Heart failure & cardiomyopathies
Postdevelopment Performance and Validation of the Artificial Intelligence-Enhanced Electrocardiogram for Detection of Cardiac Amyloidosis
Authors' conclusion
The AI-ECG for the detection of CA maintained an overall strong performance with respect to patient age, sex, race, and amyloid subtype.
Study context
- Population
- Patients with cardiac amyloidosis diagnosed after algorithm development and matched controls with a 12-lead electrocardiogram (ECG)
- Participants
- 440 vs 6,600
- Comparison
- AI-ECG for detection of cardiac amyloidosis
- Amyloidn = 440
- Controlsn = 6,600reference
- Outcome
- AUC of AI-ECG for detection of CA
Methodological detailsvalidation study, 15:1 age- and sex-matched controls
- Validation study
- 15:1 age- and sex-matched controls
Performance
As reported in the abstract
AUC, overall cohort
AI-ECG AUC, overall cohort and between amyloid subtypes (95% CI: 0.82-0.86); slight decrease from original study (AUC 0.91)
0.84
Further findings
AUC by race/ethnicity (subgroup)
- White, Black and "other" races: similar performance
- AUC >0.81
- Hispanic patients: decreased performance
- AUC 0.66
Performance shift over time
- Algorithm performance shift over time not observed
AUC by ECG pattern (subgroup)
- Low ECG voltage and infarct pattern
- AUC >0.90
- LV hypertrophy
- AUC 0.75
- Left bundle branch block
- AUC 0.76
Original publication
Postdevelopment Performance and Validation of the Artificial Intelligence-Enhanced Electrocardiogram for Detection of Cardiac Amyloidosis. JACC Adv. Published 14 September 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →