Use of Khorana score to predict VTE in patients undergoing chemotherapy for uterine cancer.

Abstract

5589 Background: Gynecologic cancers are associated with a high risk of venous thromboembolism (VTE). The Khorana score is a clinically-validated tool to assess risk of VTE in cancer patients (using disease site, BMI and blood counts). Recent ASCO clinical practice guidelines have recommended patients with a Khorana score of 2 or greater be offered pharmacologic thromboprophylaxis during systemic chemotherapy. For women with uterine cancer, the utility of the Khorana score is still unknown. Methods: A retrospective cohort study was performed from January 2016 to January 2020. All patients with uterine cancer were screened. Patients receiving chemotherapy, both neoadjuvant (NACT) and adjuvant (ACT), were included. VTE was evaluated for 12 months following the first cycle of chemotherapy. The Khorana score was calculated for each patient using both a high risk score of ≥2 and ≥3 and the patients were stratified based on NACT vs ACT. Logistic regression and chi-square were used to evaluate the prognostic utility of the Khorana score as well as other clinico-pathologic criteria on development of VTE. Results: A total of 265 patients were included. The majority of patients were obese (160, 60.4%) and 60 years or older (188, 70.9%). The most common histology was endometrioid (107, 40.4%) followed by serous (71, 26.8%) and the majority were advanced-stage (169, 63.8%). Most women underwent hysterectomy during treatment (243, 91.7%) followed by ACT (228, 86.0%). 14% (37) had NACT. 24 patients developed VTE (9.1%), which was higher, but not statistically different, with NACT vs ACT (13.5% vs 8.3%, p = 0.35). Demographics including age, race and BMI nor pathologic data including histology, grade or stage significantly correlated with development of VTE. Similarly, treatment factors including undergoing hysterectomy and radiation treatment were not statistically significant in regards to VTE. The proportion of patients with high Khorana score (both ≥2 and ≥3) was similar between groups. In the whole cohort, high Khorana score (defined either as ≥2 or ≥3) did not significantly predict VTE; however, the model using ≥3 was more predictive (OR 1.154, 95%CI 0.402-2.907, p = 0.7326). In the NACT cohort, neither model was predictive of VTE (both with OR < 1). In the ACT group, Khorana ≥3 was a better prediction model, but was still not statistically significant (OR 1.557, 95%CI 0.480-4.343, p = 0.4213). Conclusions: Although validated in other cancer types, the Khorana score was found to be a poor predictor of VTE in this population. A defined high risk Khorana score of ≥3 (per the original validation study) better predicted VTE than a score of ≥2 (per guidelines). Independent of the Khorana score, demographic and pathologic data were poor predictors of VTE. At this time, use of the Khorana score to guide routine thromboprophylaxis in patients undergoing chemotherapy for uterine cancer should be used with caution.