Valvular & structural heart disease
Automated Aortic Valve Calcification Scoring: Multicenter External Validation of a Deep Learning Algorithm
Main finding · Pearson correlation and Bland-Altman analysis of algorithm-estimated vs manual Agatston scoresResult — go to the resultMean difference 5.2 AU 95% CI −7.6 to 17.9Authors' conclusion
This automated deep learning algorithm quantifies AVC from routine chest CT with expert-comparable performance, potentially enabling opportunistic AS screening. Prospective studies are needed to determine whether it improves clinical outcomes.
Study context
- Population
- Patients undergoing nongated, noncontrast chest computed tomography (CT)Separate training, validation and holdout test sets · training 1,446 · validation 361 · holdout test set 239
- Comparison
- algorithm-estimated vs manual Agatston scores
- Outcome
- Pearson correlation and Bland-Altman analysis of algorithm-estimated vs manual Agatston scores
Methodological detailsmulticenter consortium of 33 sites spanning 8 health systems in the United States and Brazil, separate holdout test set acquired from 2021 to 2024
- Multicenter consortium of 33 sites spanning 8 health systems in the United States and Brazil
- Separate holdout test set acquired from 2021 to 2024
Main finding
As reported in the abstract
Pearson correlation and Bland-Altman analysis of algorithm-estimated vs manual Agatston scores
Mean difference
5.2 AU
95% CI −7.6 to 17.9
Algorithm vs Manual (expert reference standard)
mean difference 5.2 AU (95% CI −7.6 to 17.9)
- Pearson correlation r
- 0.99
Secondary outcomes
| Outcome | Algorithm | Manual (expert reference standard) | |
|---|---|---|---|
| Sensitivity and specificity for moderate-to-severe AS (sex-specific thresholds) · secondary outcome | |||
| Sensitivity | 0.92 (95% CI: 0.80-0.97) | Not reported in the abstract | Manual (expert reference standard): not reported in the abstract |
| Specificity | 0.98 (95% CI: 0.95-0.99) | Not reported in the abstract | Manual (expert reference standard): not reported in the abstract |
Performance across demographic subgroups and CT technical parameters
- Performance remained consistent across demographic subgroups and CT technical parameters
Original publication
Automated Aortic Valve Calcification Scoring: Multicenter External Validation of a Deep Learning Algorithm. JACC Adv. Published 23 September 2026. doi:10.
Summarised from the published abstract; every number is checked against it