High incidence and early onset of nivolumab-induced pneumonitis: four case reports and literature review

Nivolumab—an immune checkpoint inhibitor—is the humanized IgG4 monoclonal antibody that targets the programmed cell death 1 (PD-1) protein, which is an immunoinhibitory receptor expressed on T cells [1]. It promotes immune responses by blocking the binding of PD-1 and its ligands, PD-L1 and PD-L2 [1]. Nivolumab produces durable objective responses in patients with malignant tumors with a favorable safety profile, and is currently used for the treatment of patients with malignant melanoma (MM), non-small cell lung cancer (NSCLC), renal cell carcinoma, Hodgkin’s lymphoma, and head and neck cancer (HNC) [2‐7].

Considering the paradigm shift towards alternative therapeutic strategies for the treatment of the above-mentioned tumors, immune checkpoint inhibitors are known to cause a variety of unique toxicities that have not been well characterized. These effects termed as “immune-related adverse events (irAEs)” that include endocrine dysfunction, neurological disorder, hepatitis, nephritis, skin toxicity, cardiac insufficiency, colitis, and pneumonitis. Of these irAEs, pneumonitis is a relatively uncommon, but potentially life-threatening adverse event. A previous meta-analysis revealed that the incidence of pneumonitis associated with a monotherapy of PD-1 inhibitors, either nivolumab or another inhibitor, pembrolizumab, was 2.7% for all-grade and 0.8% for grade ≥ 3 [8]. Furthermore, this report showed that pneumonitis related to nivolumab monotherapy occurred in 4.1% (1.4–8.5%) of patients with NSCLC with all grades and 1.7% (0–3.4%) with ≥ grade 3 toxicity, while pneumonitis occurred in 1.5% (0–1.9%) of patients with MM at all grades and 0.1% (0–0.3%) with ≥ grade 3 toxicity. Although the prevalence of nivolumab-induced pneumonitis varies across tumor types, particular caution should be taken regarding this issue in NSCLC patients. A recent retrospective study demonstrated diverse computed tomography (CT) image patterns and clinical courses of nivolumab-induced pneumonitis, including predominant cryptogenic organizing pneumonia (COP) pattern, median onset time of 2.6 months, and successful treatment with corticosteroids for most patients [9].

Of 20 patients with malignant tumor who were treated with nivolumab from October 2014 to July 2017 at Tokyo Medical University Hachioji Medical Center, we encountered four cases of nivolumab-induced pneumonitis: two patients with NSCLC, one patient each with MM and HNC. Interestingly, these patients showed the different clinical characteristics of pneumonitis from those reported previously. In our cases, nivolumab-induced pneumonitis was identified as earlier onset and higher incidence, and affected patients had a history of heavy anti-tumor treatment [8‐11]. In order to renew clinicians’ attention regarding pneumonitis, this report presented CT image findings and clinical course of pneumonitis following nivolumab monotherapy in four cases involving NSCLC, MM, and HNC, respectively.

Treatment with nivolumab induces a variety of adverse events, including irAEs which can sometimes be serious or fatal, albeit infrequent. Among them, pneumonitis is one of the most life-threatening adverse events. Four patients discussed in the present report were diagnosed with nivolumab-induced pneumonitis, mainly based on the image findings and clinical courses. All four patients had no pre-existing interstitial pneumonia as a potential risk factor for drug-induced pneumonitis. Furthermore, in statistical analyses (data not shown), there were no significant differences in clinicopathological characteristics between patients with and without nivolumab-induced pneumonitis.

In contrast, these four patients had some characteristics that were different from patients with nivolumab-induced pneumonitis in the previously published reports (Table 1). First, the meta-analysis performed by Nishino et al. results revealed that the incidence of the pneumonitis was 2.7% for nivolumab monotherapy and pneumonitis-related death was relatively rare [8]. Another study reported that the incidence of pneumonitis associated with nivolumab monotherapy was 2.9% and 11.8% for a combination therapy with nivolumab and other immune checkpoint inhibitors [9]. Meanwhile, four (20%) out of 20 patients, who underwent nivolumab treatment in our institute developed pneumonitis (22% for NSCLC, 17% for MM, and 25% for HNC); although the number of cases included in this study was small, a strikingly larger proportion of patients developed nivolumab-induced pneumonitis in our study than that reported previously. In a previous study on Japanese patients, male gender and smoking history were suggested to be potential risk factors for nivolumab-related pneumonitis [11]. In our cases, all the patients who developed pneumonitis were Japanese males. As the previous study suggests higher incidence of drug-induced pneumonitis in Japanese patients, the ethnicity also may impact on nivolumab-induced pneumonitis [12]. Another study reported that all lung cancer patients were administered 3–5 lines of preceding systemic therapy till the onset of nivolumab-induced pneumonitis, although no significant differences in trends regarding gender were observed [9]. In our institute, all patients were also heavily treated with pharmaceutical or radiation therapy before nivolumab treatment, and it is possible that accumulated treatment histories in male patients may be associated with higher incidence of pneumonitis. Second, the median time from initiation of nivolumab treatment to onset of pneumonitis in our cases was 15 days (12.5 days in NSCLC, 17 days in MM, and 18 days in HNC), whereas the median time to onset of pneumonitis in previous reports was 1.2 months in NSCLC patients, 3.6 months in MM patients, and 40 days in Japanese patients with NSCLC [9‐11]. Furthermore, Naidoo et al. reported that time to onset of pneumonitis induced by anti-PD-1/PD-L1 antibodies ranged from 9 days to 19.2 months [13]. The reason for the earlier onset in our cases remains unknown. Pneumonitis in our report was promptly ameliorated in all patients, who were non-fatal. In this context, the early onset in our cases may be a result of an early detection of the patients’ symptoms, although it must be mentioned that this study comprised of only four cases.

Importantly, the pneumonitis induced by one drug can present various disease and image patterns, and conversely, numerous drugs have been reported to develop the pneumonitis with a similar pattern. Furthermore, previous retrospective studies and meta-analyses of nivolumab-induced pneumonitis have also shown a similarly wide range of image patterns, with COP pattern being the most common [8, 9]. Interstitial lung disease with COP and NSIP patterns commonly has a better response to corticosteroids than those with usual interstitial pneumonia (UIP)/idiopathic pulmonary fibrosis (IPF) pattern [14, 15]. Image findings of COP and NSIP patterns may reflect a favorable outcome in the pneumonitis of our cases. In fact, many cases of nivolumab-induced pneumonitis were also responsive to corticosteroids, which may be attributed to predominant COP pattern in these cases. However, nivolumab-induced pneumonitis with COP or NSIP pattern have also been shown to sometimes result in diffuse alveolar damage, further leading to lethality [11, 13, 16].

Although currently no optimal management guideline exists for the management of pneumonitis, considering the widespread use of immune checkpoint inhibitors, it is necessary to take particular caution to avoid or manage drug-related pneumonitis. Our case series showed that nivolumab had a high incidence of drug-induced pneumonitis with early onset, supporting the need for renewed attention to nivolumab-induced pneumonitis, particularly in patients with a history of heavy anti-tumor treatment.