Real-world safety and effectiveness of nivolumab in Japanese patients with unresectable advanced or recurrent gastric/gastroesophageal junction cancer that has progressed after chemotherapy: a postmarketing surveillance study

Background This postmarketing surveillance study evaluated the real-world safety and effectiveness of nivolumab as salvage (after ≥ 2 lines) therapy in Japanese patients with unresectable advanced or recurrent gastric/gastroesophageal junction (G/GEJ) cancer. Methods This multicenter, observational study was conducted at 158 centers in Japan. Patients with unresectable advanced or recurrent G/GEJ cancer were registered between Nov 1, 2017, and Oct 31, 2018, and observed for 6 months after treatment initiation with nivolumab. Correlation of background characteristics with treatment-related adverse events (TRAEs) and tumor response was explored. Results Overall, 654 patients were registered (safety analysis set, n = 650; effectiveness analysis set, n = 636; response evaluation set, n = 516). The incidences of all TRAEs and grade ≥ 3 TRAEs were 31.5 and 11.2%, respectively. TRAEs significantly correlated with the absence of peritoneal metastasis; C-reactive protein level < 1; prior G/GEJ cancer surgery; and past or concomitant pulmonary, thyroid, or renal disease (each p < 0.05). The incidence of TRAEs was significantly lower in patients with higher Glasgow prognostic scores (p < 0.05). No new safety signals were observed. Complete response, partial response, stable disease, and progressive disease were observed in 1.2, 10.1, 27.1, and 58.3% of the response evaluation set, respectively. Patients aged ≥ 65 years (13.9 vs 5.3%, p = 0.0083) and ≥ 75 years (18.8 vs 9.2%, p = 0.0036) showed a higher response rate than their younger counterparts. Conclusions The real-world safety and effectiveness of nivolumab as salvage (after ≥ 2 lines) therapy in Japanese patients with unresectable advanced or recurrent G/GEJ cancer were consistent with those observed in the phase 3 ATTRACTION-2 study. Supplementary Information The online version contains supplementary material available at 10.1007/s10120-021-01244-y.


3
Background Gastric cancer is the fifth most common cancer and the third leading cause of cancer-related deaths worldwide [1]. Gastric/gastroesophageal junction (G/GEJ) cancer accounts for ~ 783,000 deaths, with over 1,000,000 new cases of G/ GEJ cancer reported worldwide in 2018 [1]. The cumulative risk of developing gastric cancer from birth to age 74 years is higher in men (1.87%) than in women (0.79%) [1]. In Japan, gastric cancer was the third most common cause of cancer-related deaths in 2018 [2].
Nivolumab, an immune checkpoint inhibitor and a fully humanized anti-programmed cell death protein 1 (PD-1) monoclonal antibody, has been approved for various indications in more than 65 countries, including the United States, European Union, and countries in Asia [3]. In the ATT RAC TION-2 study, nivolumab, when compared with placebo in patients with unresectable advanced or recurrent G/GEJ cancer after failure of ≥ 2 lines of chemotherapy [4], improved the overall survival (OS; median, 5.26 vs 4.14 months), with a better 1-year OS rate (26.2 vs 10.9%) [4]. Consequently, the results of ATT RAC TION-2 [4] led to the approval of nivolumab as a third-or later-line therapeutic option for patients with unresectable advanced or recurrent G/GEJ cancer in several countries, including Japan [5]. In Japan and Korea, nivolumab is recommended as third-or later-line therapy in the treatment guidelines of gastric cancer [5,6].
However, as ATT RAC TION-2 was limited by its sample size and stringent inclusion/exclusion criteria, the results from this study do not sufficiently reflect the real-world scenario, where not a few patients to be excluded from the clinical trials are treated with nivolumab. Accordingly, the Ministry of Health, Labour and Welfare (MHLW) required a postmarketing surveillance (PMS) study after drug approval. Therefore, this PMS study was conducted in Japan to evaluate the real-world safety and effectiveness of nivolumab in patients with unresectable advanced or recurrent G/GEJ cancer having diverse background factors. To the best of our knowledge, this is the first real-world evaluation of an immune checkpoint inhibitor in over 600 Japanese patients with unresectable advanced or recurrent G/GEJ cancer.

Methods
This PMS was a multicenter, open-label, observational study conducted at 158 centers in Japan. Patients with unresectable advanced or recurrent G/GEJ cancer to be treated with nivolumab (n = 654) were registered between Nov 1, 2017, and Oct 31, 2018. These registered patients were observed for 6 months or until death after treatment initiation with nivolumab.

Patients
Patients with unresectable advanced or recurrent G/GEJ cancer who newly received nivolumab after failure of ≥ 2 lines of chemotherapy were registered for this PMS study. Patients were excluded from the effectiveness analysis if they received nivolumab as off-label use (treatment for an indication other than unresectable advanced or recurrent G/ GEJ cancer after failure of ≥ 2 lines of chemotherapy before the initiation of nivolumab).

Treatment
Nivolumab at the approved dose as described in the prescribing information was administered. Of note, the approved dose was changed from the body weight-equivalent dose (3 mg/kg) to a fixed dose of 240 mg/body that received marketing authorization in August 2018. Treatment was discontinued based on the physician's discretion owing to progressive disease (PD) according to the Response Evaluation Criteria in Solid Tumors (RECIST) v1.1, lack of effectiveness even if the criteria for PD were not met, intolerability, or patients' refusal.

Safety
Adverse events (AEs) and treatment-related adverse events (TRAEs) were reported; their relation to nivolumab was judged by each attending physician. The clinical course of TRAEs of special interest was monitored. All AEs were graded using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE v4.0 or v5.0).

Effectiveness
The question, "Was the response evaluated in accordance with the RECIST v1.1?," was included in the case report form (CRF). If the answer was "Yes," patients were included in the response evaluation set, in which tumor response [complete response (CR), partial response (PR), stable disease (SD), and PD] was assessed, and the objective response rate (ORR; CR or PR) was obtained. The 6-month survival rate was evaluated using the Kaplan-Meier method.

Statistical analysis
The sample size of the safety analysis set was planned at 500, which provided a power of 77.74% to detect ≥ 1 AE with a frequency of 0.30%, corresponding to the lowest frequency of AEs that occurred in the ATT RAC TION-2 trial (0.30%; 1/330 patients).
Exploratory subgroup analyses were performed on the incidence of TRAEs and the ORR; the comparison between subgroups was performed using Fisher's exact test, Wilcoxon rank sum test, or chi-squared test. Association between each patient characteristics, excluding the factors which allowed double count such as tumor location and histological differentiation, and the incidences of TRAEs and ORR were analyzed using multivariate logistic regression to adjust the number of nivolumab doses (≥ 5 vs ≤ 4).

Patient disposition, demographics, and baseline characteristics
A total of 654 patients were registered, of whom 650 comprised the safety analysis set and 636 comprised the effectiveness analysis set. Patient disposition is shown in Supplementary Fig. 1. Patient demographics and baseline characteristics of the safety analysis set are shown in Table 1.
In the safety analysis set, 73% of patients were men; the median (range) age was 69 (20-87) years; 90% of patients had an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1; and the median number of prior treatment regimens was 2 (0-6). The number of patients with poorly differentiated adenocarcinoma and well-differentiated adenocarcinoma was 385 and 343, respectively, allowing a double count if a patient had histological components of both types. In addition, the proportion of patients with a large amount of ascites was 17.7% and those with past or present comorbidities was 65.8%.

Treatment
A total of 460 (70.8%) and 69 (10.6%) patients received nivolumab at doses of 3 and 240 mg/body intravenous infusion per cycle, respectively, throughout the study period. In 77 (11.8%) patients, the dose was changed from 3 to 240 mg/body during the study period at the physician's discretion, while 44 (6.8%) patients received other doses. The mean ± standard deviation (median; range) number of nivolumab administration was 5.7 ± 4.1 (4.5; 1-14) in the safety analysis set.
A total of 106 (16%) patients continued treatment with nivolumab throughout the 6-month observation period. Among the remaining 544 (84%) patients, the major reasons for nivolumab discontinuation within 6 months were disease progression (including death; 71.8%), AEs (10.9%), lack of effectiveness (8.9%), and hospital transfer (1.1%). One patient (0.2%) discontinued treatment with nivolumab due to achieving a complete response as determined by the physician according to RECIST.

Safety
TRAEs (all grades) were reported in 205 (31.5%) patients ( Fig. 1). The median (range) duration from initiating nivolumab to occurrence of TRAEs was 32 (1-178) days. The most common TRAEs (> 2%) were hypothyroidism (4.2%), diarrhea (3.7%), decreased appetite (2.9%), rash (2.5%), and malaise (2.3%). Grade ≥ 3 TRAEs were observed in 11.2% of patients, and the most common grade ≥ 3 TRAEs were decreased appetite and diarrhea (0.9% each).  Table 2). Detailed information of past and present renal, pulmonary, and thyroid comorbidities categorized using system organ class and preferred terms are summarized in Supplementary Tables 3, 4, and 5. Although the incidence of TRAEs was statistically significantly associated with number of treatment lines before initiating nivolumab (p = 0.0281) and number of doses (p = 0.0337), a clear linear relationship was not observed; the incidence showed a peak at 3 treatment lines and 9-12 doses. The multivariate analysis for incidence of TRAEs according to background factors adjusted by the number of nivolumab doses (≥ 5 vs ≤ 4) showed a similar significant increase in the incidence of TRAEs among patients with vs without peritoneal metastasis (p = 0.0494); and patients with any past or present comorbidities (p = 0.0259), including pulmonary (p = 0.0055), thyroid (p = 0.0287), and renal (p = 0.0189) diseases (Table 2).

Clinical course of patients with TRAEs of special interest
The clinical course of patients with TRAEs of special interest is summarized in Table 3. Among different TRAEs, the CRP C-reactive protein, ECOG PS Eastern Cooperative Oncology Group performance status, FISH fluorescence in situ hybridization, HER2 human epidermal growth factor receptor 2, IHC immunohistochemistry, NLR neutrophil to lymphocyte ratio a The total number exceeds the number of the whole population because more than one category for tumor location and histological differentiation were checked in the case report form in some patients b Well-differentiated adenocarcinoma includes papillary adenocarcinoma and tubular adenocarcinoma, whereas poorly differentiated adenocarcinoma includes signet ring cell carcinoma and mucinous adenocarcinoma c IHC3 + or IHC2 + and FISH + Fig. 1 Incidence of treatmentrelated adverse events categorized by grade

Discussion
This large PMS study was conducted to assess the real-world safety and effectiveness of nivolumab after failure of ≥ 2 lines of chemotherapy in Japanese patients with unresectable advanced or recurrent G/GEJ cancer. The incidence of all TRAEs and grade ≥ 3 TRAEs was 31.5 and 11.2%, respectively. The occurrence of TRAEs categorized by system organ class was similar to that reported in the ATT RAC TION-2 study [4], in which TRAEs of any grade were reported in 141 (43%) patients in the nivolumab group, including 34 (10%) with grade 3-4 TRAEs [4]. Clinical trials mandate a more detailed monitoring of adverse reactions compared with PMS studies, which potentially results in slightly higher incidences compared with those observed in clinical practice. Patients with poor medical conditions, such as ascites and comorbidities, were excluded from the ATT RAC TION-2 study [4], and, therefore, the safety of nivolumab in these patients could not be evaluated. In this PMS, multivariate analysis showed that patients with past and present comorbidities such as pulmonary disease, thyroid disease, and renal disease, experienced significantly more TRAEs, Table 3 Outcomes of TRAEs of special interest suggesting that comorbidities likely affect the overall incidence of TRAEs in patients treated with nivolumab.
Although it is difficult to explain how past and present renal comorbidities affect renal TRAEs, TRAEs in the lungs and endocrine system were observed in a higher proportion of patients with vs without past or present pulmonary disease and thyroid disease, respectively, suggesting that pulmonary and thyroid comorbidities might be risk factors for pulmonary and endocrine TRAEs, respectively. These data from patients with comorbidities in this PMS may be useful to call attention to TRAEs of nivolumab in clinical practice. On the contrary, patients with a good condition, such as no peritoneal metastasis, Glasgow prognostic score of 0, CRP level < 1, and prior surgical treatment for G/GEJ cancer, are likely to experience significantly more TRAEs. Moreover, higher incidences of TRAEs in patients with no peritoneal metastasis and a Glasgow prognostic score of 0 and CRP level < 1 were associated with an increased exposure period of nivolumab. The multivariate analysis adjusted by the treatment period identified the absence of peritoneal metastasis and any past or present comorbidities as factors associated with TRAEs. Both the absence of peritoneal metastasis and a Glasgow prognostic score of 0 are wellknown favorable prognostic factors. It is considered that the active immune status in patients with a good condition may be related to TRAEs. About 54-82% of patients who experienced TRAEs of special interest were resolved/resolving by pharmacotherapy such as corticosteroids and hormone replacement or by observation only. As some TRAEs were not resolved, leading to sequelae and/or death, early diagnosis and optimal treatment for TRAEs are very important in clinical practice.
Tumor response among the response evaluation set (CR, 1.2%; PR, 10.1%; SD, 27.1%; and PD, 58.3%) and the 6-month survival rate (52.0%) were consistent with those reported in the ATT RAC TION-2 trial [4]. Reportedly, the occurrence of TRAEs was associated with better efficacy of nivolumab in patients with metastatic melanoma [7] and advanced G/GEJ cancer [8]. In the CheckMate 141 study in patients with recurrent or metastatic head and neck cancer receiving nivolumab, the median OS (14.3 vs 5.2 months; hazard ratio, 0.10; 95% confidence interval, 0.03-0.39) and 2-year OS rates (31.3 vs 0%) were higher in patients with TRAEs, particularly skin-related disorders, than in those without TRAEs [9]. Similarly, the ORR (18.5 vs 6.7%) was higher in patients who reported TRAEs than in those without TRAEs in this PMS study. When ORR was compared according to patient background characteristics in this PMS, ORR increased with increasing age from ≥ 50 years. However, in the ATT RAC TION-2 trial, there was no clear relationship between age and response. Thus, the reason for a higher ORR in the elderly remains unknown. Biological and/or immunological background, such as tumor mutation burden, might be confounding in this PMS study. Furthermore, the limited number of patients in the categories by age may have impacted data interpretation. It is expected that biomarkers for efficacy and TRAEs will be established in the future.
As is the nature of PMS studies, clinical information could not be collected as precisely as in the clinical trials, and the observation period was rather short. Additionally, no biomarker analysis was conducted. A comparison between subgroups could be affected by factors that were not collected or evaluated. Also, multivariate analysis could not be performed for factors such as tumor location and histological differentiation because some patients were counted repeatedly for each tumor location and differentiation. However, this study could cover a wide population of patients with unresectable advanced or recurrent G/ GEJ cancer receiving nivolumab, and the obtained results can provide useful information reflecting the real-world scenario.

Conclusions
The results of this PMS study showed that the safety and effectiveness of nivolumab as salvage (after ≥ 2 lines) therapy in Japanese patients with unresectable advanced or recurrent G/GEJ cancer was consistent with that observed in the phase 3 ATT RAC TION-2 study. No new safety concerns were reported over the 6-month observation period.
Ethical approval All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. This was a noninterventional study conducted in routine clinical practice; therefore, in accordance with good postmarketing study practice regulations, written informed consent of patients was not required.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.