Final overall survival analysis of the phase 3 HERITAGE study demonstrates equivalence of trastuzumab-dkst to trastuzumab in HER2-positive metastatic breast cancer

Full inclusion and exclusion criteria have been previously reported [16]. Eligible patients were adults with histologically confirmed HER2-positive breast cancer having ≥ 1 measurable metastatic target lesion. Key eligibility criteria included an Eastern Cooperative Oncology Group performance status of 0 to 2 and left ventricular ejection fraction (LVEF) within normal range [16]. Patients must not have received chemotherapy or HER2-targeted therapy within 1 year of diagnosis of metastatic disease.

Details of the study design and dosing schedules have been previously reported (Online Resource 1) [16]. Briefly, patients were randomly assigned 1:1 to receive taxane of institutional choice (docetaxel or paclitaxel) plus trastuzumab-dkst or trastuzumab for 8 cycles (24 weeks). Patients with at least stable disease as defined by Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST v1.1) at week 24 could continue with monotherapy mAb treatment according to their original randomization until progression, discontinuation due to unacceptable toxicity, or death. At the end of treatment, patients were followed every 3 months for 36 months from the date of randomization or death to assess survival. The secondary objective of this phase 3 study was to assess OS at 36 months or after 240 deaths, whichever occurred first, as observed from the time of randomization of the last patient.

Overall survival was defined as the time from randomization to date of death due to any cause and was cumulative through 36 months of follow-up. The endpoints for the primary and secondary study objectives (ie, ORR, PFS) were analyzed at week 24 for the combination therapy phase and at week 48 for the monotherapy phase, and have been previously reported [16, 19]. In this analysis, efficacy endpoints were reported in all patients who enrolled after the second protocol amendment (intention-to-treat population [ITT]), which excluded patients who had already received first-line therapy.

The safety population included all patients who received ≥ 1 dose of trastuzumab-dkst or trastuzumab. Assessment of treatment-emergent adverse events (TEAEs) included type, incidence, severity (graded by the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03), timing, seriousness, and relatedness. Safety analyses also included assessment of LVEF values. Safety during combination and monotherapy has been previously reported [16, 19]. This report includes assessment of any updates observed through the final analysis, including accumulated data on AEs of special interest.

Details on sample size have been previously reported [16]. Clinical activity was evaluated by assessing progression of disease, defined according to RECIST v1.1. Descriptive statistics were used to summarize patient disposition, baseline characteristics, and treatment administration, and SAS® software version 9.2 or later (SAS, Cary, NC) was used for analysis. For OS and PFS, Kaplan–Meier plots by treatment group were presented, and the log-rank test of the 2 groups unadjusted for covariates was performed.

Of the 500 randomized patients, 249 were randomized to receive trastuzumab-dkst and 251 were randomized to receive trastuzumab, each in combination with taxane. The ITT population, used to evaluate efficacy, was composed of 458 female patients (230 randomized to trastuzumab-dkst and 228 randomized to trastuzumab) who had not previously received first-line therapy (Fig. 1). The mean (SD) age of patients was slightly lower in the trastuzumab group (52.9 ± 11.2) than in the trastuzumab-dkst group (54.3 ± 11.0). Demographics and baseline characteristics for both treatment groups were similar with respect to age, race, height, weight, body surface area, exposure to prior therapies, and time since diagnosis and were unchanged compared with the week 48 analyses (Table 1) [19]. The safety population was composed of 493 patients, defined as those who received ≥ 1 dose of study drug, regardless of whether they had received prior first-line therapy.

After combination therapy, 343 patients entered the monotherapy phase for safety analysis (trastuzumab-dkst, N = 179; trastuzumab, N = 164). The demographic profile for patients with safety monotherapy data is consistent with that for the ITT efficacy population. At the time of the final OS analysis, 169 patients in the safety population had received further lines of therapy (Online Resource 2), with similar distribution of HER2-targeted treatments (15.6% [n = 28] vs 18.9% [n = 31]), endocrine therapies (9.5% [n = 17] vs 14.6% [n = 24]), and chemotherapies (30.2% [n = 54] vs 25.0% [n = 41]) in the trastuzumab-dkst and trastuzumab groups, respectively.

At the time of final analysis, 242 patients had died during the study: 116 patients treated with trastuzumab-dkst, 124 patients treated with trastuzumab, and 2 untreated patients (1 patient from each group). At final assessment, 121 (52.6%) patients were alive in the trastuzumab-dkst group, and 114 (50.0%) patients were alive in the trastuzumab group (including 79 [trastuzumab-dkst] and 75 [trastuzumab] patients who received monotherapy during the study). The survival curves did not significantly differ between treatment groups (P = 0.427; Fig. 2). The median OS by Kaplan–Meier analysis was 35.0 months in the trastuzumab-dkst group and 30.2 months in the trastuzumab group. Overall median follow-up time was 32.6 months (35.7 months in the trastuzumab-dkst group and 31.1 months in the trastuzumab group). The 95% CI of the hazard ratio (HR) for OS included “1” for both trastuzumab-dkst and trastuzumab at the time of the final analysis, indicating no relevant differences in survival benefit between treatment groups.

At final assessment, 82 (35.7%) patients in the trastuzumab-dkst group were free of progression compared with 86 (37.7%) patients in the trastuzumab group. The time-to-event curves for both treatment groups were not statistically significantly different (P = 0.864; Fig. 3). As with OS, the 95% CI of the OS ratio (trastuzumab-dkst to trastuzumab) included “1” for all subgroups at the time of the final analysis, and hence, no relevant differences between subgroups were observed. However, tumor endocrine status (negative vs positive; HR 1.40; P = 0.017), race (Black vs White and Asian vs White; HR 3.71 and 1.38, respectively; P = 0.011), previous adjuvant/neoadjuvant chemotherapy/HER2-targeted therapy (yes vs no; HR 1.39; P = 0.016), and number of metastatic sites (2 vs 1, 3 vs 1, and ≥ 4 vs 1; HR 1.44, 1.51, and 2.46, respectively; P < 0.001) had an effect on OS. The treatment HR (95% CI) adjusted for these factors was 0.85 (0.653–1.116; P = 0.248). Median PFS was 11.1 months in both treatment groups. Duration of response was also not statistically different between treatment groups (P = 0.771), with 123 (64.1%) patients in the trastuzumab-dkst group having tumor progression or dying before the final analysis, compared with 119 (64.7%) patients in the trastuzumab group. Median DR was 9.9 months in the trastuzumab-dkst group and 9.8 months in the trastuzumab group. Sensitivity analyses of all enrolled patients showed that OS, PFS, and DR results were similar to those observed in the ITT population that included patients who had received prior first-line therapy.

Safety and tolerability analyses for the combination therapy and monotherapy phases of the trial were previously reported [16, 19]. From the initiation of the monotherapy phase through final analysis, the number of patients with ≥ 1 TEAE was similar between the trastuzumab-dkst (69.3% [n = 124]) and trastuzumab groups (72.6% [n = 119]). Most TEAEs occurred in similar numbers in the trastuzumab-dkst and trastuzumab groups, except for patients experiencing anemia (3.4% [n = 6] and 10.4% [n = 17], respectively), and were generally grade 1 or 2 in severity (Table 2). Overall incidence of serious AEs (SAEs) reported from the start of the monotherapy phase through the end of the trial was 5.8% (n = 20). Incidence of SAEs was similar between the trastuzumab-dkst and trastuzumab groups (5.6% [n = 10] and 6.1% [n = 10], respectively). Rates of treatment discontinuation due to TEAEs were similar between the trastuzumab-dkst and trastuzumab groups (3.9% [n = 7] and 7.3% [n = 12], respectively). Overall, 8 (2.3%) patients withdrew from the study because of 1 or more TEAEs, 4 (2.2%) in the trastuzumab-dkst group and 4 (2.4%) in the trastuzumab group. Throughout the entire study, 57 patients (23.1%) in the trastuzumab-dkst group and 58 patients (23.6%) in the trastuzumab group reported TEAEs of special interest, including infusion reactions (6.9% [n = 17], 4.9% [n = 12]) and hypersensitivity events (2.4% [n = 6], 3.7% [n = 9]). Few patients in each treatment group had observed LVEF values of < 50% at least once post-baseline (trastuzumab-dkst, 4.5% [n = 11]; trastuzumab, 3.3% [n = 8]).

As previously reported, only 2 fatal TEAEs were reported in the monotherapy phase, both unrelated to the study drug [19]. There were no additional fatal TEAEs after week 48 and no new safety concerns reported through the end of the survival follow-up.