Machine learning for diagnosis of myocardial infarction using cardiac troponin concentrations
Study context
- Population
- Patients assessed for possible MIexternal validation, 7 cohorts
- Participants
- 10,286 patients
- Comparison
- Collaboration for the Diagnosis and Evaluation of Acute Coronary Syndrome (CoDE-ACS) score vs fixed cardiac troponin thresholds
- Outcome
- Discrimination for MI
Methodological detailsmachine learning model development, external validation in seven cohorts
- Machine learning model development
- External validation in seven cohorts
Performance
As reported in the abstract
MI diagnosis · at presentation
| Test or model | AUC | Identified as low probability | Identified as high probability |
|---|---|---|---|
| CoDE-ACS | 0.953 (95% CI 0.947-0.958) | 61% | 10% |
| Fixed cardiac troponin thresholds | 27% | 16% |
AUC
CoDE-ACS0.953 (95% CI 0.947-0.958)
Fixed cardiac troponin thresholds— not stated in the abstract
Identified as low probability
CoDE-ACS61%
Fixed cardiac troponin thresholds27%
Identified as high probability
CoDE-ACS10%
Fixed cardiac troponin thresholds16%
Further findings
Cardiac death by CoDE-ACS probability category · 30 days and 1 year
30 days
1 year
- Cardiac death by CoDE-ACS probability category: P < 0.001 for both
Original publication
Machine learning for diagnosis of myocardial infarction using cardiac troponin concentrations. Nat Med. Published 11 May 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →