Effectiveness of the AJCC 8th edition staging system for selecting patients with T1–2N1 breast cancer for post-mastectomy radiotherapy: a joint analysis of 1986 patients from two institutions

Patients with pathologically confirmed T1–2N1 breast cancer who underwent mastectomy and axillary dissection at two institutions in China between January 2000 and December 2014 were recruited for the study. Those who met the following criteria were included: no neoadjuvant systemic therapy, sufficient information on histological grade, estrogen receptor- (ER), progesterone receptor- (PR), or human epidermal growth factor receptor-2 (HER2) status, sufficient information on whether PMRT was provided, receipt of adjuvant chemotherapy, receipt of targeted therapy if HER2-positive and receipt of hormone therapy if ER or PR positive. A total of 1986 patients were included in the analysis (Fig. 1).

All patients were restaged according to the newly published AJCC 8th ed. staging system for breast cancer, which incorporated biomarkers into the classical TNM system to improve discrimination. The biomarkers used included histologic grade, ER, PR and HER2-status.

LRR was defined as any tumor recurrence in the ipsilateral chest wall or axillary, supraclavicular, or internal mammary nodes during follow-up regardless of their relation in time to DM. Recurrence at any other site was considered DM. Disease-free survival (DFS) was calculated from the date of mastectomy until the date of LRR, DM, death or last follow-up, whichever came first. Overall survival (OS) was calculated from the date of mastectomy to the date of death or last follow-up [11].

The association between PMRT and patient characteristics was assessed using Pearson’s χ² test. LRR and DM were calculated with a competing risk model and the differences were compared using Gray’s test. DFS and OS were calculated using the Kaplan-Meier method and the differences were compared using the log-rank test. Significant variables (p <  0.05) from the univariate analysis were included in multivariate analysis, which was performed using the Cox proportional hazards model. Prognostic factors for LRR and DM were identified in patients without PMRT. Points for the recurrence scores were assigned according to the Hazards ratios of the prognostic factors, which ranged from 0 points (age > 40 years, other quadrant tumor location, 1 positive node, absence of lympho-vascular invasion [LVI], stage IA), 1 point (age ≤ 40 years, inner quadrant tumor location, 2–3 positive nodes, presence of LVI, stage IB-IIA), to 2 points (stage IIB-IIIA). The points were added to determine the recurrence score, which ranged from 0 to 6 points. Patients were divided into low-risk (recurrence score = 0–1 points), intermediate-risk (recurrence score = 2 points) and high-risk (recurrence score ≥ 3 points) groups. The role of PMRT in the LRR, DM, DFS and OS of the different risk groups was evaluated. Statistical analyses were performed using cmprsk (https://​cran.​r-project.​org/​web/​packages/​cmprsk/​) and SPSS version 24.0 (IBM Corp., Armonk, NY, USA). All p values were two-tailed, with a value < 0.05 considered to be significant.

Table 1 shows the demographic, tumor and treatment characteristics of the entire cohort, and of the subgroups with and without PMRT. The median age was 49 years (range, 23–82), the median number of axillary nodes dissected was 18 (range, 2–59) and the median number of positive nodes was 1 (range, 1–3). Invasive ductal carcinoma was diagnosed in 1953/1986 patients (98.3%). All patients received adjuvant chemotherapy, with a median of 6 cycles (range, 1–12); of them, 1377 patients (69.3%) received anthracycline and taxane-based regimens, 332 (16.7%) received a anthracycline-based regimen, 125 (6.3) received a taxane-based regimen, 55 (2.8%) received other regimens and the regimens of 49 (2.5%) were unknown. A total of 182 out of 1983 patients (9.2%) had HER2-positive disease and all of them received anti-HER2 targeted therapy with trastuzumab. All 1575/1986 patients (79.3%) who had ER and/or PR positive disease received hormone therapy, with a median duration of 48 months (range, 1–132). And 465/1986 patients (23.4%) received PMRT. Among 406 patients who had detailed RT information available, the chest wall was irradiated in 406 (100%) of them, the supra-infraclavicular nodal region was irradiated in 404 (99.5%) patients, the axilla was irradiated in 9 (2.2%) patients and the internal mammary chain was irradiated in 4 patients (1.0%). The median total dose was 50 Gy (range, 46–56) for conventional fractionation in 383 (94.3%) patients, and 40 Gy (range, 40–43.5) in 15 fractions in 23 (5.7%) patients. Compared with the no-PMRT group, the PMRT group had significantly more patients with risk factors, such as ≤40 years, AJCC 8th ed. stage IIIA, 2–3 positive nodes, < 10 nodes dissected, presence of LVI, T2, ER negative, PR negative and HER2- positive disease.

After a median follow-up of 68.5 months (range, 1–182 months), 147 patients died; of them, 126 (84.6%) died from breast cancer, 1 (0.7%) from treatment complications, 16 (10.7%) from other causes and 6 (4.0%) from unknown causes; 142 had LRR and 257 had DM. Compared with the no-PMRT group, the PMRT group had a lower 5-year LRR (3.6% versus 6.6%, p = 0.005), but a similar DM (13.3% versus 9.3%, p = 0.350), DFS (85.2% versus 87.3%, p = 0.948) and OS (94.8% versus 94.9%, p = 0.394) (Fig. 2). The 10-year LRR rates of the PMRT group and no-PMRT group were 5.0 and 11.2%, respectively. After adjusting for age, tumor location, T stage, number of positive nodes, number of nodes dissected, histological grade, LVI and ER-, PR- and HER2-status, the multivariate analysis showed PMRT significantly reduced LRR (hazard ratio [HR] = 0.30, 95% CI: 0.17–0.53, p <  0.001), and increased DFS (HR = 0.71, 95% CI: 0.53–0.96, p = 0.028), but had no significant influence on DM (HR = 0.83, 95% CI: 0.59–1.15, p = 0.258), or OS (HR = 0.64, 95% CI: 0.40–1.03, p = 0.066), compared with no PMRT.

The median follow-up duration for the 1521 patients who did not receive PMRT was 71 months (range, 1–175 months). Tables 2 and 3 summarize the results of the univariate and multivariate analyses of the prognostic factors for LRR and DM, respectively. Because T stage, histological grade, ER-, PR- and HER2-status were used to define AJCC 8th ed. stages, they were not included in the multivariate analysis. As a result, age ≤ 40 years, inner quadrant tumor location, 2–3 positive nodes and higher AJCC stages were independent prognostic factors for LRR. Age ≤ 40 years, inner quadrant tumor location and higher AJCC stages were independent risk factors for DM. The presence of LVI approached significance in predicting DM (p = 0.074).

Recurrence scores were assigned by age, tumor location, AJCC 8th ed. stage, the number of positive nodes and LVI. There were 653 patients in the low-risk group, 504 patients in the intermediate-risk group and 351 patients in the high-risk group. A significant difference in LRR, DM, DFS and OS between the three groups was found (Fig. 3). The 5-year LRR and DM rates were 2.5, 5.4 and 16.2% (p <  0.001), and 4.9, 8.4 and 18.6% (p <  0.001), respectively. The 5-year DFS and OS rates were 93.9, 87.7 and 77.1% (p <  0.001), and 97.4, 95.5 and 90.3% (p <  0.001), respectively.

We stratified all the patients into three risk groups based on their risk scores and compared OS, LRR and DM between the patients who did or did not receive PMRT. Among the 742 patients in the low-risk group, 89 (12.0%) received PMRT and of the 669 patients in the intermediate-risk group, 165 (24.7%) received PMRT. Among the 1411 patients in the low- and intermediate-risk groups, PMRT had no impact on LRR (p = 0.268), DM (p = 0.252), DFS (p = 0.608) or OS (p = 0.986) (Fig. 4).

Among the 551 patients in the high-risk group, 200 (36.3%) received PMRT, which significantly reduced LRR (p <  0.001) and improved DFS (p = 0.006) and OS (p = 0.037), but had no impact on DM (p = 0.079) (Fig. 5). In multivariate analysis, after adjusting for age, tumor location, number of positive nodes, LVI and AJCC 8th ed. stage, PMRT significantly reduced LRR (HR = 0.23, 95% CI: 0.11–0.49, p <  0.001), and DM (HR = 0.63, 95% CI: 0.40–0.99, p = 0.044), and it improved DFS (HR = 0.55, 95% CI: 0.36–0.83, p = 0.004), and OS (HR = 0.48, 95% CI: 0.25–0.93, p = 0.029 in the high-risk group.