MAIA platform for routine clinical testing: an artificial intelligence embryo selection tool developed to assist...

Area of Science:

• Reproductive Medicine
• Artificial Intelligence in Healthcare
• Embryology

Background:

• Reducing multiple gestations in assisted reproductive care necessitates improved embryo selection.
• Time-lapse incubators increase data volume and subjectivity in traditional embryologist assessments.
• Objective, standardized embryo evaluation tools are needed to support clinical decision-making.

Purpose of the Study:

• To develop and evaluate an Artificial Intelligence (AI) model, Morphological Artificial Intelligence Assistance (MAIA), for embryo selection in Brazil.
• To assess MAIA's accuracy in predicting clinical pregnancy.
• To provide embryologists with a user-friendly tool for real-time embryo assessment.

Main Methods:

• Development of the MAIA AI model through collaboration between a university and a fertility clinic.
• Training the AI model on 1,015 embryo images.
• Prospective clinical testing of MAIA on 200 single embryo transfers.

Main Results:

• MAIA achieved an overall accuracy of 66.5% in clinical testing.
• For elective embryo transfers, MAIA demonstrated 70.1% accuracy in predicting clinical pregnancy.
• The AI tool features a user-friendly interface designed with embryologist input.

Conclusions:

• The MAIA AI model offers a potential solution for objective and standardized embryo assessment in assisted reproductive care.
• AI-based tools like MAIA can support embryologists in selecting embryos with higher potential for clinical pregnancy.
• Tailored AI solutions can address specific demographic and ethnic profiles in fertility treatments.