Chest Wall Recurrences

In a pooled analysis of 5352 participants in International Breast Cancer Study Group (IBCSG) trials I through VII, 21.3% experienced a LRR event as a first relapse, and 53% of these were local or chest wall failures. Another series spanning the years 1980 to 2004, reported a 10-year cumulative incidence of LRR for T1 and T2 node-negative cancers of 5.2% with 73% of these recurrences involving the chest wall. However, in the absence of systemic treatments when risk factors such as larger tumor size, lymphovascular invasion, and positive margins are present, the incidence of LRR was 19.7%.

As mastectomy techniques have evolved to less disfiguring skin and nipple-sparing mastectomies, higher rates of LRR could have been expected but have not been found. Specifically, in a study-level meta-analysis of nipple-sparing mastectomies, nipple recurrence were found to be acceptably low at 1.4% after 3 to 5 years. These operative procedures are being extended to BRCA1 and BRCA2 mutation carriers and to patients who have received neoadjuvant therapy. Importantly, there is no indication that LRR rates are measurably higher with these more natural appearing mastectomy procedures.

Invasive locoregional skin or chest wall recurrences are uncommon in mastectomy-treated ductal carcinomas in situ. In a series of 10 patients, recurrences were diagnosed 1 to 16 years after mastectomy. Not surprisingly, residual breast tissue was identified in half of these LRR.

Nodal Recurrences

The extent of axillary nodal surgery has changed greatly since the era of radical mastectomy, when level I to III axillary node dissection was the norm. Regional recurrences ranged from 4% to 8%, with the bulk of these events occurring in the first 5 years. In their pooled analysis of modified radical mastectomy-treated women in seven randomized trials, Wallgren and colleagues found that 26% of LRR involved supraclavicular/infraclavicular nodes compared with 13% in axillary nodes.

Before sentinel node biopsy, recurrence rates were significantly higher when fewer axillary nodes are removed or in the presence of extranodal extension. Weir and colleagues reported that when fewer than 6 nodes were removed during axillary dissection the rate of recurrence was 5% compared with 1%, when more than 10 lymph nodes were excised. LRR are lower in node positive breast cancers as the number of uninvolved lymph nodes resected is higher.

The management of the axilla has shifted toward a selective approach with the implementation of lymphatic mapping and sentinel node biopsy. Local recurrence rates of less than 2% have been reported in clinically node-negative breast cancers after sentinel node biopsy, which indicates that the incidence of recurrences in the axilla is comparable to that observed with axillary node dissections. Lymphatic mapping and sentinel node biopsy after neoadjuvant therapy is gaining acceptance largely based on similar false-negative sentinel node identification rates.

Extraaxillary nodal recurrences refer to supraclavicular, infraclavicular (level III or interpectoral nodes), or internal mammary locations. In newer studies, supraclavicular nodal recurrences are the second most common site after the axilla. Until recent revision of clinical staging by the American Joint Committee on Cancer, supraclavicular nodes were considered synonymous with stage IV disease. Clinical detection of infraclavicular and internal mammary nodal recurrences are rare (see Fig. 61.2 ). However, with the increasing use of new generation computed tomography (CT) scanners and the incorporation of magnetic resonance imaging (MRI) and positron emission tomography (PET)/CT subclinical detection is becoming more prevalent.

Prior Radiation Therapy

Local failure rates vary with the extent of adjuvant treatments administered for the primary cancer. The Early Breast Cancer Trialists’ Collaborative Group studied the effects of comprehensive adjuvant radiotherapy, defined as supraclavicular, axillary, and internal mammary nodal basins, using individual patient level data from 22 randomized trials. There was no statistically significant effect on 10-year cumulative risk of ILRRs after mastectomy and axillary dissection for women with pathologically negative nodes. However, among the 3131 patients with positive nodes, postmastectomy radiation therapy decreased the overall rate of ILRRs from 26.0% (no RT) to 8.1% (RT), log-rank 2 p < .00001. Radiotherapy significantly decreased first LRRs from 20.3% to 3.8% in the group with one to three positive nodes, and from 32.0% to 13.0% in those with four or more. Similarly, radiation improved local control in women who had axillary sampling instead of dissection, reducing the risk of LRR by 0.61.

In a retrospective analysis of 1000 patients from the Danish Breast Cancer Cooperative Group 82 b&c clinical trials, the greatest absolute reduction in LRRs was in the subgroup of patients with unfavorable prognostic characteristics. Notably, in HER2-positive breast cancers, the rate of LRR with trastuzumab and postmastectomy radiation therapy was 0% at 5 years, indicating remarkable synergism between the two treatment modalities ( Table 61.1 ).

Prior Systemic Therapy

The risk of recurrence is determined by tumor biology and stage at presentation. The natural course of disease can be favorably altered by the use of adjuvant therapies tailored to tumor molecular markers. Chest wall recurrences occurred in 13% of patients who received adjuvant chemotherapy regimens but no radiation. However, the incidence was only 7.5% for those with one to three positive nodes, whereas it was 14% in those with four or more involved nodes. In a combined analysis of 5758 patients with node-positive cancers treated with adjuvant therapies without postmastectomy radiation, 12.2% of patients presented with an ILRR compared with 43%, whose first failure event was distant metastasis. The proportion of skin or chest wall recurrences was 56.9% compared with 22.6% supraclavicular, 11.7% axillary, and less than 1% parasternal and infraclavicular nodal recurrences. Predictive factors for local failure consisted of age less than 50, larger tumors, greater number of positive lymph nodes, and number of resected nodes. In this same study, the 10-year cumulative incidence of isolated regional failures was 3.5% or less, for women with one to three positive nodes, compared with 5.4% to 10.9% for those with four or more nodes, irrespective of tumor size.

Additional risk factors identified in a review of 12 studies involving 12,961 node-negative patients treated by mastectomy alone were less than 40 years of age, lymphovascular invasion, positive or close margins, and tumor size greater than 2 cm. Additionally, tumor biology influences recurrence events. Specifically, higher rates of distant metastasis and LRR are reported for triple-negative breast cancers compared with non–triple negative tumors.

Ma and colleagues reported different time intervals to LRR based on tumor receptors. The shortest median time interval to LRR was 18.2 months for HER2-positive cancers (largely without trastuzumab therapy), followed by 21.8 months for triple-negative and more than twofold longer for luminal A or B cancers. In a pooled analysis of 162 nonirradiated mastectomy-treated T1–2 N0 HER2-positive cancers from the British Columbia Cancer Agency and Massachusetts General Hospital, 5-year LRR-free survival was 99.4% with anti-HER2 therapy compared with 97.4% for those who did not receive trastuzumab. Similarly, patients with HER2 positive tumors experienced higher rates of LRR in the pre-trastuzumab era compared with those who received trastuzumab, wherein chest wall and nodal recurrence rates were cut by half (see Table 61.1 ). Interestingly, the combination of adjuvant trastuzumab with postmastectomy radiation in this nonrandomized retrospective series was associated with 0% LRR at 5 years.

Prior Neoadjuvant Chemotherapy

Patients with palpable breast cancers at presentation are increasingly directed toward neoadjuvant chemotherapy based on tumor molecular characteristics and disease stage. An LRR after neoadjuvant chemotherapy is highly associated with the pathologic response, with lower rate of relapse when a pathologic complete response (pCR) is attained. In a combined analysis of NSABP prospective randomized trials B-18 and B-27, tumor size greater than 5 cm (hazard ratio [HR] 1.58, p = .0095), clinical node-positive status (HR 1.53, p = .017), and pathologically residual nodal disease (HR 4.48, p = .01) were the most significant predictive factors for LRRs among 1071 mastectomy-treated patients without adjuvant radiation therapy. The 10-year cumulative incidence of chest wall recurrences ranged from 0% to 2.2% for patients who had a breast pCR and negative nodes, regardless of their clinical presentation. In clinically node-negative cancers, regional recurrences were slightly higher, ranging from 0% to 6.2%. Among patients who were classified as clinically node-negative and were then found to have residual disease in axillary nodes, the overall rate of LRR ranged from 11.2% to 22.4%.

Genomic Characterization

Another proposed approach to predict locoregional failures among endocrine-responsive cancers is through genomic analysis. Mamounas and coworkers used the 21-gene Recurrence Score (RS) panel to examine LRRs among estrogen receptor–positive, node-negative mastectomy-treated patients on NSABP protocols B-14 and B-20. Of these 505 nonirradiated mastectomy cases, 6.1% had LRRs: 18 chest wall or scar recurrences and 13 nodal or regional recurrences. A high RS was associated with a LRR in more than 10% of this subgroup compared with intermediate or low RS groups. Another study explored the utility of the 21-gene RS for predicting LRRs using the hormone receptor positive cohort of the NSABP B-28 protocol wherein anthracycline-based chemotherapy with or without paclitaxel was randomly assigned. The risk of a LRR for mastectomy-treated subjects with one to three positive nodes was 6% or less in the low RS category versus 9.6% and 23.6% for the intermediate and high RS categories respectively. The predictive value of the RS has been described in other small series. Similarly, EndoPredict, another genomic test, was investigated in the Austrian Breast Cancer Study Group trial 8. This test was prognostic for LRRs, with a 10-year local recurrence–free survival of 97.5% in the EndoPredict low-risk group versus 91% in the high-risk group.