Rate and Risk Factors of Recurrent Immune Checkpoint Inhibitor-Related Pneumonitis in Patients With Lung Cancer

Haitao Tao; Fangfang Li; Dongxiao Wu; Shiyu Ji; Qingyan Liu; Lijie Wang; Bo Liu; Francesco Facchinetti; Tracy L. Leong; Francesco Passiglia; Yi Hu


Transl Lung Cancer Res. 2022;11(3):381-392. 

In This Article


Patient Selection

This retrospective study was conducted on patients who were diagnosed with lung cancer and received ICIs treatment in the First Medical Center of Chinese People's Liberation Army (PLA) General Hospital between January 2016 and January 2021. Patients who were ultimately diagnosed with CIP were included in analysis.

The inclusion criteria included patients who were pathologically diagnosed with locally advanced/advanced lung cancer (including NSCLC and SCLC); patients who received at least 1 cycle of PD-1/PDL-1 inhibitors and/or CTLA-4 inhibitors, whether monotherapy or combined with other drugs; and patients who developed typical clinical features and/or new typical imaging changes after ICIs treatment and were ultimately diagnosed with CIP after evaluation by a multidisciplinary team. The exclusion criteria included patients with other lung diseases with a clear alternative etiology, such as carcinomatous lymphangitis or active lung infection.

For patients who received corticosteroids for CIP treatment, a cumulative prednisolone equivalent dose was calculated. Since equivalent prednisolone 15 mg/day is the maintenance dose commonly used in rheumatic diseases,[19] this time point is marked as a key node of corticosteroid therapy. CIP-R was defined as recurrent CIP after initial CIP improved with regard to both chest CT and clinical symptoms after proper corticosteroid treatment according to NCCN guideline and the dose of corticosteroid tapered to less than equivalent prednisolone 15 mg/day (methylprednisolone 12 mg), regardless of complete discontinuation or rechallenge with immunotherapy lung.

Data Collection

For all the patients, detailed clinical data were retrospectively collected from medical records, including demographic characteristics, histologic subtype, medical history, molecular pathology and PD-L1 expression status, previous and concurrent cancer treatments, duration of ICIs therapy, antitumor efficacy of ICIs, clinical manifestations and outcomes of CIP, rechallenge with ICIs and recurrence of CIP. A chest CT scan obtained 3 months before the initiation of ICIs treatment was recorded as baseline. Baseline chest CT scan findings, including the presence of emphysema and ILA, were also recorded.

ILA were defined by at least one of the following CT findings in no less than two lobes of the lung: ground glass appearance (GGA), reticular opacity, honeycombing or traction bronchiectasis, diffuse centrilobular nodularity, or nonemphysematous cysts on CT images.[20,21] Patients with a history of drug-induced pneumonitis or those with radiation-related fibrotic lesions alone in the lung were excluded from the ILA group. Composite obstructive lung disease includes chronic obstructive pulmonary disease (COPD) by history, obstruction on spirometry, or emphysema on baseline CT images.[22] CT images were evaluated by two pulmonologists who were unaware of patient outcomes by a sequential reading method. Pneumonitis grading was evaluated by the treating investigators according to the Common Toxicity Criteria for Adverse Events (CTCAE version 4.0).

Follow-up was up to August 15, 2021, allowing sufficient time to observe pneumonitis recurrence.

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by ethics board of Chinese PLA General Hospital (No. S2021-562-01). Individual consent for this retrospective analysis was waived.

Statistical Analysis

Time to the onset of CIP was defined as the time from the first dose of an ICIs to the first occurrence of CIP-related symptoms or imaging findings in asymptomatic patients. The time to CIP-R was defined as from the time of initial CIP to the time of CIP-R. Continuous variables are described as the median [interquartile range (IQR)] and were compared using t-tests or the Mann-Whitney U test. Categorical variables were compared using the chi-square test or Fisher's exact test. Univariate and multivariate logistic regression analyses were performed on risk factors for the odds of CIP-R, including patient demographics (e.g., sex), medical histories (e.g., ILA), cancer-related factors (e.g., tumor histology), CIP-related factors (e.g., CTCAE grade). The multivariate analysis included predictors associated with CIP risk at the 0.15 significance level from the univariate analysis. Six patients who did not receive corticosteroids were excluded from the analysis of risk factors.

Kaplan-Meier estimates for median progression-free survival (PFS) and overall survival (OS) after ICIs initiation were produced for patients; log-rank and Wilcoxon tests were conducted to assess differences. The hazard ratio (HR) was estimated by a Cox regression model. All reported p values were two sided, and a significant difference was considered at the 5% level. Statistical analyses were conducted using SPSS Statistics for Windows (Version 25.0; IBM), and the figures were made by GraphPad Prism 9 (GraphPad Software Inc.).