Heart failure & cardiomyopathiesRandomised controlled trial
Using Artificial Intelligence to Adjudicate Major Adverse Cardiovascular Events in Clinical Trials
- Validation study of an artificial intelligence (AI)-based adjudication system within the PARADISE-MI randomised trial
Authors' conclusion
AI-based adjudication of MACE showed high agreement with human CEC adjudication, especially for CV death and stroke, and where the model was confident.
Study context
- Population
- Patients with myocardial infarction (MI) complicated by systolic dysfunction or pulmonary congestion in the PARADISE-MI global trial (sacubitril/valsartan vs ramipril)PARADISE-MI
- Participants
- 5,661 patients
- Comparison
- Auto-MACE vs clinical events committee (CEC) adjudication
Primary result
As reported in the abstract
Agreement of Auto-MACE with CEC adjudication
- Confident adjudication: deaths
- 315/455 (69%)
- Confident adjudication: potential MIs
- 301/659 (46%)
- Confident adjudication: potential strokes
- 136/167 (81%)
- Agreement with CEC, confident events: deaths
- 97%
- Agreement with CEC, confident events: potential MIs
- 89%
- Agreement with CEC, confident events: potential strokes
- 88%
- Agreement with CEC, all events: deaths
- 86%
- Agreement with CEC, all events: potential MIs
- 76%
- Agreement with CEC, all events: potential strokes
- 84%
Further findings
| Outcome | HR (95% CI) | |
|---|---|---|
| Effect of sacubitril/valsartan vs ramipril on composite MACE, by adjudication method — Sacubitril/valsartan vs Ramipril · Auto-MACE adjudication | 0.91 (0.78–1.07) | |
| Effect of sacubitril/valsartan vs ramipril on composite MACE, by adjudication method — Sacubitril/valsartan vs Ramipril · CEC adjudication | 0.90 (0.77–1.05) | |
| 0.512Hazard ratio (log scale) |
Effect of sacubitril/valsartan vs ramipril on composite major adverse cardiovascular events (MACE), by adjudication method
- Similar estimated effect with Auto-MACE and CEC adjudication
Original publication
Using Artificial Intelligence to Adjudicate Major Adverse Cardiovascular Events in Clinical Trials. J Am Coll Cardiol. Published 9 November 2025. doi:10.1016/j.jacc.2025.10.055
Summarised from the published abstract; every number is checked against it