Are There Patients with T1-T2, Node-Negative Breast Cancer Who Are “High-Risk” for Locoregional Recurrence?

Abstract

BACKGROUND

Indications for postmastectomy radiotherapy (PMRT) in T1-T2, node negative (N0) breast cancer with “high-risk” features are controversial. The EORTC 22922 and MA20 trials reporting improved 10-year disease-free survival with nodal irradiation included high-risk N0 patients, but benefits in patients receiving modern systemic therapy are uncertain.

METHODS

We retrospectively identified patients with T1-T2N0 disease treated with mastectomy from 1/2006–12/2011. High-risk features included age <40 years, multifocality/multicentricity, lymphovascular invasion (LVI), medial/central tumor location, and high nuclear grade.

RESULTS

Among 672 eligible patients, only 15 received PMRT and were excluded. Of the remaining 657, 187(28%) had 1 risk factor and 449(67%) patients had ≥2 high-risk features. 36 patients with unknown grade were excluded from risk analysis. 98% had sentinel node biopsy alone; 86% had adjuvant systemic therapy. At median 5.6 years follow-up, the LRR rate was 4.7% (n=31). Increasing tumor size was associated with LRR (HR 1.70, p=0.006), while other high-risk features were not (all p>0.05). Receipt of systemic therapy decreased LRR (HR 0.40, p=0.03). Although crude LRR rates increased from 3.8% to 9.4% with 1vs ≥4 high-risk features, risk factor number was not significantly associated with LRR (p=0.54).

CONCLUSIONS

A low crude LRR rate (4.7%) was seen in this large unselected cohort of T1-T2N0 cancers with high-risk features treated by mastectomy and systemic therapy without PMRT. While increasing tumor size and systemic therapy omission were predictive, other features did not confer a higher LRR risk either independently or together, and do not by themselves mandate PMRT use in this population.

INTRODUCTION

Patients with T1-T2, node-negative (N0) breast cancer undergoing mastectomy are typically considered to be at low-risk for locoregional recurrence (LRR), and postmastectomy radiotherapy (PMRT) is not a routine part of treatment.¹ Indications for PMRT in T1-T2N0 patients thought to be at increased risk for LRR are controversial, due to a lack of consensus on what constitutes “high-risk” in the absence of nodal metastasis.^1-7 The most consistently described characteristics associated with LRR in retrospective studies include age less than 40 years, the presence of lymphovascular invasion (LVI), tumor size, and high nuclear grade, with reported rates of LRR as high as 20% in women with multiple risk factors.^{1, 4} Multifocal or multicentric cancer has also been associated with increased risk of LRR.^{8, 9} While all tumor locations drain primarily to the axilla, tumors in the medial or central breast and those with involvement of internal mammary lymph nodes have been suggested to have a worse prognosis.^{10, 11} Notably, these studies include patients primarily treated from the 1980s to the mid-2000s and do not reflect the use of modern systemic therapies.^1-7

More recently, 10-year results of the National Cancer Institute of Canada Clinical Trials Group MA20 and European Organisation for Research and Treatment of Cancer (EORTC) 22922-10925 randomized controlled trials showed improved disease-free survival with the addition of regional nodal irradiation to whole-breast radiotherapy for early-stage breast cancers in patient populations which included “high-risk” N0 patients.^{12, 13} The MA20 trial randomized patients undergoing lumpectomy for T1-T3 disease, including both node-positive patients and node-negative patients with “high-risk” features, defined as tumor size > 5 cm, grade 3 histology, estrogen receptor (ER) negativity, or LVI.¹² The EORTC 22922-10925 trial randomized patients having either lumpectomy or mastectomy for T1-T3 cancers, including both node-positive patients and node-negative patients deemed high-risk due to central or medial tumor location.¹³ Node-negative patients comprised 10% and 44% of the MA20 and EORTC study populations, respectively. Although only 24% of patients in the EORTC trial were treated with mastectomy, there is no reason to believe that the benefit of nodal irradiation is related to the use of lumpectomy or mastectomy.^{12, 13}

The impact of these high-risk disease characteristics on LRR in patients with T1-T2, node-negative cancers undergoing mastectomy and treated with modern systemic therapy remains uncertain. We sought to evaluate locoregional control in a contemporary study cohort.

METHODS

Following Institutional Review Board approval, all women diagnosed with T1-T2, node-negative (T1-T2N0) invasive breast cancer with at least one high-risk feature who underwent mastectomy at Memorial Sloan Kettering Cancer Center between January 2006 and December 2011 were identified. Node negativity was defined as the absence of any tumor, including isolated tumor cells or micrometastases, in the sentinel nodes. Sentinel node processing was uniform throughout the study period and included serial sections at 2 levels 50 microns apart in all sentinel nodes which were negative on initial H&E staining. Immunohistochemical staining for cytokeratin and an additional H&E stain were performed on deeper levels. High-risk features were defined as age less than 40 years at diagnosis, multifocal or multicentric tumor, presence of LVI, central or medial tumor location, and high nuclear grade. Patients with recurrent disease, bilateral tumors, and those who received neoadjuvant chemotherapy were excluded.

Standard clinicopathologic data were collected, including patient age at diagnosis, clinical presentation of disease, tumor size, nuclear grade, subtype, and final tumor pathologic findings. Rates of administration of radiotherapy and systemic therapies, and regimens utilized, were assessed. The primary outcome of interest was rate of LRR. Locoregional recurrence was defined as a biopsy-proven recurrence of disease in the ipsilateral chest wall, or in the ipsilateral regional lymph nodes including the axillary, internal mammary or supraclavicular lymph nodes. Development of metastatic disease at any other site was considered a distant recurrence.

Continuous variables were compared using the Wilcoxon test, and categorical variables were compared using the Chi square or Fisher’s exact tests. Crude events are reported; however, all analyses are conducted using time-to-event methods that account for censoring and follow-up time. Kaplan-Meier methods were used to analyze rates of LRR, and the log-rank test used to examine the associations between categorical risk covariates and LRR while Cox regression was used for continuous variables and to estimate hazard ratios. The number of high-risk features ranged from 0 to 5. The risk score was calculated by adding the presence or absence of five risk factors in patients with complete data with equal weight for each factor. For all competing risk analyses that differentiate between types of recurrence, the first event was counted as the primary event, and deaths due to any cause were analyzed with the distant recurrence category as per the Standardized Definitions for Efficacy End Points (STEEP) system for breast cancer trials.¹⁴ The Gray’s test was used to evaluate associations. All statistical analysis was performed using SAS 9.2 (SAS Institute, Cary, NC, USA) and R version 3.1.1 (https://www.r-project.org/). Any P value less than 0.05 was considered statistically significant.

RESULTS

From January 2006 to December 2011, 672 women with T1-T2N0 invasive breast cancer and at least one high-risk feature underwent mastectomy. Of these patients, 187 (28%) had a single high-risk feature: 21 (3%) were age < 40 years, 132 (20%) had multifocal or multicentric tumors, and 34 (5%) displayed LVI. More than one high-risk feature was seen in 449 (67%) patients, with two, three, and four or more risk factors seen in 268 (40%), 146 (22%), and 35 (5%), respectively. The remaining 36 (5%) of patients were excluded from the risk analysis due to unknown nuclear grade.

Postmastectomy radiotherapy was received by only 15 (2%) of the high-risk node negative patients during the 6-year study period. The remaining 657 patients treated without PMRT comprised the study cohort. Clinicopathologic characteristics are summarized in Table 1. Median patient age was 49 years, median tumor size was 1.4 cm, and the majority of patients had ductal histology (86%) and estrogen receptor (ER) positive (78%) disease. Sentinel lymph node biopsy (SLNB) alone was performed in 641 (98%) patients, with a median of 4 (range 1–15) sentinel lymph nodes removed. The remaining 16 patients had axillary dissection for failed mapping (n = 3) or physician choice (n = 13). Adjuvant systemic therapy was received by 565 (86%) patients: 268 (41%) received both chemotherapy and endocrine therapy, 90 (14%) received chemotherapy alone, and 207 (32%) received endocrine therapy alone. Eighty-six (13%) patients received no systemic therapy, and receipt of treatment was unknown in 6 patients. Of the 358 patients who received chemotherapy, 181 (51%) had doxorubicin, cyclophosphamide, and paclitaxel (ACT), 137 (38%) had cyclophosphamide, methotrexate, and 5-fluorouracil (CMF), 38 (11%) had docetaxel and cyclophosphamide (TC), and 2 patients received other regimens on trial protocols. Eighty-six (70%) of 123 human epidermal growth factor 2-overexpressing (HER2+) patients also received trastuzumab. Among 513 ER positive (ER+) patients, 466 (91%) received endocrine therapy.

TABLE 1.

Clinicopathologic Characteristics of Patients with T1-T2 Node-Negative Invasive Breast Cancer Treated by Mastectomy Without PMRT

	No PMRT n = 657
Age, years	49 (24–89)
Histology
Ductal	566 (86%)
Lobular or mixed	90 (14%)
Other	1 (0.2%)
Pathologic tumor size, cm	1.4 (< 0.1–5.0)
Associated DCIS
< 25%	404 (61%)
> 25%	245 (37%)
Unknown	8 (1%)
Grade
Low	31 (5%)
Intermediate	326 (50%)
High	266 (40%)
Unknown	34 (5%)
Receptor status
ER+/HER2–	438 (67%)
ER+/HER2+	75 (11%)
ER–/HER2+	48 (7%)
ER–/HER2–	70 (11%)
Unknown*	26 (4%)
Lymphovascular invasion	232 (35%)
Multifocal/multicentric	447 (68%)
Tumor location
Medial or central	226 (34%)
Lateral	431 (66%)
# of risk factors
1	183 (28%)
2	265 (40%)
3	143 (22%)
4	28 (4%)
5	4 (1%)
Unknown**	34 (5%)

Data reported as median (range) or n (%)

^*Cases with insufficient sample to test ER and/or HER2

^**Those with unknown nuclear grade excluded from risk factor analysis

Abbreviations: PMRT, postmastectomy radiotherapy; DCIS, ductal carcinoma in situ; ER, estrogen receptor; HER2, human epidermal growth factor 2

After a median follow-up period of 5.6 years (range 0–115 months), 31 (4.7%) patients in the study cohort have experienced LRR, of whom 22 had isolated LRR and 9 had simultaneous (diagnosed within one month) locoregional and distant recurrence. Twenty-three (3.5%) patients have experienced distant recurrence alone. The median time to LRR among the 31 patients who experienced an LRR was 33 months (range 1–79 months). Location of LRR was the chest wall alone in 14 (45%), synchronous chest wall and axillary nodes in 3 (10%), axillary nodes alone in 9 (29%), supraclavicular nodes alone in 2 (6%), internal mammary nodes alone in 1 (3%), synchronous axillary and supraclavicular nodes in 1 (3%), and synchronous axillary and internal mammary nodes in 1 (3%) case. Overall, local recurrence was seen in 17 (2.6%), and regional recurrence in 17 (2.6%) patients, respectively. Table 2 summarizes clinicopathologic characteristics of the 31 patients who experienced LRR. These patients had a median age of 49 years, median tumor size of 1.6 cm, and the majority had ER+ (81%) ductal (84%) cancers. Systemic therapy as recommended was not received by 10 (32%) of patients with LRR, including chemotherapy for 3 patients (2 HER2 positive), and endocrine therapy for 7 patients.

TABLE 2.

Clinicopathologic Characteristics of Patients with T1-T2 Node-Negative Invasive Breast Cancer Treated by Mastectomy Without PMRT Who Experienced Locoregional Recurrence

	Experienced LRR n = 31
Age at diagnosis, years	49 (33–86)
Tumor size, cm	1.6 (< 0.1–4.8)
Ductal histology	26 (84%)
High nuclear grade	15 (48%)
Receptor status
ER+/HER2–	21 (68%)
ER+/HER2+	4 (13%)
ER–/HER2+	3 (10%)
ER–/HER2–	3 (10%)
Lymphovascular invasion	14 (45%)
Multifocal/multicentric	19 (61%)
Medial or central tumor location	11 (35%)
# of risk factors
1	7 (23%)
2	14 (45%)
3	7 (23%)
4	3 (10%)
5	0 (0%)

Data reported as median (range) or n (%)

Abbreviations: PMRT, postmastectomy radiotherapy; LRR, locoregional recurrence; ER, estrogen receptor; HER2, human epidermal growth factor 2

Crude rates of LRR among those with one, two, three, or four or more high-risk features were 3.8%, 5.3%, 4.9%, and 9.4%, respectively. The hazard ratios and P values that evaluate the relationship of patient and tumor characteristics with LRR are provided in Table 3. Increasing tumor size was associated with LRR (hazard ratio [HR] 1.70, 95% confidence interval [CI] 1.26–2.29, P = 0.006), while age, histology, nuclear grade, receptor status, LVI, multifocal or multicentric tumor, and location of tumor were not associated with LRR (all P > 0.05). Receipt of systemic therapy decreased LRR (HR 0.40, 95% CI 0.17–0.92, P = 0.03). A Kaplan-Meier curve of LRR by number of high-risk features (Figure 1) showed no significant relationship between LRR and number of risk factors when comparing patients with 1 vs 2 vs 3 vs ≥4 high-risk features (P = 0.54). The 5-year LRR-free survival for patients with 1 risk factor was 96% (95% CI 0.91–0.98), for 2 risk factors 95% (95% CI 0.91–0.97), 3 risk factors 95% (95% CI 0.90–0.99) and for 4 or 5 risk factors 92% (95% CI 0.71–0.98). There was no difference in cumulative incidence of events with an increasing number of risk factors among any type of recurrence: isolated LRR (P = 0.77), combined LRR and distant (P = 0.6), and distant events (P = 0.11), respectively (Figure 2).

TABLE 3.

Locoregional Failures by Patient and Tumor Characteristics

Characteristic	Number of patients n = 657	Number of LRR n (%)	Hazard Ratio (95% CI)	P‡
Age at diagnosis
< 40 years	149	9 (6.0%)	.94 (.88, .97)	0.38
≥40 years	508	22 (4.3%)	.96 (.93, .97)
Tumor size (continuous)
≤2 cm (T1)	532	19 (3.6%)	HR† (95% CI)	0.006
2–5 cm (T2)	125	12 (9.6%)	1.70 (1.26, 2.29)
Histology*
Ductal	566	26 (4.6%)	.96 (.93, .97)	0.90
Lobular or mixed	90	4 (4.4%)	.94 (.86, .98)
Nuclear grade**
Low	31	0 (0%)	1.00
Intermediate	326	16 (4.9%)	.95 (.92, .97)	0.41
High	266	15 (5.6%)	.94 (.91, .97)
Receptor status**
ER+/HER2–	438	21 (4.8%)	.95 (.86, .98)
ER+/HER2+	75	4 (5.3%)	.94 (.82, .98)	0.90
ER–/HER2+	48	3 (6.3%)	.96 (.93, .97)
ER–/HER2–	70	3 (4.3%)	.93 (.82, .97)
Lymphovascular invasion
Yes	232	14 (6.0%)	.96 (.93, .98)	0.24
No	425	17 (4.0%)	.95 (.90, .97)
Multifocal/multicentric
Yes	447	19 (4.3%)	.95 (.90, .97)	0.43
No	210	12 (5.7%)	.96 (.95, .97)
Tumor location
Medial or central	226	11 (4.9%)	.95 (.92, .97)	0.97
Lateral	431	20 (4.6%)	.96 (.92, .98)

Abbreviations: LRR, locoregional recurrence; CI, confidence interval; HR, hazard ratio; ER, estrogen receptor; HER2, human epidermal growth factor 2

^*n = 1 with metaplastic squamous histology excluded from analysis

^**Patients with unknown variable data excluded from grade (n = 34) and receptor status (n = 26) analyses

^†Hazard ratio calculated using tumor size as a continuous variable

^‡P values are from the Log-Rank test with the exception of continuous tumor size, which is from the Wald Chi-Square test

Figure 1.

Kaplan-Meier estimates of locoregional recurrence by number of high-risk features.

1 vs 2 vs 3 vs 4–5 risk factors: Log-Rank P = 0.54

Abbreviations: LRR, locoregional recurrence, RF, risk factor

Figure 2.

Cumulative incidence of events by number of risk factors.

Isolated LRR: 22 events, Gray’s test P = 0.77

Combined LRR and distant: 9 events, Gray’s test P = 0.6

Deaths due to any cause included in ‘Distant’ category, as per the Standardized Definitions for Efficacy End Points (STEEP) system for breast cancer trials.¹⁴

Abbreviations: LRR, locoregional recurrence; CIF, cumulative incidence function

DISCUSSION

Our study demonstrates a low rate of locoregional recurrence of 4.7% at 5.6 years of median follow-up in a large, contemporary cohort of T1-T2, node-negative patients with high-risk disease features treated with mastectomy and systemic therapy, without PMRT. Tumor size was significantly associated with LRR (HR 1.70), and LRR was decreased with the receipt of systemic therapy (HR 0.40); however, no significant increase in LRR was observed with an increasing number of high-risk features (P = 0.54). The observed LRR rate is lower than past reported rates of 5.2–9.2% in similar populations of node-negative patients treated without PMRT (Table 4).^{1, 4, 6, 15} While PMRT has been shown to reduce rates of LRR, the Early Breast Cancer Trialists’ Collaborative Group meta-analysis showed an absolute reduction in LRR of only 4% among node-negative patients treated with mastectomy and PMRT, and no improvement in 15-year survival.¹⁵

TABLE 4.

Results of past studies of LRR among T1-T2, node-negative breast cancers treated with mastectomy without PMRT^{1, 3-6}

Study	N	% received systemic therapy	Treatment regimen	Median follow-up (yrs)	LRR rate (%)
Abi-Raad et al	1136	38.4%	Not stated	9	5.2%
Truong et al	1994	80.5%	Trastuzumab used in 53% of HER2+ patients, otherwise not stated	4.3	3.1%
Yildirim et al	502	56% chemotherapy, 43% tamoxifen	Among those who received chemotherapy, 67% had CMF	6.4	2.8%
Wallgren et al	1275	33% chemotherapy	One cycle of CMF	15.3	16%, 19%*
Sharma et al	753	Cannot be calculated	Not stated	7.5	2.1%
Mamtani et al (Current study)	657	86%	Among those who received chemotherapy, 51% had ACT; Trastuzumab used in 70% of HER2+ patients	5.6	4.7%

^*Reported as 16% among pre-menopausal patients and 19% among post-menopausal patients

The benefit of PMRT in patients with T1-T2N0 cancer with one or multiple defined high-risk features remains controversial. The potential for a higher risk of LRR conferred by these characteristics was demonstrated in retrospective studies of patients treated in older time periods prior to the widespread use of systemic therapy in node-negative breast cancer.^1-7 Sharma et al found that age ≤40 years was an independent predictor of LRR in a study of 753 patients with T1-T2N0 cancer treated with mastectomy without PMRT from 1997 to 2002.³ In a population of 1136 similar patients from 1980 to 2004, Abi-Raad et al found young age, tumor size > 2 cm, and LVI to be predictive of LRR.¹ Truong et al demonstrated higher risk with LVI, young age, high grade and tumor size > 2 cm among 1505 such patients treated from 1989 to 1999.⁴ Similarly, Yildirim et al found tumor size > 2 cm, LVI, and high grade in a cohort of 502 patients from 1990 to 2004 to be predictive of LRR.⁵ However, conflicting results have also been reported. Voogd et al found no influence of age on LRR in patients undergoing mastectomy for node-negative disease¹⁶, while Janni et al showed no relationship between LRR and grade or tumor size among T1-T2N0 patients having mastectomy.¹⁷ While age < 40 years was associated with a lower rate of relapse-free and overall survival for HER2 positive patients treated between 1986 and 1992, this difference was not seen in the cohort treated between 2004 and 2007, and no difference in outcome based on age was noted in the study of Sheridan et al¹⁸, suggesting that effective systemic therapy many diminish the effect of the poorer prognostic profile often seen in younger women. The only disease feature found to be predictive of LRR in this study was tumor size (HR 1.70, 95% CI 1.26–2.29, P = 0.006), while all other studied high-risk features were not predictive of LRR (all P > 0.05).

The low rate of LRR observed in this study most likely reflects results of present-day practice, in which the vast majority of breast cancer patients receive systemic treatment tailored to individual tumor biology. This is supported by the observed association between systemic therapy and disease-free survival in this study. Adjuvant endocrine and/or chemotherapy were received by 86% of patients in this study, as compared to 38–77% in previous studies examining this question.^{1, 3-5} Systemic targeted therapies have been demonstrated in randomized trials to reduce both local and systemic disease recurrence.^19-21 In a study of 86,598 patients treated in phase 3 trials between 1990 and 2011, Bouganim et al found that LRR as a proportion of all recurrences decreased from 30% to 15%, independent of patient age, use of radiotherapy, or treatment with lumpectomy or mastectomy.²² Importantly, as improvements in systemic therapy lead to improvements in distant disease-free or overall survival, parallel reductions in local recurrence are seen. For example, while chemotherapy alone reduces LRR by ≥30%, the addition of trastuzumab in HER2 overexpressing patients leads to a further 25–53% reduction.^{20, 23, 24} This concept is supported by our finding that while the rate of LRR did increase to 9.4% in patients with four risk factors, this rate is significantly lower than the historically reported rate of 20% in patients with multiple risk factors in older study cohorts.^{1, 4} The importance of systemic therapy was noted in our study, with a significant reduction in LRR with the receipt of adjuvant systemic treatment. Among patients who experienced LRR, recommended chemotherapy was not received by 3 (10%) patients, and 7 (28%) of 25 ER+ patients with LRR declined endocrine therapy. In an era where tumor biology is recognized as a principal determinant of disease recurrence and indications for systemic therapy in low-risk patients have expanded, re-evaluation of criteria for PMRT use is appropriate.

The majority of recurrences in this cohort involved the chest wall, consistent with previously demonstrated patterns.¹ However, 8 of 22 isolated LRR events involved the axillary nodes alone. This is a higher proportion than anticipated, compared to a very low 0.12% axillary relapse observed in a study including 2340 patients with negative sentinel nodes who underwent SLNB alone.²⁵ We hypothesize that this may be, in part, due to false-negative SLNBs, as 7 of the 8 patients underwent SLNB alone. A meta-analysis including 9220 patients with pathologically negative sentinel nodes followed by completion axillary dissection reported a crude overall false-negative rate (FNR) of 8.6%.²⁶ However, the FNR has been shown to decrease to less than 5% with the retrieval of multiple nodes,²⁷ and a median of 4 lymph nodes (range 2–17) were obtained among the 8 patients with isolated axillary recurrence in our cohort.

Recent support for comprehensive nodal irradiation has been provided from randomized controlled trials demonstrating a reduction in regional (nodal) recurrence with this approach, although observed in study populations comprised largely of node-positive patients.^{12, 13} The 10-year results of the EORTC 22922 trial showed a significant improvement in disease-free survival with the addition of internal mammary and medial supraclavicular irradiation to whole-breast radiotherapy.¹³ T1-T2N0 patients with high-risk features undergoing mastectomy comprised only a small minority of the heterogeneous study population. While location of LRR was not provided in this subset, in the study arm treated with comprehensive irradiation, the rate of regional nodal recurrence was reduced to 2.7%, from 4.2% in the control arm treated with whole-breast irradiation only.¹³ We observed a regional nodal recurrence rate of 2.6% in our study, suggesting limited benefit to adding nodal irradiation in this population, given the already very low rates of nodal recurrence without any radiotherapy, despite the presence of high-risk disease features.

There are limitations to our study. It is a retrospective review and potentially subject to selection bias. However, only 15 patients (2%) treated during the study time period underwent PMRT, making it unlikely that the study population is a particularly favorable cohort. In addition, the number of patients with 4 or more risk factors was small, limiting the ability to assess LRR in this subgroup. However, this reflects the distribution of risk factors in a node-negative population. A group of high-volume, dedicated breast surgeons performed the surgeries on these patients, and this may also have contributed to the low rates of LRR that were seen, so the generalizability of the results is uncertain.

CONCLUSIONS

The overall low crude locoregional recurrence rate of 4.7%, with a median of 5.6 years of follow-up in our large, contemporary cohort of T1-T2 node-negative patients with high-risk features treated with mastectomy demonstrates that excellent locoregional control can be obtained in this population without PMRT. This is inarguably affected by improvements in, and increased utilization of systemic therapies tailored to individual tumor biology, with an observed significant association between systemic therapy and LRR-free survival. Given the known possible complications of radiation, appropriate risk stratification to identify patients who would benefit most from PMRT is crucial. While an increasing tumor size did confer a higher risk for LRR in these patients, all other high-risk disease features assessed, either independently or together, did not confer a higher risk of locoregional relapse. These results strongly suggest that features traditionally thought to convey an increased risk of LRR, in the setting of appropriate adjuvant systemic treatment, do not by themselves support the use of PMRT in T1-T2N0 patients undergoing mastectomy. Results from the ongoing Selective Use of Postoperative Radiotherapy after Mastectomy (SUPREMO) trial, which includes a subset of node-negative, high-risk patients, will provide important information regarding the risks and benefits of PMRT in this population.