Deep-learning analysis of 3D microarchitectural remodeling in hypertrophic cardiomyopathy.

Summary

Using CaMVIA-3D, the authors quantified genotype-specific 3D microarchitectural remodeling in HCM, showing greater concentric hypertrophy and disarray in pathogenic variant cases and predominant fibrosis in variant-negative cases. Longitudinal profiling in a pig HCM model revealed early fibrosis preceding cardiomyocyte hypertrophy, and multi-omic integration nominated genes linked to cellular and extracellular remodeling.

Key Findings

• Developed CaMVIA-3D, a deep-learning volumetric imaging/analysis pipeline for cardiac microarchitecture.
• HCM hearts exhibited genotype-specific remodeling: pathogenic variants showed greater concentric hypertrophy/disarray; variant-negative cases showed predominant fibrosis.
• In a longitudinal pig HCM model, fibrosis emerged before cardiomyocyte hypertrophy.
• Integrated transcriptomic and morphologic data identified genes linked to cellular and extracellular remodeling.

Clinical Implications

3D microarchitectural signatures and early fibrosis detection could refine phenotyping, timing of intervention, and genotype-tailored therapy development in HCM; identified genes may inform biomarker and target discovery.

Limitations

• Translational generalizability to in vivo clinical imaging and outcomes remains to be established.
• Sample sizes and tissue sourcing heterogeneity were not detailed in the abstract.

Future Directions

Prospective clinical validation linking 3D microarchitectural metrics to outcomes; development of noninvasive proxies; testing antifibrotic, genotype-tailored therapies guided by identified pathways.