Clinical Outcomes in Patients With Metastatic Lung Cancer Treated With PD-1/PD-L1 Inhibitors and Thoracic Radiotherapy

Immune checkpoint inhibitors (CPIs) have become integral in the treatment of patients with advanced lung cancer but are associated with immune-related adverse events (IRAEs).^1,2 Many of these patients also receive thoracic radiotherapy (TRT). A better understanding of potential toxic effects when combining CPIs with TRT is imperative.

We performed a retrospective cohort study (2013-2016) of patients with metastatic lung cancer treated with PD-1/PD-L1 (programmed cell death 1/programmed cell death 1 ligand 1) inhibitors at a US academic medical center. The Massachusetts General Hospital institutional review board approved the study, waiving patient written informed consent. Primary outcomes were IRAEs, as specified in the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE), version 4.0—including pneumonitis—and survival, measured from the initiation of CPI therapy. Features for multivariate analysis were selected clinically, without bias from the univariate analysis.

Among the 164 patients included in the analysis (95% with non–small-cell lung cancer [n = 158], 5% with small-cell lung cancer [n = 6]), baseline characteristics were similar (Table 1) between patients who received TRT (n = 73) and those who did not (n = 91), except that fewer patients in the TRT cohort had adenocarcinoma (49% [n = 36] vs 75% [n = 68]; P = .001) and targetable mutations (EGFR/ALK/ROS1; 4% [n = 3] vs 16% [n = 15]; P = .01). Rates of grade 2 or higher IRAEs (13.7% [n = 10] vs 15.4% [n = 14]; P = .83), all-grade pneumonitis (8.2% [n = 6] vs 5.5% [n = 5]; P = .54), and grade 2 or higher pneumonitis (4.1% [n = 3] vs 3.3% [n = 3]; P > .99) were not significantly different between the TRT and non-TRT cohorts. Median TRT dose was similar between those patients who developed pneumonitis and those who did not (52.8 vs 50.4 Gy; P = .76). Most patients (57 of 73) were treated with a median TRT dose of 60 Gy (range, 44-79.1 Gy) before CPI initiation (median interval, 8.6 months; range, 0.1-69.0 months). None of the 16 patients who received TRT (median/range: 40 Gy, 8-54 Gy) between CPI cycles (n = 5), after CPI (n = 5), or more than 1 course of TRT (n = 6) developed symptomatic pneumonitis. Median overall survival was 12.1 months. Multivariate analysis (Table 2) demonstrated that all-cause mortality was significantly lower in patients with grade 2 or higher IRAEs (HR, 0.45; 95% CI, 0.22-0.93; P = .03) or in those treated with fewer chemotherapy lines (HR, 1.21; 95% CI, 1.05-1.40; P = .01). Receipt of TRT showed reduced all-cause mortality, although the reduction was not significant (HR, 0.66; 95% CI, 0.42-1.01; P = .06).

This study demonstrates that IRAEs, including pneumonitis, are not more common in patients with metastatic lung cancer who received both CPIs and TRT. The overall incidence of IRAEs in this series was comparable to rates observed in randomized trials of CPIs for metastatic lung cancer.^1,2 Development of grade 2 or higher IRAEs was associated with improved survival, possibly reflecting that patients responding to CPIs likely received more cycles of therapy, which may have in turn predisposed them to the toxic effects. Interestingly, response rates to CPIs in melanoma are significantly higher among patients with treatment-related adverse events, even after adjusting for CPI doses,³ suggesting that augmentation of the antitumor response may be coupled to anti-self activity.

The trend toward increased overall survival with TRT despite significantly fewer patients harboring targetable mutations is intriguing. It has been postulated that lack of antigenic mutations and other mechanisms of immune evasion may underlie resistance to CPIs.⁴ In certain situations, radiotherapy may potentiate the efficacy of immunotherapy, even restimulating a durable systemic response in disease that had become refractory to CPIs.⁵ A recent single-institution secondary analysis of KEYNOTE-001⁶ found that patients who received any radiotherapy prior to pembrolizumab had significantly longer progression-free and overall survival.⁶ While the few patients in that study who received TRT before pembrolizumab (n = 24; median interval, 11.5 months) had a higher incidence of low-grade pulmonary toxic effects, there was no difference in grade 3 or higher pneumonitis, consistent with our findings.

Limitations of this analysis are that we report a single-institution series, and median time between TRT and CPI was 8.6 months. Prospective study is needed, and studies investigating both consolidative CPI after definitive chemoradiotherapy (NCT02768558 and NCT02125461) and concurrent radiotherapy with CPI (NCT03035890 and NCT02599454) are ongoing. Nonetheless, pending prospective validation, our results suggest that TRT does not significantly increase the risk of symptomatic IRAEs, including pneumonitis, compared with CPIs alone.

Corresponding Author: Florence K. Keane, MD, Department of Radiation Oncology, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (fkeane@partners.org).

Accepted for Publication: September 6, 2017.

Published Online: September 27, 2017. doi:10.1001/jamaoncol.2017.3808

Author Contributions: Dr Keane had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: W. L. Hwang, Gainor, Keane.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: W. L. Hwang, Keane.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: W. L. Hwang, Niemierko, Keane.

Administrative, technical, or material support: W. L. Hwang, Hubbeling, Schapira, Keane.

Study supervision: Gainor, Keane.

Conflict of Interest Disclosures: Dr Gainor has served as a compensated consultant and/or received honoraria from Bristol-Myers Squibb, Merck, Novartis, Genentech/Roche, Loxo, Clovis, Boehringher Ingelheim, Incyte, and Theravance. No other disclosures are reported.

Meeting Presentation: This article was presented at the ASTRO 2017 Annual Meeting; September 27, 2017; San Diego, California.

Additional Contributions: We are grateful to Henning Willers, MD (Massachusetts General Hospital), for his invaluable contributions in the conception and design of the study as well as data interpretation and critical revision of the manuscript. We also thank Melin Khandekar, MD, PhD (Massachusetts General Hospital), and Noah C. Choi, MD (Massachusetts General Hospital), for their input and guidance. They were not compensated for their contributions.