One size does not fit all: rethinking the 15-mm threshold in hypertrophic cardiomyopathy

Background

Hypertrophic cardiomyopathy (HCM) diagnosis continues to rely on a fixed maximum wall thickness (MWT) threshold ≥15 mm, a universal cut-off that disregards sex and body size and may introduce systematic diagnostic bias, particularly in women and individuals with lower body surface area (BSA). Although demographic-adjusted thresholds derived from large cardiovascular magnetic resonance (CMR) datasets have been proposed, real-world validation against genotype remains limited.

Purpose

To assess whether demographic-adjusted CMR-derived MWT improves diagnostic classification and genotype alignment in patients with presumed HCM.

Methods

We performed a cross-sectional analysis of 80 consecutive patients clinically managed as presumed HCM, all undergoing CMR and genetic testing.

End-diastolic MWT was measured using artificial intelligence–based automated segmentation across the 16-segment model. For external validation, we applied the Demographic-adjusted Left Ventricular Hypertrophy thresholds from the PRECISION-HCM study, using the upper 95% prediction interval (≈2 SD) as the abnormality cut-off. Agreement and reclassification versus the ≥15 mm criterion were assessed (Cohen’s κ).

Genotype (pathogenic/likely pathogenic variant, P/LP) served as biological reference. Discriminative performance of absolute MWT, indexed MWT/BSA, the ≥15 mm criterion, and the demographic-adjusted criterion was evaluated using logistic regression and ROC analysis. An optimal MWT/BSA cut-off was derived using the Youden.

Results

Demographic-adjusted thresholds showed good agreement with the ≥15 mm criterion (κ = 0.65), with 86.2% concordant classifications. Reclassification occurred in 13.8% of patients, exclusively as upclassification. Reclassified patients were predominantly female (81.8% vs 30.2%) and had lower BSA (1.71 ± 0.20 vs 1.92 ± 0.23 m²).

Using genotype as biological reference, the adjusted criterion identified all P/LP carriers (20/20; sensitivity 100%), whereas the ≥15 mm criterion missed four (sensitivity 80%). Specificity was modest in this enriched cohort (25% vs 36.7%).

Discriminative performance for genotype was moderate across metrics: AUC 0.66 for indexed MWT, 0.64 for absolute MWT, 0.63 for demographic-adjusted criterion, and 0.58 for ≥15 mm. Differences were not statistically significant.

The optimal MWT/BSA cut-off was ≥8.33 mm/m² (sensitivity 75%, specificity 55%). Bootstrap validation confirmed stability (median 8.33 mm/m²; 95% CI 6.97–12.48).

Conclusion

A fixed 15-mm threshold introduces measurable sex- and size-related bias. Demographic-adjusted wall thickness improves sensitivity for genetically defined HCM and corrects systematic under-recognition in women and smaller individuals. Although wall thickness alone shows only moderate genotype discrimination, personalised thresholds represent a biologically aligned step towards more precise and equitable HCM diagnosis.