Venous thromboembolism incidence and risk assessment in lung cancer patients treated with immune checkpoint inhibitors

Background There are scarce data on venous thromboembolism (VTE) rates among non-small cell lung cancer (NSCLC) patients treated with immune-checkpoint inhibitors (ICI). The Khorana Score (KS), used to guide thromboprophylaxis in cancer patients, was validated in patients receiving chemotherapy. Objective To assess VTE rates and KS performance among NSCLC patients treated with ICI or chemotherapy. Methods We performed a retrospective cohort study of NSCLC patients starting either ICI or platinum-based chemotherapy. The 6-month cumulative incidence of VTE in the ICI and chemotherapy cohorts and hazard ratios (HR) with 95% confidence intervals (CI) were calculated, using death as a competing risk. Subgroup analysis of low (0-1) and high (>= 2) KS risk groups was performed. Results The study included 345 NSCLC patients receiving single agent ICI (n = 176) or chemotherapy (n = 169). The 6-month cumulative incidence of VTE was 7.1% in the chemotherapy cohort and 4.5% in the ICI cohort (HR for chemotherapy = 1.6, 95% CI 0.66-3.9). Among chemotherapy treated patients, the high-risk KS group had a trend toward a higher VTE incidence, compared with patients with a low-risk KS (HR 3.04, 95% CI 0.82-11.22). Among ICI-treated patients, the high-risk KS group had a trend toward a lower VTE incidence compared with the low-risk group (HR 0.17, 95% CI 0.02-1.36). Conclusions VTE rates were higher among NSCLC patients treated with platinum-based chemotherapy than those treated with ICI alone, though the precision of the relative estimate is low. The KS did not identify high-risk ICI-treated patients, suggesting that an ICI-specific risk model is warranted.

Automated summary

VTE rates were higher among NSCLC patients treated with platinum-based chemotherapy compared to those treated with ICI. The Khorana Score did not effectively identify high-risk ICI-treated patients, suggesting the need for an ICI-specific risk model.