Scoring systems to predict thrombotic complications in solid tumor patients

To explore the use of large datasets in predicting and managing cancer-associated venous thromboembolism (CAT) by stratifying patients into risk groups. This includes evaluating current predictive models and identifying potential improvements to enhance clinical decision-making.

Cancer patients are at an elevated risk of developing venous thromboembolism (VTE), which significantly impacts mortality and quality of life. Traditional approaches to risk assessment fail to account for the procoagulant changes associated with cancer, making individualized risk prediction a challenge. Current clinical guidelines as per ASCO recommend risk assessment before chemotherapy and endorse thromboprophylaxis as a standard preventive measure. Since any cancer population is highly heterogeneous in terms of VTE risk, predicting the risk of CAT is an oncological challenge. To address this, different predictive models have been developed to stratify patients by risk, enabling targeted thromboprophylaxis. However, these models vary in accuracy and utility. The present review discusses the pros and cons of these different models.

The review examines existing CAT risk prediction models, highlighting their strengths, limitations, and diagnostic performance. It also identifies additional variables that could enhance these models to improve their effectiveness in guiding clinicians toward better risk stratification and treatment decisions for cancer patients.