Assessment of embryo implantation potential with a cloud-based automatic software.

Area of Science:

• Reproductive Medicine and Embryology
• Artificial Intelligence in Healthcare
• In Vitro Fertilization (IVF)

Background:

• Standard embryo evaluation in IVF relies on subjective morphological assessment.
• Objective and automated methods are needed to improve IVF success rates.
• The Dana system offers automated embryo evaluation using video analysis.

Purpose of the Study:

• To determine if automated embryo selection by the Dana system increases implantation rates in IVF.
• To validate the Dana system's accuracy in predicting embryo implantation potential.
• To assess the system's applicability across different IVF clinics.

Main Methods:

• A three-phase study involving a large dataset of known implantation data (KID) embryos.
• Embryo evaluation using the Dana system, incorporating manual annotations and developmental videos.
• Validation of Dana's automatic selection without embryologist intervention across multiple centers.

Main Results:

• Dana system established four embryo grades (A-D) with distinct implantation rates (34% to 19%).
• A classification ranking based on Unit Average Distance (UAD) showed significant differences in pregnancy rates (59% to 28%).
• Automatic categorization confirmed higher implantation rates for top-ranked embryos (UAD ≤1.03) compared to low-ranked ones (25% vs 46%).

Conclusions:

• Embryo selection using Dana's automated ranking significantly enhances IVF treatment success, especially in oocyte donation programs.
• The multicenter study validates Dana's applicability and standardization of embryo interpretation across clinics.
• Dana's accuracy improves over time as its database of embryo development grows.