Related Experiment Video
Updated: Jan 16, 2026
09:09
Gene Expression Analyses in Human Follicles
Published on: February 17, 2023
1.5K
Modelling Follicular Growth During Ovarian Stimulation Using Agent-based Artificial Intelligence.
Artsiom Hramyka1,2, Thomas W Kelsey1, Simon Hanassab3,4,5
1School of Computer Science, University of St Andrews, St Andrews KY16 9SX, UK.
The Journal of Clinical Endocrinology and Metabolism
|September 29, 2025
Summary
This study models individual ovarian follicle growth rates during fertility treatments using artificial intelligence. The AI model accurately predicts follicle size, potentially reducing the need for multiple monitoring scans.
Area of Science:
- Reproductive Endocrinology
- Medical Artificial Intelligence
- Ovarian Physiology
Background:
- Ovarian stimulation is crucial in medically assisted reproduction (MAR), using follicle-stimulating hormone (FSH) to promote multi-follicular growth.
- Limited data exist on individual follicular growth rates during fertility treatments.
- Understanding these dynamics is key to optimizing MAR protocols.
Purpose of the Study:
- To develop an agent-based artificial intelligence (AI) model to predict individual ovarian follicle growth rates during ovarian stimulation in MAR cycles.
- To enable prediction of final follicle sizes based on early growth patterns.
Main Methods:
- An observational cohort study involving 11,572 patients across 11 European assisted conception clinics (2005-2023).
- Predictive modeling utilized 39,698 scans and 434,082 follicles from 12,950 MAR cycles.
- Daily growth rates of individual follicles were modeled to predict end-of-stimulation sizes.
Main Results:
- The mean follicle growth rate was 1.35mm/day, significantly associated with antral follicle count and FSH dose changes (p < 0.001).
- The AI model predicted follicle sizes within 2mm with 75.0% accuracy using the first scan, increasing to 80.1% with two scans.
- Model performance was consistent across clinics, with a mean accuracy of 78.0% in random splits and 77.1% in cross-validation.
Conclusions:
- Advanced AI techniques enhance the understanding of follicle growth dynamics during ovarian stimulation.
- The developed model reliably predicts follicle size profiles, potentially optimizing monitoring frequency.
- This predictive capability can aid in managing ovarian stimulation protocols in MAR.
Related Concept Videos
01:20Folliculogenesis
2.1K
Folliculogenesis is the development of ovarian follicles, the specialized structures within the ovarian cortex where oogenesis, or egg development, occurs. This process is essential for female reproductive health and begins during fetal development when primordial follicles are formed. Each primordial follicle comprises a primary oocyte in the center, surrounded by a single layer of squamous pre-granulosa cells. These follicles remain dormant in late prophase I of meiosis until triggered by...
2.1K
01:30Hormonal Control of the Ovarian Cycle
6.4K
The ovarian cycle is meticulously regulated by the hypothalamic-pituitary-gonadal axis. This cycle orchestrates the release of a mature oocyte, essential for reproduction.
Before puberty, the hypothalamus releases GnRH in a low frequency, low amplitude pulsatile manner. This along with the immature hypothalamic-pituitary-gonadal axis activity, results in low estrogen levels and the absence of a fully functional ovarian cycle. At puberty, GnRH secretion increases in both frequency and...
Before puberty, the hypothalamus releases GnRH in a low frequency, low amplitude pulsatile manner. This along with the immature hypothalamic-pituitary-gonadal axis activity, results in low estrogen levels and the absence of a fully functional ovarian cycle. At puberty, GnRH secretion increases in both frequency and...
6.4K