External validation of the COMPASS-Cancer Associated Thrombosis Study: A predictive score to identify patients with solid tumors on treatment who are at risk for venous thromboembolism.

Abstract

e18005 Background: A predictive model to identify patients on active treatment for solid tumors at risk for venous thromboembolism (VTE) remains an important clinical need. The COMPASS-CAT Risk Assessment Model (RAM) for VTE in a Eurasian patient population with solid tumors undergoing active treatment displayed good discrimination to identify patients into low/intermediate versus high VTE risk groups. We sought to externally validate this RAM in a North American cancer population. Methods: Patients age > 17 years who presented to the Monter Cancer Center between 1/12014 to 12/31/2016 with breast, ovarian, lung, or colorectal cancers were included. The COMPASS CAT RAM included 8 scored variables that were identified by ICD-9 and -10 codes and chart review. The primary endpoint at 6-month study follow-up was symptomatic VTE. Area under the ROC Curve (AUC) was calculated. We calculated a binary cut off to define VTE risk groups into low/intermediate and high VTE risk categories and calculated sensitivity, specificity, and negative and positive predictive values (NPV, PPV). Results: 3814 patients were included with a mean age of 64, of whom 49%, 5%, 29%, and 17% had breast, ovarian, lung, and colorectal cancer, and 52% had localized and 48% had advanced stage disease. In 94% of patients, cancer was diagnosed within the last 6 months prior to inclusion and 46% of patients were on active therapy when evaluated. Symptomatic VTE at 6-month follow-up occurred in 5.85% of patients. The AUC was 0.62. Using model cut-offs of 0-6 or > 7 points, patients stratified into low/intermediate and high-risk groups had VTE rates of 2.27% and 6.31%, respectively. The sensitivity, specificity, NPV, and PPV of the RAM were 95%,12%, 97.73%, and 6.31%, respectively. Conclusions: In this large retrospective external validation study of the COMPASS-CAT RAM for VTE in cancer patients undergoing active treatment, model discrimination and calibration was fair. Further prospective validation studies are needed before the model can be implemented into routine clinical practice for primary thromboprophylaxis of high VTE risk cancer patients with solid tumors.