Clinicians at Toho University in Japan created an AI model to predict male infertility risk without needing semen analysis. This model uses hormone levels from a blood test and could replace semen tests. About half of infertility cases are due to men, but semen analysis is not widely available.
The AI model, developed by Associate Professor Hideyuki Kobayashi and reported in Scientific Reports, was accurate 74% of the time and 100% for predicting severe male infertility.
The study analyzed data from 3,662 men who had semen and hormone tests for infertility between 2011 and 2020. It measured semen volume, sperm concentration, motility, and hormone levels. The AI model, which uses this data, was about 74% accurate overall and 100% correct in predicting severe male infertility (non-obstructive azoospermia).
Validated with data from 2021 and 2022, its accuracy was around 58% in 2021 and 68% in 2022. Associate Professor Kobayashi noted that while this AI model is a useful initial screening tool, it should not replace semen tests but can help identify cases that need further specialist evaluation.
CreaTact, Inc., based in Mito City, Japan, is developing an AI prediction model to screen for male infertility. They hope this model will make infertility testing more accessible at clinical labs and health checkup centers. Associate Professor Kobayashi shared this goal. The study was published in Scientific Reports on July 31, 2024.
The study concludes that an AI model can predict male infertility risk using a blood test, eliminating the need for semen analysis. This approach offers a more accessible screening method. It accurately identifies severe forms of infertility, providing a valuable tool for initial evaluations before more detailed testing.
Journal reference :
- Kobayashi, H., Uetani, M., Yamabe, F., et al. A new model for determining the risk of male infertility from serum hormone levels, without semen analysis. Scientific Reports. DOI: 10.1038/s41598-024-67910-0.