Racial disparities in the use of programmed death-1 checkpoint inhibitors.

Abstract

3068 Background: There are concerns about racial disparities in access to trials of new cancer drugs, including the programmed death 1 checkpoint inhibitors (anti-PD1s). It is unknown whether these disparities extend to anti-PD1 treated patients in real-world practice. Methods: We used retrospective data from Flatiron Health’s electronic health record database, which includes more than 250 cancer clinics and 1.5 million patients with cancer. We identified patients diagnosed after January 1, 2011 who underwent systemic therapy for: advanced non-small cell lung cancer (aNSCLC; n = 13,473), metastatic renal cell carcinoma (mRCC; n = 1,537), and advanced melanoma (n = 1,221). Within each cohort, we identified treatment type (anti-PD1 versus non-anti-PD1). Therapy lines containing study drugs were excluded. We used logistic regressions to model the use of anti-PD1s by race, adjusting for factors such as age, sex, stage at diagnosis and line of therapy. Results: Of 16,231 patients in our sample, 4,643 (28.6%) were treated with anti-PD1s. Racial distributions differed for anti-PD1 treated patients compared to non-anti-PD1 treated patients in the aNSCLC cohort (Table: p < 0.01), but not in the mRCC cohort (p = 0.84) or the advanced melanoma cohort (p = 0.96). In bivariate analyses of patients with aNSCLC, anti-PD1 treatment was associated with other race, male sex, stage II at diagnosis, squamous histology, smoking history and line of therapy (all p < 0.05). Adjusted models showed there were no significant differences in likelihood of receiving anti-PD1s when comparing black and white patients undergoing systemic therapy for aNSCLC (aOR for black vs. white: 0.86, 95% CI 0.72-1.02), mRCC (aOR 0.90, 95% CI 0.53-1.49), or melanoma (aOR 2.02, 95% CI 0.42-14.59). Conclusions: Among patients undergoing systemic therapy in a large national network of cancer clinics, we found no significant racial disparities in the use of anti-PD1s. [Table: see text]