Genetic test yield in non-dilated left ventricular cardiomyopathy
A Faggiano, M Sanfilippo, M Calcagnigno, F Moda, C Giordano, A Bernareggi, M Ruscica, A Pollina, M Saviano, M Zilocchi, E Gherbesi, P Castronovo, P Finelli, S CarugoAbstract
Background
The 2023 ESC guidelines introduced nondilated left ventricular cardiomyopathy (NDLVC) as an umbrella entity including a spectrum of sub-phenotypes ranging from isolated non-ischemic scar/fat infiltration to forms associated with left ventricular (LV) dysfunction. Although genetic testing is recommended for all patients, milder presentations—such as isolated minor scar, particularly when asymptomatic—may represent bystander or post-myocarditic findings rather than true cardiomyopathy.
Purpose
We aimed to evaluate the diagnostic yield of genetic testing across NDLVC sub-phenotypes and to determine whether the Madrid-DCM score (validated for predicting genetic positivity in dilated cardiomyopathy) and the 2025 NDLVC-risk score (5-year arrhythmic risk) could identify patients with actionable variants.
Methods
We retrospectively included patients diagnosed with NDLVC according to ESC criteria who underwent NGS-based virtual panel testing in 2024–2025 at a single Italian tertiary center. Pathogenic or likely pathogenic variants were considered actionable, whereas variants of uncertain significance or benign variants were classified as non-actionable. Patients were stratified into four NDLVC sub-phenotypes: isolated minor scar, isolated LV dysfunction, extensive scar/fat infiltration, and combined LV dysfunction plus scar/fat. Genetic yield was compared among groups, and the predictive performance of the Madrid-DCM and NDLVC scores was assessed using logistic regression and ROC analysis.
Results
Among 50 consecutive patients (48±19 years, 69% male), actionable variants were identified in 14 (28%), predominantly involving DSP (6/14, 43%) and LMNA (3/14, 21%). Genetic yield varied substantially across sub-phenotypes (Figure 1): isolated minor scar (2/18, 11.1%, reference), isolated LV dysfunction (2/9, 22.2%; OR 2.3, p=0.053), LV dysfunction plus scar/fat (7/17, 41.2%; OR 5.6, p=0.043), and extensive scar/fat (3/6, 50.0%; OR 8.00, p=0.046). The NDLVC-risk score was significantly higher in patients with actionable variants (10.3 ± 7.1% vs 3.6 ± 2.3%, p=0.041) and showed good discriminative ability (Figure 2, AUC 0.73; OR per unit increase 1.2, p=0.043). Conversely, the Madrid-DCM score was not associated with actionable variants (35.2±12% vs 32.3±15%, p=0.57; AUC 0.57; OR 1.02, p=0.73).
Conclusions
The yield of genetic testing across the NDLVC spectrum is highly heterogeneous, being very low in the common phenotype of isolated minor scar and higher in more complex presentations. A uniform (“one-size-fits-all”) strategy for genetic testing may therefore have limited utility in NDLVC, whereas careful phenotypic characterization combined with clinical risk stratification appears crucial to optimize patient selection for genetic testing.