New AI Model Achieves 74% Accuracy in Predicting Male Infertility Without Semen Testing

A World Health Organization (WHO) study revealed that half of infertility is male-related and sex analysis is crucial for diagnosing male infertility. But it is not readily available in medical institutions yet. There is also a high threshold for receiving it. This study was conducted in 2017. But researchers have recently developed an AI model that predicts male infertility risk without semen analysis by measuring hormone levels in a blood test. This model uses AI creation software without programming.  

In this study, researchers used clinical data from 3,662 men. These men underwent semen and hormone testing for male infertility between 2011 and 2020. Semen tests measured sperm concentration, semen volume, and motility. On the other hand, hormone tests measured testosterone, FSH, LH, PRL and E2 with T/E2 added. Total motile sperm count was calculated from semen test results. 

The WHO laboratory manual defines a total motile sperm count of 9.408 X 106 as the lower limit of normal for semen testing. If the count is above this threshold, the patient is assigned a value of “0” and a value of “1”. This AI model was around 74% accurate. 

Then researchers validated this AI model by using data of hormone and semen test from 2021 and 2022. It showed an accuracy of 58% with 188 patients in 2021 and 68% with 166 patients in 2022. But it accurately predicted non-obstructive azoospermia in both years with 100% accuracy. 

One of the researchers said that this AI prediction model used in a study is intended as a screening step before semen testing, not a replacement for it. It was particularly accurate in predicting non-obstructive azoospermia. If abnormal values are detected, patients may need to undergo detailed testing at a specialist infertility clinic for appropriate treatment.  

CreaTact, Inc. is developing an AI prediction model for male infertility and wants to make testing more accessible in clinical laboratories and health checkup centres. The model will be developed through software development and data analysis. It will overcome hurdles and make the process more accessible. 

The results of this study were recently published in Scientific Reports. They show that this AI model has an accuracy rate of 74%. It can predict non-obstructive azoospermia with an accuracy of 100%. Hence this study will help healthcare professionals to diagnose those men who are infertile.