Machine learning in predicting infertility treatment success: A systematic literature review of techniques.

Abstract

Assisted reproductive technology (ART) is one of the major developments that has had a significant impact on infertility treatment. A predictive model of ART success based on machine learning (ML) techniques can provide a robust basis for estimating treatment success. This study aimed to identify predictive models of ART success and their determinants. A systematic search was conducted in PubMed, Web of Science, Scopus, and Embase. Data extraction involved collecting data in studies on dataset characteristics, ML techniques, and predictive model performance indicators. The search resulted in 3655 records, of which 27 papers were selected for analysis. ML publications in ART prediction have been in the past 5 years. In general, 107 various features were reported in all reviewed studies. Female age was the most common feature used in all identified studies. Most studies (96.3%) applied a supervised approach to develop predictive models. Among all, support vector machine (SVM) was the most frequently applied technique (44.44%). Nineteen different indicators have been used in studies to evaluate the model performance. 74.07% of the reviewed papers reported area under the receiver operating characteristic (ROC) curve (AUC) as their performance indicator. Accuracy (55.55%), sensitivity (40.74%), and specificity (25.92%) were also commonly reported. ML has the potential to bring hope to infertile couples and to facilitate making challenging decisions. Considering relevant contributing factors and ML techniques is critical for reliable predictive modeling.