Anti-PD1–Induced Pneumonitis: Capturing the Hidden Enemy

In this issue of Clinical Cancer Research, Nishino and colleagues describe the different radiologic patterns found on CT scan images in patients who developed pneumonitis during anti-programmed death receptor 1 (PD1) treatment and their different clinical outcomes (1).

Pneumonitis is a rare immune-related adverse event observed in <5% of patients receiving anti-PD1 agents. Although the prevalence seems to be lower, there have also been observed cases of pneumonitis in patients receiving anti-PDL1 agents (2). In a recent meta-analysis assessing anti-PD1 agents in different histologies, Nishino and colleagues found that the prevalence of all grade pneumonitis was 2.7% (with 0.8% prevalance of grade ≥3 pneumonitis; ref. 3). According to published data, pneumonitis has led to death in some patients in different clinical trials (4–7). For this reason, it is of vital importance to detect pneumonitis as early as possible, so that appropriate treatment can be administered promptly.

Anti-PD1 pneumonitis may appear in patients more prone to developing lung injury due to different etiologies, such as oncologic lung conditions (lung tumors, carcinomatous lymphangitis, smoking exposure), previous radiotherapy (mediastinal radiotherapy), previous treatments (mTOR inhibitors, tyrosine kinase inhibitors), or previous primary lung diseases (interstitial pathologies, chronic obstructive pulmonary disease, infectious pathogens). This hypothesis is based on the fact that the pulmonary damage present in the aforementioned pathologies may trigger host immune response when exposed to anti-PD1 therapies.

One of the main problems in diagnosing pneumonitis is that symptoms may be rather subtle. Cough, mild fever, chest pain, or shortness of breath may be the only clinical findings of an underlying pneumonitis. These symptoms are very unspecific, and, therefore, a potential infection or disease progression should be considered as a priority, although drug-induced pneumonitis should be systematically ruled out early as well, considering the severity of its outcome. Thus, patients and general practitioners should be widely informed that in case of pulmonary symptoms, a thorough examination should be carried out, including imaging and pulmonary function tests.

CT scan is a useful technique to diagnose anti-PD1–induced pneumonitis. Some findings, such as multiple and extensive consolidations or ground-glass opacities, may help us to establish a pneumonitis diagnosis (ref. 8; Fig. 1). In this issue of Clinical Cancer Research, Nishino and colleagues have classified CT scan images from patients with anti-PD1 pneumonitis according to the American Thoracic Society/European Respiratory Society (ATS/ERS) classification of interstitial pneumonia and correlated them with patients' outcomes (9). They show that there may be a relation between the CT scan patterns and the pneumonitis severity. This relation is in line with the results presented by Naidoo and colleagues at the 2015 European Cancer Congress (10). Although further research should be done, CT scan images may help us not only to diagnose but also to classify pneumonitis cases and to guide our therapeutic approach. Bronchoalveolar lavage (BAL) may be required to distinguish pneumonitis from other pathologies, such as opportunistic infections or disease progression. Nevertheless, findings of BAL pointing to pneumonitis may be scarce, and the usefulness of this test mainly results in its ability to exclude differential diagnoses (e.g., presence of erythrocytes may suggest alveolar hemorrhage, infectious organisms may point to a lower respiratory infection, and malignant cells may lead to a diagnosis of disease progression). A deeper review on BAL findings is summarized in the ATS clinical practice guideline for interstitial lung disease (11). Lung biopsy may be needed in cases in which both CT scan images and BAL are inconclusive. It would be interesting to validate the findings reported by Nishino and colleagues in this issue, as both BAL and lung biopsy are more invasive, and there is no consensus about when to perform them. In many cases, BAL and biopsy results may be deficient, but if pneumonitis is highly probable based on clinical suspicion and CT scan findings, especially if patients have a recent history of anti-PD1 therapy exposure, treatment should not be delayed.

Regarding treatment, corticosteroids remain the cornerstone for immune-related adverse events (12). Antibiotic prophylaxis should be contemplated, as a high dose of corticosteroids may be used for a prolonged time, and the risk of opportunistic infections may be high. The lack of a consensus in treatment of pneumonitis makes it difficult to establish a clear algorithm. Generally, grade 1 pneumonitis may be treated with anti-PD1 withholding and image control in 2 to 3 weeks. Regarding grade ≥2 pneumonitis, anti-PD1 should be discontinued, and the patient should be monitored very closely. In case of grade 2, corticosteroids are recommended, and patients may be treated without hospitalization. Nevertheless, strong instructions should be given to patients to detect any sign of worsening, such as increase of pulmonary symptoms, fever, or aggravation of dyspnea. In the same way, a close follow-up every 2 to 3 days should be carried out to avoid any possible complication. In case of grade 3 to 4 pneumonitis, hospitalization is strongly recommended due to its high risk of fatal event development. After ruling out a possible infection, intravenous corticosteroids should be given. Generally, if a benefit is not observed within 72 hours, use of other immunosuppressors, such as infliximab or mycophenolate, may be considered. It is important to uncover any possible underlying infection or previous exposure to hepatitis B and C virus as well as tuberculosis before starting these immunosuppressors to avoid unexpected infectious reactivations. In case of previous exposure to those pathogens, advice from infectious teams should be asked to establish the real risk of reactivation and to evaluate starting antiviral or antituberculosis prophylaxis before immunosuppressor administration.

As reported by Nishino and colleagues in this issue, some patients had to be treated with infliximab (1). Interestingly, one of these patients presented a radiologic pattern compatible with cryptogenic organizing pneumonia and the other two an acute interstitial pneumonia/acute respiratory syndrome pattern. These data might suggest that those specific patterns may be related to corticosteroid-refractory pneumonitis, and, thus, an early, aggressive therapeutic approach should be encouraged.

Although Nishino and colleagues insist these findings should be taken cautiously due to the small sample size as well as the retrospective nature of the study (1), ATS/ERS classification may complement the Common Terminology Criteria for Adverse Events (CTCAE) scoring system to distinguish different grades of pneumonitis. Although the current article is based on pneumonitis that developed in patients receiving anti-PD1 agents, it would be interesting to know whether patients treated with anti-PDL1 agents present different radiologic patterns in comparison with anti-PD1 agents.

The lack of standardized treatment guidelines as well as the low prevalence of anti-PD1–related pneumonitis makes its diagnosis and therapeutic approach a real challenge. Therefore, multidisciplinary teams should be encouraged to create expert networks that may handle both frequent and especially rare immune-related adverse events. As an illustration of such initiative, an interactive mobile application has been developed at the Gustave Roussy Cancer Center (Villejuif, France) based on a network of organ specialists. Moreover, a French national register, referred to as “RESIAMIC” (Registre des Effets Indésirables Sévères des Anticorps Monoclonaux Immunomodulateurs en Cancérologie), is aimed at collecting all the immune-related severe events to get a perspective of the real prevalence of these events and to facilitate their therapeutic approach. Once a CTCAE grade ≥2 immune-related event is observed, practitioners are asked to contact the RESIAMIC database through a Web-based platform. Monthly, a Web conference is performed for doctors outside the Gustave Roussy Cancer Center who would like to discuss their cases with reference specialists. The development of such networking will allow reaching the sufficient sample size to perform robust future retrospective studies on rare but severe events, such as anti-PD1–induced pneumonitis. This will also allow validating in independent datasets the radiologic patterns reported by Nishino and colleagues (1) and, ultimately, developing a robust radiologic classification as a bedside tool for clinicians in pneumonitis diagnosis and management.

No potential conflicts of interest were disclosed.