Effects of Multimodal AI-Assisted Noninvasive Euploid Embryo Selection on Clinical Outcomes After Single-Blastocyst Transfer in Women of Advanced Maternal Age: A Prospective Randomized Controlled Trial

After routine ovarian stimulation, oocyte retrieval and IVF/ICSI, embryos will undergo closed time-lapse culture. Patients will be recruited and provide informed consent on day 3. Eligible patients will be randomized by the random envelope method in a 1:1 ratio on day 5. The control group uses EmbryoInsights™ for morphological assessment. The experimental group uses EmbryoInsights™ Plus to predict euploidy probability and rank morphologically transferable blastocysts, prioritizing those classified as predicted euploid and then using morphology. Senior embryology professionals and study clinicians make the final transfer decision. If AI results are unavailable or the system fails, routine manual assessment is used. An AI result alone does not automatically cancel transfer. The investigational euploidy-prediction model has already been developed. No additional embryo biopsy or biological sampling is required. Follow-up includes hCG at 14 days after transfer, ultrasound at 35 days, assessment at 12 gestational weeks (±7 days), fetal anatomy assessment at 20-24 weeks, delivery, and 42 days postpartum. The primary analysis of ongoing pregnancy will use the intention-to-treat (ITT) population, with all randomized participants analyzed in their originally assigned groups. A per-protocol (PP) analysis will also be performed.