AI Model Offers More Accurate Fetal Age Estimates for Indian Women

Chennai: Indian Institute of Technology Madras (IIT Madras) and the Translational Health Science and Technology Institute (THSTI) have developed the first-ever artificial intelligence (AI) model specifically designed to determine the age of a fetus in pregnant Indian women during the second and third trimesters. This development holds significant potential for improving pregnancy care and reducing maternal and infant mortality rates in India.

Addressing Discrepancies: Currently, doctors rely on formulas developed using data from Western populations to estimate fetal age, also known as gestational age (GA). However, these formulas may not be accurate for Indian women due to differences in fetal growth patterns. This new AI model, named “Garbhini-GA2,” aims to address this gap by using data specific to the Indian population.

Improved Accuracy: Garbhini-GA2 utilizes data collected from pregnant women in India and employs advanced machine learning techniques to achieve significantly higher accuracy compared to existing methods. This translates to a reduction in the median error rate by more than three times compared to the commonly used Hadlock model.

This new model has the potential to:

·         Enhance the quality of care: By providing more accurate fetal age estimates, Garbhini-GA2 can help doctors make better decisions regarding prenatal care, interventions, and delivery timing.

·         Reduce maternal and infant mortality: Improved pregnancy care can contribute to a decrease in complications and deaths associated with childbirth.

·         Refine epidemiological data: Accurate fetal age estimates are crucial for conducting reliable research on pregnancy outcomes in India.

The researchers are currently working towards validating Garbhini-GA2 in larger studies across diverse regions in India. Once validated, the model has the potential to be integrated into clinical practice, benefiting pregnant women and healthcare professionals nationwide.