AI May Estimate Gestational Age

An adapted AI model estimated gestational age with accuracy comparable to the clinical standard when novice operators performed blind sweep ultrasonography in two clinical settings.

By
Andrea Surnit
September 15, 2026
3 min

MDSpire News

Overview

A study evaluated an AI model's ability to estimate gestational age using blind sweep ultrasonography in diverse clinical settings. The AI demonstrated a mean absolute error of 4.2 days, comparable to the clinical standard.

Background

Accurate gestational age estimation is crucial for monitoring fetal growth and identifying potential complications during pregnancy. Traditional ultrasound methods require skilled operators and expensive equipment, limiting access in low-resource settings. The development of AI systems that utilize portable ultrasound devices may enhance accessibility.

Data Highlights

Setting	Mean Absolute Error (MAE)
Overall	4.2 days
Chicago	4.1 days
Nairobi	4.3 days

Key Findings

The AI model achieved a mean absolute error of 4.2 days, meeting noninferiority criteria compared to the clinical standard.
Performance was consistent across both clinical settings, with MAEs of 4.1 days in Chicago and 4.3 days in Nairobi.
Performance remained stable across different reference dating windows.
In Nairobi, the clinical standard outperformed the AI model in late-term pregnancies.
Exploratory analyses suggested improved accuracy for fetuses above the 90th weight percentile.

Clinical Implications

Further refinement of the model may be necessary for optimal performance in late-term pregnancies.

Conclusion

This study demonstrates the accuracy of AI in estimating gestational age using blind sweep ultrasonography, achieving comparable results to traditional methods in varied clinical settings.