Risk of CNS adverse events (CNS-AEs) for patients with non-small cell lung cancer (NSCLC) and melanoma brain metastases (BM) treated with CNS radiation (CNS-RT) and immune checkpoint inhibitors (CPIs).

Abstract

2010 Background: CPIs are widely used in the treatment of both metastatic melanoma and NSCLC. BM frequently occur and are treated with CNS-RT. Since both CNS-RT and CPIs can cause neuro-inflammation, we tested the hypothesis that concomitant treatment with CPIs and CNS-RT results in an increased risk of CNS-AEs. Methods: We identified patients with melanoma and NSCLC with BM treated with CNS-RT and seen at our institution between 2014 and 2016. Concomitant treatment with CPIs and CNS-RT was defined as administration of CPIs within 3 months before or after CNS-RT. CNS-AEs were defined as new or worsening edema on brain MRI without disease progression, new or worsening neurological deficit, or need to start or increase corticosteroids. A generalized linear model incorporated significant variables from a univariate analysis to model the incidence of CNS-AEs. Variables considered included the use of CPIs within 3 months of CNS-RT, cancer type, type of CNS-RT (gamma knife [GKRS] versus whole brain radiation therapy [WBRT]), number of metastases, and maximum metastasis size. Results: We identified 213 cases of CNS-RT (NSCLC 167 [78%], GKRS 147 [69%], WBRT 63 [30%], median 2 BM [1 to > 20], median 17 mm max diameter [2 mm-74 mm]). Patients were 52% female with median age 61 (range 21-87), and ECOG 0-2 in 93% at time of CNS-RT. CNS-AEs occurred in 40 (19%) cases. Receipt of CPIs within 3 months of CNS-RT was the only factor associated with an increased risk of CNS-AEs (odds ratio 3.9, 95% CI 1.6-9.2, p-value 0.002). The rates of CNS-AEs were 11 of 28 (39%) in cases which received CPIs within 3 months of CNS-RT and 29 of 184 (16%) in cases which did not. The characteristics of the 11 cases with CPI exposure and CNS-AEs were: 73% underwent GKRS, 45% were NSCLC, 18% received CTLA4 alone, 55% PD-(L)1 alone, 27% combined CTLA4/PD-1, and 55% had a neuro deficit as part of their CNS-AE. Conclusions: This retrospective analysis demonstrates that the use of CPIs within 3 months of CNS-RT is associated with an increased risk of CNS-AEs. CNS-RT modality, cancer type, and metastasis size or number were not associated with an increased risk of CNS-AEs.