Fetal ultrasound randomised trial of AI-assisted workflow for anomaly detection with health economic assessment

ISRCTN	ISRCTN11223725
DOI	https://doi.org/10.1186/ISRCTN11223725
Integrated Research Application System (IRAS)	355136
Central Portfolio Management System (CPMS)	71420
Protocol serial number	R&D 1388
Sponsor	King's College London
Funder	National Institute for Health and Care Research

Submission date: 30/07/2025
Registration date: 19/02/2026
Last edited: 05/06/2026

Recruitment status: Recruiting
Overall study status: Ongoing
Condition category: Pregnancy and Childbirth

Prospectively registered

Protocol

Statistical analysis plan

Results

Individual participant data

Record updated in last year

Plain English summary of protocol

Background and study aims
This study is exploring whether new computer-based tools can improve pregnancy ultrasound scans. In the UK, pregnant women are usually offered a detailed scan halfway through pregnancy to check how their baby is developing. However, some serious health conditions can be missed. Researchers have developed artificial intelligence (AI) tools that can help spot problems during the scan and support expert review afterwards. The aim is to find out if using both tools together helps detect more problems early and makes clinics run more efficiently.

Who can participate?
Around 9,500 pregnant people will be invited to take part in the study. Participation is entirely voluntary.

What does the study involve?
Participants will be randomly assigned to one of two groups. One group will receive the usual scan, while the other group will have a scan supported by AI and reviewed by an expert. Some participants may also be asked to complete a short survey or take part in an interview to share their views.

What are the possible benefits and risks of participating?
Taking part could help improve how pregnancy scans are done in the future and may lead to earlier detection of health problems in babies. There are no known risks beyond those of a standard scan. Participation will not affect the care participants receive.

Where is the study run from?
King’s College London (UK)

When is the study starting and how long is it expected to run for?
July 2025 to June 2027

Who is funding the study?
National Institute for Health and Care Research (NIHR) (UK)

Who is the main contact?
Professor Reza Razavi, reza.razavi@kcl.ac.uk

Contact information

Prof Reza Razavi
Public, Scientific, Principal investigator

School of Biomedical Engineering and Imaging Sciences
King’s College London
9th floor, Becket House
Lambeth Palace Road
London
SE1 7EU
United Kingdom

ORCID ID	0000-0003-1065-3008
Phone	+44 20 784 89587
Email	reza.razavi@kcl.ac.uk

Study information

Primary study design	Interventional
Study design	Balanced two-arm explanatory randomized controlled trial with parallel process evaluation
Secondary study design	Randomised controlled trial
Scientific title	Fetal ultrasound randomised trial of AI-assisted workflow for anomaly detection with health economic assessment
Study acronym	FRAIYA
Study objectives	Current study objectives as of 05/06/2026: Gather stakeholder opinions on the acceptability of the AI-supported technology, and how this can be integrated into current clinical workflows. Also evaluate the usability of AI disease detection algorithms in routine clinical workflow. Primary Outcomes: 1. Scan duration 2. Detection rate of congenital anomalies Secondary Outcomes: 1. Qualitative assessment of user and patient acceptability 2. The accuracy of AI disease detection models in helping with detection rate of congenital anomalies 3. Diagnostic accuracy metrics (true positives, false positives, true negatives, false negatives, sensitivity, specificity, positive predictive value, and negative predictive value) 4. Referral to specialist fetal medicine or fetal cardiologist 5. Local follow-up scanning (including recalls for repeat scanning) after the 20-week scan 6. Screen-positive rates 7. Positive predictive value of referral Previous study objectives: Can an AI-enabled workflow for ultrasound screening of prenatal abnormalities, including second review, improve detection rates in a cost-effective way? Primary Outcomes: 1. Scan duration 2. Detection rate of congenital anomalies Secondary Outcomes: 3. Qualitative assessment of user and patient acceptability At the end of the study, we’ll know whether this technology helps detect serious conditions earlier, whether it’s acceptable to patients and staff, and whether it could be affordable for the NHS. We’ll share our findings through public summaries, press releases, and scientific conferences.
Ethics approval(s)	Approved 16/12/2025, Yorkshire & The Humber - Leeds East Research Ethics Committee (Health Research Authority, 2 Redman Place, Stratford, London, E20 1JQ, United Kingdom; -; leedseast.rec@hra.nhs.uk), ref: 25/YH/0243
Health condition(s) or problem(s) studied	Ultrasound screening of prenatal abnormalities
Intervention	Pregnant women will be recruited to participate in the trial at the time of their mid-trimester ultrasound anomaly scan. Each participant will be randomised (1 to 1) to have the scan either with AI assistance (FraiyaScan), or in standard unassisted fashion. Those who have been randomised to AI assistance will also have an AI-assistance secondary review of the scan (FraiyaDetect). The scan duration and detection rates of congenital anomalies between the two scanning methods will be compared. A parallel process evaluation will employ a convergent parallel design based on survey and interview data from service users (main trial service users) and the healthcare workforce (end user stakeholders). Interventional and control arms are both going to last 26 weeks from the initial participation (20 week antenatal scan) to final question about the newborn outcome at 6 weeks of age. Randomisation (1-1) is via a digital application provided by the clinical trials unit (KCL CTU) that runs on the tablet attached to the ultrasound machine.
Intervention type	Device
Phase	Not Applicable
Drug / device / biological / vaccine name(s)	FraiyaScan, FraiyaDetect
Primary outcome measure(s)	Current primary outcomes as of 05/06/2026: 1. Scan duration: Time (minutes) to complete 20-week anomaly scan, measured using ultrasound machine timestamps/scan logs at 20-week scan 2. Detection rate of congenital anomalies: Proportion identified on ultrasound, measured against postnatal diagnosis and/or specialist confirmation at 20-week scan and postnatal follow-up Previous primary outcomes: 1. Scan time in minutes (collected by the computer connected to the ultrasound machine that is recording the examination) 2. Correct diagnosis of a congenital abnormality in the newborn (confirmed by the follow-up call at 6 weeks to parents by a member of the research team and recorded on to the eCRF). The initial diagnosis of an abnormality if present is recorded at the time of the scan on the eCRF by the scanning sonographer
Key secondary outcome measure(s)	Added 05/06/2026: 1. User and patient acceptability: Acceptability scores and qualitative feedback, measured using questionnaires/interviews at 20-week scan 2. Accuracy of AI-assisted anomaly detection: Proportion detected with AI vs standard assessment, measured against postnatal diagnosis and/or specialist confirmation at 20-week scan and postnatal follow-up 3. Diagnostic accuracy metrics: calculated from ultrasound findings vs reference standard (postnatal diagnosis/specialist confirmation) at 20-week scan and postnatal follow-up 4. Referral to specialist services: Proportion referred to fetal medicine/cardiology, measured from clinical records at 20-week scan and antenatal follow-up 5. Follow-up scanning (recalls): Proportion requiring additional scans post 20-week scan, measured from imaging records during pregnancy 6. Screen-positive rate: Proportion of scans indicating suspected anomalies requiring further assessment, measured from ultrasound reports at 20-week scan 7. Positive predictive value of referral: Proportion of referred cases confirmed as anomalies, measured against specialist diagnosis and/or postnatal outcomes at referral and postnatal follow-up
Completion date	30/06/2027

Eligibility

Participant type(s)	Patient
Age group	Mixed
Lower age limit	18 Years
Upper age limit	80 Years
Sex	Female
Target sample size at registration	9566
Key inclusion criteria	Current inclusion criteria as of 05/06/2026: For pregnant participants: 1. Booked for routine NHS mid-trimester anatomy ultrasound scan 2. Viable singleton pregnancy 3. Appointment booked in a participating ultrasound room  For healthcare workforce professionals: 1. NHS staff members at participating hospital sites and staff members at industry partner Previous inclusion criteria: 1. Pregnant 2. All cases who are due for a routine mid-trimester (18-22 week) antenatal scan
Key exclusion criteria	Current exclusion criteria as of 05/06/2026: For pregnant participants: 1. Under 18 years of age 2. Multiple pregnancy  3. Confirmed presence of fetal structural anomalies  4. Confirmed presence of fetal chromosomal/genetic anomaly  5. High-risk pregnancy (e.g., under care of fetal medicine/fetal cardiology)  6. Insufficient English language skills to provide informed consent (even with available translation materials)  For healthcare workforce professionals: 1. Staff groups or professionals not involved in ultrasound services; IT or network security; procurement; service or industry managers.  Previous exclusion criteria: 1. Any identified structural abnormality 2. Any genetic or chromosomal abnormality identified 3. Participant withdrawal 4. Refusal of consent 5. Insufficient English language skills to provide informed consent 6. Multiple pregnancies
Date of first enrolment	08/04/2026
Date of final enrolment	05/04/2027

Locations

Countries of recruitment

United Kingdom
England

Study participating centres

Guys and St Thomas's NHS Foundation Trust

Westminster Bridge Road
London
SE1 7EH
England

Lewisham and Greenwitch NHS Trust

Lewisham High St
London
SE13 6LH
England

Liverpool Women's NHS Foundation Trust

Liverpool Womens Hospital
Crown Street
Liverpool
L8 7SS
England

William Harvey Hospital

Kennington Road
Willesborough
Ashford
TN24 0LZ
England

Queen Elizabeth the Queen Mother Hospital

St. Peters Road
Margate
CT9 4AN
England

Results and Publications

Individual participant data (IPD) Intention to share	Yes
IPD sharing plan	The datasets generated and/or analysed during the study will be published as a supplement to the results publication.

Study outputs

Output type	Details	Date created	Date added	Peer reviewed?	Patient-facing?
Study website		11/11/2025	11/11/2025	Yes	Yes

Editorial Notes

05/06/2026: The following changes were made to the study record:
1. The study objectives, primary and secondary outcomes, inclusion and exclusion criteria, study participating centres and ethics approval details and study website were updated.
2. The date of first enrolment was changed from 01/10/2025 to 08/04/2026.
3. The date of final enrolment was changed from 01/10/2026 to 05/04/2027.
30/07/2025: Trial's existence confirmed by the National Institute for Health and Care Research (NIHR) (UK).