Evaluation of cancer-associated thrombosis risk in patients with lung adenocarcinoma using the COMPASS-CAT risk assessment model

Abstract

Introduction: A variety of risk assessment models (RAMs) are being developed for the early identification of lung cancer patients at increased risk for Cancer-Associated Thrombosis (CAT).

Aim: We herein aimed to evaluate the discrimination capacity of COMPASS-CAT RAM for the prediction of venous thromboembolism (VTE) risk in lung adenocarcinoma.

Methods: The medical records of 129 patients with lung adenocarcinoma, diagnosed and treated at the Oncology Unit of “Sotiria” Athens General Hospital between June 2014 and May 2016, were retrospectively reviewed. VTE events were recorded and diagnostic accuracy measures of COMPASS-CAT RAM were calculated. The discrimination capacity of the model was tested with receiver operating characteristics (ROC) analysis.

Results: The majority of patients were male (75%), with stage IV disease (69%). Incidence of CAT was 10.08%. Based on COMPASS-CAT score values, 62% of patients were stratified in the low/ intermediate CAT risk group and had CAT rate of 8.75%, while the remaining 38% were patients with high CAT risk and CAT rate of 12.24%. Sensitivity, specificity, positive predictive value and negative predictive value of COMPASS-CAT RAM were 46.15%, 62,93%, 12.24% and 91.25%, respectively, and the area under the curve (AUC) of ROC analysis was 0.611.

Conclusions: COMPASS-CAT Score showed moderate discrimination and predictive capacity in our patient population. Further validation studies may be warranted prior to implementation of COMPASS-CAT RAM into routine clinical practice for patients with lung adenocarcinoma.