Figure 2.

Performance of our AI system in identifying ploidy (euploids/non-euploids)

(A) Receiver-operating characteristic (ROC) curves for a binary classification using the clinical metadata-only model, the embryo video-only model, and the combined model in the internal test set. The videos of embryo development are captured using time-lapse system. AUC, area under the curve.

(B) Illustration of features contributing to the progression to euploids by SHAP values. Features on the right of the risk explanation bar pushed the risk higher, and features on the left pushed the risk lower.

(C and D) Performance comparison between our AI model and eight practicing embryologists in embryos’ euploidy scoring and ranking. (C) Correlation analysis between the euploidy rate and the score groups for PGT-A ranking based on AI or embryologist score. The dashed line is the overall euploidy rate of 46.1%. AI score groups were defined by binning AI-predicted probability. (D) Performance comparison between our AI model and eight practicing embryologists in embryos’ euploidy ranking.