Locoregional Recurrence of Breast Cancer

After Mastectomy

The mastectomy experience provides a reference point regarding the natural history of disease in an era of local therapy alone. Of women treated by radical mastectomy alone, in the landmark National Surgical Adjuvant Breast Project (NSABP) B-04 trial, the 25 year LRR and distant metastasis rates were 9% and 28% in node negative disease and 16% and 41% in women with node positive cancers. Contrastingly, in the Danish Breast Cancer Group (DBCG) 82b trial conducted in the early 1980s, 26% of patients treated by modified radical mastectomy and partial axillary dissection developed an ILRR.

In older series, recurrences after mastectomy more commonly involved the skin and/or chest wall, ranging from 49% to 83%, compared with 15% to 52% in nodal basins. Rates of chest wall recurrences were lower in patients with one to three positive nodes (7.5%) than for those with four or more involved nodes (14%). The International Breast Cancer Study Group (IBCSG) conducted a pooled analysis of 5352 participants in trials I to VII (1978–1993). LRR occurred in 21.3% of patients (with or without distant failure), 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 the recurrences involving the chest wall. These recurrences may represent progression of residual disease in lymphovascular spaces, subcutaneous tissues, as well as de novo transformation in residual glandular tissue.

Rates of local recurrence after mastectomy alone or mastectomy and breast reconstruction are similar. Higher rates of LRR after skin-sparing mastectomies were anticipated, but have not been found to be significant. In a systematic Cochrane review of 14 therapeutic mastectomy series, the rates of LRR after nipple-sparing versus skin-sparing mastectomy were 3.9% and 3.1%, respectively. Nipple recurrences are variable depending on the follow-up time interval, but overall acceptably low, ranging from 0% to 4.1%.

Surgical margin status has been linked to LRR events and studied far more extensively after lumpectomy than after mastectomy. Briefly, in a systematic review of 34 studies comprising 34,833 patients, the risk of a local recurrence with close or positive mastectomy margins for both invasive and ductal carcinoma in situ (DCIS) was approximately two- to threefold higher than those with negative margins.

Of note, invasive locoregional skin or chest wall recurrences are rare, at less than 1%, in mastectomy-treated DCIS. In a small series of 10 patients, recurrences were diagnosed 1 to 16 years after mastectomy, and half of these were notably identified within residual breast tissue.

After Lumpectomy

Early clinical trial experience with breast-conserving surgery provides a framework on the risk of IBTR with and without breast irradiation. The 12-year follow-up report of the NSABP B-06 trial found an overall cumulative incidence of IBTR among patients treated by lumpectomy alone of 35%, compared to 10% for those who received breast irradiation. Similarly, in a meta-analysis by the Early Breast Cancer Trialist’s Collaborative Group (EBCTCG), local recurrences were reduced by the use of breast irradiation from 29.2% to 10% in node-negative patients and from 46.5% to 13.1% in node-positive disease. The addition of adjuvant systemic therapies has proven to further decrease the rates of these local failure events. Specifically, the 10-year cumulative incidence of IBTR among 2669 women randomized in five node-positive NSABP clinical trials was 8.7%. Contrastingly, the 12-year cumulative incidence of IBTR was 6.6% among 3799 women randomized in five node-negative clinical trials who received systemic therapy, and 12.3% for those who did not receive adjuvant therapy. Patients undergoing breast-conserving surgery are also at risk for other locoregional recurrences (oLRRs) involving the skin, chest wall, or nodal basins. In these same trials, the cumulative risk of oLRRs was 6.0% among patients presenting with node-positive versus 1.9% for those with node-negative cancers. Age also plays a role. IBTR rates were 11.1% for women <50 years versus 6.1% for those ≥60 years, independent of the number of positive nodes. However, the rate of oLRR events were 4.4% in patients with one to three positive nodes versus 14.6% for those with ≥10 nodes.

Lumpectomy margin status has been extensively analyzed in countless series. The odds ratio of IBTR was 1.96 for close/positive margins compared to 1.0 for those with negative margins in a systematic review of 33 studies encompassing 28,162 patients treated by breast-conserving therapy.

The use of adjuvant whole-breast radiation has also been demonstrated to decrease local recurrences in patients with DCIS treated by breast-conserving surgery. The 15-year cumulative risk of IBTR for lumpectomy with radiotherapy was 8.9% to 10.0%, compared to 19.4% when radiation was omitted. Approximately, 50% of IBTR events after lumpectomy surgery for DCIS are invasive. Omission of radiation is associated with a 24.6% in-breast recurrence rate for high-grade DCIS compared to 14.4% for low or intermediate-grade DCIS.

The question of whether mutation carriers with an inherited disposition to develop breast cancer have a higher incidence of in-breast recurrences remains somewhat unclear. In a multinational study, BRCA1/2 mutation carriers were found to have a 23.5% 15-year cumulative risk of IBTR. Interestingly, unlike in sporadic breast cancers, most of the recurrences appeared at longer time intervals or in other quadrants or had different histologies, suggesting second primaries within the breast rather than true recurrence. Contrastingly, a large multicenter Dutch study compared outcomes between 191 BRCA1 mutation carriers and 5820 noncarriers, finding the 10-year risk of in-breast recurrences after breast-conserving therapy to be comparable, at 7.3% and 7.9%, respectively.

Nodal Recurrences

The most common sites of nodal recurrence are the axilla and supraclavicular fossa. Infraclavicular (level III axillary nodes), interpectoral, or internal mammary locations are less prevalent. Regional nodal recurrences have markedly decreased in patients with node-negative and node-positive disease over time. Regional recurrences in the era of routine level I-III ALND, ranged from 4% to 8%, with the bulk of these events occurring in the first 5 years. Supraclavicular/infraclavicular LRR predominated over axillary nodes in a pooled analysis of women treated by modified radical mastectomy. Similarly, supraclavicular recurrences were more prevalent than axillary failures among 2669 women with node-positive disease treated by breast-conserving therapy from 1984 to 1994 and 80.6% of these recurrences appeared within the first 5 years. In a randomized trial of 4004 patients treated between 1996 and 2004, radiation of the internal mammary node chain and supraclavicular basin decreased the recurrences from 5% in the control group to 3.2% in the radiation cohort, as well as improved breast cancer–related mortality with a median follow-up of 15.7 years.

The ipsilateral axilla was most common site of nodal recurrences in an analysis of 3083 patients treated by mastectomy and RT. The extent of axillary nodal resection also impacts the risk of axillary relapses. Weir and colleagues reported axillary recurrences in 5% of cases when fewer than six nodes were removed compared to 1% for 10 or more nodes. Similarly, in a meta-analysis of 22 randomized trials, LRR after mastectomy and axillary dissection in node-negative patients was 1.4%, but was markedly higher, at 16.1%, if only axillary sampling was performed. Recurrences in the node-positive subgroup were considerably higher, specifically 19.4% and 29%, respectively.

The management of the axilla has shifted toward a selective approach, with the implementation of lymphatic mapping and sentinel node biopsy for patients with clinically node-negative disease undergoing surgery first and in select patients after neoadjuvant chemotherapy. Local recurrence rates of less than 2% have been reported in patients with clinically node-negative breast cancer after sentinel node biopsy, comparable to observed rates with axillary node dissections. For example, in a retrospective analysis of 1060 patients with stage I and II disease with 95% undergoing sentinel node biopsy, only 2% experienced nodal recurrences, and 57% of these involved axillary nodes. Most recurrences were symptomatic and discovered on patient self-examination or physician examination. Only 10% were discovered on imaging studies. The false negative sentinel node identification rate in patients after neoadjuvant chemotherapy is similar to that of patients undergoing surgery first. At this time it is not clear whether axillary nodal recurrences following sentinel node resection alone have a more favorable outcome than such recurrences after axillary node dissections.

Contralateral nodal metastases are uncommon events involving the internal mammary, supraclavicular, or axillary nodes. These occurrences have been defined as distant failure events by the Maastricht Delphi Consensus on Event Definitions in 2014. Contralateral axillary nodal metastases have been reconsidered to be akin to LRR, based on possible aberrant lymphatic drainage following nodal surgery for the index cancer and are associated with better outcomes than in patients with metastatic disease. Distant disease-free survival (DFS) and overall survival (OS) rates of 62.5% and 82.6% at a median of 50.3 months correspond to more favorable outcomes observed in women experiencing IBTR after node-negative primary breast cancer. In another series, contralateral axillary node metastasis was associated with distant metastasis in 52% of patients within 2 years of the recurrence event. In a study by Magnoni et al, axillary node dissection was used in 91% of the contralateral axillary node metastases, and a DFS of 61% and an OS of 72% at 5 years were reported. Although radiotherapy was administered to the axillary basin in 40% of these recurrences, and included the contralateral breast in 14 of the 19 cases, use of this treatment did not improve outcomes.

Prior Adjuvant Radiation and Systemic Therapy

Postmastectomy radiation improves rates of local control. In a retrospective analysis of patients with one to three positive nodes treated between 1978 and 1997, LRR occurred in 14.5% after mastectomy alone compared with 6.1% who received postmastectomy RT. However, the benefit was not observed among patients treated during the years 2000 to 2007, when chemohormonal therapies were systematically used. Extraaxillary nodal recurrences in this cohort comprised the second most common event after chest wall recurrences.

Adjuvant radiation and systemic therapy have been shown to work synergistically and decrease rates of LRR for both lumpectomy- and mastectomy-treated patients. Molecular marker targeted therapies and chemotherapy have positively impacted locoregional events, as well as reduced distant failures. One of the first studies to demonstrate the synergism between adjuvant radiotherapy and systemic therapy was the NSABP B-06 study. Women who presented with node-negative disease experienced a 32% cumulative incidence of IBTR versus 12% for those who received breast irradiation. In contrast, women with node-positive disease received adjuvant chemotherapy and experienced a 12-year cumulative incidence of IBTR of 41% without radiation compared to 5% with radiation. Synergism between adjuvant tamoxifen therapy with postmastectomy radiation was also shown in the DBCG 82c trial, whereby LRR events decreased from 35% to 8%.

The EBCTCG studied the effects of comprehensive adjuvant radiotherapy, directed to supraclavicular, axillary, and internal mammary nodal basins, using individual patient–level data from 22 randomized trials. Systemic therapies were administered to 47% of node-negative and 91% of node-positive patients. Radiotherapy had no statistically significant effect on the 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 significantly decreased 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 involved nodes. In 8 of these 22 trials, RT was administered to regional nodes alone without the chest wall. RT reduced LRR events from 20.9% to 6.8% for the node-positive cohort but was not beneficial in node-negative disease. The effects of systemic therapies alone in 5758 patients with node-positive cancers resulted in a 12.2% incidence of ILRR, while 43% of patients experienced distant metastasis as a first failure event. Skin or chest wall recurrences were most commonly found and accounted for 56.9% of ILRRs, followed by 22.6% supraclavicular, 11.7% axillary, and less than 1% parasternal or infraclavicular nodal recurrences. Predictive factors for local failure consisted of the following—age less than 50 years, larger tumors, greater number of positive lymph nodes, and lower number of resected nodes. Specifically, the 10-year cumulative incidence of isolated regional failures was 3.5% or less for women with one to three positive nodes, compared to 5.4% to 10.9% for those with four or more nodes, irrespective of tumor size.

Prior to routine HER2 testing, 3% to 32% of patients with stage I to III breast cancer developed an LRR within 10 years of mastectomy, either as a first event or as accompanying distant disease. The use of targeted anti-HER2 therapies in combination with adjuvant chemotherapy has further decreased the number of LRRs. The rate of LRR in HER2-positive cancers was reduced to 0% at 5 years with trastuzumab and postmastectomy radiation therapy, indicating remarkable synergism between the two treatment modalities. In a joint analysis of NCCTG N9831 and NSABP B-31 trials encompassing 4046 women, there were 119 locoregional events after chemotherapy alone compared to 82 with chemotherapy and trastuzumab.

Systemic treatment with endocrine therapy enhances local control of disease. The SOFT and TEXT trials included premenopausal women with estrogen-dependent cancers, 57% of whom also received chemotherapy in addition to endocrine therapy. For this cohort the overall LRR was 4.2%, and was even lower among those who received ovarian suppression and exemestane compared to tamoxifen alone.