Valvular & structural heart disease
CT-Based Deep Learning Assessment of Aortic Valve Fibrosis and Calcification: Development and Multicenter External Validation Against Experts, Histology, and Outcomes
Authors' conclusion
Deep learning enables rapid quantification of aortic valve fibrotic and calcific tissue from CTA, demonstrating excellent agreement with both expert CTA and histological assessments. Quantitative analysis of aortic tissue composition provides additional prognostic value for all-cause mortality after transcatheter aortic valve replacement in multicenter external evaluation.
Study context
- Population
- Patients who underwent computed tomography angiography (CTA), including patients undergoing transcatheter aortic valve replacement for severe aortic stenosispatients who underwent CTA at 9 sites · training 468 · external test set 1,206 · with valve histology 29 · with serial CTA 12 · with clinical outcomes 1,165
- Participants
- 1,674 patients
- Outcome
- Analysis time
Methodological details9 sites, external test set
- 9 sites
- External test set
Main finding
As reported in the abstract
Deep learning
1.35 ± 0.4 min
Experts
7.2 ± 1.2 min
Further findings
Independent predictors of mortality
| Estimate | P value | |
|---|---|---|
| Fibrotic tissue volume | aHR 1.42 per IQR increase | P < 0.001 |
| Calcific tissue attenuation | aHR 1.02 per 25 HU increase | P = 0.005 |
| Left ventricular ejection fraction | aHR 1.10 per 5% decrease | P < 0.001 |
| Life-threatening bleeding | aHR 1.96 | P = 0.003 |
| Major vascular complication | aHR 1.51 | P = 0.042 |
| Postprocedural peak pressure gradient | aHR 1.03 per 1 mm Hg increase | P < 0.001 |
| Outcome | aHR | P value |
|---|---|---|
| Risk of death by fibrotic tissue volume and calcific tissue attenuation2-fold higher risk of death; comparator not stated | 2.0 | P < 0.001 |
- Risk of death by fibrotic tissue volume and calcific tissue attenuation — Dense calcific tissue attenuation (>885 HU): risk of death
- 1.5-fold higher
- Independent predictors of mortality: adjusting for clinical, echocardiographic, and procedural covariates
Agreement of deep learning tissue composition analysis
- vs expert assessment r
- 0.915
- vs expert assessment ICC
- 0.913
- vs histology r
- 0.909
- vs histology ICC
- 0.905
- Scan-rescan reproducibility r
- 0.999
- Scan-rescan reproducibility ICC
- 0.999
Mortality in the external cohort · 4.1 ± 1.1 years of follow-up
- Patients who died
- 469 patients (40%)
Original publication
CT-Based Deep Learning Assessment of Aortic Valve Fibrosis and Calcification: Development and Multicenter External Validation Against Experts, Histology, and Outcomes. JACC Cardiovasc Imaging. Published 4 November 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →