurred in the chest wall after mastectomy. The remaining marginal LRRs occurred in targeted draining lymph node basins that were not entirely encompassed by the original XRT fields: 2 supraclavicular, 1 infraclavicular, and 2 level I/II nodal recurrences. All patients who had a marginal recurrence within a lymph node basin had axillary evaluation at the time of initial diagnosis, by sentinel lymph node biopsy (nz1) or by axillary dissection (nz4). Of these patients, 3 patients had positive lymph nodes with a mean positive nodal ratio of 0.17 (range, ) and lymph nodes removed. Out-of-field recurrences Out-of-field LRRs consisted of 6 IM and 6 supraclavicular lymph nodal recurrences. All 6 IM LRRs occurred in women with inner quadrant primary breast tumors, with the exception of 1 patient who had a centrally located tumor. IM LRRs were the first sites of failure in 4 of 6 patients. Among 6 patients with supraclavicular nodal recurrences, 2 had stage N0 disease, and 4 had stage N1 disease at initial diagnosis. Four of these 6 patients had less than 10 lymph nodes removed. None of the N1 patients had 3 positive lymph nodes, but 2 patients had two positive lymph nodes. Six out-of-field recurrences occurred simultaneously with distant metastasis. Factors associated with LRR Among the initial cancer cases (nz21), 7 were ER/PR-positive and HER2-negative, 10 were triple-negative, 2 were triplepositive, and 2 were ER/PR-negative and HER2-positive. Among the 7 patients with hormone receptor-positive and HER2-negative tumors, 1 patient developed an in-field recurrence, 1 developed a marginal recurrence, and the remaining patients developed outof-field recurrences. Among the 10 patients with triple-negative tumors, none recurred within field, 5 were marginal, and 5 were out-of-field. In our series, the only 2 patients with tumors that were ER/PR-negative and HER2-positive had recurrences within field. Finally, the 2 patients with ER/PR-positive and HER2-positive tumors both developed out-of-field LRRs in the IM nodes. None of the patients in our study had low-grade disease at presentation or at recurrence. Eighteen of the 21 patients had highgrade tumors, while 3 had intermediate-grade tumors at initial diagnosis. Grade did not predict for type of recurrence, whether it was in-field, marginal, or out-of-field. Only 4 patients had lymphovascular space invasion (LVSI) at presentation and 3 patients had recurrence out-of-field (one supraclavicular and two IM nodal recurrences), while the fourth patient had marginal recurrence in the supraclavicular region. No patient had multifocal or multicentric disease at the original diagnosis of breast cancer. Three patients had close surgical margins (<2 mm) following lumpectomy without re-excision, and among these patients, 2 developed marginal recurrences, while 1 developed an out-of-field recurrence. In addition to grade, age, triple-negative receptor status, LVSI, and use of IMRT did not predict for LRR type (Table 2). Outcome All patients who developed distant metastasis with a synchronous LRR went on to receive systemic therapy. Among the 12 patients with isolated LRRs, 3 received chemotherapy, 4 received XRT, 2 underwent surgery, and 2 underwent surgery and postoperative XRT. Three of these 12 patients subsequently developed distant metastasis. With a median follow-up of 4.8 years, the 4-year disease-free survival (DFS) was 59.4% and OS was 69.7% among all patients. No difference was seen in survival by LRR type (Fig. 4). For patients with isolated LRR, the 4-year DFS was 67.5%, and OS was 90.0%, while among patients with synchronous LRR and distant metastasis, the 4-year DFS and OS were both 51.9%. Discussion Local control in breast cancer not only increases a patient s chances for breast conservation following lumpectomy but also Table 2 Factors associated with LRR No. of occurrences (%) Factor Out-of-field (nz12) Infield/marginal (nz12) Overall (nz24) P value* Age (y) 50.8 (12.9) 55.1 (12.8) 52.9 (12.7).336 Triple 8 (66.7%) 5 (41.7%) 13 (54.2%).414 negativity LVSI 1 (14.3%) 3 (33.3%) 4 (25.0%).585 IMRT 4 (33.3%) 6 (50.0%) 10 (41.7%).680 Abbreviations: IMRT Z intensity-modulated radiation therapy; LSVI Z lymphovascular space invasion. * Analysis was performed with Wilcoxon signed rank test for continuous variables and the Fisher exact test for categorical variables. Significance was assessed at a P value of <.05.

5 Volume 85 Number Recurrence after definitive RT for breast cancer 313 are potentially limited by small patient numbers and the deformable image registration algorithm used to fuse the image sets. We were unable to include all 35 patients at Emory University who experienced LRR due to missing XRT treatment plans and PET/CT images documenting LRR in DICOM format. However, among our 21 patients, most appeared to experience LRR due to limited XRT field design and dose coverage. In addition, while [ 18 F]FDG PET/CT and XRT planning images were carefully aligned, the deformable registration performed to adjust for anatomical and positional variations may theoretically have distorted the true location of an LRR. For this reason, a board-certified nuclear radiologist (DMS) was asked to confirm all LRR locations on the fused image sets after viewing the LRR on the original [ 18 F] FDG PET/CT. Conclusions Fig. 4. Kaplan-Meier OS by LRR type (log rank P valuez.463). influences survival in select patients after either lumpectomy or mastectomy (4). Prior research has indicated that tumor and treatments factors, primarily surgical and systemic therapies, are associated with LRR. Our study is one of the first to critically evaluate XRT technique and its impact on breast cancer outcome by assessing the relationship between LRR location LRR and the planned dose with the original XRT fields. We found that most LRRs occurred in areas not completely covered by the full XRT dose, as a consequence of XRT field design. In the treatment of breast cancer, National Comprehensive Cancer Network guidelines recommend daily delivery of at least 1.8 Gy to the breast (15). In our series, most of our subjects developed recurrences in areas where the XRT dose per fraction was below standard fraction sizes due to the LRR occurring at the field edge or out-of-field. This suggests that field design and daily setup play important roles in preventing LRR in breast cancer patients. Indeed, in our study, 9 patients developed marginal recurrences. Moreover, 2 of the in-field recurrences were marginal misses of the additional boost fields. Although tangential breast or chest wall XRT incidentally treats a portion of the axilla, only 55% of the axillary lymph node levels I and II will receive 95% of the prescribed dose (16). During the study period, it was not our institutional policy to routinely contour lymph node basins in breast XRT planning, which may have contributed to the five marginal nodal recurrences, as there were potentially undertreated nodal regions not fully encompassed by the prescribed XRT dose. Several investigators have indicated that high-grade and triplenegative tumors are more likely to recur locally than those that are low grade and hormone receptor positive (17). However, while most patients in our series had high-grade tumors, none of the triple-negative patients developed in-field recurrences. In fact, those patients had marginal or out-of-field recurrences, suggesting that our patients with triple-negative tumors developed recurrences due to inadequate XRT dose. In contrast, 2 of the 3 in-field recurrences in our series were in patients with ER/PR-negative, HER2/neu-positive disease, suggesting that being HER2 positive may be a predictor of LRR despite adequate XRT dose, a finding supported by other studies (18). While our study is the first to indicate that LRR may be prevented by adequate XRT field design and dose, our findings In conclusion, our study indicates that LRRs may be directly related to XRT field design and subsequent dose delivered to areas at risk. These two factors may play a pivotal role in preventing breast cancer recurrence in addition to biological, patient, and tumor factors. As systemic therapy continues to improve and prevent distant metastasis, both XRT field design and dose will play an increasing role in breast cancer outcomes, including survival. Indeed, modern treatment planning with CT simulation images currently allows clinicians to identify variability in patient anatomy and extent of disease (16, 19). Implementation of systematic contouring of regional lymph nodes with the use of the RTOG breast contouring atlas may also improve outcome by assuring adequate XRT dose coverage to the breast and draining lymphatics (10). Our findings indicate that detailed attention to XRT treatment technique and delivery may directly impact disease control in women with breast cancer. References 1. Darby S, McGale P, Correa C, et al. Effect of radiotherapy after breastconserving 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: Jones HA, Antonini N, Hart AA, et al. Impact of pathological characteristics on local relapse after breast-conserving therapy: a subgroup analysis of the EORTC boost versus no boost trial. J Clin Oncol 2009; 27: Freedman G, Fowble B, Hanlon A, et al. Patients with early stage invasive cancer with close or positive margins treated with conservative surgery and radiation have an increased risk of breast recurrence that is delayed by adjuvant systemic therapy. Int J Radiat Oncol Biol Phys 1999;44: 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: Edwards BK, Howe HL, Ries LA, et al. Annual report to the nation on the status of cancer, , featuring implications of age and aging on U.S. cancer burden. Cancer 2002;94: Dunne C, Burke JP, Morrow M, et al. Effect of margin status on local recurrence after breast conservation and radiation therapy for ductal carcinoma in situ. J Clin Oncol 2009;27: Early Breast Cancer Trialists Collaborative Group (EBCTCG). Effects of chemotherapy and hormonal therapy for early breast cancer

6 314 Chen et al. International Journal of Radiation Oncology Biology Physics on recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005;365: 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 trial. J Clin Oncol 2007; 25: Hepel JT, Evans SB, Hiatt JR, et al. Planning the breast boost: comparison of three techniques and evolution of tumor bed during treatment. Int J Radiat Oncol Biol Phys 2009;74: Li XA, Tai A, Arthur DW, et al. Variability of target and normal structure delineation for breast cancer radiotherapy: an RTOG multi-institutional and multiobserver study. Int J Radiat Oncol Biol Phys 2009;73: Beadle BM, Jhingran A, Yom SS, et al. Patterns of regional recurrence after definitive radiotherapy for cervical cancer. Int J Radiat Oncol Biol Phys 2010;76: Int J Radiat Oncol Biol Phys Aug 1;80(5): Epub 2010 Jul Huang E, Buchholz TA, Meric F, et al. Classifying local disease recurrences after breast conservation therapy based on location and histology: new primary tumors have more favorable outcomes than true local disease recurrences. Cancer 2002;95: Soret M, Bacharach SL, Buvat I. Partial-volume effect in PET tumor imaging. J Nucl Med 2007;48: Carlson RW, Allred DC, Anderson BO, et al. Invasive breast cancer. J Natl Compr Canc Netw 2011;9: Reed DR, Lindsley SK, Mann GN, et al. Axillary lymph node dose with tangential breast irradiation. Int J Radiat Oncol Biol Phys 2005; 61: Nguyen PL, Taghian AG, Katz MS, et al. Breast cancer subtype approximated by estrogen receptor, progesterone receptor, and HER-2 is associated with local and distant recurrence after breast-conserving therapy. J Clin Oncol 2008;26: Arvold ND, Taghian AG, Niemierko A, et al. Age, breast cancer subtype approximation, and local recurrence after breast-conserving therapy. J Clin Oncol 2011;29: Garg AK, Frija EK, Sun TL, et al. Effects of variable placement of superior tangential/supraclavicular match line on dosimetric coverage of level III axilla/axillary apex in patients treated with breast and supraclavicular radiotherapy. Int J Radiat Oncol Biol Phys 2009;73: