Abstract

Introduction Background: Cancer-associated thrombosis (CAT) is a major complication in oncology, often leading to increased morbidity and mortality. Current risk stratification models, such as the Khorana Risk Score (KRS), have shown limited predictive value, particularly in patients with metastatic disease and high risk tumours. D-dimer, a biomarker of coagulation activation, has been proposed as a potential tool for improving risk assessment. Objective: To evaluate the predictive performance of baseline (DD1) and one-month post-treatment (DD2) D-dimer levels for CAT in ambulatory cancer patients initiating systemic therapy, and to compare their performance against KRS and the Vienna CATScore (CATS). Methods This prospective cohort study included ambulatory patients with metastatic gastrointestinal, lung, or ovarian cancers at UC-CHRISTUS Cancer Center (Santiago, Chile). Patients were identified by oncologists prior to initiating systemic therapy. Exclusion criteria included prior anticoagulation. D-dimer levels were measured at baseline (DD1) and one month post-initiation (DD2) using the ELFA technique (Vidas®, Biomerieux). The primary outcome was CAT at six months. ROC curves assessed predictive performance; the Net Reclassification Index (NRI) evaluated the impact of adding KRS to DD1. Results A total of 67 patients were recruited. CAT incidence was 16% (n=11). DD1 (AUC 0.70, 95% CI 0.52–0.89) and DD2 (AUC 0.75, 95% CI 0.55–0.95) outperformed KRS (AUC 0.52, 95% CI 0.33–0.71) and CATS (AUC 0.67, 95% CI 0.48–0.86). The optimal DD1 cutoff (1062 ng/mL) showed 100% sensitivity and 46% specificity. The adjusted NRI (-30.9%) indicated that adding KRS to D-dimer did not improve predictive accuracy. Conclusion D-dimer demonstrated superior predictive value compared to KRS and CATS for CAT risk stratification in ambulatory cancer patients. Its high sensitivity and accessibility make it a promising standalone biomarker, potentially refining thromboprophylaxis strategies. Future studies should validate these findings and explore the integration of D-dimer into clinical decision-making to optimize thrombosis prevention in oncology.