Prediction of gestational trophoblastic neoplasia based on trend of β-hCG during the first 21 days after molar pregnancy evacuation: an application of growth mixture model

Gestational trophoblastic neoplasia (GTN) is a broad spectrum of benign and malignant tumors derived from human placenta. Although they are rare in incidence, they have the potential to become rapidly fatal diseases. Therefore, predicting this disease in early stages is important. The aim of this study was to gain access to an appropriate marker for early prediction of GTN based on a trend of &beta-hCG in patients with molar pregnancy. In the present study we analyzed the available data from the medical files of 201 patients with hydatidiform mole, according to their pathological results, who referred to the educational and health care centers affiliated to Shahid Beheshti University of Medical Sciences from 2003 to 2013. We used the growth mixture model for longitudinal data analysis in order to assess early prediction of GTN. The findings from the growth mixture model showed that there were two different patterns (classes) of trend in &beta-hCG logarithm in these women. So that, in the first pattern with a modest decreased slope in &beta-hCG logarithm, the post-molar GTN was observed in 91 percent of the women. While, in the second pattern with a steeper decreased slope in &beta-hCG logarithm, the post-molar GTN was not observed in the women under study. The slope of trend in &beta-hCG logarithm (estimated from the growth mixture model) three weeks after molar pregnancy evacuation could be considered as an appropriate criterion for predicting the GTN.