Postmastectomy Radiotherapy

Abstract

Postmastectomy radiotherapy (PMRT) is an integral component of the multimodal treatment of invasive breast cancer in patients with sufficient risk of harboring a reservoir of locoregional disease despite mastectomy and systemic therapy. Microscopic residual disease in the chest wall and/or regional lymph nodes, if left untreated, may seed or reseed distant metastases after initial clearance of distant disease by highly effective contemporary systemic therapies. In addition, occult disease may also serve as a source of locoregional recurrence, which is often highly morbid to the patient in the postmastectomy setting. Therefore, the goal of PMRT is to eliminate residual occult locoregional disease and reduce the risk of both locoregional and distant recurrences. This chapter focuses on the role of PMRT in the treatment of patients with invasive breast cancer and the evidence regarding the impact of PMRT on locoregional control and overall survival. It begins by reviewing the evidence from randomized trials, followed by a discussion of how best to generalize from these trials and select patients appropriately for treatment, especially among those with N1 and N0 disease, for whom treatment is more controversial. It then turns to special considerations, including axillary management postmastectomy in patients with limited volume node-positive disease treated without axillary node dissection, the role of PMRT in patients who were treated with neoadjuvant chemotherapy, implications of differences in tumor biologic subtype, and the optimal integration of PMRT and breast reconstruction.

Keywords

postmastectomy, radiotherapy, locoregional recurrence, node-positive breast cancer, node-negative breast cancer, neoadjuvant chemotherapy, biological subtypes, breast reconstruction

This chapter focuses on the role of PMRT in the treatment of patients with invasive breast cancer and the evidence regarding the impact of PMRT on locoregional control and overall survival. It begins by reviewing the evidence from randomized trials, followed by a discussion of how best to generalize from these trials and select patients appropriately for treatment, especially among those with N1 and N0 disease, for whom treatment is more controversial. It then turns to special considerations, including axillary management postmastectomy in patients with limited volume node-positive disease treated without axillary node dissection, the role of PMRT in patients who were treated with neoadjuvant chemotherapy, implications of differences in tumor biologic subtype, and the optimal integration of PMRT and breast reconstruction.

Randomized Trials of Postmastectomy Radiotherapy

Although even early randomized trials of PMRT consistently revealed considerable reduction in the risk of locoregional recurrence, initial studies did not demonstrate an improvement in overall survival. This is likely due to treatment-related toxicity caused by antiquated radiation techniques such as the classic anterior “hockey stick” photon field, which exposed large volumes of the heart to substantial doses of radiation, resulting in cardiac deaths that offset the salutary effects of PMRT on breast cancer control.

Conclusive evidence for a substantial survival benefit from PMRT was eventually demonstrated in randomized trials from Denmark and British Columbia, initiated in the early 1980s, with landmark publications in the late 1990s. In the British Columbia trial, 318 pathologically node-positive premenopausal patients were randomized to cyclophosphamide, methotrexate, and fluorouracil (CMF) chemotherapy alone or chemotherapy and PMRT to the chest wall and the regional lymph nodes after modified radical mastectomy with axillary node dissection. This study demonstrated both a significant reduction in locoregional failure (from 28% to 10%) and improvement in overall survival (from 37% to 47% at 20 years, p = .03) with addition of PMRT. These findings were consistent with the larger Danish 82b trial, which randomized 1708 premenopausal women with high-risk breast cancer (defined as axillary node involvement, tumor size greater than 5 cm, or invasion of the skin or pectoral fascia) to CMF chemotherapy alone or chemotherapy and PMRT after total mastectomy and axillary dissection. This trial also demonstrated both a reduction in locoregional failure (from 32% to 9%) and an improvement in overall survival (from 45% to 54% at 10 years, p = .001). The Danish 82c trial, which evaluated the benefits of PMRT in postmenopausal women, randomized 1375 patients younger than 70 years of age with high-risk breast cancer to PMRT and tamoxifen versus tamoxifen alone, after total mastectomy and axillary dissection. This study also demonstrated that PMRT reduced locoregional failure (from 35% to 8%) and improved overall survival (from 36% to 45% at 10 years, p = .03). Together the findings from these studies paved the way for a change in paradigm in favor of PMRT in patients with substantial risk of locoregional recurrence after surgery and initial systemic therapy.

Controversies Surrounding Application of Randomized Trial Data in N1 Disease

Some concerns remain, however, regarding the applicability of the findings of these studies to patients who have undergone more comprehensive clearance of axillary levels I and II, which is often the case for patients treated in the United States. The median number of lymph nodes removed in the Danish study (only seven) was lower than what one would expect from a standard axillary dissection of levels I and II performed in the United States. This could have contributed to increased incidence of locoregional failures and amplified the benefit of PMRT; it could also have led to understaging of patients treated on these trials. Of note, locoregional recurrence rates after mastectomy alone in patients with one to three positive lymph nodes were much lower in retrospective American studies of patients treated without PMRT than those noted in the patients treated without radiotherapy in the Danish and British Columbia studies. Analyses of patients treated on Eastern Cooperative Oncology Group and National Surgical Adjuvant Breast and Bowel Project (NSABP) trials suggested a rate of only 13%; other studies, including series from MD Anderson and the International Breast Cancer Study Group (IBCSG), also demonstrated similar findings. Therefore consensus guidelines developed soon after the landmark publications from these trials did not recommend for or against PMRT for patients with one to three positive lymph nodes, and practice patterns studies showed divided opinion regarding the adoption of PMRT in the United States. The ambiguity remained unresolved after the premature closure in 2003 due to accrual failure of a SWOG/Intergroup randomized control trial designed to assess role of PMRT in patients in the United States with one to three positive lymph nodes.

To address the concerns related to limited axillary dissection, the Danish studies were reanalyzed after pooling the data from both the 82b and 82c studies and including only those patients with 8 or more lymph nodes removed (n = 1152). Remarkably, even though the rates of locoregional failures were lower among patients with one to three involved lymph nodes than those with four or more involved lymph nodes, the same absolute magnitude of survival benefit was observed among these two groups of patients. This was important because a landmark analysis of the Oxford Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) comprehensive meta-analysis had suggested that there might be a consistent ratio between prevention of locoregional recurrence and improvement in survival, with one life saved at 15 years for every four locoregional recurrences prevented at 5 years.

More specifically, although initial meta-analyses had raised concerns that any benefits of PMRT were more than offset by cardiac and other toxicity, the 2005 publication reflected the maturation of the large Danish trials and demonstrated a benefit. That study analyzed data from 8340 women with node-positive disease who, after mastectomy and axillary clearance, were randomized to PMRT or no further treatment on trials initiated through 1995. PMRT reduced 5-year local recurrence risk from 22.8% to 5.8%, 15-year breast cancer mortality risk from 60.1% to 54.7% (5.4% absolute risk reduction, 2 p = .0002) and overall mortality from 64.2% to 59.8% (4.4% absolute risk reduction, 2 p = .0009). Yet because this meta-analysis was itself strongly influenced by the inclusion of the large number of patients from the Danish trials, the concerns about inadequate axillary surgery that hampered the interpretation of those trials also extended to this iteration of the meta-analysis.

To address the uncertainty regarding the benefit of PMRT in patients treated with mastectomy and adequate axillary dissection, the EBCTCG more recently performed an updated meta-analysis of 22 trials initiated before 2000 consisting of individual data for 8135 women randomly allocated to adjuvant radiotherapy to the chest wall and regional lymph nodes (supraclavicular or axillary fossa or both, and internal mammary nodes) after mastectomy and axillary surgery versus the same surgery alone. Patients were classified as having axillary dissection (defined as removal of axillary lymph nodes in at least levels I and II or, if this information was not available, removal of at least 10 nodes) or axillary sampling (when less extensive axillary surgery was performed). Among women with axillary dissection and no involved nodes (n = 700), PMRT did not have significant impact on locoregional recurrence, overall recurrence, or breast cancer–specific mortality. Unsurprisingly, for women with four or more involved nodes on axillary dissection, PMRT reduced both locoregional recurrence (from 32.1% to 13.0% at 10 years, 2 p = .00001) and overall recurrence (from 75.1% to 66.3% at 10 years, 2 p = .003), as well as breast cancer–specific mortality (from 80.0% to 70.7% at 20 years, 2 p = .04). More notably, among 1314 women with axillary dissection and one to three involved nodes, for whom the benefit of PMRT had previously not been clearly established, radiotherapy was again observed to reduce locoregional recurrence (from 21% to 4.3% at 10 years, 2 p = .00001), overall recurrence (from 45.5% to 33.8% at 10 years, 2 p = .00009), and breast cancer–specific mortality (from 49.4% to 41.5% at 20 years, 2 p = .01). The vast majority (1133 of 1314) of patients in this group received systemic therapy (CMF or tamoxifen), but the proportional reductions in overall recurrence rates and breast cancer–specific mortality rates did not differ based on whether systemic therapy was given. Furthermore, within the group of 1133 women who had one to three involved nodes on axillary dissection, who received systemic therapy, and for whom additional information regarding the number of positive nodes was available (n = 683), the proportional reduction in locoregional recurrence rate or breast cancer–specific mortality rate did not differ significantly based on whether one, two, or three nodes were involved. Overall, PMRT appeared to prevent one breast cancer death at 20 years for every 1.5 recurrences avoided at 10 years.

These findings have together provided significant impetus for use of PMRT in patients with one to three involved nodes after mastectomy and axillary dissection. For instance, the current National Comprehensive Cancer Network guidelines recommend strongly considering radiation therapy to “chest wall + infraclavicular region, supraclavicular area, internal mammary nodes, and any part of the axillary bed at risk” in this group of patients. Nonetheless, some degree of controversy continues regarding the use of PMRT for patients with one to three involved nodes because the absolute risk of locoregional failure is likely lower with the advent of more effective systemic regimens such as anthracycline-based contemporary chemotherapy regimens, longer courses of endocrine therapy with aromatase inhibitors, and particularly the introduction of targeted agents such as trastuzumab, which were not routinely available in earlier eras. Furthermore, with more screening-detected cancers and surgical and pathologic advances that allow detection of small amounts of nodal involvement, the risk of locoregional recurrence is likely to be even lower among contemporary patients, who have been “upstaged” by more sensitive nodal assessment procedures. Particularly in patients with favorable biology, risks may be considerably lower in the modern era than the 20% documented in the most recent EBCTCG meta-analysis.

Indeed, in an analysis of estrogen receptor (ER)-positive patients enrolled in the NSABP B-28 trial who underwent mastectomy and did not receive PMRT, the 10-year cumulative incidence of locoregional recurrence was less than 10% for those with N1 disease, even among patients with high recurrence score on 21-gene assay. An MD Anderson study presented a retrospective analysis of locoregional recurrence rates in two cohorts of patients with T1–2N1 breast cancer treated with mastectomy and adjuvant chemotherapy, with 19% of patients treated in an early era (1978–1997) and 25% of patients treated in a more recent era (2000–2007) having received PMRT. It revealed that the risk among patients treated in the earlier cohort without PMRT was 9.5% at 5 years and 14.5% at 15 years versus 3.4% at 5 years and 6.1% at 15 years among those who received PMRT; in the later cohort, the 5-year risks were only 2.8% without PMRT and 4.2% with PMRT, leading the authors to conclude that modern treatment advances and the selected use of PMRT for those with high-risk features has allowed for the identification of a cohort at very low risk for LRR in the absence of PMRT. Still, although the lower absolute rates of locoregional recurrence in the absence of PMRT in some of these patient groups suggest that they may not benefit substantially from treatment, it is important to note recent findings from the MA20 and EORTC 22922-10925 studies that suggest that in some patients with early-stage breast cancer, an important benefit in overall disease control can be attained from comprehensive radiotherapy, even though the risk of regional recurrence itself is not high. There may not be a straightforward ratio between the prevention of locoregional recurrence and the potential for survival benefit.

Special Considerations

Patients With Node-Positive Disease and Undissected Axillae After Sentinel Lymph Node Biopsy

The near universal adoption of sentinel lymph node biopsy as the preferred approach for staging the axilla in recent years has further complicated the management of patients by identifying more patients with limited nodal involvement, including micrometastases and isolated tumor cells. These patients were unlikely to have been included on the randomized trials of PMRT because those patients were treated before sentinel node biopsy techniques made the detection of lower burden axillary nodal metastases more common.

Several recent studies have raised the question of whether select patients with low-volume nodal burden treated in an era of modern systemic therapy require any directed axillary treatment, including subsequent axillary dissection or directed axillary radiation. The IBCSG 23-01 trial randomized patients with primary tumors of 5 cm or less, who on sentinel lymph node biopsy had one or more micrometastatic (defined as less than or equal to 2 mm) sentinel lymph nodes without extracapsular extension, to axillary dissection or no further surgery. The study revealed extremely low axillary recurrence rates and did not show any statistically significant differences in disease-free survival between the two groups (5-year disease-free survival 87.8% in the no axillary dissection group and 84.4% in the axillary dissection group, p = .16). However, only 9% of patients in this study underwent a mastectomy, so it is somewhat difficult to draw strong conclusions about the omission of both axillary dissection and PMRT in patients with micrometastatic disease.

The EORTC 10981-22023 AMAROS trial showed no statistically significant difference in axillary recurrence in patients with clinically negative axillae, T1–T2 primary breast cancers, and a positive sentinel lymph node who were randomized to axillary dissection versus axillary radiotherapy. This may increase referral of patients with positive sentinel lymph nodes and no further axillary surgery to radiation oncologists, who must then face complex management decisions in mastectomy patients with incomplete axillary staging information.

Postmastectomy Radiotherapy in Node-Negative Breast Cancer

The use of PMRT in node-negative breast cancer patients is a particularly controversial area. Although the Danish trials included patients with T3–T4 node-negative disease and demonstrated a benefit, recent retrospective pooled analyses, including of five NSABP trials, have suggested that the risk of isolated locoregional failures as a first event in T3N0 patients may be lower (7.1%) than previously expected with omission of radiotherapy after mastectomy. Therefore, many radiation oncologists do not routinely recommend adjuvant radiotherapy in all T3N0 breast cancer patients after mastectomy. Instead a personalized treatment decision based on a number of risk factors such as tumor size, receipt of an inadequate axillary dissection, close or positive margins and the presence of lymphovascular invasion is often warranted.

This is supported by the findings of other studies in node-negative patients with smaller primary tumors, which have suggested that consideration of these other risk factors may predict for locoregional recurrence in the absence of RT. In one large retrospective analysis of 1505 T1–2N0 breast cancer patients treated between 1989 to 1999 with mastectomy with clear margins but no adjuvant radiotherapy, histologic grade, lymphovascular invasion, T stage and systemic therapy use were significant predictive factors for locoregional recurrence. For instance, patients with grade 3 disease and concomitant lymphovascular invasion had a locoregional recurrence risk of 21.2% at 10 years. Other retrospective studies have identified tumor size greater than 2 cm, margin less than 2 mm, premenopausal status, lymphovascular invasion, patient age, and nonreceipt of systemic therapy as significant risk factors for locoregional recurrence. The 10-year risk of rates of locoregional recurrence ranged from 1.2% for node-negative women with no risk factors to 40.6% for women with three risk factors treated in these historical series. These results suggest that the risk of locoregional failure may be substantial in a subset of node-negative patients when PMRT is omitted. However, in the setting of more modern systemic therapy and recognition of the importance of tumor biology, the true risks remain uncertain. Margin status, in particular, has been the subject of recent observational studies, which have suggested that not all patients with close or even positive margins require radiation therapy. In sum, the decision regarding PMRT in node-negative patients, particularly those with smaller primary tumors, must be made after comprehensive consideration of the many factors identified in these studies to provide individualized risk assessment for each patient.

Postmastectomy Radiotherapy in Patients Treated With Neoadjuvant Chemotherapy

Although there is considerable evidence from randomized clinical trials evaluating the benefits of PMRT in patients with node-positive breast cancer, none of the patients treated in these trials received preoperative (neoadjuvant) chemotherapy. It has been unclear how best to apply the evidence from the historical PMRT trials to patients treated in the neoadjuvant setting. Therefore retrospective analyses have been undertaken to better guide patient selection in that setting.

Much of the initial information regarding the locoregional outcomes of patients treated with neoadjuvant chemotherapy came from the MD Anderson Cancer Center. Retrospective analyses suggested that patients with locally advanced disease on clinical presentation had a substantial risk of locoregional recurrence in the absence of radiation, even if they had a pathologic complete response to the chemotherapy, with much lower rates when PMRT was administered (33% vs. 3%). However, among the small number of patients with clinical stage I and II disease who achieved a pathologic complete response, there were no locoregional failures at all even in the absence of PMRT.

These findings were further supported by an analysis of patterns of failure after doxorubicin/cyclophosphamide-based neoadjuvant chemotherapy and mastectomy in patients enrolled in the large NSABP B18 and B27 trials (in whom PMRT was prohibited). This analysis revealed that 10-year locoregional recurrence rates in patients with a complete pathologic response in both the nodes and breast were less than 10%, regardless of the tumor size or whether the axilla was clinically positive or negative. On the other hand, patients with pathologically positive nodes after neoadjuvant chemotherapy had more substantial risk of locoregional recurrence after mastectomy without radiotherapy.

More recently, the CTNeoBC investigators combined the pathologic complete response evaluation with information with tumor biology to optimize locoregional management decision in patients receiving neoadjuvant chemotherapy. Their analysis included more than 13,000 patients enrolled in 12 trials of neoadjuvant chemotherapy and most patients received anthracycline and/or taxane-based regimens, with a few trials including trastuzumab. All patients with hormone receptor–positive tumors were supposed to receive at least 5 years of endocrine therapy. Radiotherapy was not administered in a randomized fashion and criteria for administration differed across the studies.

The pathologic complete response was evaluated based on three definitions:

1.

absence of invasive cancer and in situ disease in both the breast and nodes (ypT0 ypN0),
2.

absence of invasive cancer in the breast and nodes irrespective of the presence of in situ disease in the breast (ypT0/is ypN0), and
3.

absence of invasive cancer in the breast irrespective of in situ disease in the breast or disease in the nodes (ypT0/is)

The study has shown that more aggressive subtypes had higher rates of complete pathologic response: 50.3% in HER2-positive, hormone receptor–negative tumors treated with trastuzumab; 33.6% for triple-negative tumors; and 7.5% for hormone receptor–positive, HER2-negative grade 1/2 tumors. Patients with HER2 positive, hormone receptor–negative tumors and triple-negative tumors had the strongest association between pathologic complete response (CR) and long-term outcomes. Triple-negative patients with a pathologic CR had a low 5-year locoregional recurrence of 6.2% after mastectomy, but recurrence risk was significantly elevated at 22.1% when residual nodal disease was present. Similarly, HER2-positive, hormone receptor–negative patients who had mastectomy after neoadjuvant chemotherapy and had complete pathologic CR had a locoregional recurrence risk of only 4.1%, whereas it was 24% with residual nodal disease. Although treatments were not uniform across the patient populations, these data suggest that information regarding breast cancer subtypes and differences in pathologic response to neoadjuvant chemotherapy may be used to design clinical trials to address omission of PMRT in patients with low risk of locoregional recurrence and dose intensification in patients with high risk of locoregional failure. In fact, the NSABP B-51/Radiation Therapy Oncology Group (RTOG) 1304 trial, which opened in August 2013, is evaluating whether omission of PMRT may be acceptable in patients with T1–3N1 disease based on palpation or imaging (ultrasound, computed tomography, magnetic resonance imaging, or positron emission tomography) and pathologic confirmation of axillary involvement without a sentinel node biopsy, who achieve complete pathologic response in the axilla after neoadjuvant chemotherapy.

Only gold members can continue reading. Log In or Register to continue