Response to immune checkpoint inhibition as monotherapy or in combination with chemotherapy in metastatic ROS1-rearranged lung cancers

9049 Background: ROS1 fusions are oncogenic drivers in various cancers types, including 1-3% of non-small cell lung cancers (NSCLCs). Immunotherapy approvals for NSCLC include ROS1-rearranged carcinomas, but the activity of immune checkpoint inhibition (ICI) as monotherapy or in combination with chemotherapy (chemo-ICI) therapy, as well as the immunophenotypic characteristics of these tumors, have not been described in a large data set. Methods: In this multi-institutional study, patients with ROS1-rearranged NSCLC were identified retrospectively. Tumor PD-L1 expression and tumor mutational burden (TMB) were assessed as part of routine clinical care. In patients who received ICI monotherapy or chemo-ICI in the metastatic setting, time to treatment discontinuation (TTD) and objective response rate (ORR; RECIST v. 1.1) were calculated. TTD was assessed with Kaplan-Meier methods; patients remaining on treatment were censored at last follow up. Results: 184 patients with ROS1-rearranged NSCLC were identified. Among 146 PD-L1 evaluable cases, PD-L1 expression was < 1% in 60 (41%), 1-49% in 35 (24%) and ≥50% in 51 (35%) tumors. Ninety-two of 100 (92%) TMB-evaluable tumors had < 10 mutations/megabase (mut/Mb). TMB was significantly lower for ROS1-rearranged NSCLCs (n = 97) vs. ROS1-wild type tumors (n = 5,380) evaluated with next-generation sequencing using MSK-IMPACT (median 2.6 vs. 5.9 mut/Mb, p < 0.001). Twenty-eight patients received ICI monotherapy and 11 patients received chemo-ICI. The median TTD was 2.1 months (95% CI: 1.0-4.2; n = 28) for single-agent ICI therapy and 10 months (95% CI: 4.7-14.1; n = 11) for chemo-ICI therapy. The ORR was 13% (2/16 RECIST-evaluable; 95% CI: 2-38%) for ICI monotherapy and 83% (5/6 RECIST-evaluable; 95% CI: 36-100%) for chemo-ICI therapy. There was no difference in PD-L1 tumor expression (p = 0.9) or TMB (p = 0.8) between responders and non-responders and no correlation between PD-L1 tumor expression (rho = 0.16, p = 0.6) or TMB (rho = 0.03, p = 0.9) and maximum change in sum of target lesions. Conclusions: Most ROS1-rearranged NSCLCs have low or no PD-L1 expression and low TMB. The activity of checkpoint inhibitor monotherapy is disappointing in ROS1-driven NSCLC. In contrast, combination chemoimmunotherapy can achieve clinically meaningful activity.