PB0918 - Comparison of the Khorana, PROTECHT and ONKOTEV risk assessment models to predict venous thromboembolism in patients with pancreatic cancer

Background: Several risk assessment models (RAMs) have been developed to identify cancer patients at high-risk of venous thromboembolism (VTE) when undergoing chemotherapy who may benefit from thromboprophylaxis. The clinical usefulness of these RAMs remains debated in pancreatic cancer (PC) patients.

Aims: To analyze the discriminatory performance of the Khorana, PROTECHT, and ONKOTEV RAMs for VTE in patients with newly diagnosed PC undergoing ambulatory chemotherapy enrolled in the BACAP prospective multicenter cohort study (2014-2020).

Methods: The 6-month cumulative risk of VTE was estimated using death as a competing event. Sub-distribution hazard ratios (SHR) for dichotomized RAMs were estimated using the Fine and Gray competing risks regression model to assess VTE risk difference between low/intermediate- and high-risk patients. The time-dependent concordance index (C-index) was calculated to evaluate the RAMs discriminatory performance, with confidence intervals (CI) calculated by repeating analyses in 250 bootstrap samples.

Results: 762 PC patients were enrolled prior to chemotherapy, of whom 125 (16.4%) developed VTE during a 6-month follow-up period. The rate of PC patients predicted at “high-risk of VTE” was 28.1% using the Khorana score, 99.0% using the PROTECHT score and 34.1% using the ONKOTEV score. Whichever score was used, the 6-month cumulative incidence of VTE did not differ between patients predicted at “high-risk of VTE” and those predicted at “low/intermediate risk of VTE” (Table 1). The RAM C-index value was 0.54 (95%CI 0.49-0.59) for the Khorana score, 0.53 (95%CI 0.49-0.59) for the PROTECHT score and 0.56 (95%CI 0.50-0.61) for the ONKOTEV score.

Conclusion(s): In the BACAP cohort, the Khorana, PROTECHT and ONKOTEV RAMs failed to accurately predict VTE in patients with newly diagnosed PC receiving chemotherapy. Effective RAM are needed to predict VTE risk in PC patients, meanwhile primary pharmacological prophylaxis of VTE in patients with locally advanced or metastatic pancreatic cancer receiving chemotherapy and having a low bleeding should be implemented.