ICMR develops AI tool to predict IVF outcomes in men with infertility

New Delhi: The Indian Council of Medical Research (ICMR) in collaboration with Amity University has developed an artificial intelligence-based tool to detect the type of 'Y' chromosome microdeletion -- a genetic cause of male infertility -- and predict IVF outcomes.

The study regarding the AI tool was published in the Journal of Assisted Reproduction and Genetics last week.

In nearly 50 per cent of couples experiencing infertility, the problem is with the male partner, said Dr Deepak Modi, senior scientist at ICMR's National Institute for Research in Reproductive and Child Health (NIRRCH) and the lead author of the study.

"In them there could be problems with sperm production. One of the leading causes, Y chromosome microdeletion (YCMD) is observed in one in every 10 men with infertility. Because of this genetic defect, the testes are unable to make enough sperm leading to infertility," Dr Modi said.

Men with YCMD cannot benefit from medical treatment for improving the sperm count.

To become fathers, such men require assisted reproductive techniques such as in-vitro fertilization (IVF) for parenthood, he said.

The AI-based tool -- 'Fertility Predictor' -- developed by the ICMR-NIRRCH in collaboration with Amity University, Noida can predict sperm retrieval rates and success rates of assisted reproductive technology (ART) in men having this genetic problem.

It also predicts the rate of fertilization, clinical pregnancy and live birth rates based on the type of Y chromosome microdeletion, Dr Modi said. This helps the couples make informed decisions, he said.

However, Dr Modi cautioned that the male babies born through IVF from men with YCMD will inherit the same defect and will be infertile as it is 100 per cent transmitted from fathers to their sons.

Developing this tool took about two years by collating data of more than 500 men having YCMD and undergoing ART.

After applying an artificial intelligence algorithm based on machine learning on this data, the tool could predict the outcomes. This was then validated on another sub-set and it was found to have accuracy of about 80 per cent, Dr Stacy Colaco, a scientist at NIRRCH and the first author of the study, said.

"Fertility Predictor also provides a numerical output for chance of clinical pregnancy and live birth in men with YCMD. The validation studies indicated its robustness and high accuracy in predicting both these parameters solely based on the type of YCMD," said Dr Abhishek Sengupta of Amity University, Noida.