Our latest scientific study evaluating FOLLISCAN, an AI-powered solution for ovarian follicle counting and measurement, has been published in the Journal of Assisted Reproduction and Genetics (JARG).

JARG is a well-established and internationally recognised peer-reviewed journal, widely respected in the fields of reproductive medicine and assisted reproductive technologies.

The publication presents robust clinical evidence for FOLLISCAN, our flagship artificial intelligence platform developed to support accurate, standardised, and time-efficient ovarian follicle assessment during hormonal stimulation in IVF cycles. Accurate follicle measurement is a crucial component of controlled ovarian stimulation, as it directly influences clinical decision-making, treatment timing, and patient outcomes.

In this study, we investigated whether FOLLISCAN can accurately and efficiently automate ovarian follicle counting and measurement from ultrasound scans, reducing manual workload while maintaining high clinical precision. The results demonstrate that FOLLISCAN achieves performance comparable to that of expert sonographers, highlighting its potential to support clinicians in daily practice by improving the consistency, efficiency, and scalability of follicular monitoring during IVF treatment.

Why We Looked at Follicle Measurement

Every IVF cycle relies on careful ultrasound monitoring. Clinicians measure follicle size and count follicles to decide how stimulation is progressing, when to adjust medication, or when to trigger ovulation.

But anyone who performs or reviews IVF ultrasounds knows the reality:

• Manual measurements take time
• Results can vary between operators
• And growing IVF volumes make scalability a real challenge

We asked a simple but important question:

Can AI match expert sonographers and make IVF monitoring faster and more consistent at the same time?

 

What went into the study?

• 5,508 transvaginal ultrasound scans
• 1,689 patients undergoing ovarian stimulation
• 4 IVF centers across Poland, Argentina, Colombia, and the USA.

The AI’s performance was evaluated using standard clinical metrics (precision, recall, F1 score), compared with expert annotations, and validated prospectively in real-world use.

FOLLISCAN is an AI-powered clinical software platform designed to automatically detect, count, and measure ovarian follicles on transvaginal ultrasound images during controlled ovarian stimulation in IVF cycles.

Read more about FOLLISCAN: https://mimfertility.ai/folliscan/ 

What the Results Showed

For follicles that matter most clinically ( ≥ 10 mm), the AI performed at a level comparable to experienced sonographers:

• 98.2% precision
• 88.9% recall
• 93.3% F1 score

Even when looking at all follicles, precision remained high (94.2%).

Importantly, performance stayed stable across different ultrasound machines in various clinics and countries.

Where AI Really Shines: workflow impact

Accuracy is essential, but efficiency is where clinicians feel the difference.

With AI assistance:

• Ultrasound annotation time was reduced 2.5× (p < 0.01)
• Experts needed to make just 0.54 corrections per scan on average
• Performance remained stable during prospective, day-to-day clinical use

Read more about FOLLISCAN: https://mimfertility.ai/folliscan/ 

What This Means for IVF Clinics

FOLLISCAN doesn’t replace clinicians and sonographers. It supports them.

It allows experts to focus on clinical decision-making, scale IVF monitoring without compromising quality, and maintain high standards even as patient volumes grow.

 

The Takeaway

This JARG published study provides strong evidence supporting the clinical reliability, robustness, and standardisation potential of FOLLISCAN in routine ovarian stimulation monitoring across diverse clinical settings. By demonstrating performance comparable to experienced sonographers, FOLLISCAN shows how artificial intelligence can meaningfully support clinicians without compromising accuracy or clinical trust.

Beyond efficiency gains, the findings highlight the role of AI in reducing inter- and intra-operator variability, improving consistency in follicle assessment, and supporting scalable IVF care as patient volumes continue to grow. Standardised follicle measurement and counting can help clinics optimise workflows, support data-driven clinical decisions, and ultimately enhance patient experience throughout treatment.

We view this publication as an important milestone in our commitment to responsible, evidence-based AI in reproductive medicine. Clinical validation and peer-reviewed research remain central to how we develop, evaluate, and deploy our solutions—ensuring that innovation is guided by scientific rigor and real clinical needs.

We invite you to read the full article in JARG and follow our journey as we continue building AI tools that make IVF monitoring smarter, more consistent, and more accessible for clinicians and patients worldwide.

Read the full article: 

An artificial intelligence platform for automated measurement and count estimation of ovarian follicles during ovarian stimulation and IVF: a multicenter study

JARG, 2026 | DOI: 10.1007/s10815-025-03777-y 

https://link.springer.com/article/10.1007/s10815-025-03777-y