Automatic CT-based quantification of myocardial extracellular volume fraction and high ECV burden and their prognostic value for death and heart failure in patients with aortic stenosis: a real-world data analysis.

BACKGROUND

Computed tomography (CT)-based myocardial extracellular volume (ECV) assessment during transcatheter aortic valve implantation (TAVI) planning is prognostically informative but limited by operator-dependent workflows. We developed a fully automated, standardized pipeline for volumetric ECV quantification and a derived metric reflecting the burden of myocardial extracellular matrix expansion (hECV-B) and evaluated their association with clinical outcomes after TAVI.

METHODS

In a single-centre study of consecutive pre-TAVI cardiac CT examinations (10/2020-12/2023), a standardized pipeline performed automated cardiac segmentation, reorientation, co-registration and subtraction of late post-contrast and pre-contrast scans, and volumetric ECV mapping. hECV-B was defined as the fraction of left-ventricular voxels above prespecified ECV thresholds. Agreement between automated and manual (two radiologists) ECV was assessed with Bland-Altman analysis. Associations with a composite of all-cause death or heart-failure hospitalization at 12 months were evaluated using Kaplan-Meier/log-rank and multivariable Cox models adjusted for clinical and echocardiographic covariates.

RESULTS

Of 664 screened patients, 438 were analysed (221 women and 217 men; median age 82 years). End-to-end processing was ≤4.5 ​min/patient. The composite outcome occurred in 74 patients (16.9%). Bland-Altman analysis showed negligible mean bias (<1%), with wider limits of agreement (-11.3%-13.1%). Elevated automated whole-LV ECV and hECV-B were independently associated with adverse outcomes, with risk predominantly concentrated at higher ECV values.

CONCLUSION

Fully automated CT-derived ECV and hECV-B assessment is feasible at scale and provides operator-independent markers of myocardial extracellular expansion associated with adverse outcomes after TAVI, supporting its potential role in standardized pre-procedural risk stratification in aortic stenosis.