A research team led by Associate Professor Hideyuki Kobayashi from Toho University School of Medicine has developed an AI model that predicts male infertility risk based on hormone levels from a blood test.

Researchers have developed a new blood test that can predict male infertility with approximately 74% accuracy.

The model was particularly effective in predicting non-obstructive azoospermia, a severe form of infertility, with a 100% accuracy rate.

Data from 3,662 patients was used to develop the model, which achieved an accuracy rate of approximately 74%.

The AI model was further validated with data from 188 patients in 2021 and 166 patients in 2022, showing lower accuracy rates of 58% and 68%, respectively, but maintained 100% accuracy for predicting non-obstructive azoospermia.

Semen analysis is essential for diagnosing male infertility but is often not readily available outside specialized clinics.

The test is designed to be accessible and can be performed in GP surgeries without requiring specialized laboratories.

The researchers hope that their AI model will facilitate male infertility screening in clinical laboratories and health check-up centers.

Allan Pacey, a professor of andrology at the University of Manchester, noted that the blood test could streamline the initial detection of male infertility, which usually requires semen analysis at specialized labs.

Kobayashi emphasized that the AI model serves as a preliminary screening tool before semen testing and is not a substitute for it.

The study was published on July 31, 2024, in the journal Scientific Reports.