An artificial intelligence platform to optimize workflow during ovarian stimulation and IVF: process improvement and...

Area of Science:

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

Background:

• Ovarian stimulation in IVF requires frequent monitoring, increasing patient burden and impacting laboratory workflow.
• Optimizing monitoring and retrieval scheduling is crucial for efficient IVF cycle management.

Purpose of the Study:

• To develop and evaluate an AI-driven algorithm to streamline IVF workflows.
• To reduce the number of monitoring visits during ovarian stimulation to a single day.
• To enable level-loading of oocyte retrievals in the embryology laboratory.

Main Methods:

• A dataset of 1591 autologous IVF cycles was analyzed.
• An algorithm was designed to predict the optimal monitoring day, trigger day, and oocyte yield.
• Patient data included age, hormone levels (FSH, estradiol, anti-Müllerian hormone), follicle counts, and BMI.

Main Results:

• The algorithm predicted the single best monitoring day with a mean error of 1.355 days.
• Oocyte retrieval prediction accuracy was 0.76 (baseline) and 0.80 (with monitoring data).
• Sensitivities for predicting total and mature oocytes ranged from 0.76 to 0.81.

Conclusions:

• A novel AI algorithm can facilitate IVF workflow by enabling single-day monitoring.
• The algorithm aids in scheduling trigger days for level-loading retrievals.
• This approach minimizes patient visits and optimizes embryology lab operations.