Combining Machine Learning with Metabolomic and Embryologic Data Improves Embryo Implantation Prediction

Abstract

This study investigated whether combining metabolomic and embryologic data with machine learning (ML) models improve the prediction of embryo implantation potential. In this prospective cohort study, infertile couples (n=56) undergoing day-5 single blastocyst transfer between February 2019 and August 2021 were included. After day-5 single blastocyst transfer, spent culture medium (SCM) was subjected to metabolite analysis using nuclear magnetic resonance (NMR) spectroscopy. Derived metabolite levels and embryologic parameters between successfully implanted and failed groups were incorporated into ML models to explore their predictive potential regarding embryo implantation. The SCM of blastocysts that resulted in successful embryo implantation had significantly lower pyruvate (p<0.05) and threonine (p<0.05) levels compared to medium control but not compared to SCM related to embryos that failed to implant. Notably, the prediction accuracy increased when classical ML algorithms were combined with metabolomic and embryologic data. Specifically, the custom artificial neural network (ANN) model with regularized parameters for metabolomic data provided 100% accuracy, indicating the efficiency in predicting implantation potential. Hence, combining ML models (specifically, custom ANN) with metabolomic and embryologic data improves the prediction of embryo implantation potential. The approach could potentially be used to derive clinical benefits for patients in real-time.

Keywords: ANN; Blastocyst; Machine learning; Metabolomics; NMR spectroscopy.

Fig. 1

Flowchart of the machine learning (ML) model training and testing procedures

Fig. 2

A Comparison of the metabolite levels in spent culture medium (SCM) samples from successfully implanted embryos (n=23) and embryos that failed to implant (n=33) relative to the levels in the medium control samples (n=44). B) Principal component analysis (bi-plot) of the region-wise integrals of the three groups. Gray represents medium control, orange represents SCM from successfully implanted embryos, and blue represents SCM from embryos that failed to implant

Fig. 3

Performance evaluation of classical ML model (random forest) combined with metabolite dataset. A Confusion matrix for random forest classifier, B receiver operating characteristic (ROC) curve, and C precision-recall curve

Fig. 4

Accuracy and loss curves obtained for the custom ANN model with the training and testing datasets. A Accuracy curve (demonstrating 100% accuracy for the training and testing dataset) for ANN model with the metabolite dataset for 50 epochs and B loss curve. C Accuracy curve for ANN model with embryologic dataset for 30 epochs and D loss curve