Protocol for a prospective accuracy study on an artificial intelligence-based ultrasound system for gestational age estimation among pregnant women in Ghana, Kenya and South Africa
Authors
Affiliations (1)
Affiliations (1)
- Burnet Institute
Abstract
BackgroundRisk screening for pre-eclampsia relies on accurate gestational age assessment, but routine access to ultrasound-based gestational dating remains challenging in many low- and middle-income countries (LMICs). As part of the formative work for the "Preventing pre-eclampsia: Evaluating AspiRin Low-dose regimens following risk Screening" (PEARLS) trial, we aim to validate and implement an Artificial Intelligence (AI)-based algorithm for estimation of gestational age, using blind sweeps done with a handheld ultrasound device. This study protocol outlines the accuracy cohort for AI-based gestational age estimation in participating facilities in Ghana, Kenya, and South Africa. MethodsThis multi-country prospective cohort study will recruit 969 pregnant women at 13 health facilities across Kenya, Ghana and South Africa. The eligible population are pregnant women presenting for antenatal visit from 11+0 to 13+6 weeks gestation. Eligible women will have a gestational age assessment by a trained sonographer using fetal biometry (reference standard), followed by gestational age estimation conducted by a trained midwife using the AI-based Intelligent Ultrasound ScanNav FetalCheck system (experimental). Both conventional and AI-based gestational age scans will be conducted with the General Electric (GE) VScanTM Air platform. Women will return for a second visit between 14+0 and 27+6 weeks gestation (week of visit is randomly selected) for an assessment with both conventional and AI-based ultrasound. The primary objective is to determine the accuracy and precision of gestational age estimation using an AI ultrasound system in first and second trimesters, as compared to gestational age estimation using crown-rump length (CRL) measurement by conventional ultrasound in first trimester (11+0 to 13+6 weeks). DiscussionThe study will provide critical evidence on the accuracy of a point-of-care, AI-based gestational age estimation ultrasound algorithm in sub-Saharan African settings. This study will inform the design of the PEARLS trial, as well as provide vital evidence for expanding implementation of ultrasound-based gestational age assessment for women in Africa.