Impact of Post-Mastectomy Radiation Therapy for Sentinel Lymph Node Micrometastases in Early-Stage Breast Cancer Patients.

BACKGROUND The association of radiotherapy with breast cancer survival in patients who underwent a mastectomy and had micrometastases in the sentinel lymph node is unclear. MATERIAL AND METHODS The survival benefit of radiotherapy was examined in patients with T0/1-T2N1mi breast cancer undergoing mastectomy plus sentinel lymph node biopsy (SLNB). Kaplan-Meier curves were employed for survival analysis and competing risk analysis, and a propensity score matching (PSM) cohort was enrolled to investigate whether such patients benefit from radiotherapy. RESULTS We identified 2864 patients in the SEER database from 2004 to 2015. All eligible patients were divided into the radiotherapy and the no-radiotherapy cohorts. With the median follow-up of 53 months, 5-year breast cancer-specific survival (BCSS) was 94.4% vs 95.2% (P=0.135), and 5-year overall survival (OS) was 91.2% vs 90.1% (P=0.466) in the radiotherapy cohorts and no-radiotherapy cohorts, respectively. The results of the competing risk analysis showed a comparable 5-year cumulative incidence of breast cancer-specific death (BCSD) in the radiotherapy and no-radiotherapy groups (5.5% vs 4.7%, P=0.107) but a higher 5-year cumulative incidence of other causes of death (OCD) in the no-radiotherapy cohort (3.3% vs 5.3%, P=0.011). No significant difference was observed for BCSS or OS in the PSM cohort. CONCLUSIONS Radiotherapy has no benefit for patients with T0/1-T2 breast cancer undergoing mastectomy with N1mi disease on SLNB. This analysis provides evidence that radiotherapy may safely be omitted in this group of patients.

Background

Axillary lymph node status is one of the most important prognostic factors for breast cancer survival [1]. Axillary lymph node dissection (ALND) was the criterion standard in patients with regional nodal metastasis in past decades. An axillary lymph node examination dissected by the largest diameter was based on hematoxylin and eosin (H&E) staining. However, when lymph node slices are thicker than 2 mm, micrometastases of cancer within the slice can be missed and underestimate the risk of lymph node involvement [2-10].

Sentinel lymph node biopsy (SLNB) has replaced ALND as the surgical standard staging procedure for breast cancer patients clinically considered to be node-negative over the last decade [11]. Since the introduction of SLNB, a more comprehensive examination of a smaller number of sentinel lymph nodes, routine use of a step-sectioning procedure, and immunohistochemistry resulted in increased detection of micrometastases [7]. Although several prospective studies have been published suggesting that ALND can be safely omitted in breast-conserving surgery patients with micrometastases in the sentinel lymph node (SLN), these studies included only a few patients who underwent mastectomy surgery [12-14]. The role of radiation therapy in post-mastectomy patients with micrometastases in the SLN is uncertain. The question remains whether further axillary radiotherapy is indicated in patients with micrometastases in the SLN of T0/1-T2 tumors who underwent a mastectomy.

The present study aimed to determine whether radiation therapy had any impact on the breast cancer-specific survival (BCSS) and overall survival (OS) of breast cancer patients undergoing mastectomy with SLN micrometastases in whom a completion ALND was omitted, based on data from the Surveillance, Epidemiology, and End Results (SEER) database.

Material and Methods

Data Source and Cohort Selection

Data were obtained from the SEER 18-regions database (Incidence-SEER 18 Regs Research Data [with additional treatment fields], Nov 2018 Sub [1975–2016 varying]) using SEER*Stat version 8.3.9 (http://seer.cancer.gov/seerstat). The SEER registries routinely collect data on patient demographics, age at diagnosis, primary tumor size, stage at diagnosis, survival months, and follow-up vital status. The International Classification of Diseases for Oncology and histology codes identified women older than 18 and under 80 years diagnosed with primary breast cancer from 2004 to 2015.

Patients were included based on the following criteria: histologically diagnosed with staged T0/1-T2 breast cancer who had a mastectomy, unilateral early breast cancer as the first primary tumor, evaluation of lymph node status determined by SLNB as pN1mi, M0, the number of SLN examined <6 involved, and primary surgical procedure with mastectomy plus SLNB without completion ALND. The SEER database also provides information regarding chemotherapy and radiotherapy; thus, these data are also included. Information regarding the HER2 status and molecular subtypes was only available after 2010 and was excluded from our analysis. Patients were excluded if they underwent ALND (with or without SLNB), did not undergo a lymph node evaluation, or did not undergo primary surgery.

Clinicopathological Variables

The following clinicopathological variables were collected: age at diagnosis, race, marital status, median household income, T stage, nuclear grade, estrogen-receptor (ER) status, progesterone-receptor (PR) status, and results of SLNB pathological examination, as well as administration of chemotherapy and radiotherapy. Age was categorized as <50 or >50 years. Ethnicity was classified into White or non-White. Patients were divided into married, single, or divorced, and widowed were combined as unmarried based on marital status. Socioeconomic status was divided into Quartile 1 (<$52 620), Quartile 2 ($52 621–$60 260), Quartile 3 ($60 261–$74 610), and Quartile 4 (>$74 610). Tumor grade was classified into I+II and III. T staging was classed as T0/1 or T2, and micro-metastatic disease (staged as pN1mi) was defined as deposits >0.2 to ≤2.0 mm according to the 6th edition of the American Joint Committee on Cancer (AJCC). Chemotherapy was categorized as “yes” or “no/unknown.” For the reasons previously stated, information concerning hormone therapy and anti-HER2 treatment was not collected.

The primary surgical procedure included a total mastectomy and SLNB, and patients were further categorized into 2 groups by whether they received radiotherapy (radiotherapy group) or not (no-radiotherapy group).

Statistical Analyses

Continuous variables were compared using the t test, and the chi-square test or Fisher’s exact test were used to compare categorical variables to describe the patients’ and tumors’ characteristics. The primary aim of this study was to compare patients with staged T0/1-T2 breast cancer undergoing mastectomy plus SLNB to determine if there was any benefit in BCSS and OS when N1mi patients received radiotherapy. Survival probabilities for BCSS and OS were carried out by plotting survival curves using the Kaplan-Meier method for various patients. The log-rank test was used to analyze associations between clinicopathologic variables and survival of groups based on whether they received radiotherapy or not in N1mi patients. Subgroup analyses were also performed to determine whether radiotherapy could benefit different groups.

Univariable and multivariable Cox proportional hazards regression models were performed to assess clinical-pathological features. A log-rank test was used for univariate comparison; variables with a statistically significant P value in the univariable model were included the multivariable model. The Cox proportional hazards regression model was applied to assess the relative risk of death according to radiotherapy. Hazard ratios (HRs) with 95% CIs were calculated as estimated risks of death.

PSM can minimize selection bias and mimic randomized controlled trials [15]. To balance patient characteristics between radiotherapy and no-radiotherapy groups, PSM was performed using 1: 1 nearest-neighbor matching with a caliper of 0.01. Patient cohorts were matched for marital status, age, race, median household income, grade, T stage, ER status, PR status, and chemotherapy. In addition, we use a competing risk model that treated deaths without evidence of breast cancer recurrences as the competing risk, and cumulative risk curves of different outcomes were plotted by cumulative incidence function (CIF).

All statistical tests were two-sided, statistical significance was set at the 0.05 level for all tests, and R software (version 4.0.3, https://www.r-project.org/) was used to conduct all statistical analyses. R packages, including tableone, rms, survival, survminer, ggplot2, cmprsk, and MatchIt, were used.

Results

Descriptive Statistics

In the present study, based on the eligibility criteria, 2864 patients diagnosed with T0/1-T2 breast cancer were treated with a mastectomy who had node micrometastases determined by SLNB were available for analysis from 2004 to 2015. Of these patients, the radiotherapy and no-radiotherapy cohorts included 588 (20.5%) and 2276 (79.5%) patients, respectively. There were a series of significant differences between the groups. Patients in the no-radiotherapy group vs the radiotherapy group were more likely to be older (mean age, 58 vs 53.9 years, respectively). Among the subgroups of age, there was also a significant difference in age <50 or >50 years between the groups; patients in the radiotherapy group were more likely to be <50 years (40.6% vs 31.0%, P <0.001). A higher proportion of patients with radiotherapy had more T2 tumors compared to those without radiotherapy (57.8% vs 42.4%, P<0.001), while the no-radiotherapy group had fewer grade III tumors (27.5% vs 33.5%, P<0.001). In the no-radiotherapy group, 79.9% were PR-positive, compared to 75.5% in the radiotherapy group (P=0.024). In terms of adjuvant chemotherapy, 69.2% of patients in the radiotherapy group received chemotherapy compared to only 45.1% in the no-radiotherapy group (P<0.001). After propensity score matching, a total of 1120 patients (radiotherapy 560 [50.0%] versus no-radiotherapy 560 [50.0%]) were matched. All variables were balanced between the radiotherapy group and the no-radiotherapy group. The baseline characteristics of the patients before and after propensity score matching are summarized in Table 1.

Table 1

Baseline characteristics before and after propensity score matching.

Characteristics	Before PSM			After PSM
	No radiotherapy 2276	Radiotherapy 588	P	No radiotherapy 560	Radiotherapy 560	P
Age (mean (SD))	58.00 (13.59)	53.93 (13.11)	<0.001	55.08 (12.77)	54.46 (13.01)	0.423
Age
<50	705 (31.0)	239 (40.6)	<0.001	215 (38.4)	215 (38.4)	1.000
>50	1571 (69.0)	349 (59.4)		345 (61.6)	345 (61.6)
Marital status
Married	1438 (63.2)	389 (66.2)	0.197	378 (67.5)	376 (67.1)	0.949
Unmarried	838 (36.8)	199 (33.8)		182 (32.5)	184 (32.9)
Race
Non White	424 (18.6)	105 (17.9)	0.711	95 (17.0)	96 (17.1)	1.000
White	1852 (81.4)	483 (82.1)		465 (83.0)	464 (82.9)
Median household income
Quartile 1	537 (23.6)	136 (23.1)	0.560	132 (23.6)	133 (23.8)	0.997
Quartile 2	521 (22.9)	121 (20.6)		121 (21.6)	118 (21.1)
Quartile 3	578 (25.4)	162 (27.6)		148 (26.4)	148 (26.4)
Quartile 4	640 (28.1)	169 (28.7)		159 (28.4)	161 (28.7)
Tumor size	35.85 (123.31)	34.90 (97.95)	0.862	34.79 (108.29)	35.05 (100.35)	0.967
T stage
T0/1	1311 (57.6)	248 (42.2)	<0.001	245 (43.8)	246 (43.9)	1.000
T2	965 (42.4)	340 (57.8)		315 (56.2)	314 (56.1)
Grade
I+II	1649 (72.5)	391 (66.5)	0.005	389 (69.5)	379 (67.7)	0.562
III	627 (27.5)	197 (33.5)		171 (30.5)	181 (32.3)
ER status
Negative	252 (11.1)	79 (13.4)	0.127	59 (10.5)	73 (13.0)	0.228
Positive	2024 (88.9)	509 (86.6)		501 (89.5)	487 (87.0)
PR status
Negative	458 (20.1)	144 (24.5)	0.024	116 (20.7)	132 (23.6)	0.280
Positive	1818 (79.9)	444 (75.5)		444 (79.3)	428 (76.4)
Chemotherapy
No/unknown	1250 (54.9)	181 (30.8)	<0.001	179 (32.0)	179 (32.0)	1.000
Yes	1026 (45.1)	407 (69.2)		381 (68.0)	381 (68.0)

Effects of Radiotherapy on BCSS and OS Before PSM

After a median follow-up of 53 months (ranging from 0 to 155 months, 44 and 53 months in the radiotherapy and no-radiotherapy cohorts, respectively), the BCSS and OS of patients with T0/1-T2N1mi breast cancer treated with a mastectomy who had node micrometastases determined by SLNB were assessed using Kaplan-Meier analysis.

There were 34 (5.78%) breast cancer-related death events observed in the radiotherapy group and 124 (5.44%) in the no-radiotherapy group. The 5-year cancer-specific survival was 94.4% in the radiotherapy group and 95.2% in the no-radiotherapy group (P=0.135). Kaplan-Meier analysis showed that patients who received radiotherapy had a similar BCSS compared with those who did not; the log-rank test P value was 0.13 (Figure 1A). No significant differences in 5-year OS were observed between the groups, with the 5-year OS was 91.2% vs 90.1% in the radiotherapy group compared to those with no-radiotherapy, respectively (P=0.466). Kaplan-Meier analysis showed that patients with radiotherapy had similar outcomes as patients without radiotherapy; the log-rank test P value was 0.47 (Figure 1B).

Figure 1

Breast cancer-specific survival (A) and overall survival (B) curves of breast cancer patients according to radiotherapy based on the overall population before PSM.

In consideration of competing risks (death from other causes), we constructed cumulative incidence plots in the overall population, indicating comparable 5-year cumulative incidence of breast cancer-specific death (BCSD) in the radiotherapy and no-radiotherapy groups (5.5% vs 4.7%, P=0.107). However, there was a higher 5-year cumulative incidence of other causes of death (OCD) in the no-radiotherapy cohort compared to the radiotherapy cohort (5.3% vs 3.3%, P=0.011) (Figure 2A).

Figure 2

Cumulative incidence curves for breast cancer-specific deaths and deaths from other causes before and after PSM. (A) cumulative incidence curves before PSM; (B) cumulative incidence curves after PSM.

Univariate Cox and Multivariate Cox Analysis

The results of the univariate Cox analyses revealed that radiotherapy was not associated with improved BCSS (P=0.135) (Table 2) or OS (P=0.466) (Table 3). In addition, the univariate analysis demonstrated that marital status, T stage, tumor grade, ER, and PR status were significantly associated with BCSS (Table 2) and OS (Table 3) (all P<0.05). Moreover, our univariate analysis showed that older age, median household income in Quartile 1, and did not receive chemotherapy were significantly associated with shorter OS (Table 3) (all P<0.05).

Table 2

Univariate and multivariate analyses of breast cancer-specific survival in unmatched and matched cohort.

Characteristics	Before PSM						After PSM
	5 year BCSS	Univariate		Multivariate			5 year BCSS	Univariate		Multivariate
		Chisq	P	HR	95%CI	P		Chisq	P	HR	95%CI	P
Age		2.609	0.106					0.244	0.622
<50	0.96			Reference			0.949			Reference
>50	0.946			1.17	0.81–1.7	0.397	0.961			0.67	0.39–1.18	0.165
Marital status		6.221	0.013					0.397	0.529
Married	0.957			Reference			0.961			Reference
Unmarried	0.939			1.46	1.06–2.02	0.02	0.948			1.15	0.66–2.03	0.617
Race		1.227	0.268					2.055	0.152
Non White	0.955			Reference			0.962			Reference
White	0.95			1.32	0.83–2.1	0.247	0.955			2.48	0.96–6.39	0.06
Median household income		6.27	0.099					5.659	0.129
Quartile 1	0.94			Reference			0.921			Reference
Quartile 2	0.966			1.01	0.74–1.37	0.968	0.956			0.70	0.41–1.17	0.171
Quartile 3	0.946			1.08	0.78–1.49	0.653	0.962			1.16	0.68–1.99	0.579
Quartile 4	0.953			0.63	0.45–0.89	0.008	0.98			0.64	0.36–1.13	0.121
T stage		21.9	<0.001					12.22	<0.001
T0/1	0.968			Reference			0.982			Reference
T2	0.929			1.89	1.37–2.62	<0.001	0.935			3.47	1.89–6.37	<0.001
Grade		42.439	<0.001					11.19	0.001
I+II	0.97			Reference			0.968			Reference
III	0.903			1.80	1.27–2.56	0.001	0.931			1.26	0.69–2.3	0.442
ER status		63.363	<0.001					30.08	<0.001
Negative	0.831			Reference			0.824			Reference
Positive	0.968			0.67	0.42–1.07	0.09	0.973			0.41	0.17–0.97	0.043
PR status		70.351	<0.001					22.29	<0.001
Negative	0.869			Reference			0.889			Reference
Positive	0.974			0.44	0.29–0.68	<0.001	0.975			0.45	0.21–0.97	0.042
Radiation		2.234	0.135					2.488	0.115
None/unknown	0.952			Reference			0.964			Reference
Yes	0.944			1.28	0.87–1.9	0.212	0.946			1.46	0.87–2.48	0.155
Chemotherapy		1.876	0.171					1.394	0.238
No/unknown	0.957			Reference			0.978			Reference
Yes	0.945			0.94	0.66–1.34	0.742	0.947			1.04	0.54–2	0.913

BCSS – breast cancer-specific survival; PSM – propensity score matching; HR – hazard ratio; CI – confidence interval.

Table 3

Univariate and multivariate analyses of overall survival in unmatched and matched cohort.

Characteristics	Before PSM						After PSM
	5 year OS	Univariate		Multivariate			5 year OS	Univariate		Multivariate
		Chisq	P	HR	95%CI	P		Chisq	P	HR	95%CI	P
Age		50.288	<0.001					8.045	0.005
<50	0.954			Reference			0.944			Reference
>50	0.879			2.03	1.48–2.78	<0.001	0.918			1.39	0.87–2.22	0.168
Marital status		79.005	<0.001					9.054	0.003
Married	0.938			Reference			0.942			Reference
Unmarried	0.84			2.22	1.78–2.77	<0.001	0.896			1.48	0.98–2.22	0.061
Race		0.803	0.37					1.47	0.225
Non White	0.915			Reference			0.936			Reference
White	0.901			1.09	0.8–1.48	0.603	0.925			1.43	0.75–2.71	0.277
Median household income		8.782	0.032					12.242	0.007
Quartile 1	0.861			Reference			0.862			Reference
Quartile 2	0.919			0.87	0.7–1.07	0.197	0.931			0.64	0.44–0.94	0.022
Quartile 3	0.91			1.15	0.92–1.43	0.212	0.953			1.39	0.92–2.08	0.115
Quartile 4	0.92			0.86	0.68–1.08	0.188	0.954			0.86	0.55–1.34	0.503
T stage		25.135	<0.001					15.59	<0.001
T0/1	0.925			Reference			0.964			Reference
T2	0.876			1.69	1.35–2.11	<0.001	0.897			2.78	1.79–4.31	<0.001
Grade		17.861	<0.001					7.583	0.006
I+II	0.917			Reference			0.94			Reference
III	0.867			1.44	1.12–1.84	0.005	0.899			1.37	0.87–2.16	0.169
ER status		25.656	<0.001					21.061	<0.001
Negative	0.788			Reference			0.785			Reference
Positive	0.919			0.77	0.54–1.09	0.141	0.946			0.44	0.22–0.87	0.018
PR status		45.864	<0.001					16.306	<0.001
Negative	0.814			Reference			0.858			Reference
Positive	0.928			0.60	0.44–0.8	0.001	0.947			0.66	0.37–1.17	0.154
Radiation		0.531	0.466					1.056	0.304
None/unknown	0.901			Reference			0.939			Reference
Yes	0.912			1.07	0.78–1.45	0.686	0.912			1.27	0.85–1.89	0.243
Chemotherapy		31.872	<0.001					13.02	<0.001
No/unknown	0.877			Reference			0.914			Reference
Yes	0.929			0.55	0.43–0.7	<0.001	0.933			0.41	0.26–0.62	<0.001

OS – overall survivall; PSM – propensity score matching; HR – hazard ratio; CI – confidence interval.

The multivariate analysis results were consistent with the result of the univariate analysis, except for chemotherapy. In multivariate Cox regression analysis, radiotherapy did not significantly improve the BCSS (HR=1.28, 95% CI=0.87–1.9, P=0.212) or OS (HR=1.07, 95% CI=0.78–1.45, P=0.686) for breast cancer patients, but chemotherapy improved the OS (HR=0.55, 95% CI=0.43–0.7, P<0.001). All subgroup analyses are summarized in Tables 2 and 3.

After matching, the univariate analysis of BCSS showed similar results with before matching, except for marital status. The multivariate analysis indicated that a T2 tumor, ER, and PR-negative were poor prognosticators for BCSS (Table 2). Furthermore, a T2 tumor, ER-negative, and no-chemotherapy administration remained poor prognosticators for OS (Table 3).

Effects of Radiotherapy on BCSS and OS After PSM

After 1: 1 PSM, the standardized difference (SD) of all baseline features was less than 0.1, which indicated a good agreement between the no-radiotherapy and radiotherapy groups. The 5-year BCSS was 94.6% in the radiotherapy group and 96.4% in the no-radiotherapy group (P=0.115). Kaplan-Meier analysis showed that patients who received radiotherapy had a similar BCSS compared with those who did not receive radiotherapy, and the log-rank test P value was 0.11 (Figure 3A). The 5-year OS was 91.2% in the radiotherapy group and 93.9% in the no-radiotherapy group (P=0.304); Kaplan-Meier analysis showed that patients with radiotherapy had similar outcomes as patients without radiotherapy; the log-rank test P value was 0.3 (Figure 3B).

Figure 3

Breast cancer-specific survival (A) and overall survival (B) curves of breast cancer patients according to radiotherapy based on overall population after PSM.

After using this approach, we observed that radiotherapy did not decrease BCSD or OCD in patients with T0/1-T2N1miM0 breast cancer treated with mastectomy and who had micrometastases in SLNB. The 5-year cumulative incidences of BCSD were 5.3% and 3.5% (P=0.115) in the radiotherapy and no-radiotherapy cohorts, respectively. No significant difference was observed in the 5-year cumulative incidence of OCD between the 2 cohorts (3.5% vs 2.6%, P=0.745) (Figure 2B).

We conducted an exploratory subgroup analysis in the matched cohort to identify the patient or tumor features that may benefit from radiotherapy among a specific population. There were no significant BCSS benefits to radiotherapy seen for subgroups with high risk (age <50 years, staging T2, grade III, ER-negative, PR-negative, and receiving chemotherapy) (log-rank P>0.05 for all) (Figure 4A–4F) in survival analysis. Notably, a significantly shorter BCSS was identified when radiotherapy was administrated in ER-positive patients (log-rank P=0.022) (Figure 4E) and PR-positive patients (log-rank P=0.022) (Figure 4F). The stratified survival analysis also showed a close to statistically significantly decreased BCSS for the radiotherapy group among married (log-rank P=0.051) and grade I+II patients (log-rank P=0.067).

Figure 4

Breast cancer-specific survival for breast cancer patients according to radiotherapy in subgroups stratified by age (A), grade (B), T staging (C), chemotherapy (D), ER status (E), and PR status (F) after PSM.

Discussion

Macro-metastasis in SLN is associated with worse disease-free survival [16], and current guidelines recommend further axillary surgery or radiotherapy when there are macro-metastases in the SLN [17]. With the increasing use of SLNB, identification rates of micrometastases in sentinel nodes have increased, but the prognostic significance and clinical management remain controversial [18,19].

Several studies have demonstrated that micrometastases in sentinel nodes are associated with a worse prognosis [20-28], and other studies found that the presence of sentinel node micrometastases in breast cancer patients was not significantly associated with a more inferior OS or DFS compared to node-negative disease [29-35]. For those patients undergoing breast-conserving surgery Who have SLN micrometastases, whole-breast irradiation was administered to the breast, in which the dose is delivered to the lower part of the axilla and helps control this region [36-38]. Based on the results of the ACOSOG Z0011 [17] and IBCSG 23-01 trials [12], supplemental ALND has been abandoned for breast-conserving patients only with micrometastases in the SLN in some centers, but in patients who underwent a mastectomy, it was not comprehensively analyzed. Whether radiotherapy can be safely omitted in patients with micrometastases in the SLN who undergo mastectomy without a completion dissection remains unclear. Less than 10% of patients undergoing mastectomy had positive SLNs in the randomized prospective data. In the Z0011 trial, which randomized patients with positive SLNs to ALND vs no further surgery, approximately 40% of patients had micrometastases in the SLNs, but only patients who received breast-conserving therapy were included [17]. The International Breast Cancer Study Group (IBCSG) 23-01 trial randomized 934 patients with T1–T2 tumors with isolated tumor cells and micrometastases in lymph nodes to SLNB or ALND, but only 9% underwent mastectomy [12]. The AATRM trial similarly randomized 233 patients with micrometastases to SLNB or ALND, but only 8% underwent mastectomy [13].

Wu et al performed a retrospective study based on The National Cancer Database (NCDB) identified, in which 14 019 patients diagnosed with pT1-2N1mi breast cancer between 2004 and 2014 were treated with mastectomy [39]. The study aimed to investigate the impact of post-mastectomy radiation therapy (PMRT) on OS for patients with early-stage breast cancer post-mastectomy with micrometastases in the axillary nodes, leaving 2043 patients for subset analyses in those patients who only received an SLNB without ALND, with 1490 patients in the SLNB alone arm and 553 patients in the SLNB+PMRT arm. In this subgroup, there was a trend to better OS in the SLNB+PMRT arm compared with the SLNB alone arm, with 2 (0.9%) deaths in the SLNB+PMRT arm vs 21 (2.9%) deaths in the SLNB-alone arm (log-rank P=0.053). As the authors mentioned, the work lacks LRR information, and it did not evaluate the effect of radiation on BCSS, which would be a more effective measure than OS for the cancer patient population [40].

Some previous studies have mentioned the prognostic impact of radiotherapy for breast cancer and sentinel node micrometastases on patients after mastectomy, but the number of patients is small. Of the 566 patients with breast cancer and micrometastases in the SLNs for analysis in the SENOMIC trial, only 67 (30.9%) patients who underwent mastectomy were treated with radiation to the chest wall and/or regional lymph nodes [41]. In the SERC trial, 134 PMRT was delivered in 82.7% (134/162) of patients, but only 34/55 (61.8%) for ITC or micrometastases [42].

To date, no trials with a large sample of patients have assessed the BCSS and OS benefit of radiation in patients undergoing mastectomy with pN1mi disease, which only have an SLNB. This study evaluated the effect of radiotherapy on BCSS and OS by the PSM method in patients with stage T0/1-T2 breast cancer who underwent mastectomy only with N1mi disease found on SLNB. There was a similar 5-year BCSS and OS in patients with or without radiotherapy in the overall population, and radiotherapy showed no specific survival benefits among the before or after PSM population.

The subgroups with high-risk patients were further examined in the PSM population; no significant benefit to radiotherapy was seen on survival analysis. When subdivided by T stage, there was no difference for BCSS in patients with T0/1 tumor (log-rank P=0.63) or T2 tumors (log-rank P=0.1); and patients with radiotherapy did not have significantly different OS compared with those without radiotherapy in patients with T0/1 tumor (log-rank P=0.23), but OS was significantly worse in those with T2 tumors (log-rank P=0.041). We further performed cumulative incidence plots in the T2 tumors population by cumulative incidence function, showing a comparable 5-year cumulative incidence of BCSD (8.6% vs 4.6%, P=0.119) and OCD (5.0% vs 3.1%, P=0.248) in the radiotherapy and no-radiotherapy groups.

Interestingly, in the ER-positive and PR-positive subgroups, patients receiving radiotherapy were associated with a reduced BCSS. In addition, there was a trend toward improved BCSS in married patients and in the grade I+II subgroup who did not receive radiotherapy. It is difficult to explain the adverse effects of radiotherapy on BCSS of low-risk features such as hormone receptor-positive status and low-to-intermediate grade tumors. We speculate that the presence of other negative prognostic primary tumor characteristics (such as larger tumor size, positive margins, and histologic type) in the ER-positive and PR-positive populations who received radiotherapy significantly influenced the breast cancer outcome. Furthermore, the SEER database does not contain information regarding the HER2 status, as it was only available after 2010. Other tumor characteristics, including Ki-67, lymphovascular invasion, and information on adjuvant treatment, including anti-HER2 targeted therapy and hormonal therapy, are lacking. In addition, the SEER database does not provide a detailed protocol for the use of chemotherapy or radiotherapy, which may have weakened the reliability of the results.

Our study has certain limitations. It was a retrospective study, with an inevitable selection bias and a relatively short follow-up period. The study lacks data on detailed pathologic examination of sentinel nodes, including the number of sentinel nodes removed and the number of sentinel nodes with metastases, and there was no detailed information on radiotherapy and chemotherapy in the SEER database. Nevertheless, PSM was performed to balance the patient characteristics, and this study provides insight into the under-sampled population in clinical trials of mastectomy with SLN micro-metastatic disease. To the best of our knowledge, this is the first study to assess the effect of radiotherapy on BCSS and OS in a large-cohort population in patients who underwent a mastectomy with micro-metastatic disease confirmed by SLN.

Conclusions

In summary, the current findings demonstrated that in patients with T0/1-T2 breast cancer undergoing mastectomy with N1mi disease found on SLNB, no differences in BCSS or OS were found between patients with and without radiotherapy. This study supports the recommendation that radiotherapy for this specific group of the population should be individualized. Furthermore, according to the risk stratification, individualized radiotherapy strategies should be further investigated in the subgroup with lower-risk populations.