Risk factors for the prediction of treatment failure in gestational trophoblastic tumors treated with EMA/CO regimen

Objective. The purpose of this study was to determine the independent risk factors for resistance to EMA/CO chemotherapy in gestational trophoblastic tumor (GTT), to develop a more accurate scoring system for the evaluation of high-risk GTT patients. Methods. We performed a retrospective study and reviewed 165 GTT patients who received the EMA/CO regimen. Among these patients, 27 showed resistance to EMA/CO. Results. According to the univariate analysis of risk factors, tumor age, initial human chorionic gonadotropin level, metastatic site, number of metastatic organs, unplanned operation, gravidity, and inadequate previous chemotherapy all showed statistical significance. However, the results of the stepwise-Cox proportional hazards regression of prognostic factors using multivariate analysis showed statistical significance for tumor age, number of metastatic organs, metastatic site, and previously inadequate chemotherapy. The results of the stepwise logistic regression of prognostic factors showed tumor age and number of metastatic organs as significant. According to the performance of the fitted logistic regression model, the accuracy of predicted death and survival was 80.5%. Conclusions. We determined that the following factors indicate a poor prognosis: (1) tumor age greater than 12 months, (2) number of metastatic organs greater 2, (3) incomplete previous treatment, including unplanned operation and inadequate chemotherapy. In patients with two and three of these factors, the death rates were 17.7 and 56.6%, respectively.