Selecting the embryo with the highest implantation potential using a data mining based prediction model

摘要: Embryo selection has been based on developmental and morphological characteristics. However, the presence of an important intra-and inter-observer variability standard scoring system (SSS) reported. A computer-assisted (CASS) potential to overcome most these disadvantages associated with SSS. The aims this study were construct a prediction model, data mining approaches, compare predictive performance models in SSS CASS evaluate whether using model would impact embryo for transfer. total 871 single transferred embryos between 2008 2013 included evaluated two systems: CASS. Prediction developed multivariable logistic regression (LR) multivariate adaptive splines (MARS). externally validated test set 109 transfers January June 2014. Area under curve (AUC) training validation was compared determine utility models. In models, AUC declined significantly from (p < 0.05). No significant difference detected derived Two final obtained LR MARS, which showed moderate discriminative capacity (c-statistic 0.64 0.69 respectively) data. that introduction improved generalizability combination modeling is promising approach improve highest implantation potential.