Abstract
(1) Background: The presence of metastatic disease significantly increases the risk of venous thromboembolism (VTE) in breast cancer, particularly during chemotherapy. Although not categorized as a highly thrombogenic malignancy, the elevated global prevalence of this cancer places a substantial number of patients at risk of thrombosis, which cannot yet be accurately predicted by validated risk assessment models (RAMs), highlighting the need for a dedicated model. (2) Aim: This study aims to develop a RAM for VTE in newly diagnosed metastatic breast cancer patients enrolled in a prospective, observational, and multicenter study. (3) Methods: A cohort of 189 patients beginning antitumor therapy were enrolled and prospectively monitored for VTE and mortality. Blood samples collected at enrollment were tested for D-dimer, fibrinogen, FVIII, prothrombin fragment 1 + 2 (F1 + 2), and thrombin generation (TG). Competing risk analyses were performed to identify significant predictors. (4) Results: Within one year, the cumulative incidences of VTE and mortality were 7.0% and 12%, respectively. Univariable analysis identified high Ki-67, D-dimer, FVIII, fibrinogen, and TG levels, along with low hemoglobin levels, as independent predictors of VTE. Only Ki-67, fibrinogen, FVIII, and hemoglobin were retained as significant predictors in multivariable analysis. These variables were further examined by multiple linear regression, which revealed Ki-67 and fibrinogen as the most significant parameters. A continuous RAM was then developed based on Ki-67 and fibrinogen (c-statistics 0.78), categorizing patients into low-risk and high-risk groups for VTE (2% vs. 13%; SHR 3.6, p = 0.018). This stratification could not be achieved using currently validated models for VTE risk. (5) Conclusions: We developed an accurate RAM for VTE that enables the identification of metastatic breast cancer patients at high risk for VTE, which supports clinicians in personalized thromboprophylaxis strategies if externally validated.