AI predicts embryonic health through image analysis in non-invasive study

A recent study in eClinicalMedicine highlights the potential of artificial intelligence (AI) to predict embryonic health non-invasively by analyzing embryo images. This method aims to assess chromosome conditions without the need for invasive testing, which is currently used in IVF. The study reviewed 20 relevant studies, finding that AI algorithms showed a sensitivity of 67% and specificity of 58% in predicting embryonic ploidy. Improvements were noted when using the highest accuracy data, with sensitivity rising to 71% and specificity to 75%. While AI shows promise, the current models do not yet match the accuracy of invasive preimplantation genetic testing. Researchers emphasize the need for further development to enhance the reliability of AI in embryo selection.