RESUMO
INTRODUCTION: Venous thromboembolism (VTE) is common in non-small cell lung cancer (NSCLC) patients undergoing systemic chemotherapy. The usefulness of Khorana score (KRS) to predict risk in lung cancer patients is limited, and the identification of patients who would benefit most from thromboprophylaxis is challenging. We aimed to identify variables whose values before chemotherapy helped in predicting VTE occurrence, and build a model to assess VTE risk. MATERIALS AND METHODS: A cohort of newly diagnosed NSCLC patients to undergo outpatient chemotherapy, not under anticoagulant treatment, was recruited. Pre-chemotherapy demographic, clinical, analytical and tumor-specific variables were collected. Patients were prospectively followed-up for 12â¯months to record VTE events. Bivariate and multivariate analyses were performed to identify VTE-associated variables, and a prediction model was built and compared with KRS. RESULTS: 90 patients were recruited, 18 of whom had a VTE event during follow-up. High baseline levels of factor VIII (FVIII) and, especially, soluble P-selectin (sP-selectin), were independently associated with VTE risk (hazard ratio [HR] 4.15, 95% confidence interval [CI] 1.17-14.71, and 66.40 [8.70-506.69], respectively). Our so-called Thrombo-NSCLC risk score, which assigns 1 and 3 points to high FVIII and sP-selectin values, respectively, was significantly better than KRS in predicting VTE (area under the curve [AUC] 0.93 vs. 0.55, sensitivity 94.4 vs. 35.0%, specificity 93.1 vs. 60.0%). Our prediction model showed significant discriminating capacity between high risk vs. intermediate/low risk patients, while KRS did not. CONCLUSIONS: The Thrombo-NSCLC risk score may be useful to identify those NSCLC patients who would benefit most from thromboprophylaxis.
