Postmarketing surveillance of nivolumab plus ipilimumab combination therapy in Japanese patients with unresectable malignant melanoma

Although malignant melanoma is relatively rare in Japan, it is often diagnosed at a later stage than in Western countries. Nivolumab and ipilimumab are immune checkpoint inhibitors targeting programmed death 1 and cytotoxic T‐lymphocyte‐associated protein 4, respectively. Owing to their complementary anticancer effects, nivolumab and ipilimumab combination therapy (N + I) has been studied and approved for treating malignant melanoma in various countries including Japan. Real‐world postmarketing surveillance was implemented to record treatment‐related adverse events (TRAEs) in patients treated with N + I following its approval in Japan. Patients were eligible for registration if they had unresectable malignant melanoma and started N + I between September 2018 and August 2019. The observation period was 13 weeks from starting N + I. Only safety information was collected and evaluated. The final case report form lock was March 2021. Overall, 173 patients (median age, 66.0 years; performance status 0–1, 88.4%; skin: 53.2%; mucosal: 32.4%) were eligible for the analyses. Overall, 34.1% of patients completed 4 doses of N + I. N + I was discontinued by 63.0% (due to adverse events in 67.9% and disease progression/death in 22.9%). Any grade and grade ≥3 TRAEs were reported in 73.41% and 52.02%, respectively. TRAEs in ≥10 patients were hepatic function abnormal (any grade/grade ≥3: 23.12%/13.29%), pyrexia (10.40%/0.58%), diarrhea (9.25%/2.89%), rash (8.67%/0.58%), hypophysitis (5.78%/5.20%), interstitial lung disease (5.78%/2.89%), and liver disorder (5.78%/4.62%). TRAEs were classified as recovered (36.99% of patients), recovering (44.51%), unrecovered (13.29%), recovered with sequelae (2.31%), and death (1.73%). Overall, 24 of 34 patients (70.59%) with gastrointestinal‐related and 53 of 65 (81.54%) liver‐related TRAEs received treatment, such as a steroid with/without an immunosuppressant; most patients recovered within 1 to 2 months. In conclusion, this postmarketing surveillance of N + I in patients with unresectable malignant melanoma revealed no new safety concerns compared with results of prior studies. Immune‐related TRAEs were generally manageable by appropriate treatment including a steroid.


| INTRODUC TI ON
Malignant melanoma is a common cancer in Western countries, but it is rare in Japan, with an estimated incidence of only 1.75 per 100 000 people, 1 and an estimated 2000 new cases of melanoma diagnosed each year. 1,2 The predominant subtype of melanoma in Japan is acral lentiginous melanoma (40.4%), while superficial spreading melanoma (20.5%), nodular melanoma (10.0%), lentiginous melanoma (8.1%), and mucosal melanoma (9.5%) are less common. 2,3 These frequencies differ from those reported in Western countries (acral lentiginous melanoma: 4.6%; superficial spreading melanoma: 54.1%; nodular melanoma: 13.8%; lentiginous melanoma: 3.9%; mucosal melanoma: 0.4%). 4 The higher frequencies of acral lentiginous melanoma and mucosal melanoma in Japan may also contribute to differences in the distribution of clinical stages at diagnosis; therefore, patients in Japan are often given a diagnosis at a more advanced stage (stage III: 24.1%; stage IV: 6.2%) compared with European or American patients (stage III: 7.5%; stage IV: 2.5%). 2,3 Since 2010, several immune checkpoint inhibitors (ICIs) have been developed and introduced into the clinical recommendations for metastatic or unresectable melanoma. 5 One of these, nivolumab, is a human monoclonal antibody of the IgG4 subclass that targets human programmed death 1. [6][7][8] However, the objective response rates for nivolumab monotherapy in clinical trials are ~40%; thus, alternative treatments were required. Another ICI is ipilimumab, a human monoclonal antibody of the IgG1 subclass that targets cytotoxic T-lymphocyte-associated protein 4 (CTLA-4). 9,10 Both nivolumab and ipilimumab are approved in Japan and other countries for treating various cancers such as unresectable malignant melanoma. 6,9 Owing to their complementary anticancer effects, nivolumab and ipilimumab combination therapy (N + I) was considered to have additive or synergistic effects based on animal studies. 11 The clinical efficacy of combination therapy was investigated in the global (excluding Japan), phase III CheckMate 067 (CM 067) study, in which N + I significantly prolonged survival time compared with ipilimumab monotherapy. 12 The phase II ONO-4538-17  study subsequently demonstrated that the efficacy (progressionfree survival and overall survival) and safety of N + I in Japanese patients were similar to those reported in the CM 067 study. 13,14 Those findings led to the approval of N + I in Japan, where this combination is frequently used as first-line therapy for unresectable metastatic melanoma.
In clinical practice, it is important to consider the safety of therapies and common treatment-related adverse events (TRAEs).
For example, almost all of the patients treated with N + I in CM 067 experienced TRAEs of any grade, with diarrhea, rash, fatigue, and pruritus being the most common, and about half experienced TRAEs of grades 3 or 4. 12 Moreover, all of the patients in ONO-17 experienced TRAEs of any grade and three-quarters experienced grade 3 or 4 TRAEs. 13 This high rate of TRAEs, including immune-related adverse events (irAEs), highlights the importance of careful observation of patients treated with N + I in clinical practice. It is also important to consider that the data from clinical trials such as ONO-17 may not fully reflect real-world settings mainly due to the small number of patients enrolled. In a retrospective study that investigated the safety of N + I in 100 patients in Japan, Takahashi et al. reported that TRAEs occurred in 89% of patients, including grade 3 or 4 TRAEs in 56%. 15 However, that retrospective study was based on data from a single facility with a limited number of patients, and the authors did not report the management or outcomes of the TRAEs. Thus, more information on the frequencies of TRAEs as well as their management and outcomes is needed.
Following the approval of N + I for patients with unresectable malignant melanoma in Japan, postmarketing surveillance (PMS) was implemented to collect safety information in real-world clinical use for a large number of patients. Information on TRAEs recorded in the PMS included the grade, timing of onset, management, and outcomes. In CM 067 and ONO-17, several gastrointestinal-(GI; colitis, diarrhea, and GI perforation) and liver-related (hepatic function abnormal and cholangitis sclerosing) adverse events (AEs), often considered irAEs, were among the most frequent grade ≥3 TRAEs that led to treatment discontinuation. As a consequence, these TRAEs were included in the safety specification for N + I and the proper use guide for the combination. 16 However, there are few reports about the management or outcomes of these TRAEs in clinical practice. Therefore, in discussions with the Japanese Pharmaceutical and Medical Devices Agency (PMDA), this PMS specifically recorded information regarding the incidence, timing of onset, management, outcome, and time to recovery, with an emphasis on GI-and liver-related TRAEs as priority items, to understand the impact and management of these irAEs in clinical practice. Because these events generally occurred within 3 months of starting treatment, this PMS recorded TRAEs occurring within 13 weeks from the start of N + I.

| Ethics
We performed an observational, open-label, noninterventional PMS of patients who started N + I in clinical practice at 38 institutions in Japan. As an observational PMS, clinical trial registration was not necessary. This PMS adhered to the Japanese regulatory requirements stipulated in Good Postmarketing Study Practice. Ethics committee approval and informed consent are not mandated under Japanese ministerial ordinance for PMS.

| Patients
Patients were eligible for registration in the PMS if they had unresectable malignant melanoma and started N + I between September 2018 and August 2019. Confirmation that critical missing data had been entered was declared at the final case report form (CRF) lock in March 2021. Patients who had received nivolumab, ipilimumab (not in combination), or other prior treatments were eligible for registration. It was planned to register 100 patients for the safety analysis to detect approximately 43 patients with colitis/diarrhea/GI perforation and 25 with hepatic dysfunction/cholangitis sclerosing, based on the reported frequencies of 42.8% and 24.9%, respectively, in CM 067. To account for potential dropouts, the target registration was 120 patients.
In terms of the dosing schedule of N + I for melanoma, the usual dose of nivolumab is 80 mg per body, administered intravenously every 3 weeks for the first 4 doses and then 240 mg per body every 2 weeks. The usual dose of ipilimumab is 3 mg/kg (body weight) administered as a single intravenous infusion every 3 weeks (a total of 4 times) ( Figure 1).
The registered patients were to be observed for 13 weeks after the start of N + I, corresponding to the usual treatment process ( Figure 1). For patients who terminated or discontinued N + I before week 13, their outcomes and TRAEs were to be followed up as much as possible until the end of the observation period.

| Assessments and outcomes
All information was collected using CRFs, which captured information about patient characteristics, duration of treatment, and TRAEs during the 13-week observation period. The physicians assessed the causality and graded the TRAEs. TRAEs were defined as AEs for which a causal relationship with treatment could not be excluded. As TRAEs of special interest, the CRFs specifically collected information on GI-(colitis, enteritis, diarrhea, and GI perforation) and liver-related events (hepatic failure, hepatic function disorder, hepatitis, cholangitis sclerosing, and liver-related laboratory tests). For these TRAEs, the CRFs recorded the status of drug use at the time of onset (discontinuation and interruption) and actions taken (e.g., administration of steroids, antitumor necrosis factor α agents, and other treatments). The time to onset, outcome, and time to recovery were to be recorded for all TRAEs. All TRAEs were tabulated by system organ class and preferred term using MedDRA (Medical Dictionary for Regulatory Activities)/J version 24.0. If any patient experienced multiple TRAEs of a specific type, it was counted only once, unless the events were of different grades, in which case they were tallied by grade. The PMS did not record information on the effectiveness or cancer responses.
For the purposes of data analyses, we used the safety population, which consisted of patients from whom the CRF was collected, excluding any ineligible patients (patients registered outside the contract period, patients with a history of N + I treatment, patients registered outside the surveillance period). The data were analyzed using SAS release 9.4 (SAS Institute Inc).

| Patients
Of 194 registered patients, CRFs were collected for 181. Of these, 8 were excluded (6 outside the contract period, 1 with history of N + I therapy, and 1 outside the surveillance period) from the safety analysis set, which therefore comprised 173 patients with unresectable malignant melanoma who newly started N + I ( Figure 2).
The median age was 66.0 years and 54.3% of patients were aged ≥65 years ( Table 1) F I G U R E 1 Dosing schedule of nivolumab and ipilimumab. iv, intravenous.

| Overall TRAEs
The timing of onset of each type of TRAE is summarized in Tables

| GI-and liver-related TRAEs
The GI-related TRAEs occurred in week ≤3 in 9 of 173 patients (5.20%), in week >3 to ≤6 in 11 of 172 patients (6.40%), in week >6 to ≤9 in 7 of 160 patients (4.38%), and in week >9 to ≤12 in 8 of 154 patients (5.19%); no GI-related TRAEs were reported after week 12 (Table S2). The time to onset of GI-related TRAEs is presented in detail in Table 4. The median time (range) to onset of diarrhea, enterocolitis, colitis, and immune-mediated enterocolitis were 31.0    Table 5). There were no deaths due to liver-related TRAEs. The outcome was unknown for 5 events (6.76%). Overall, 62 of 74 liverrelated events were classified as recovered or recovering.  Table S1. increased, and hepatic enzyme increased, respectively ( Table 6). Figure 3 shows the treatments used for the 39 GI-related TRAEs, of which 69.23% were treated, mostly with a steroid ± other treatment (for 51.28%). Table 7 shows the outcomes of GI-related TRAEs according to their management. Of 34 patients with GI-related TRAEs ( Table 7), 24 (70.59%) received an intervention, which comprised a steroid + immunosuppressant ± other treatment in 5 patients and a steroid ± other treatment in 17 patients. The majority of patients were classified as recovered/recovering, regardless of treatment, with the exception of 2 patients classified as unrecovered (enterocolitis and GI perforation) and 1 patient who died (enterocolitis).

| Any grade TRAEs
The treatments used for the 75 liver-related TRAEs are presented in Figure 3. Most of these TRAEs were treated (81.33%); a steroid ± other treatment was the most common approach (for 57.33%) for managing these TRAEs. TA B L E 4 Incidence and time to onset of GI-and liver-related TRAEs (N = 173). treatment, 9 recovered, 21 were recovering, 3 were unrecovered, and 4 had an unknown outcome. Of 9 patients who received other treatments, 5 were recovered, 4 were recovering, and 1 was recovering with sequalae.

| DISCUSS ION
This PMS was implemented to obtain more insight into the safety (TRAEs) associated with the use of N + I for unresectable metastatic melanoma in real-world clinical practice in Japan. TRAEs TA B L E 6 Time to recovered/recovering (days) of GI-and liverrelated TRAEs.
No. In CM 067, the most frequent TRAEs in the N + I group were diarrhea (any grade/grade 3 or 4: 44.1%/9.3%), rash (40.3%/4.8%), fatigue (35.1%/4.2%), and pruritus (33.2%/1.9%). 12 In the Japanese ONO-17 study, rash (60%/7%), diarrhea (57%/3%), pyrexia (43%/3%), and lipase increased (40%/23%) were the most common TRAEs. 13 In a small real-world study in Japan involving 57 patients treated at a single and ONO-17 (by system organ class/preferred term) were as follows (Table S1)  in some cases, an immunosuppressant (infliximab). The majority of cases of GI-related TRAEs recovered or were recovering, suggesting appropriate management of these TRAEs. A similar pattern was also observed for liver-related TRAEs, for which comparable recommendations are made in the proper use guide, 16 and most events were also classified as recovered or recovering. As indicated in Tables 7 and 8, the administration of immunosuppressants was reserved for grade ≥3 GI-and liver-related events and the majority of cases were classified as recovered or recovering. This highlights the importance of early detection and appropriate management of TRAEs, particularly irAEs in clinical practice.

GI-related TRAEs
Early detection and timely interventions are important to support continuation of N + I. The frequency of TRAEs during N + I treatment tended to peak within 3-6 weeks, and subsequently decreased over time, and many resolved during the 13-week observation period. Furthermore, the GI-and liver-related TRAEs were often manageable with a steroid and, in some cases, an immunosuppressant. Continuation of ICI therapy after a TRAE is expected to be advantageous in terms of the prognosis according to the results of some but not all studies. [22][23][24][25][26] Further studies are therefore needed to understand the relationship between AEs and prognosis in this setting.

| Limitations
Some limitations of this PMS include its single-arm design, wherein patients were selected by the physicians, who also made the decisions to discontinue treatment. The PMS recorded safety information but not effectiveness data. Furthermore, safety information was collected for up to 13 weeks after the start of N + I; TRAEs occurring beyond this time were not recorded. The TRAEs of special interest were defined according to MedDRA preferred terms at the time of planning this PMS in accordance with Japanese regulations. 27 In this article, the TRAEs were analyzed and aggregated based on the definitions set forth in the PMS protocol; further aggregation of similar TRAEs was not planned.

| Conclusion
Based on the results of this PMS with a follow-up of 13 weeks after the start of N + I for unresectable malignant melanoma in Japan, we observed no new safety concerns compared with results of clinical trials conducted in Japan or other countries. TRAEs identified as being immune-related were manageable by appropriate treatment, including a steroid.

ACK N OWLED G M ENTS
The authors thank all of the physicians and institutions that participated in this surveillance. This surveillance was funded by Ono Pharmaceutical Co., Ltd. and Bristol-Myers Squibb K.K. The authors thank Hiroaki Ito (Bristol-Myers Squibb K.K., Tokyo, Japan) for assistance with drafting the manuscript. The authors also acknowledge Nicholas D. Smith (EMC K.K.) for medical writing support, which was funded by Ono Pharmaceutical Co., Ltd. and Bristol Myers Squibb K.K.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data generated during and/or analyzed in the current study are not publicly available for the protection of personal data.