Comparing Artificial Intelligence and Standard Ultrasound Methods for Estimating Fetal Weight in Pregnancy. Patients Eligible for Inclusion Are Women With a Gestational Age Between 24-42 Weeks Undergoing a Growth Scan. The Image Data From the Scan Are Used to Calculate Fetal Weight.

A Prospective Silent Trial of Artificial Intelligence for Fetal Weight Estimation

Status

Completed

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT06314178

Enrollment

283

Registered

2024-03-15

Start date

2024-07-01

Completion date

2025-12-30

Last updated

2026-06-12

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Pregnancy Complications

Brief summary

The primary aim of this observational study is to compare the accuracy of two artificial intelligence (AI) models with the traditional Hadlock formula for estimating fetal weight from ultrasound scans performed in pregnant women between 24 and 42 weeks of gestation. The secondary aim is to investigate potential demographic bias in the AI models. The demographic factors examined include body mass index (BMI), parity, gestational age, maternal age, fetal sex, and the presence of preeclampsia. Participants' ultrasound scans will be pseudonymized and securely stored on password-protected removable drives to ensure the protection of their identity and privacy. The ultrasound data will subsequently be transferred to the Technical University of Denmark (DTU), where the AI models will analyze the images to estimate fetal weight.

Interventions

None listed

Study design

Observational model

COHORT

Time perspective

PROSPECTIVE

Eligibility

Sex/Gender

FEMALE

Healthy volunteers

Yes

Inclusion criteria

* Women with gestational age between 24-42 weeks undergoing a third-trimester growth scan.

Exclusion criteria

* Women with multiple pregnancies.

Design outcomes

Primary

Measure	Time frame	Description
Comparing the accuracy of the Hadlock formula and the AI model	From enrollment to the birth of the child	The primary objective is to compare the accuracy of fetal weight estimation between the Hadlock formula and two deep learning models in clinical practice

Secondary

Measure	Time frame	Description
Demographic biases	From enrollment to the birth of the child	The secondary objective is to investigate whether the deep learning models show any demographic biases when estimating fetal growth in clinical practice. This is assessed by comparing the accuracy of the Hadlock formula and the deep learning models against the fetal weight at the time of the scan, which is estimated from the birth weight using the Marsal growth curve.

Countries

Denmark

Outcome results