Multiple older randomized trials compared mastectomy alone to mastectomy with adjuvant RT. PMRT consistently reduced the risk of LRR but in early trials there was no demonstrated benefit with regard to overall survival. An early meta-analysis, first published in 1987 and updated in 1994, examined 7,941 patients treated across 10 trials and showed a 7% decrement in overall survival when comparing patients who did and did not receive PMRT; however this difference was nonsignificant (2,3). It appeared that patients who received PMRT had an excess of cardiac mortality compared to patients not receiving PMRT. Interestingly, the meta-analysis showed that for patients receiving RT on trials with a later start date did significantly better than those patients on older trials, likely reflecting improvements in more modern radiation planning and delivery and resultant decrease in cardiac toxicity.

Given these findings, there was significant concern about the relative value of PMRT, particularly in the context of data showing that systemic therapy not only reduced the risk of distant recurrence but also significantly decreased local recurrence (4). As a result, multiple studies examined whether or not there was a benefit of adjuvant RT for patients receiving chemotherapy. Two notable trials were conducted by the Danish Breast Cancer Cooperative Group. In the 82b trial, a total of 1,708 high-risk premenopausal women with breast cancer who had undergone mastectomy randomly were assigned to receive cyclophosphamide, methotrexate, and fluorouracil chemotherapy with or without PMRT to the chest wall and regional lymph nodes (5). Patients were defined as high risk in this trial if they had pathologically involved lymph nodes, primary tumors greater than 5 cm in size, or invasion of the skin or pectoralis fascia. Patients receiving PMRT had a significantly lower risk of locoregional recurrence, either in isolation or with synchronous distant disease, than those who did not receive radiotherapy. Similarly, PMRT resulted in improved disease-free survival and overall survival at 10 years. Protocol 82c was a complementary trial examining outcomes in high-risk postmenopausal patients with breast cancer (6). Patients were deemed high risk by the same criteria as those in 82b and were randomized to either adjuvant tamoxifen alone or adjuvant tamoxifen and PMRT. RT again was found to significantly decrease the risk of locoregional recurrence and improve disease-free survival and overall survival. A similar trial examining the role for PMRT in premenopausal patients with lymph node–positive breast cancer was conducted in British Columbia around the same time (7). At a median follow-up of 249 months, the addition of RT to chemotherapy was associated with significant improvement in all endpoints examined including
decreased 20-year risk of isolated LRR (26% to 10%), improved breast cancer–specific survival (28% to 53% at 20 years), and improved overall survival (37% to 47% at 20 years). When the investigators stratified patients by number of involved lymph nodes, patients with one to three positive lymph nodes and those with four or more positive lymph nodes benefited from adjuvant RT; there was an overall survival benefit also demonstrated for both subgroups.

Together these trials demonstrated that there was a role for PMRT in selected patients deemed high risk by study investigators although by slightly different criteria. They also demonstrated a relationship between a reduction in local recurrence and improved overall outcomes, likely because of persistent or recurrent locoregional disease acting as a source of distant metastases, which subsequently impacted survival. However, there have been several criticisms of these trials. One criticism is regarding the adequacy of systemic therapy; some feel that the chemotherapy given to patients in these trials was not as efficacious as moderate chemotherapy and that patients received too short a tamoxifen treatment course. Another criticism is regarding the adequacy of surgical treatment in the Danish studies given that only a mean of seven lymph nodes were recovered during ALND. Additionally, the LRR rate was felt to be higher than anticipated in the patients who did not receive PMRT. A subsequent analysis intended to address this criticism retrospectively analyzed outcomes for patients with eight or more lymph nodes removed (8). For the 1,152 patients who met this criterion, the overall 15-year survival rate was improved with adjuvant PMRT from 29% to 39% (p = 0.015). A stratified subgroup analysis of patients with one to three and those with four or more positive lymph nodes was performed and showed that LRR was significantly decreased and overall survival was improved with PMRT for both subgroups.

Based on these trials and others, in 2000 the National Institutes of Health published a consensus statement regarding adjuvant therapy for breast cancer (9). In regard to PMRT, they expressed a need to balance the potential benefits in terms of locoregional control and overall survival with the potential side effects of treatment, particularly the previously documented excess of non—breast cancer deaths. Although they acknowledged that these deaths were likely related to outdated radiation techniques contributing to high doses received by the heart and great vessels, they felt PMRT should only be recommended for women with high-risk disease, which they defined as having four or more positive lymph nodes or an advanced primary tumor. They felt the role for PMRT in women with one to three positive lymph nodes remained uncertain. PMRT for patients with one to three positive lymph nodes was the subject of a randomized trial run by the Southwest Oncology Group (SWOG), trial s9927. Patients with pathologically staged T1–2, N1 tumors with one to three positive axillary lymph nodes who had undergone modified radical mastectomy and axillary dissection with a minimum of 10 lymph nodes recovered were randomized to PMRT or no PMRT. The targeted accrual for the trial was 2,500 patients; unfortunately, it closed after fewer than 100 patients enrolled and no conclusions could be drawn.

In 2005 the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) published a meta-analysis examining outcomes for women with early-stage breast cancer enrolled on randomized trials in the 1960s–1980s (10). For 8,500 women with lymph node–positive breast cancer treated with mastectomy and axillary clearance, PMRT improved local recurrence at 5 years from 23% to 6% (absolute reduction of 17%) and 15-year breast cancer mortality from 60.1% to 54.7% (absolute reduction 5.4%, p = 0.0002). This proportional reduction in local recurrence was seen across multiple subgroups analyzed. An overall mortality benefit was also demonstrated with a 4.4% absolute improvement at 15 years, with radiotherapy decreasing the risk from 64.2% to 59.8%. In contrast, for women with lymph node–negative disease, the 5-year local recurrence risk was only 6% in the absence of radiotherapy. PMRT reduced this risk to 2% (absolute benefit 4%) and there was no significant reduction seen in 15-year breast cancer mortality.

This meta-analysis was updated in 2014 at a median follow-up of 10 years (11). It showed that for women with lymph node–negative disease there was no significant improvement in LRR, any first recurrence, or breast cancer mortality. In comparison, for women with one to three positive lymph nodes, PMRT improved 10-year isolated locoregional recurrence (20.3% to 3.8%; p < 0.00001) and 20-year breast cancer mortality (50.2% to 42.3%; p = 0.01). PMRT showed similar benefits for women with four more positive lymph nodes (10-year isolated locoregional recurrence: 32.1% to 13.0%; p < 0.00001; 20-year breast cancer mortality: 80.0% to 70.7%; p = 0.04). Although these data were significant and inform modern practice, the LRR rates in this analysis were higher than more recently published trials and likely the absolute benefit of PMRT is smaller in the modern era. The results of several more modern trials examining locoregional failure rates in the absence of PMRT are presented in Table 17-1.

Some radiation oncologists feel that not all patients with one to three positive lymph nodes benefit from PMRT. In particular, the absolute number of lymph nodes involved seems to be predictive of local recurrence. A number of studies examined the impact of having only one positive lymph node versus two or three positive lymph nodes and found that the risk of LRR is significantly higher for those with two or three positive lymph nodes (16,17). Therefore, in addition to the number of lymph nodes involved with metastatic carcinoma, other risk factors are considered when deciding whether
or not to offer PMRT for patients with one to three positive lymph nodes. Additional factors that predict for LRR include age, tumor size, grade, presence of lymphovascular invasion (LVI), margin status, tumor subtype, and response to neoadjuvant therapy. Younger patients (14,18,19,20), those with larger tumors (21), higher-grade tumors (19,22,23), and with evidence of LVI (14,20,23) are at higher risk of LRR. Tumor subtype is another predictive variable for LRR. Retrospective evaluation of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2) status of patients on the Danish Breast Cancer Cooperative Group Trials 82b and 82c showed that on multivariable analysis, patients with ER−/PR−/HER-2− tumors and those with ER-/PR-/HER-2+ tumors were most likely to have a LRR (24). Interestingly, in this analysis, there was only a survival benefit with PMRT for patients with ER+, PR+, or HER-2− tumors; a survival benefit was not seen among ER−, PR−, or HER2+ patients.

In an effort to again evaluate the benefit of PMRT in patients with one to three positive lymph nodes, a United Kingdom Medical Research Council trial, the Selective Use of Postoperative Radiotherapy aftEr Mastectomy (SUPREMO), was initiated. It enrolled patients undergoing mastectomy with pT1-2N1 disease, pT3N0 disease, or pT2N0 disease with grade III histology and/or LVI. All patients had to have undergone axillary node clearance (minimum 10 lymph nodes removed if lymph node positive; minimum of 4 nodes removed if lymph node negative). Patients were randomized to receive PMRT or no PMRT. The study has completed accrual. Quality-of-life (QOL) results at 2 years have been published (25). Patients receiving PMRT had significantly more local symptoms but the difference between groups was small. There were no other differences in other 2-year QOL scores. Results regarding the primary outcome of the trial, overall survival, have not yet been published.

Other attempts have been made to try to identify those patients at highest risk of LRR and therefore perhaps most likely to benefit from PMRT. The 21-gene recurrence score (RS), as provided by the Oncotype DX assay developed by Genomic Health, Inc., has been widely adopted to try to help understand a patient’s individual benefit from chemotherapy. For patients with ER+ cancers, the question is whether the RS can also predict locoregional outcomes. Two retrospective analyses, one of patients enrolled on National Surgical Adjuvant Breast and Bowel Project (NSABP) B-28 (26) and the other of patients on NSABP B-14 and B-20 (27) have examined the role of the RS in potentially helping to guide local therapy recommendations. Patients enrolled on NSABP B-28 had resected, node-positive breast cancer and the trial compared two different chemotherapy regimens. Patients who underwent mastectomy did not receive RT. In an analysis of 1,065 patients with lymph node–positive disease, the RS significantly predicted LRR, with a 10-year cumulative incidence of LRR of 3.3%, 7.2%, 12.2% for patients with low (<18), intermediate (18 to 30), and high (≥31) RSs, respectively (p < 0.001) (28). On multivariable analysis, RS remained an independent predictor of LRR as did tumor size and pathologic nodal status for four or more versus one to three positive lymph nodes.

Patients on NSABP B-14 had lymph node–negative disease and were randomized to tamoxifen versus placebo. Patients on NSABP B-20 had lymph node–negative breast cancer and were randomized to tamoxifen-alone or tamoxifen plus chemotherapy. Patients from these two trials were grouped together and the predictive value of the RS on LRR was examined (27). For patients treated with tamoxifen, the RS was predictive, with LRR rates at 10 years of 4%, 7.2%, and 15.8% for those with low, intermediate and high RS, respectively. Similar findings were seen in patients treated with placebo and those treated with chemotherapy plus tamoxifen. Taken together, these data suggest that the RS may be beneficial in helping to determine who might benefit from PMRT.

To further understand the prognostic impact of the RS on locoregional outcomes, a prospective, randomized trial run by the Canadian Clinical Trials Group (CCTC) TailorRT (MA.39) (clinicaltrials.gov Identifier: NCT03488693) is currently enrolling patients with a RS of less than 18, a pathologic T1 or T2 tumor, and pN1 nodal involvement. Eligible patients can have one to three positive axillary lymph nodes (macrometastases) if treated with ALND, one to two positive axillary lymph nodes if treated with breast-conserving surgery and SLNB
alone, and one positive lymph node if treated with mastectomy and SLNB alone. Patients undergoing lumpectomy are randomized to whole breast RT with or without regional lymph node RT. Patients undergoing mastectomy are randomized to PMRT or no PMRT. The primary outcome is breast cancer recurrence-free survival. The study is currently anticipated to be completed in 2027.

For patients with a higher lymph node burden, the definition of which is variable but most often defined as having four or more positive lymph nodes, PMRT is routinely recommended given the high risk of LRR seen in this patient population. For patients with one to three positive lymph nodes, a variety of risk factors are considered as described above when considering whether or not to offer PMRT. In 2016, the American Society of Clinical Oncology (ASCO), American Society for Radiation Oncology (ASTRO), and Society of Surgical Oncology (SSO) published a focused guideline update on Postmastectomy Radiation (29). They stated that “the available evidence shows that PMRT reduces the risk of locoregional failure, any recurrence, and breast cancer mortality for patients with T1–2 breast cancer and one to three positive lymph nodes. However, some subsets of these patients are likely to have such a low risk of locoregional failure that the absolute benefit of PMRT is outweighed by its potential toxicities. Thus, the decision to recommend PMRT or not requires a great deal of clinical judgement.”

As described above, in the EBCTCG meta-analysis there was no significant improvement in LRR, any first recurrence, or breast cancer mortality with the use of PMRT in patients without pathologic axillary lymph node involvement. However, several investigators have wondered if patients with risk factors for LRR, even in the absence of axillary lymph node metastases, would benefit from PMRT. One such study retrospectively examined nearly 9,000 lymph node–negative patients enrolled in five NSABP studies (30). The analysis specifically examined patients with tumors 5 cm or larger. At a median follow-up of 15.1 years, 28 out of 313 patients had experienced locoregional failure. Cumulative incidence for isolated LRR was 7.0% for patients with 5-cm tumors and 7.2% for patients with tumors larger than 5 cm. The investigators concluded that as locoregional failure was low for these patients, PMRT should not be routinely recommended for patients with tumors 5 cm or larger.

In contrast, two studies performed at Massachusetts General Hospital similarly tried to address the risk of LRR in high-risk lymph node–negative patients. In a retrospective analysis of nearly 900 patients who were diagnosed between 1980 and 2000 who did not receive adjuvant PMRT, they identified four significant and independent risk factors for LRR: size greater than 2 cm, margin less than 2 mm, premenopausal status, and LVI (31). Ten-year LRR was 1.2%, 10%, 17.9%, and 40.6% for those with zero, one, two or three risk factors, respectively. The investigators concluded that patients with multiple risks factors despite negative axillary lymph nodes therefore would benefit from PMRT given the high LRR rates seen in this population. A subsequent analysis specifically examined patients with no axillary lymph node involvement but tumors 5 cm or larger (32). At a median follow-up of 85 months, the risk of locoregional failure was only 7.6% (96% confidence interval [CI], 3% to 16%); however, for those patients with evidence of LVI, the 5-year actuarial local failure rate was 21% (CI = 5% to 45%) compared to those without LVI (4%, 1% to 11%). The presence of LVI was also associated with overall survival and disease-free survival. The authors suggested, therefore, that for those patients with pT3N0 breast cancer the presence of LVI may help identify a population of lymph node–negative patients most likely to benefit from PMRT.