Heart tEam Decision-making Assisted by Revaz AI Trial

NCT07837765 | Not Yet Recruiting

• 1 other identifier: REB26-1546
• Study Type: interventional
• Target: 712
• Locations: 1 country
• Sites: 2

Brief Summary

Coronary artery disease, the narrowing of the blood vessels that supply the heart, is the leading cause of death worldwide. About 2.4 million Canadian adults live with it. There are three main ways to treat it: opening the narrowed artery with a stent, heart bypass surgery, or medications alone. For many patients the best choice is clear, but for people with complicated disease or multiple health conditions, it can be genuinely difficult to know which treatment will lead to the best outcome. For these complex cases, hospitals bring together a "heart team", a group of heart specialists including cardiologists and heart surgeons, to discuss the case and recommend the best treatment. Even so, studies show that different doctors, teams, and hospitals often make different decisions for similar patients, which suggests there is room to make these decisions better. The investigators developed a computer program called Revaz AI. It uses artificial intelligence (AI) that has learned from the health records of more than 38,000 past patients with coronary artery disease in Alberta. When given a new patient's health information, Revaz AI estimates that person's risk of dying or having a major heart problem, such as a heart attack, stroke, heart failure, or the need for another procedure, over the next 90 days to 5 years, under each of the three treatment options. In testing on past patient records, Revaz AI's predictions were more accurate than the risk scores doctors currently use. A separate study estimated that using it could improve treatment choices and reduce health care costs. However, Revaz AI has never been tested in real time with real patients. The central question of this trial is: Do patients do better over 5 years when the heart team can see Revaz AI's risk estimates during its discussion, compared with the usual way of deciding? The investigators will also look at whether the tool changes treatment decisions, whether it makes the team more confident in its recommendations, whether patients feel better in daily life, and whether it saves the health care system money. The study will take place at two Canadian hospitals: Foothills Medical Centre in Calgary and the University of Ottawa Heart Institute. Together, their heart teams discuss about 900 patients each year. The investigators plan to enroll 712 adult patients whose cases have been referred to the heart team. Each patient who agrees to join will be assigned by chance, like a coin flip, to one of two groups. In one group, Revaz AI's risk estimates for that patient will be shown on screen during the heart team's virtual meeting, alongside all the usual information. In the other group, the heart team will discuss the case in the usual way, without Revaz AI. Assigning patients by chance ensures the two groups are alike, so any difference in results can be traced to Revaz AI. Importantly, Revaz AI does not make decisions: in both groups, the heart team makes the recommendation, and each patient and their doctor make the final treatment choice together, just as they do now. The investigators will then follow every patient for 5 years. Most of the follow-up information - hospital stays, procedures, and major heart problems - will come from existing health databases, so patients do not need extra hospital visits. The investigators will also contact patients four times (at about 3 months, 1 year, 3 years, and 5 years) with short questionnaires about their symptoms and quality of life. The main measure is the proportion of patients who have a major heart problem or die within 5 years of treatment. Based on past data, about 40% of similar patients experience one of these events within 5 years. The study is designed to detect whether Revaz AI can reduce this to 30%. Independent statisticians will analyze the results, and an independent safety board of experts not involved in the study will review the data regularly and can stop the study if any safety concern appears. The study is led by university researchers with heart specialists at both hospitals. The lead researcher who co-founded the company that makes Revaz AI will stay at arm's length from the study's conduct and analysis, which will be handled by independent team members. All patient information will be coded and stored on secure hospital and university servers. No AI tool for choosing coronary artery disease treatment has ever been tested in a rigorous trial like this. If Revaz AI improves outcomes, patients with complex heart disease could receive treatment decisions tailored to them, and the tool could be adopted by hospitals across Canada and beyond. If it does not, the study will still teach the medical community valuable lessons about how doctors and AI tools work together - knowledge that will shape the safe use of AI in health care.

Conditions

Coronary Artery Disease

Keywords

artificial intelligence; clinical decision support; coronary artery disease treatment selection

Outcome Measures

Primary Outcomes (1)

• 5-year post-treatment 5-point MACE

  5-point MACE includes myocardial infarction, heart failure, stroke, repeat revascularization, and all-cause mortality.

  5 years post-treatment

Secondary Outcomes (8)

• 90 days, 1 year, and 3 years post-treatment 5-point MACE

  90 days, 1 year, and 3 years post-treatment

• Patient-reported Health-related Quality of Life (SAQ-7) at 90 days, 1 year, 3 years, and 5 years post-treatment

  90 days, 1 year, 3 years, and 5 years post-treatment

• Patient-reported Health-related Quality of Life (EQ-5D-5L) at 90 days, 1 year, 3 years, and 5 years post-treatment

  90 days, 1 year, 3 years, and 5 years post-treatment

• The lead clinician's level of confidence in the heart team treatment recommendation

  Immediately after the heart team meeting

• Concordance between the heart team's final treatment recommendation and the treatment predicted by Revaz AI to yield the lowest 5-year MACE risk

  Immediately after the heart team meeting

Study Arms (2)

Revaz AI-supported coronary artery disease treatment decision-making

EXPERIMENTAL

Coronary artery disease treatment decision-making will be supported by Revaz AI insights

Behavioral: Revaz AI

Control

NO INTERVENTION

Usual coronary artery disease treatment decision-making without Revaz AI support

Interventions

Revaz AI BEHAVIORAL

Revaz AI is an AI-enabled coronary artery disease treatment decision support tool that predicts short-term and long-term MACE and mortality outcomes conditioned on treatment.

Revaz AI-supported coronary artery disease treatment decision-making

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• adult patients (≥18 years) referred for heart team discussion of coronary artery disease treatment at a participating center
• diagnosis of obstructive coronary artery disease, defined as any stenosis ≥50% in the left main coronary artery and/or ≥70% in any other coronary artery, as confirmed by diagnostic angiography

You may not qualify if:

• any subsequent heart team referral for patients who are already enrolled in the trial
• Heart Team Clinician Eligibility:
• any staff clinician (interventional cardiologist, cardiac surgeon, or non-invasive cardiologist) who participates in heart team discussions and contributes to coronary artery disease treatment recommendations at the participating centers
• trainees such as fellows or residents

Sponsors & Collaborators

• University of Calgary: lead
• University of Ottawa Heart Institute, Canada: collaborator
• Foothills Medical Centre: collaborator
• Alberta Health services: collaborator

Study Sites (2)

Foothills Medical Centre

Calgary, Alberta, T2N 4Z6, Canada

Location

University of Ottawa Heart Institute

Ottawa, Ontario, K1Y 4W7, Canada

Location

MeSH Terms

Conditions

Coronary Artery Disease

Condition Hierarchy (Ancestors)

Coronary Disease; Myocardial Ischemia; Heart Diseases; Cardiovascular Diseases; Arteriosclerosis; Arterial Occlusive Diseases; Vascular Diseases

Central Study Contacts

Joon Lee

CONTACT

403-220-2968; joon.lee@ucalgary.ca

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: OTHER
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: September 11, 2026
• First Posted: September 24, 2026
• Study Start (Estimated): April 1, 2027
• Primary Completion (Estimated): September 1, 2033
• Study Completion (Estimated): September 1, 2033
• Last Updated: September 24, 2026

Record last verified: 2026-09

Data Sharing

• IPD Sharing: Will not share

Locations

CA (2)