Cancer-associated Thrombosis clinical trials

Development and Validation of a Risk Prediction Model for Venous Thromboembolism in Gynecological Cancer Patients Undergoing Systemic Antineoplastic Treatment: The Gynecological Cancer Associated Thrombosis (GynCAT) Study

Cancer patients are burdened by an increased risk of venous thromboembolism (VTE), which has a significant impact on morbidity and mortality. Existing Risk Prediction Models (RPMs), including the widely accepted Khorana Risk Score (KRS), have some limitations when used in certain tumor site populations, such as gynecological cancers. Notably, gynecological patients exhibit a variable risk of VTE based on their specific tumor sites, with ovarian cancer representing the highest risk. Moreover, currently available RPMs lack validation in a broad gynecological population and may fail to effectively stratify VTE risk. GynCAT is a prospective cohort study that will be conducted on female patients with gynecologic malignancies scheduled for systemic antineoplastic treatment. During the screening phase, symptomatic VTE will be excluded, and KRS will be assessed. Pharmacological thromboprophylaxis will be considered and prescribed at clinical judgement, for patients with a KRS score of 3 or higher. Clinical, hematological, biochemical, coagulation, and genetic variables will be collected. Follow-up will last for the entire duration of the antineoplastic treatment line, and VTE events, bleeding events, and mortality will be recorded. The primary objective is the development and validation of an RPM for VTE in gynecologic cancer patients undergoing systemic antineoplastic treatment. Secondary objectives are evaluation of the predictive value of the identified model, comparing it with existing general oncology RPMs; assessment of its performance in predicting mortality; evaluation of VTE incidence in patients with KRS≥3 receiving thromboprophylaxis; identification of risk factors for bleeding in this patient population. The sample size calculation is based on an estimated VTE incidence of 5% over a mean follow-up of 12 months. Hence, a sample size of at least 1,200 patients in the derivation cohort is considered sufficient for the determination of a risk prediction model incorporating up to six predictor variables. A split-sample method will be used, with two-thirds of the study participants randomly assigned to the model derivation cohort (n=1,200) and one-third (n=600) to an independent validation cohort. The total number of patients recruited in the study will thus be of 1,800. A competing risk survival analysis with Fine & Gray model will be used to study the association between prognostic variables and VTE occurrence, considering death as a competitive risk. The RPM will be identified through a bootstrap approach to reduce the risk of overfitting. Discrimination power of the RPM will be assessed using time-dependent Receiving Operating Characteristic curve, and model calibration will be evaluated graphically and with the calculation of relative calibration slopes. In conclusion, this prospective cohort study aims to overcome the limitations of current RPMs in gynecologic cancer patients, improving the accuracy of VTE risk stratification in this population.