Valvular & structural heart disease
An Artificial Intelligence Algorithm for Detection of Severe Aortic Stenosis: A Clinical Cohort Study
Authors' conclusion
Without relying on left ventricular outflow tract measurements, an AI-DSA used echocardiographic reports to reliably identify the phenotype of severe AS.
Study context
- Population
- US Medicare beneficiaries with routine TTE reports at an academic medical center, 2003–2017
- Participants
- 31,141 beneficiaries
- Outcome
- Detection of the phenotype associated with AVA<1 cm² (severe AS)
Methodological detailssingle academic medical centre
- Single academic medical centre
Performance
As reported in the abstract
Detection of phenotype associated with AVA<1 cm²
| Test or model | Sensitivity | Specificity | NPV | C-statistic |
|---|---|---|---|---|
| Artificial intelligence decision support algorithm (AI-DSA) | 82.2% | 98.1% | 9.2% | 0.986 |
Artificial intelligence decision support algorithm (AI-DSA)
Sensitivity82.2%
Specificity98.1%
NPV9.2%
C-statistic0.986
Further findings
| Outcome | Values |
|---|---|
| Additional individuals identified: guideline-defined moderate AS with severe-AS-like phenotype | AI-DSA 3.3% · Aortic valve replacement 6.6% |
| Five-year mortality · 5 years | Known severe AS 75.9% · Similar phenotype to severe AS 73.5% · Without severe AS 44.6% |
Performance in depressed LVEF
- AI-DSA continued to perform well to identify severe aortic stenosis (AS) with depressed LVEF
Aortic valve replacement in aortic valve area (AVA)<1 cm²
- Overall AVR rates remained low, even with AVA<1 cm²
- 21.9%
Original publication
An Artificial Intelligence Algorithm for Detection of Severe Aortic Stenosis: A Clinical Cohort Study. JACC Adv. Published 25 September 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →