A RECENT study, led by Hideyuki Kobayashi, Toho University School of Medicine, Tokyo, has developed an AI model capable of predicting male infertility risk using hormone levels from a blood test. The novel screening method for male infertility uses serum hormone levels tested for various infertility conditions such as non-obstructive azoospermia, obstructive azoospermia, cryptozoospermia, oligozoospermia, asthenozoospermia, and normal fertility.
Infertility affects approximately 72.4 million people globally, with male factors involved in approximately half of the cases. Traditional semen analysis is critical for diagnosing male infertility, reflecting spermatogenesis, seminal duct patency, and glandular secretory activity. However, it is labour-intensive and often stigmatised, leading to low testing rates.
The cohort comprised 3,662 patients, whose serum was analysed with the AI model and was tested to detect infertility conditions or normal fertility. The study utilised a no-code AI creation software to construct the model which notably achieved a perfect accuracy rate in predicting non-obstructive azoospermia, a severe form of male infertility.
The study explored whether AI could predict male infertility based solely on serum hormone levels. Machine learning, an applied AI technology, enables computers to learn from data without explicit programming. By comparing models generated by two different sources, researchers aimed to find the most accurate approach.
The AI models accurately predicted non-obstructive azoospermia cases, with follicle-stimulating hormone emerging as the most significant marker. The results suggest that an AI-based screening tool using serum hormone levels could identify men at risk for infertility without requiring semen analysis thus, allowing timely referrals to specialist clinics for further examination and treatment. The model provides a more reliable and less variable method for preliminary screening that, if adopted widely, could revolutionise male infertility screening.
Katie Wright, EMJ
Reference
Kobayashi H et al. A new model for determining risk of male infertility from serum hormone levels, without semen analysis. Sci Rep. 2024;14(1):17079.
