Using Data Mining to Assist in Predicting Reproductive Outcomes Following Varicocele Embolization.

Abstract

We carried out a retrospective analysis of infertile couple data using several methodologies and data analysis techniques, including the application of a novel data mining approach for analyzing varicocele treatment outcomes. The aim of this work was to characterize embolized varicocele patients by ascertaining the improvement of some of their clinical features, predicting the success of treatment via pregnancy outcomes, and identifying data patterns that can contribute to both ongoing varicocele research and the more effective management of patients treated for varicocele. We retrospectively surveyed the data of 293 consenting couples undergoing infertility treatment with male varicocele embolization over a 10-year period, and sperm samples were collected before and at 3, 6, and 12 months after varicocele embolization treatment and analyzed with World Health Organization parameters—varicocele severity grades were assessed with medical assessment and scrotal ultrasound, patient personal information (e.g., age, lifestyle, and embolization complications) was collected with clinical inquiries, and varicocele embolization success was measured through pregnancy outcomes. Varicocele embolization significantly improved sperm concentration, motility, and morphology mean values, as well as sperm chromatin integrity. Following this study, we can predict that a male patient without a high varicocele severity grade (with grade I or II) has a 70.83% chance of conceiving after embolization treatment if his partners’ age is between 24 and 33 with an accuracy of 70.59%. Furthermore, male patients successful in achieving pregnancy following embolization are mostly characterized by having a normal sperm progressive motility before treatment, a normal sperm concentration after treatment, a moderate to low varicocele severity grade, and not working in a putatively hazardous environment.