Exploratory subgroup analysis of patients with prior trastuzumab use in the ATTRACTION-2 trial: a randomized phase III clinical trial investigating the efficacy and safety of nivolumab in patients with advanced gastric/gastroesophageal junction cancer

Background Data on immune checkpoint inhibitor efficacy in patients with human epidermal growth factor receptor 2-positive (HER2+) advanced gastric/gastroesophageal junction (G/GEJ) cancer are lacking. Because HER2 status was not captured in the ATTRACTION-2 trial, we used patients with prior trastuzumab use (Tmab+) as surrogate for HER2 expression status to evaluate the efficacy and safety of nivolumab as third- or later-line therapy in these patients. Methods In ATTRACTION-2, a randomized, double-blind, placebo-controlled, phase 3 multicenter trial, patients were randomized (2:1) to receive nivolumab (3 mg/kg) or placebo every 2 weeks until disease progression or toxicity requiring study discontinuation. Overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and safety were assessed. Results Of 493 enrolled patients, 81 (nivolumab, n = 59; placebo, n = 22) were Tmab+ and 412 (nivolumab, n = 271; placebo, n = 141) were Tmab−. In both groups, patients receiving nivolumab showed a longer median OS vs placebo (Tmab+, 8.3 [95% confidence interval, 5.3–12.9] vs 3.1 [1.9–5.3] months, hazard ratio, 0.38 [0.22–0.66]; P = 0.0006; Tmab−, 4.8 [4.1–6.0] vs 4.2 [3.6–4.9] months, 0.71 [0.57–0.88]; P = 0.0022). PFS was longer in both groups receiving nivolumab vs placebo (Tmab+, 1.6 [1.5–4.0] vs 1.5 [1.3–2.9] months, 0.49 [0.29–0.85]; P = 0.0111; Tmab−, 1.6 [1.5–2.4] vs 1.5 [1.5–1.5] months, 0.64 [0.51–0.80]; P = 0.0001). Conclusions Nivolumab was efficacious and safe as third- or later-line therapy regardless of prior trastuzumab use in patients with advanced G/GEJ cancer.


Introduction
Gastric/gastroesophageal junction (G/GEJ) cancer is the fifth most common cancer and the third leading cause of cancer-related deaths globally, according to the World Health Organization's latest estimates [1]. Nearly 1 million new cases and ~ 725,000 deaths attributable to G/GEJ cancer occurred in 2012, with the highest incidence and mortality rates reported in East Asia [1].
First-line standard of care in patients with human epidermal growth factor receptor 2-negative (HER2−), unresectable, or recurrent G/GEJ cancer includes a two-drug combination of platinum-based agents and fluoropyrimidines, e.g., cisplatin, oxaliplatin, S-1, capecitabine, or 5-fluorouracil. In some medically fit patients with good performance status, a third drug, e.g., docetaxel or epirubicin, may be added [2][3][4]. Depending on the method of assessment, ~ 6 to ~ 37% of all G/GEJ cancers show HER2 overexpression [5,6]. Based on the results of the ToGA study [7], trastuzumab is now recommended in combination with chemotherapy as first-line therapy in patients with HER2-positive (HER2+) metastatic or advanced G/GEJ cancer [2][3][4].
The relationship between HER2 and PD-L1 expression has been assessed in a few studies recently; however, there are conflicting conclusions regarding their correlation [15][16][17][18]. In a study evaluating the expression and impact of PD-L1/PD-1 in gastric cancer in Caucasian patients, almost 50% of PD-L1+ patients were HER2+ [15]. In in vitro studies in HER2-overexpressing cell lines, PD-L1 expression decreased in a dose-and time-dependent manner after EGFR/HER2-targeted treatment [17]. In contrast, in a retrospective analysis of resected, treatment-naïve gastric cancers, PD-L1 expression was observed more frequently in the HER2 − group than in the HER2+ group (39.0% vs 24.2%; P = 0.020) [18]. Thus, discrepancies concerning the relationship between HER2 and PD-L1 expression exist and data related to the combined or sequential use of anti-HER2 and anti-PD-1/PD-L1 agents are also lacking. In the ATT RAC TION-2 trial, we did not observe a clear relationship between PD-L1 expression and efficacy of nivolumab; furthermore, there is no evidence in literature regarding the outcome of immune checkpoint inhibitor treatment, including nivolumab, on HER2+ patients. Although HER2 status was not captured in ATT RAC TION-2, the study did include patients who had previously used or not used trastuzumab. Therefore, exploratory analysis of ATT RAC TION-2 data based on prior trastuzumab-use status could serve as surrogate for HER2 expression status.
The objective of this post hoc analysis of the ATT RAC TION-2 trial was to assess the efficacy and safety of nivolumab for advanced G/GEJ cancer in patients who had previously used trastuzumab.

Study design
In this post hoc analysis of the ATT RAC TION-2 trial [12], a randomized, double-blind, placebo-controlled, phase 3 trial conducted at 49 sites in Japan, South Korea, and Taiwan between October 2014 and August 2016, the efficacy and safety of nivolumab were studied according to previous use of trastuzumab. ATT RAC TION-2 (ClinicalTrials. gov ID: NCT02267343) was conducted in accordance with the Declaration of Helsinki and the Good Clinical Practice guidelines developed by the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use [19]. The protocol was approved by the institutional review board or independent ethics committee for each study center. Written informed consent was obtained from all patients.

Treatment
Enrolled patients were randomized in a 2:1 ratio via an interactive web response system to receive nivolumab or placebo. Patients received 3 mg/kg nivolumab or placebo intravenously every 2 weeks for 6 weeks (one treatment cycle). Dose modification was allowed in case of a ≥ 10% change in body weight after randomization. Study treatment was continued until progressive disease (PD) evaluated by an investigator or onset of toxicity requiring permanent discontinuation of study treatment. Patients could continue study treatment after the first episode of PD if they showed evidence of investigator-assessed clinical benefit, tolerance of study drug, and stable performance status; if continuation of treatment would not delay an intervention to prevent serious complications of disease progression; and if the patient provided written informed consent before continuing the study treatment.

Endpoints
The primary endpoint was OS. Secondary efficacy endpoints included PFS, objective response rate (ORR; proportion of patients with confirmed complete response [CR] or partial response [PR]), disease control rate (DCR; proportion of patients with confirmed CR, PR, or stable disease), duration of response (DOR), time to response (TTR), best overall response (BOR), and maximum percentage change from baseline in the sum of diameters of target lesions. Safety endpoints included adverse events (AEs) and treatmentrelated AEs occurring through the study period.

Assessments
Tumor responses were assessed via computed tomography or magnetic resonance imaging according to Response Evaluation Criteria In Solid Tumors guidelines version 1.1. Tumor assessment was repeated after each cycle for ten cycles ( ~ 14 months), then after every two treatment cycles until discontinuation of study treatment because of PD, initiation of post-study treatment, or assessment of PD in patients who discontinued because of toxicity. Tumor assessments were also performed at the end of treatment. AEs were evaluated according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0 continuously during treatment and for 28 days thereafter. For patients with available tumor samples, PD-L1 tumor expression was determined retrospectively by immunohistochemistry performed at a central laboratory (28-8 pharmDx assay; Dako, Carpinteria, CA, USA). PD-L1 positivity was defined as staining in ≥ 1% of tumor cells.

Statistical analysis
A subgroup analysis by history of trastuzumab use was conducted. The Kaplan-Meier method was used to estimate the median OS, PFS, and the 95% CI for each treatment group and subgroup. HRs between treatment groups and their 95% CIs were calculated using the Cox proportional hazards model adjusted by demographic factors selected by a stepwise method for each subgroup. The P value of the HR was reported for the descriptive analysis of the difference between the two treatment groups. The P value of interaction term is reported for the model with both Tmab+ and Tmab− groups to evaluate association of Tmab use and treatment group. ORR and DCR, and their 95% CIs were estimated for each treatment group. Safety analyses were performed in the safety population (all patients who received at least one dose of study treatment). SAS software (versions 9.3 and 9.4) was used for statistical analyses.   Table 2). Notably, median PFS with nivolumab was similar in the Tmab+ and Tmab− groups. Kaplan-Meier curves for PFS of nivolumab and placebo initially overlapped in both Tmab+ and Tmab− groups but separated later (Fig. 2c, d).

Discussion
In this post hoc analysis, nivolumab improved OS, PFS, ORR, DCR, and reduction in tumor burden compared with placebo in both Tmab+ and Tmab− groups. Patients treated with nivolumab had a sustained and durable response compared with placebo in both Tmab+ and Tmab− groups. Safety of nivolumab was comparable between Tmab+ and Tmab− patients.  Trastuzumab in combination with platinum-based chemotherapy is the standard first-line treatment in HER2+ gastric cancer patients [2][3][4] based on results from ToGA [7]. Although clinical trials using new anti-HER2 agents were successful for metastatic breast cancer, the results from previous studies conducted in the first-or second-line setting for HER2+ gastric cancer patients, however, were disappointing. As the first-line palliative chemotherapy treatment in the phase 3 trial (TRIO-013/LOGiC), patients with HER2+ advanced G/GEJ cancer received lapatinib (an anti-HER2 agent) or placebo in combination with capecitabine plus oxaliplatin (CapeOX); OS was not significantly different (median OS, 12.2 vs 10.5 months; HR, 0.91 [95% CI 0.73-1.12]; P = 0.3492) [20]. Moreover, in a phase 3 trial (JACOB), metastatic G/GEJ cancer patients received pertuzumab (another anti-HER2 antibody) or placebo in combination with trastuzumab plus chemotherapy (standard cisplatin/fluoropyrimidine regimen) as first-line treatment; OS was not significantly different between the pertuzumab and placebo arms (median OS, 17.5 vs 14.2 months; HR, 0.84 [95% CI 0.71-1.00]; P = 0.0565) [21]. As the secondline treatment in a randomized phase 2 study conducted by the West Japan Oncology Group (WJOG7112G [T-ACT]), trastuzumab plus paclitaxel showed no benefit over paclitaxel alone in patients with HER2+ advanced G/GEJ cancer refractory to first-line trastuzumab plus chemotherapy [8]. Development of new active agents for HER2+ G/GEJ cancer is warranted.
Nivolumab is recommended as a third-line or later-line therapy in gastric cancer patients who are likely to have received trastuzumab. Previous reports are suggestive of the impact of anti-HER2 therapy on the expression of PD-L1 [15][16][17][18]    In this analysis, nivolumab showed similar efficacy regardless of prior trastuzumab use. However, the HR for OS was better in the Tmab+ group than in the Tmab− group (interaction P = 0.0431). Apart from prior trastuzumab use, the proportion of patients receiving post-progression pharmacotherapy was slightly higher in the Tmab+ group treated with nivolumab. Because OS is affected by many other factors, including the patient's background and postprogression pharmacotherapy, the efficacy of nivolumab is best represented by the response rate and PFS. In terms of HR for PFS, no difference was observed between the Tmab+ and Tmab− groups (interaction P = 0.3046). Therefore, it could not be deduced that nivolumab showed better efficacy in HER2+ G/GEJ cancer patients. To say the least, the results of this study suggest that nivolumab can be efficacious regardless of HER2 expression status. Also, considering the patient background, fewer patients receiving nivolumab had peritoneal metastases in the Tmab+ than in the Tmab− group (8.5% vs 21.4%, respectively). On the other hand, the proportion of patients receiving nivolumab who had received ≥ 4 previous treatment regimens was higher in the Tmab+ group compared with the Tmab− group (50.8% vs 34.7%, respectively). Furthermore, 15.3% and 9.6% of patients receiving nivolumab in the Tmab+ and Tmab− groups, respectively, had been previously treated with ramucirumab. These differences in prior chemotherapy suggest that fewer active agents were available after nivolumab for the Tmab+ group than the Tmab− group. Therefore, it is difficult to explain the longer OS in the Tmab+ group based on patient background alone.
As the relationship between HER2 and PD-L1 expression is still unknown, further investigation will be required to examine whether nivolumab provides longer survival benefit in HER2+ G/GEJ cancer and to elucidate the mechanisms involved.
The safety profile of nivolumab both in Tmab+ and Tmab− patients with advanced G/GEJ cancer was similar to that previously reported and known for nivolumab. Most AEs and treatment-related AEs were mild to moderate in severity; no unexpected AEs were observed.
There are certain limitations associated with this study. The number of HER2+ patients was small and associated with substantial differences in patient backgrounds. The HER2 status of patients immediately before enrollment and its change during treatment was not assessed. In addition, PD-L1 status was assessed using archival tissues before trastuzumab use in a limited number of patients. Therefore, it is not clear whether trastuzumab treatment may have altered PD-L1 expression, just as lapatinib and afatinib have previously been shown to suppress PD-L1 expression via the EGFR/HER2 signaling blockade [17]. Although the higher proportion of patients receiving post-progression pharmacotherapy is one of the factors that could have contributed to the significantly longer OS in the Tmab+ group between nivolumab and placebo, we did not evaluate its influence on the OS.

Conclusions
The efficacy and safety of nivolumab for advanced G/GEJ cancer were consistent regardless of prior trastuzumab use. We anticipate the future development of anti-PD-1 agents including nivolumab for HER2+, advanced G/GEJ cancer.