Development and validation of a novel bleeding risk score during direct oral anticoagulant therapy in patients with cancer-associated venous thromboembolism

Background Patients with cancer-associated venous thromboembolism (VTE) are at high risk of major bleeding during anticoagulant therapy, underscoring the need for accurate bleeding risk assessment in the era of direct oral anticoagulants (DOACs). However, few bleeding risk scores have been specifically developed and validated in this population. Purpose We developed and validated a novel bleeding risk score, the ONCO-DOAC BLEED score, for patients with cancer-associated VTE receiving DOAC therapy. Methods We derived a bleeding risk prediction model from 1,166 patients with cancer-associated VTE treated with DOACs in the COMMAND VTE Registry-2. Independent predictors were identified using Cox proportional hazards models and incorporated into a point-based score. External validation was performed in an independent cohort of 779 patients from the ONCO-DVT and ONCO-PE studies. Discriminatory performance was assessed using the area under the receiver-operating characteristic curve (AUC). Results During a median follow-up of 163 days, 127 patients (10.9%) experienced major bleeding in the derivation cohort. Multivariable analysis identified the following independent predictors: history of major bleeding, chronic kidney disease, nonsteroidal anti-inflammatory drug use, distant metastatic cancer, terminal cancer, upper gastrointestinal cancer, pancreatic cancer, and uterine cancer. Based on these risk factors, we developed the ONCO-DOAC BLEED score, assigning 1 point to each variable that was independently associated with major bleeding. Patients were stratified into three risk categories: low risk (0 points), intermediate risk (1 point), and high risk (≥2 points) (Table 1). The ONCO-DOAC BLEED score demonstrated good discrimination for major bleeding in the derivation cohort (Figure 1A). In the derivation cohort, the AUC was 0.64 (95% CI, 0.58-0.69) for the ONCO-DOAC BLEED score, compared with 0.55 (95% CI, 0.49-0.60) for VTE-BLEED, 0.56 (95% CI, 0.51-0.61) for RIETE, 0.59 (95% CI, 0.54-0.64) for CAT-BLEED, and 0.53 (95% CI, 0.47-0.58) for Perform score. In the validation cohort, 53 patients (6.3%) experienced major bleeding during DOAC therapy. A stepwise increase in the cumulative incidence of major bleeding was observed with increasing risk category (Figure 1B). The AUC was 0.62 (95% CI, 0.53-0.70) for the ONCO-DOAC BLEED score, compared with 0.61 (95% CI, 0.53-0.69) for VTE-BLEED, 0.61 (95% CI, 0.54-0.70) for RIETE, 0.55 (95% CI, 0.47-0.63) for CAT-BLEED, and 0.54 (95% CI, 0.46-0.61) for Perform score, suggesting improved risk stratification compared with existing bleeding risk scores. Conclusions The ONCO-DOAC BLEED score enables clinically meaningful stratification of major bleeding risk in patients with cancer-associated VTE receiving DOAC therapy and may serve as a useful tool to support therapeutic decision-making in clinical practice.