Efficiency of AI Reporting (EAIR). (EAIR)

Evaluation of the Efficiency and Accuracy of Artrya-Salix AI Software in the Reporting of Coronary Computed Tomography Angiography Scans in a Real-world NHS Setting. A Retrospective Study.

While there are features on a CCTA that are correlated with plaque vulnerability and can be reported on, this is not always done. This is important as it is thought that vulnerable plaques present the greatest risk of myocardial infarction. This study will assess the correlation between software and human classification of these plaques, as well as demonstrating the time efficiency of AI reporting when compared to a clinician.

Study Overview

• Status: Withdrawn
• Conditions: Coronary Artery Disease

Detailed Description

This is a retrospective observational study. It will be performed at London Northwest Healthcare NHS Trust (LNWH -Northwick Park Hospital (NPH), Harrow and Ealing Hospital, Southall).

Patients who have had a CCTA in the last 12 months will be contacted about the study. Their CCTA scans will already have been reported and the patient aware of the results. After giving informed consent, 100 patients will have their clinical CCTA images uploaded to the Artrya Salix software for analysis. Simultaneously the images will be re-reported by the physician as per standard of care. Both reports will be obtained and any differences noted. The time the images are sent to Artrya Salix and report received will be recorded. The time the clinician opens the image to analyse and the time when the report is complete will also be recorded using a stopwatch. All images will be analysed by both expert readers to take account of inter-observer variability. The first 30 patients will be treated as a pilot for the study and after their recruitment and image upload and analysis, the team will meet to discuss any issues that have arisen.

During the analysis, any patient who has plaque identified by the software or expert reader will have further analysis of their images. This is to look at the plaques in detail and to compare the sensitivity and specificity of the expert reader analysis. The time it takes for the clinician to report the scans on this group of patients will also be recorded. In addition, assessment will be made of the plaques using comparison of stenosis to calcium score.

.

Data that will be collected will include:

1. Demographic data (age, gender, ethnicity)
2. CCTA results - standard of care
3. CCTA results - Artrya Salix
4. Time from image upload to Artrya Salix to time report received from Artrya Salix
5. Time clinician takes to analyse scan
6. Clinical assessment of plaque for vulnerable characteristics.

The data will be from the electronic medical records of LNWUH NHS Trust. A level 2 or level 3 Cardiology imaging or Radiology Consultant will report the CCTA images.

Data will be captured on an eCRF. The time of image upload to Artrya Salix will be recorded by the system and transposed directly into the eCRF. The time starting and ending image analysis by the clinician will be recorded on a worksheet as source documentation and then entered into the eCRF. A score will be given by the physician as to the likelihood of a plaque being vulnerable where 1= unlikely and 5 = highly likely.

• Study Type: Observational

Contacts and Locations

Study Locations

• United Kingdom
  • Middx
    • Harrow, Middx, United Kingdom, HA1 3UJ
      • London North West University Healthcare NHS Trust

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

Patients who have been referred for CCTA over the last 12 months will be invited to take part. The rationale is to test this software in a real-world setting and this method will provide a non-selective sample. No patient will be contacted until their vital status has been checked on the national SPINE database.

Description

Inclusion Criteria:

• Clinical CCTA
• Able to provide informed consent

Exclusion Criteria:

• Unable / unwilling to give informed consent
• Previous CABG or stent insertion

Study Plan

How is the study designed?

Design Details

• Observational Models: Cohort
• Time Perspectives: Retrospective

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Speed of reporting	1. Time taken for Artrya Salix reporting of a CCTA scan is as fast or faster than clinician reporting	Through study completion, up to 26 weeks
Vulnerable plaque identification	2. In a subset of patients where the software has identified vulnerable plaques, this is in agreement with the expert clinical report.	Through study completion, up to 26 weeks

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Speed assessing plaque vulnerability	The secondary outcome of this study will be the additional time taken by the clinician to report plaque vulnerability characteristics.	Through study completion, up to 26 weeks

Collaborators and Investigators

• Sponsor: London North West Healthcare NHS Trust
• Investigators: Principal Investigator: Navtej Chahal, PhD FESC, LNWUH NHS Trust

Study record dates

Study Major Dates

• Study Start (Estimated): August 15, 2024
• Primary Completion (Estimated): February 9, 2025
• Study Completion (Estimated): March 9, 2025

Study Registration Dates

• First Submitted: May 30, 2022
• First Submitted That Met QC Criteria: October 13, 2022
• First Posted (Actual): October 18, 2022

Study Record Updates

• Last Update Posted (Actual): May 23, 2025
• Last Update Submitted That Met QC Criteria: May 21, 2025
• Last Verified: May 1, 2025

More Information

Terms related to this study

• Keywords: Coronary Artery Disease; Artificial Intelligence; vulnerable plaque; atherosclerotic plaque; CCTA
• Additional Relevant MeSH Terms: Vascular Diseases; Cardiovascular Diseases; Heart Diseases; Arteriosclerosis; Arterial Occlusive Diseases; Coronary Artery Disease; Myocardial Ischemia; Coronary Disease
• Other Study ID Numbers: CRC/2021/003

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: dat will be fully anonymous and will remain within the study team.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No