Trastuzumab Provides a Comparable Prognosis in Patients With HER2-Positive Breast Cancer to Those With HER2-Negative Breast Cancer: Post Hoc Analyses of a Randomized Controlled Trial of Post-Mastectomy Hypofractionated Radiotherapy

Background and Purpose We investigated the locoregional effect of trastuzumab, and determined whether patients with human epidermal growth factor receptor (HER)2-positive breast cancer (BC) treated with trastuzumab could achieve comparable efficacy to that of patients with HER2-negative BC. Materials and Methods This was post hoc analyses of data of 793 BC patients from a randomized controlled trial comparing post-mastectomy hypofractionated radiotherapy with conventional fractionated radiotherapy. Survival rates were analyzed by the Kaplan–Meier method and compared by the log-rank test. Results Patients were classified into three groups: HER2-negative (HER2−; n = 547), HER2-positve with trastuzumab (HER2+ + T; n = 136), and HER2-positive without trastuzumab (HER2+ − T; n = 110). The HER2+ + T group had significantly lower locoregional recurrence (LRR, 6.0% vs. 13.9%), distant metastasis (DM, 17.4% vs. 33.8%) and higher disease-free survival (DFS, 81.2% vs. 61.9%) at 5 years than that of the HER2+ − T group (P <.05). The HER2− group had significantly lower LRR (6.8% vs. 13.9%), DM (22.4% vs. 33.8%) and higher DFS (76.1% vs. 61.9%) at 5 years than that of the HER2+ − T group (P <.05). The difference in LRR, DM and DFS at 5 years was not significant between the HER2+ + T group and HER2− group (P >.05). Different annual LRR patterns was found among groups according to HR status. Conclusion Trastuzumab reduces LRR in patients with locally advanced HER2-positive BC who have received post-mastectomy radiotherapy. It provides comparable DFS to that with patients with HER2-negative BC.


INTRODUCTION
Increased expression of human epidermal growth factor receptor (HER)2, one of the most important molecular markers for breast cancer (BC), occurs in 15-25% of women with BC, and is associated with a poor prognosis (1,2). Several randomized trials have demonstrated adjuvant use of trastuzumab (Herceptin ® ; Roche, Basel, Switzerland) to have a major and persistent benefit in reducing the risk of distant recurrence and death in patients with HER2-positive BC (3)(4)(5). However, the locoregional benefit of trastuzumab has not been elucidated clearly, especially in patients who have received radiotherapy (6)(7)(8). Little is known about the locoregional effect of trastuzumab in patients with locally advanced BC treated with adjuvant radiotherapy. In addition, there are limited data on outcome comparisons between patients with HER2-positive disease in the setting of trastuzumab and patients with HER2negative disease (9,10).
We aimed to determine whether: (i) adjuvant trastuzumab reduced locoregional recurrence (LRR) in patients with stage-II and III BC (according to AJCC 7 th edition) treated with postmastectomy radiotherapy; (ii) patients with HER2-positive BC treated with trastuzumab had comparable outcomes to those of patients with HER2-negative BC.

Patients
This study was a post hoc analysis of data from a randomized clinical trial (NCT00793962) conducted between 2008 and 2016. The original trial recruited 820 women aged 18-75 years who had undergone mastectomy for high-risk BC with negative margins (i.e., ≥4 positive axillary lymph nodes or a primary tumor of stage T3-4) and compared post-mastectomy hypofractionated radiotherapy (HFRT, 43.5 Gy in 15 fractions over 3 weeks) with conventional fractionated radiotherapy (CFRT, 50 Gy in 25 fractions over 5 weeks). The cohort has been described in detail elsewhere 11 . Patients were designated as "HER2-positive" if they demonstrated a membranous staining score of 3+ on immunohistochemical (IHC) analyses, or gene amplification on fluorescence in situ hybridization (FISH). HER2 copy number that was not amplified by FISH or read as 0-1+ by IHC was designated as "HER2-negative". Patients who were IHC 2+ without a confirmatory FISH result were excluded from analyses (n = 17). Other clinical and histologic information of these patients was collected from case records. The final followup was November 2019. A total of 793 patients were included for evaluation ( Figure 1).

Statistical Analyses
The relationship between HER2 status and outcomes was analyzed. "LRR" was defined as the first recurrence after surgery in the ipsilateral chest wall or axillary, supra-/ infraclavicular, or internal mammary nodal regions regardless of the status of systemic disease. "Distant metastasis" (DM) was defined as the first recurrence after surgery in any distant site. Overall survival (OS) was calculated from the date of the surgical procedure to the date of death due to any cause or final followup. Disease-free survival (DFS) was calculated from the date of surgery to date of recurrence or death, or final follow-up.
Statistical analyses were undertaken using SPSS v23.0 (IBM, Armonk, NY, USA) and R v3.5.3 (R Project for Statistical Computing, Austria, Vienna; www.r-project.org/). Categorical variables were summarized as frequencies and percentages, and were compared using the Fisher exact test or chi-square test.
The relapse rate and survival rate were calculated using the Kaplan-Meier method and compared using the log-rank test (including pairwise comparisons). The hazard rate of LRR was computed using the "muhaz" package in R. The association of potential prognostic factors with survival outcomes was tested by univariate Cox regression analysis. Factors associated with outcome in the univariate analysis (at P <.05) were entered into multivariable proportional hazards regression analysis to identify independent predictors. Statistical significance was set at P ≤.05. Bonferroni's correction for significance of association was applied to compensate for multiple comparisons. Table 1 shows the demographic, tumor, and treatment characteristics of the 793 patients for whom HER2 status was identified. The median age of the study cohort was 49 (range, 24-74) years. All patients underwent mastectomy and radiotherapy. All patients received a median of six (range, 4-11) cycles of chemotherapy (neoadjuvant and/or adjuvant).
The proportion of HER2-positive patients treated with trastuzumab increased from 31.6% in 2008 to 77.8% in 2016 (Supplementary Figure 1). As shown in Table 1, the HER2 + − T group had more patients treated between 2008 and 2013 compared with that in the HER2 − group and HER2 + + T group (Bonferronicorrected P = .033 and Bonferroni-corrected P = .006, respectively). The proportion of patients with hormonal receptor (HR)-positive disease was significantly higher in the HER2 − group than that in the HER2 + + T group and HER2 + − T group (Bonferroni-corrected P <.001 and Bonferroni-corrected P = .003, respectively). As a result, significantly more patients in the HER2 − group received hormonal therapy than patients in the HER2 + + T group or HER2 + − T group (Bonferroni-corrected P <.001 and Bonferroni-corrected P = .018, respectively). All other characteristics were well balanced between the three groups.
Univariate and multivariate analyses of prognostic factors for LRR, DM, DFS and OS are shown in Tables 2 and 3. In multivariate analyses, the combination of HER2 status and trastuzumab treatment was an independent prognostic factor for LRR (P = .011), DM (P = .002) and DFS (P <.001). Other independent prognostic factors were the treatment period, age, stage, neoadjuvant chemotherapy, and hormonal therapy.
When HR was negative, there were no significant differences in LRR, DM, DFS or OS among HER − , HER2 + + T, or HER2 + − T groups (P = .273, P = .174, P = .219, P = .508, respectively; Figures 3E-H).  Annual Hazard Rate of LRR Figure 4 shows the annual LRR patterns of the three groups in the entire cohort, HR + cohort, and HRcohort. Visual inspection of the LRR hazard curves showed a difference in LRR patterns between the HER2group and HER2 + + T group and HER2 + -T group. In the entire cohort, the annual hazard rate of LRR in the HER2 + -T group was consistently higher than that in the HER2group or HER2 + + T group. The annual LRR curve of the HER2 + -T group displayed an obvious double-peaked pattern, with an early peak at~2.5 years and a late peak at~9 years. The annual LRR curve of the HER2 + + T group displayed a single early peak at~2.5 years. The annual LRR curve of the HER2 − group displayed a continuously low risk without an obvious peak ( Figure 4A). The observed difference in LRR among these three groups occurred mainly during the first 5 years rather than beyond 5 years. The annual LRR rate of HER2 -, HER2 + + T, and HER2 + -T groups was 1.4%, 1.3% and 3.0% during the first 5 years (P = .049), and 0.5%, 0.4% and 1.9% after 5 years (P = .070), respectively. The annual LRR pattern of the three groups in the HR + cohort was similar to that of the entire cohort ( Figure 4B). In the HR + cohort, the annual LRR rate of HER2 − , HER2 + + T, and HER2 + − T groups was 1.1%, 1.2% and 2.8% during the first 5 years (P = .042), and 0.6%, 0.6% and 2.4% after 5 years (P = .082), respectively. In the HR − cohort, all three groups showed a single early LRR peak, at~1 year for the HER2 − group, and at~2 years for HER2 + + T and HER2 + − T groups ( Figure 4C). The annual LRR rate of HER2 − , HER2 + + T, and HER2 + − T groups was 3.9%, 1.4% and 3.7% during the first 5 years (P = .240), and 0.0%, 0.0% and 0.0% after 5 years, respectively.

DISCUSSION
Previously, we demonstrated that post-mastectomy HFRT was non-inferior to CFRT in high-risk BC patients (11). The present post hoc analyses showed that: (i) trastuzumab decreased the risk of LRR and improved DFS significantly in patients with HER2positive disease who received post-mastectomy radiotherapy (PMRT); (ii) patients with HER2-positive disease treated with trastuzumab had a comparable prognosis to that of patients with HER2-negative disease.
Randomized trials have revealed survival and local-control benefits from PMRT in patients with high-risk BC (12)(13)(14). One meta-analysis concluded that PMRT reduced LRR and BC mortality, especially for patients with lymph-node involvement (15).  Therefore, patients with stage-III BC who have undergone modified mastectomy have received PMRT routinely. However, those studies were conducted in the non-trastuzumab era, and whether there is a synergistic effect between trastuzumab and PMRT is not known. As trastuzumab treatment was not covered by the national health insurance plan during the period of this study, only half of the patients with HER2-positive breast cancer received anti-HER2 therapy. It gave us an opportunity to discuss these issues. In the present study, 93.8% of patients had stage-III BC and all received PMRT. We showed that the absolute reduction in LRR from trastuzumab treatment in HER2-positive patients was 8%. The question regarding whether trastuzumab reduces LRR risk in patients with HER2-positive disease is controversial.   (16). However, a study by Keiss and colleagues also involving patients with T1-2N0 BC showed that locoregional recurrence-free survival after breast-conserving treatment was higher among the trastuzumab-treated group (99% vs. 90%, P = .01) (7). Furthermore, other studies have shown that trastuzumab reduced LRR risk among HER2-positive patients with HR-positive stage-I-III BC, whereas patients with HR-negative BC had persistently increased LRR risk despite trastuzumab use (6,17,18). The present study showed similar findings whereby trastuzumab significantly reduced LRR risk in a HR-positive cohort rather than a HR-negative cohort, however, the sample size in the HR-negative cohort was small. The small sample size might also explain the findings that the relapse and survival outcomes were similar between triple-negative and HR -HER2 + + T groups. In HER2-positive patients with stage-I-III BC who received mastectomy, Lanning et al. found that trastuzumab did not decrease LRR risk in patients who did not receive PMRT, but it decreased LRR risk among those who received PMRT, suggesting that the greatest benefit is seen in a higher-risk subset of patients (8). That finding was confirmed by Wang et al., who found that trastuzumab reduced the 5-year LRR rate (4.9% vs. 17.9%, P <.05) in HER2-positive patients treated with mastectomy and PMRT (19). Taken together, data suggest that trastuzumab may play an important part in reducing LRR risk among patients with more advanced disease and those with HR-positive disease.
Analyses of hazard function defines, in much greater detail, changes in LRR risk over time and highlights when recurrence occurs rather than simply calculating the overall recurrence risk. We observed a double-peaked pattern of LRR in HER2-positive patients who did not receive trastuzumab in the HR-positive cohort, and the second peak disappeared upon trastuzumab treatment. The treatment effect of trastuzumab on LRR was not proportional over time, and occurred mainly during the first 5 years rather than beyond 5 years. Previously, it has been reported that the hazard rate for tumor relapse of estrogen receptor (ER)-positive and ER-negative tumors displays a bimodal curve with a first, dominant early peak and a second, lower later surge followed by a long-lasting "tail" (20). A doublepeaked pattern of relapse or death has been observed in BC patients regardless of ER status or HER2 status in the absence of trastuzumab (21). Ribelles and colleagues also observed this double-peaked pattern of relapse or death in a Luminal B and HER2-positive group (88% treated without trastuzumab), whereas a single early peak was noted in a triple-negative group (22). However, reports on LRR patterns are limited. The present study is the first to show different annual LRR patterns among groups stratified by trastuzumab treatment and according to HR status. In contrast to our findings that trastuzumab did not significantly reduce the risk of LRR beyond 5 years, a combined analysis for relapse or death by Chumsri and coworkers showed that the benefit of adjuvant trastuzumab therapy was seen in years 0 to 5 (hazard ratio, 0.42; P = .001) and years 5 to 10 (hazard ratio, 0.69; P = .03) (23). This difference might be explained by the different endpoints used in our study, or  Randomized studies have established that trastuzumab improves the survival rates of patients with HER2-positive tumors (3)(4)(5)23), whereas studies are limited for comparing the outcome of HER2positive patients taking trastuzumab with that of HER2-negative patients. We found that HER2-positive patients taking trastuzumab had comparable DFS with that of HER2-negative patients. This finding has been corroborated by other studies. Fokter et al. reported no significant difference in DFS between trastuzumabtreated HER2-positive patients and HER2-negative patients (9). Qin et al found that the DFS difference at 5 years between HER2negative patients and HER2-positive patients receiving trastuzumab was not significant (71.7% vs. 77.8%, P = .503) (10). Some studies on locally advanced BC or metastatic BC have shown that trastuzumab-treated HER2-positive patients have better event-free survival or OS than that in patients with HER2-negative BC (24,25). The negative prognostic effect of HER2-positivity observed before trastuzumab treatment became available might disappear completely in the era of routine administration of trastuzumab. Therefore, the prognostic importance of HER2 status must be considered in the context of anti-HER2 treatment. In the recent eighth edition of the American Joint Committee on Cancer (AJCC) staging system for BC, the prognostic-stage groups showed that patients with HER2-positive tumors tended to be grouped into a lower stage than patients with HER2-negative tumors provided that other staging factors were identical because most of the patients used to generate the eighth edition of the AJCC staging system in the initial models received trastuzumab (26). For example, for a patient with T3N2M0, grade 3, ER-positive, and PR-positive BC, she was grouped into IIA if HER2 was positive, and IIB if HER2 was negative.
Our study had three main limitations. First, thanks to developments in diagnostic and therapeutic strategies, treatment guidelines for patients with BC have changed, and our study population may not reflect the outcomes for patients being treated currently. This study initiated originally from 2008, so we found the treatment period to be an independent predictor for DM and DFS. Nevertheless, the combination of HER2 status and trastuzumab treatment was demonstrated to be an independent predictor for LRR, DM, and DFS after adjustment for the treatment period. Second, because of the limited followup time, the findings on hazard rates over time warrant further validation. Third, this study was an unplanned post hoc analysis, patients taking and not taking trastuzumab were not assigned randomly, so a selection bias was unavoidable. We attempted to eliminate the influence of confounding factors in the multivariate analysis by including all other known prognostic factors. Nevertheless, this study is the first to assess the locoregional effect of trastuzumab in patients with locally advanced BC treated with adjuvant radiotherapy, and the findings on the LRR patterns over time merit further investigation.

CONCLUSIONS
Trastuzumab reduces the risk of LRR in patients with locally advanced HER2-positive BC who have received PMRT. It provides comparable DFS in patients with HER2-positive disease to those with HER2-negative disease.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/ Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Institutional Review Board of Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. The patients/participants provided their written informed consent to participate in this study.