Prospective Evaluation of an AI Diagnostic Ultrasound Tool for Fetal Weight Estimation
Trial statusRecruiting
Trial phaseNot listed
Trial typeObservational
Biological sexFemale
Age18+
SponsorUniversity of North Carolina, Chapel Hill
About this trial
Purpose: The primary objective of this study is to assess the diagnostic accuracy of an AI-enabled ultrasound tool for estimating fetal weight Participants: 1,000 pregnant individuals Procedures (methods): This prospective diagnostic accuracy study will enroll 1,000 pregnant individuals within one week of anticipated delivery. At a single visit, each participant will undergo two ultrasound assessments: (1) standardized sweeps for AI analysis (performed by both specialist and nonspecialist users), (2) specialist-performed fetal biometry.
Eligibility criteria
This trial does not accept healthy volunteersQualifiers
18 years of age or older
Viable intrauterine pregnancy
Delivery expected within one week of study procedures between 24 0/7 and 42 6/7 weeks, including participants with a scheduled induction or cesarean delivery on a known date, or those admitted in spontaneous labor
Ability and willingness to provide written informed consent
Disqualifiers
Maternal body mass index ≥ 40 kg/m²
Multiple gestation (i.e., twins or higher order)
Known major fetal malformation or anomaly
Any maternal condition (medical, psychological, or social) that, in the opinion of the study team, may interfere with study participation or data integrity.
Trial population
1,000 pregnant individuals
Trial design
Design model
Cohort
Time perspective
Prospective
Treatments tested in this trial
AI ultrasound diagnostic tool for fetal weight estimation
Diagnostic testParticipants will undergo study-specific transabdominal ultrasound acquisition using standardized abdominal sweeps of the gravid abdomen, guided by external maternal landmarks and saved as cineloop videos. The cineloop videos will be analyzed by a locked deep-learning AI diagnostic tool to generate an estimated fetal weight. The AI-generated estimate will be compared with specialist-performed fetal biometry and actual birth weight to evaluate diagnostic accuracy. The AI output is for research evaluation only and will not direct clinical management during the study.
Treatment groups
1,000 Participants
are divided into 1 treatment groupGroup A: Pregnant Women within One Week of Delivery1 intervention
Trial outcomes
Primary outcomes
1
Difference in Mean Absolute Percent Error (MAPE) in fetal weight estimation
Mean of pairwise differences in absolute percent error between the AI diagnostic tool (index test) and specialist biometry (clinical reference standard), compared against actual birthweight (ground truth).
Time frame
Within 1 week of delivery, 24-42 weeks of gestationSecondary outcomes
1
Proportion of fetal weight estimates within 10% of actual birthweight
Between-method difference in the proportion of estimates within 10% of actual birthweight (ground truth) for the AI diagnostic tool (index test) versus specialist biometry (clinical reference standard).
Time frame
Within 1 week of delivery, 24-42 weeks of gestationSponsors and contacts
Click on the lead sponsor to view all of their trials.
University of North Carolina, Chapel Hill
Lead sponsor
Bill and Melinda Gates Foundation
Collaborator
Source ClinicalTrials.gov