Validation of the CLOVER Machine Learning Model to Predict Hidden Cancer in Adults with Venous Thromboembolism

What this Trial Is About

Researchers are evaluating the CLOVER score, a machine learning-based tool designed to predict the risk of hidden occult cancer in patients with venous thromboembolism VTE, which includes conditions like deep vein thrombosis and pulmonary embolism. The study combines past patient data from 2000 to 2022 and newly enrolled patients from December 2025 to December 2027 to validate how well the CLOVER score identifies those at increased cancer risk. It also assesses how satisfied clinicians are with the web-based CLOVER tool designed to support clinical use. The study includes two groups a retrospective cohort of adult patients with confirmed symptomatic VTE diagnosed in hospitals across Spain, excluding those from centers involved in developing the model, and a prospective cohort recruited consecutively from December 2025 to December 2027. For the prospective group, all participants will receive routine clinical evaluations, laboratory tests, chest X-rays, and cancer screenings based on guidelines. The CLOVER score will be calculated at diagnosis using the web tool, which classifies patients as low or high cancer risk. Participants will be followed for at least two years to identify any cancer diagnosis. Participants will undergo standard exams, lab tests, and imaging as part of routine care. Researchers will collect data from electronic health records for the retrospective group and monitor cancer outcomes over two years for the prospective group. The study measures how accurately the CLOVER score predicts occult cancer by evaluating sensitivity, specificity, and predictive values. Clinicians will complete an online survey to report their satisfaction with the CLOVER-Web tool. All data will be handled securely, with informed consent obtained from prospective participants, and privacy maintained throughout the study.

External Validation of the CLOVER Score for Detecting Occult Cancer in Venous Thromboembolism Patients