Comparison of predictive value of different risk assessment models for venous thromboembolism in elderly patients with lung cancer / 中华老年医学杂志

ABSTRACT

Objective:To explore the predictive value of the COMPASS-cancer associated thrombosis(COMPASS-CAT)risk assessment model and the modified Khorana risk assessment model for the risk of venous thromboembolism(VTE)in elderly patients with lung cancer.Methods:A retrospective analysis was conducted on clinical data of 276 hospitalized lung cancer patients aged 60 years and over in the Cancer Hospital of the Chinese Academy of Medical Sciences and Peking Union Medical College from March 2013 to March 2017.Patients were divided into the VTE group(n=39)and the non-VTE group(n=237). The COMPASS-CAT model and the modified Khorana model were used to evaluate the risk of venous thromboembolism in the two groups.The sensitivity, specificity and Youden index of the two models were calculated.The receiver-operating characteristics(ROC)curves of the two evaluation models were drawn.The predictive effect and influence of two evaluation models on VTE risk in elderly lung cancer patients were compared by using Medcalc software analysis and multivariate Logistic regression analysis.Results:The incidence of VTE was 14.13%(39/276). The sensitivity, specificity, Youden index and the AUC under ROC curves of the COMPASS-CAT model and the modified Khorana model were 0.718 and 0.795, 0.861 and 0.527, 0.524 and 0.348, and 0.789 and 0.661, respectively.Using the Medcalc software to compare the two models, the area under the curve of the COMPASS-CAT risk assessment model was increased by 0.128, compared with the modified Khorana model( Z=2.676, P=0.0075). Multivariate Logistic regression analysis showed that COMPASS-CAT≥7 points and Khorana≥2 points were independent risk factors for VTE in elderly patients with lung cancer( P<0.05). Conclusions:The modified Khorana model can predict the risk of VTE in elderly patients with lung cancer, but the accuracy of prediction is low.The COMPASS-CAT model has a higher predictive value for VTE risk assessment than the modified Khorana model and is more suitable for elderly patients with lung cancer.