Evaluation of the effect on ongoing pregnancy rate of using artificial intelligence for embryo prioritization: An interim analysis of a prospective randomized control trial

OBJECTIVE:

To evaluate the ongoing pregnancy rate of embryos selected by trained embryologists with the adjunctive use of artificial intelligence (AI) compared to embryos selected using standard morphology criteria alone in a prospective randomized control trial.

MATERIALS AND METHODS:

An investigational AI system was previously developed by Alife Health that analyzes static images of day 5, 6, and 7 blastocyst-stage embryos that have already been determined to be suitable for transfer by an embryologist1. The system was developed to provide a prediction of the likelihood of leading to ongoing pregnancy, defined as achieving a fetal heartbeat at 6-8 week ultrasound. A prospective randomized controlled trial (NCT05483985) was initiated at seven clinics in the United States from June 2022 - present. Patients between 21 and 43 years old undergoing autologous IVF treatment and having retrieved at least 8 eggs were included in the study. Participants were randomized in a 1:1 ratio controlled for age and PGT status to the treatment group (AI + morphology grading for embryo selection) or the control group (morphology grading alone). The primary study endpoint was to evaluate pregnancy rate, defined as a fetal heartbeat at 6-8 weeks. Per study protocol, this was a planned interim analysis performed when 100 out of the overall 440 patients had completed the trial through the primary pregnancy assessment.

RESULTS:

There were 59 patients randomized to the treatment arm and 55 randomized to the control arm. There were no significant differences in average age (35 vs 34, p=0.63), anti-mullerian hormone (3.91 vs 4.61, p=0.15), BMI (26.89 vs 27.66, p=0.39), or PGT status (76% vs 73%, p=0.66) between the two groups. The ongoing pregnancy rate was 69% (41/59) in the treatment group and 64% (35/55) in the control group, representing a 5% difference (p = 0.5, not statistically significant). There were no device related adverse events reported for patients in the AI arm of the study. When the study is completed, these analyses will be repeated with increased statistical power to demonstrate that AI as an adjunctive tool for embryo selection is safe and effective.

CONCLUSIONS:

The selection of embryos using AI as an adjunctive tool resulted in promising clinical pregnancy rates when compared to embryo selection using standard morphology grading alone and did not result in any clinical safety concerns.

IMPACT STATEMENT:

This is an interim analysis of the largest randomized control trial on AI for embryo selection conducted in US clinics to date. Further study recruitments that were enrolled in July 2023-February 2024 will be analyzed later this year to evaluate the utilization of AI as an adjunctive tool for embryo selection.