Combined analysis of clinical features, human chorionic gonadotropin (hCG) value, and hCG ratios for early prediction of postmolar gestational trophoblastic neoplasia.

Abstract

To investigate factors predicting postmolar gestational trophoblastic neoplasia (GTN) by combined analysis of clinical features, human chorionic gonadotropin (hCG) value, and hCG ratios. This retrospective study enrolled patients with histopathologically proven molar pregnancy. Patients lost to follow-up before remission or developing postmolar GTN were excluded. Demographic and clinical characteristics and hCG data obtained before and after molar evacuation were collected. Area under the receiver operating characteristic curve (AUC) analysis was used to identify the hCG and hCG ratio cutoff values that predict postmolar GTN. Multivariate analysis was employed to identify independent predictors of GTN. There were 113 complete moles, 11 partial moles, and 52 unspecified moles included in the final analysis. Of the 176 cases, 90 achieved remission and 86 developed post-molar GTN. The incidence of postmolar GTN was 48.9%, with a median time to GTN development of 5weeks. Univariate analysis showed age, molar evacuation performed elsewhere, pre-evacuation hCG, hCG at 2ndweek post-evacuation, and ratio of hCG at 2ndweek post-evacuation to post-evacuation hCG significantly predict GTN. Multivariate analysis revealed an hCG value ≥ 1400IU/L at 2nd week post-evacuation (AUC: 0.92, aOR: 6.51, 95% CI 1.28-33.16; p = 0.024) and a ratio of hCG at 2nd week post-evacuation to post-evacuation hCG of ≥ 0.02 (AUC: 0.88, aOR: 12.27, 95% CI 2.15-70.13; p = 0.005) to independently predict GTN. An hCG value ≥ 1400IU/L at 2ndweek post-evacuation and a ratio of hCG at 2ndweek post-evacuation to post-evacuation hCG of ≥ 0.02 independently and reliably predict postmolar GTN.