Introduction

Nivolumab is a fully human immunoglobulin G4 monoclonal antibody targeted against programmed cell death protein-1 (PD-1). Nivolumab was approved in March 2017 for the treatment of recurrent or distant metastatic head and neck cancer (HNC) in Japan. This approval was based on the survival benefits and the manageable safety profile demonstrated by nivolumab in the global phase III CheckMate 141 study, also known as ONO-4538-11 [1]. CheckMate 141 showed that nivolumab significantly prolonged overall survival (OS) compared with standard therapy alone in patients with recurrent squamous cell HNC (median OS of 7.5 vs 5.1 months; p = 0.01) [1]. In addition, OS benefit was maintained with 2-year follow-up [2].

In general, patients eligible for clinical trials are highly selected [3]. To be eligible for CheckMate 141, patients were required to have platinum-refractory squamous cell HNC. CheckMate 141 also excluded patients with Eastern Cooperative Oncology Group performance status (ECOG PS) ≥ 2. In addition, the CheckMate 141 study specifically excluded patients with primary tumor sites such as the nasopharynx, nasal cavity, paranasal sinuses, salivary glands, or lip [1].

In melanoma and non-small cell lung cancer, nivolumab has shown long-term survival in patients with stable disease (SD) as well as complete response (CR) or partial response (PR) [4, 5]. In HNC patients, however, the effectiveness by the best overall response (BOR) has not been reported in real-world practice.

In Japan, nivolumab is indicated for any type of HNC, and is not restricted to only those subtypes included in the CheckMate 141 study. In addition, only 18 patients from Japan received nivolumab in CheckMate 141. This means that there is a data gap between the results of the clinical trial and how nivolumab performs in real-world clinical practice in Japan.

The aim of this observational study was to evaluate the effectiveness and safety of nivolumab in a large real-world setting in Japan, and to address the data gap between clinical trial and real-world settings. Here, we report the primary results of a retrospective chart review of patients with HNC who were treated with nivolumab in Japan.

Patients and methods

Study design and patients

This was a multicenter, non-interventional, retrospective study conducted at 23 centers in Japan in accordance with relevant regulations in Japan (Ministerial Ordinance on Good Post-Marketing Study Practice, Ministry of Health, Labour and Welfare Ordinance Number 171, December 20, 2004). The study protocol was reviewed and approved by the Institutional Review Board/Independent Ethics Committee at each study site, and the study was conducted according to the ethical principles of the Declaration of Helsinki. Although informed consent was not obtained, patients were given the opportunity to decline to have their clinical records used for research (opt-out consent provision). This study was registered at Clinicaltrials.gov (NCT03569436).

The study includes patients with recurrent or metastatic (distant sites) HNC cancer who were treated with nivolumab for the first time between July 1, 2017, and December 31, 2017. All eligible patients were included except those who had participated in a clinical trial with antineoplastic therapy.

The data were extracted from patients’ medical charts into a specific electronic case report form. Data were collected from baseline until the most recent patient visit. The data cut-off date was 1 year after the first treatment of nivolumab in each patient. Baseline was defined as the visit prior to the start of nivolumab therapy, but the chart review encompassed the period from the diagnosis of HNC to collect data on therapies received prior to nivolumab.

Endpoints

The primary objectives were to determine the overall effectiveness, including BOR, progression-free survival (PFS), and OS, and to evaluate immune-related adverse events (AEs) in real-world clinical practice. Progression and response primarily recorded by physicians were assessed according to investigator-assessment Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria [6]. Evaluation time was not set due to the nature of this study.

AEs were classified according to the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use Medical Dictionary for Regulatory Activities Japanese edition (MedDRA/J) Version 21.0.

Drug use information was collected by recording the doses of nivolumab that the patient received, the treatments received before and after administration of nivolumab and their outcomes, the duration of nivolumab treatment, the line of therapy in which patients received nivolumab, any changes in dose or dose interruptions made as a result of AEs, and eventual reason for discontinuation of nivolumab. Platinum-refractory disease in the context of primary therapy was defined as cancer progression within 6 months after the last administration of platinum [1]. Platinum-sensitive disease was defined as cancer progression from 6 months or longer after the last administration of platinum [1].

Statistical analysis

Effectiveness and safety analyses were performed with all patients who had received ≥ 1 dose of nivolumab. Demographic and baseline characteristics, response data, and AEs were summarized using descriptive statistics (number of patients, mean and standard deviation) for continuous efficacy variables, and frequency and percentage for categorical variables. OS and PFS were estimated and plotted using the Kaplan–Meier method and expressed as the proportion of patients who survived to a specific point in time and median duration, with the corresponding two-sided 95% confidence intervals (CI). For subgroup analyses, tests of statistical significance were conducted using the log-rank test.

Statistical analyses were conducted using SAS Version 9.4 (SAS Institute, Japan).

Results

Patient disposition and characteristics

Among 256 registered patients, 79% were men and the median age was 66 years (range 20–84 years; Table 1). Of the 246 patients with known ECOG PS score, 31 patients (12.6%) had a performance status of ≥ 2. Overall, 198 of the 239 patients (82.8%) with known disease stage had stage III or IV disease at the time of HNC diagnosis (Table 1). Most patients (n = 217; 84.8%) had squamous cell carcinoma, but 29 patients (11.3%) had non-squamous cell histology. Primary tumor sites were the hypopharynx (n = 64; 25.0%), oral cavity (n = 56; 21.9%), oropharynx (n = 40; 15.6%), salivary gland (n = 23; 9.0%), larynx (n = 21; 8.2%), nasopharynx (n = 19; 7.4%), maxillary sinus (n = 14; 5.5%), and other sites (n = 19; 7.4%).

Table 1 Baseline characteristics of patients
Full size table

Seventy patients (27.3%) received nivolumab as the first-line treatment for recurrent/metastatic HNC, 110 (43.0%) as second-line, and 76 (29.7%) as third-line or later treatment.

After the 12-month follow-up, 14.1% (n = 36) were still under the treatment with nivolumab. Of the 220 patients (85.9%) who completed nivolumab treatment, 170 patients (77.3%) had shown disease progression and nine patients had died. The median (min, max) duration of nivolumab treatment was 72.5 (1, 380) days, with patients receiving a median of 6.0 (range 1–27) doses of nivolumab (Table S1).

Overall effectiveness

The BOR was assessed in 223 of 256 evaluable patients. Among these 223 patients, the BOR was CR in 3 patients (1.3%) and PR in 32 patients (14.3%), for an objective response rate (ORR) of 15.7 (95% CI 11.2–21.1) % (n = 35). An additional 61 patients (27.4%) had SD, resulting in a disease control rate of 43.0% (Fig. 1a). The median duration of response was 5.1 (95% CI 2.8–NE) months. The median PFS was 2.1 (95% CI 1.8–2.7) months (Fig. 1b) and the median OS in the 256 patients treated with nivolumab was 9.5 (95% CI 8.2–12.0) months (Fig. 1c). The estimated 12-month OS rate was 43.2 (95% CI 36.7–49.5)%, with an estimated 12-month PFS rate of 14.8 (95% CI 10.5–19.7)%.

Fig. 1
figure 1

a Overall response rate b progression-free survival and c overall survival among all patients. CI confidence intervals, BOR best overall response, CR complete response, ORR objective response rate, OS overall survival, PD progressive disease, PFS progression-free survival, PR partial response, SD stable disease

Full size image

Effectiveness by subgroup

Regarding the effectiveness of nivolumab in Japanese patients with primary tumor sites specifically excluded in the CheckMate 141 study, the ORR was 21.1 (95% CI 6.1–45.6) %, 13.6 (95% CI 2.9–34.9) %, and 7.1 (95% CI 0.2–33.9) % in the nasopharynx, salivary glands, and maxillary sinus, respectively (Fig. 2a). The median PFS was 6.1 (95% CI 2.5–NE) months in the nasopharynx, 2.1 (95% CI 1.4–4.2) months in the salivary glands, and 1.9 (95% CI 1.3–3.8) months in the maxillary sinus with 12-month PFS rates of 37.2%, 17.4%, and 7.1% for the nasopharynx, salivary glands, and maxillary sinus, respectively (Fig. 2b). Finally, the median OS was not reached for the nasopharynx and salivary glands and was 7.7 (95% CI 3.0–11.9) months in the maxillary sinus (Fig. 2c). Further, the 12-month OS rate in the nasopharynx, salivary glands, and maxillary sinus was 52.8%, 61.0%, and 24.1%, respectively. This study included a small number of patients (n = 29) with non-squamous cell carcinoma (non-SCC) in addition to patients with SCC (n = 217). Effectiveness according to histological type (SCC versus non-SCC) is shown in Figure S1. ORR, PFS, and OS were similar between patients with SCC or non-SCC with no statistically significant differences noted for any comparison. The primary site and histology of patients with non-SCC are detailed in Table S2.

Fig. 2
figure 2

a Overall response rate b progression-free survival and c overall survival among patients stratified according to the primary site. BOR best overall response, CR complete response, NR not reached, ORR objective response rate, OS overall survival, PD progressive disease, PFS progression-free survival, PR partial response, SD stable disease. *Primary tumor types excluded from Checkmate 141 study

Full size image

In the present real-world study, nivolumab use was not limited to platinum-refractory but also platinum-sensitive patients. In a subgroup analysis by prior platinum responsiveness, the ORR was 16.2 (95% CI 10.3–23.6)% and 16.1 (95% CI 5.5–33.7) % in platinum-refractory patients and platinum-sensitive patients, respectively (Fig. 3a). The median PFS was 2.0 (95% CI 1.7–2.6) months in platinum-refractory and 2.8 (95% CI 1.7–6.1) months in platinum-sensitive (Fig. 3b). The median OS was 9.1 (95% CI 6.9–11.9) months and 9.1 (95% CI 6.2–NR) months for platinum-refractory and platinum-sensitive patients, respectively (Fig. 3c).

Fig. 3
figure 3

a Overall response rate b progression-free survival and c overall survival among patients stratified according to platinum-sensitive or platinum-refractory status. BOR best overall response, CR complete response, ORR objective response rate, OS overall survival, PD progressive disease, PFS progression-free survival, PR partial response, SD stable disease. *Log-rank test

Full size image

In a subgroup analysis by PS, the ORR, PFS, and OS were generally numerically greater in patients with ECOG PS 0 status, compared with patients with ECOG PS 1 or, in particular, ECOG PS ≥ 2 status (Fig. 4). The ORR was 17.9 (95% CI 11.3–26.2) %, 14.1 (95% CI 7.5–23.4) %, and 5.6 (95% CI 0.1–27.3) % in patients with ECOG PS 0, ECOG PS 1, and ECOG PS ≥ 2 status, respectively (Fig. 4a). PFS was 2.6 (95% CI 2.0–3.8) months, 2.1 (95% CI 1.7–2.8) months, and 1.4 (95% CI 0.8–2.2) months in patients with ECOG PS 0, ECOG PS 1, and ECOG PS ≥ 2 status, respectively (Fig. 4b). The median OS was not reached in patients with ECOG PS 0 status and was 6.9 (95% CI 4.9–10.4) months and 3.1 (95% CI 2.1–3.6) months in patients with ECOG PS 1 and ECOG PS ≥ 2 status, respectively (Fig. 4c).

Fig. 4
figure 4

a Overall response rate b progression-free survival and c overall survival among patients according to ECOG status. BOR best overall response, CR complete response, ECOG PS Eastern Cooperative Oncology Group performance status, NR not reached, ORR objective response rate, OS overall survival, PD progressive disease, PFS progression-free survival, PR partial response, SD stable disease. *Log-rank test

Full size image

Effectiveness by BOR at 3 months are shown in Fig. 5. Both PFS and OS were greater in patients with CR/PR or SD compared with patients who experienced progressive disease (Fig. 5a, b). Median PFS was not reached in patients with CR/PR and was 4.7 (95% CI 4.1–6.7) months and 1.6 (95% CI 1.4–1.6) months in patients with SD and PD, respectively (Fig. 5a). Median OS was not reached in patients with SD or CR/PR at 3 months and 6.2 (95% CI 4.6–7.8) months in patients with progressive disease (Fig. 5b).

Fig. 5
figure 5

a Progression-free survival and b overall survival according to the best overall response at 3 months. CR complete response, OS overall survival, PD progressive disease, PFS progression-free survival, PR partial response, SD stable disease. *Log-rank test

Full size image

Incidence of AEs

Thirty-eight patients (14.8%) developed immune-related AEs of any grade; 15 patients (5.9%) developed a grade ≥ 3 immune-related AE (Table 2). Endocrine disorders were the most common type of immune-related AEs which affected 14 patients, but only two of these events were grade ≥ 3. The most common type of grade ≥ 3 immune-related AEs were lung disorders (n = 6). One patient developed a grade 5 immune-related AE (interstitial pneumonia). The median time to onset of any immune-related AEs was 8.7 (0.1–43.7) weeks; each categorized immune-related AE widely appeared throughout the 12-month observation periods. Median time to resolution of most immune-related AEs was 3 to 6 weeks.

Table 2 Incidence of immune-related adverse events over 12 months
Full size table

Discussion

To the best of our knowledge, this is the largest study to evaluate the effectiveness and safety of nivolumab in patients with HNC in Japanese real-world clinical practice. Median PFS (2.1 vs 2.0 months), 12-month OS rate (43% vs 36%), and median OS (9.5 vs 7.5 months) in our study were similar to those in CheckMate 141 [1]. Furthermore, our data are also comparable to other reported real-world data in Japan [7, 8].

The current study addresses several data gaps between the clinical trial and real-world settings. First of all, the current study reveals the effectiveness of nivolumab in patients with primary tumors in the nasopharynx, maxillary sinus, or salivary gland, who would have been excluded from previous phase III clinical trials for HNC including the CheckMate 141, KEYNOTE-040, KEYNOTE-048, and EXTREME studies [9,10,11]. In particular, we found a patient with maxillary sinus cancer responded to nivolumab, which is the first positive result recorded for any checkpoint inhibitor monotherapies in such patients [12]. In addition, the effectiveness of nivolumab for nasopharyngeal and salivary gland tumors in the current study was consistent with the results in prior phase I/II trials [9, 13,14,15]. Historical data showed that prognosis is generally poor in patients with primary HNC in the hypopharynx, oral cavity, and nasal sinuses, whereas prognosis tends towards improvement in patients with nasopharyngeal or laryngeal primary tumors [10, 16,17,18,19]. It has been suggested that the primary HNC site is associated with poor response to chemotherapy [20]. By contrast, the current study showed that patients well responded to nivolumab irrespective of the expected prognosis.

The effectiveness of nivolumab for platinum-refractory HNC in this study was comparable to that of CheckMate 141. While most patients receiving nivolumab in real-world clinical practice have platinum-refractory disease, 18% of the patients in our study had platinum-sensitive disease, a population that was not included in CheckMate 141 [1]. In the present study, ORR and OS were numerically similar between patient populations with platinum-refractory and platinum-sensitive diseases. However, because the effectiveness of the platinum-sensitive disease has been evaluated only in a limited number of patients [9], further large-scale studies are warranted. Stratification of patients by performance status in this study showed that nivolumab was most effective, in terms of ORR, PFS, and OS, in patients with ECOG PS 0 status. Other studies utilizing anti-PD-1 inhibitors, including nivolumab, in various real-world settings including advanced melanoma, non-small cell lung cancer, and head and neck cancer have demonstrated that favorable PS is predictive of greater OS and PFS [21,22,23]. Recent longitudinal studies of anti-PD-1 inhibitors, also conducted in real-world clinical practice settings with median follow-up periods of up to 12.9 months, further verify better responses in patients with ECOG PS 0–1 status versus those with ECOG PS > 1 status [24,25,26]. Therefore, the present results provide additional evidence to support the notion that nivolumab and similar agents in this class are more effective in patients who are in good general condition at treatment initiation.

Superior PFS and OS were observed in patients whose BOR at 3 months was CR, PR, or SD. This is consistent with previous findings with nivolumab in patients with melanoma or non-small cell lung cancer [4, 5], and is also consistent with the known response profile to immune-checkpoint inhibitors. In addition, even after tumor progression, half of the patients with PD survived for more than 6 months. These patients may receive subsequent chemotherapy after discontinuation of the nivolumab therapy, which may contribute to prolonged survival because it has been reported for several cancers that chemotherapy after immunotherapy is highly effective [27,28,29,30]. Future studies with longer follow-up will reveal possibly different impacts of subsequent chemotherapy.

The incidence and timing of immune-related AEs in our study were similar to or even lower than previous reports, including CheckMate 141 [1, 31], and with Japanese clinical use of nivolumab in patients with HNC [8]. No new safety signals were identified in our study compared with CheckMate 141, despite the fact that our study included a more vulnerable patient population. These data indicate that nivolumab has a well-characterized and manageable safety profile in real-world clinical use.

Retrospective observational studies have limitations. As an observational design, there was no control group. Because completed by individual physicians during real-world clinical practice, patients’ medical records may not always contain complete and comparable information and may contain measurement errors. To recruit a large number of patients, we preferentially included centers that treat a high number of patients with HNC. This may have introduced some selection bias in the study population.

In conclusion, this retrospective observational analysis in a real-world clinical setting supports the effectiveness and safety of nivolumab in a range of Japanese patients with HNC. No new safety signals were identified, and the findings of randomized clinical trials with nivolumab are applicable to a real-world population of patients with a more diverse clinical profile in terms of performance status and primary site.