Machine learning algorithms for predicting arrhythmic events in Hypertrophic Cardiomyopathy: limited enhancement beyond late gadolinium enhancement.

Prognostic significance of artificial intelligence-quantified late gadolinium enhancement on cardiac magnetic resonance in hypertrophic cardiomyopathy

Background/Aims: This study aimed to evaluate the prognostic value of artificial intelligence (AI)-based late gadolinium enhancement (LGE) quantification using cardiovascular magnetic resonance (CMR) in patients with hypertrophic cardiomyopathy (HCM).Methods: We retrospectively analyzed 142 patients with HCM (mean age 58.5 ± 13.7 yr; 72.5% men) who underwent CMR at Yeungnam University Medical Center between 2015 and 2023. LGE was quantified using an AI-based segmentation algorithm with a 6-standard deviation method. Patients were stratified into a high LGE group (≥ 15%) and a low LGE group (< 15%). The primary outcome was a composite of cardiovascular death, including sudden cardiac death (SCD) and SCD-equivalent events.Results: Over a median follow-up of 59 months, the high LGE group experienced a higher incidence of the primary outcome compared with the low LGE group (21.2% vs. 4.6%, p = 0.0067). After adjustment for age and left ventricular ejection fraction, a high AI-quantified LGE burden (≥ 15%) remained independently associated with the primary outcome (adjusted hazard ratio 4.67, 95% confidence interval 1.43–15.30, p = 0.011). Receiver operating characteristic analysis determined an optimal LGE cutoff of 8% for predicting the primary outcome (area under the curve 0.821). Patients with LGE ≥ 8% exhibited markedly increased rates of the primary outcome (18.2% vs. 0%, p < 0.001) and SCD/SCD-equivalent events (13.6% vs. 0%, p = 0.00074).Conclusions: AI-based quantitative assessment of LGE is an independent predictor of adverse cardiovascular outcomes in patients with HCM and may represent a clinically meaningful imaging biomarker for risk stratification.

Enhancing risk prediction for sudden cardiac death: quantitative late gadolinium enhancement analysis in hypertrophic cardiomyopathy

To assess the prognostic value of multiple late gadolinium enhancement (LGE) quantification methods for sudden cardiac death (SCD) prediction in hypertrophic cardiomyopathy (HCM). In this single-center retrospective study, 617 HCM patients who underwent cardiac magnetic resonance (CMR) examinations were consecutively enrolled. LGE was quantified using the n-standard deviation (SD) technique (with multiple thresholds) and the full width at half maximum method. “Gray zone” was defined as the myocardium with signal intensity between two predefined thresholds on LGE images. The primary outcome was SCD or aborted SCD. Among 617 HCM patients, 424 (68.7%) were male, mean age was 49.1 ± 14.0 years, and LGE was identified in 438 (71.0%). During 67.9 ± 21.0 months of follow-up, 24 patients (3.9%) reached the primary endpoint. All dual-threshold gray zone measures and five single-threshold LGE quantifications (2–6 SDs) independently predicted SCD (HR range: 1.031–1.220, all adjusted p < 0.05). Gray zone (HR range: 1.091–1.220) showed stronger associations than single-threshold LGE measures (HR range: 1.031–1.046). Both approaches improved discrimination beyond the composite American Heart Association (AHA)/American College of Cardiology (ACC) model, with increased net reclassification improvement (0.321 and 0.397, respectively) and comparable C-statistics (0.794–0.871 vs 0.794–0.898, p = 0.155). While conventional single-threshold LGE quantification predicts SCD in HCM, dual-threshold gray-zone analysis demonstrates stronger associations. Both approaches provide incremental prognostic value beyond the AHA/ACC guideline-based risk model. These findings underscore the importance of standardizing LGE quantification to improve SCD risk stratification in patients with HCM. Question: The optimal LGE quantification approach for SCD risk stratification in HCM remains uncertain. Findings: Both single- and dual-threshold LGE quantification independently predict SCD in patients with HCM, while dual-threshold gray zone metrics demonstrate stronger prognostic associations. Critical Relevance: By systematically comparing multiple LGE quantification approaches, this study highlights the impact of methodological variability on SCD risk stratification in HCM, underscores the need for standardized LGE assessment, and informs clinical radiology practice. Question: The optimal LGE quantification approach for SCD risk stratification in HCM remains uncertain. Findings: Both single- and dual-threshold LGE quantification independently predict SCD in patients with HCM, while dual-threshold gray zone metrics demonstrate stronger prognostic associations. Critical Relevance: By systematically comparing multiple LGE quantification approaches, this study highlights the impact of methodological variability on SCD risk stratification in HCM, underscores the need for standardized LGE assessment, and informs clinical radiology practice.

Prognostic Value of Native T1 Mapping in Nonischemic Dilated Cardiomyopathy: A Prospective Observational Study from India

Background Native T1 mapping can detect diffuse myocardial fibrosis in nonischemic dilated cardiomyopathy (NIDCM), which may not be visible on late gadolinium enhancement (LGE) imaging. We assessed the prognostic relevance of native T1 mapping and its incremental value over LGE in risk stratification of NIDCM patients. Materials and Methods This prospective observational study enrolled 61 NIDCM patients who underwent cardiac magnetic resonance imaging. Patients were followed up for major adverse cardiac events (MACE), including heart failure hospitalization, sudden cardiac death, and all-cause mortality. Native septal and global T1 values were quantified. Receiver operating characteristic (ROC) curve analysis identified optimal T1 cut-off values for predicting MACE. Results During a median follow-up of 14 months, 23 patients (37.7%) experienced MACE. Mean native septal T1 values were significantly higher in the MACE group compared to the no-MACE group (1,106.0 ± 55.03 ms vs. 1,069.41 ± 46.8 ms, p  = 0.007). ROC analysis identified a septal T1 value of 1,082.8 ms as the optimal threshold (sensitivity: 73.91%, specificity: 63.16%). On multivariable regression analysis, native septal T1 remained an independent predictor of MACE (odds ratio = 4.2, 95% confidence interval: 1.01–18.07, p  = 0.04). Among 30 LGE-negative patients, 11 (36.7%) had elevated septal T1 values above the cut-off, indicating additional risk stratification capability. Conclusion Native septal T1 mapping provides independent prognostic information in NIDCM patients beyond LGE. It identifies high-risk patients even in the absence of focal fibrosis, offering incremental value for risk-stratification and clinical decision-making. Advances in Knowledge Native T1 mapping independently predicts MACE beyond ejection fraction and LGE. One in three LGE-negative patients demonstrate high-risk diffuse disease, highlighting complementary roles of these techniques.

Left Atrial Reservoir Strain Improves Sudden Cardiac Death Risk Stratification in Hypertrophic Cardiomyopathy.

BACKGROUND Current risk scores for sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HCM) have limited ability to identify high-risk subgroups. We aimed to investigate the prognostic value of left atrial reservoir strain (LARS) for SCD-related events and its utility for risk stratification. METHODS This retrospective cohort study included 1,761 patients with HCM from two referral centers. The primary outcome was SCD-related events, including SCD, aborted SCD, and appropriate implantable cardioverter-defibrillator shocks. Explainable machine learning approaches were used to explore the importance of LARS and identify a clinically relevant threshold. The prognostic value of LARS was evaluated using Cox regression analyses, particularly among patients classified as low-to-intermediate risk (HCM Risk-SCD score<6%). RESULTS During a median follow-up of 6.8 years (IQR: 3.0-10.7 years), 69 (3.9%) SCD-related events occurred. Decreased LARS was independently associated with a higher risk of SCD-related events (per 1% decrease, adjusted HR 1.11, 95% CI, 1.07-1.15, p<0.001). SHAP analysis identified LARS as the top-ranked predictor and a clinically relevant threshold of <21%. LARS <21% was significantly associated with a higher risk of SCD-related events (adjusted HR 5.01, 95% CI 2.77-9.03, p<0.001), even among patients without atrial fibrillation. Adding LARS to the HCM Risk-SCD score significantly improved risk discrimination in the low-to-intermediate risk population (5-year time-dependent AUC 0.73 vs. 0.63, p=0.005). Among low-to-intermediate risk patients, LARS <21% was associated with a significantly higher risk of SCD-related events (adjusted HR 6.13, 95% CI 3.13-12.03, p<0.001), whereas the original low- and intermediate-risk categories showed limited risk discrimination. CONCLUSIONS LARS was an independent predictor of SCD-related events and effectively stratified risk among patients classified as low-to-intermediate risk by the HCM Risk-SCD score. Integrating LARS into the HCM Risk-SCD score may enhance risk stratification and guide preventive strategies.

Incremental prognostic value of late gadolinium enhancement in hypertrophic cardiomyopathy in a large contemporary cohort.

PVCuRe: a machine-learning based tool to predict left ventricular systolic function recovery in patients undergoing ablation for premature ventricular contraction.

BACKGROUND Ablation of frequent premature ventricular complexes (PVCs) can improve left ventricular ejection fraction (LVEF) in patients with systolic dysfunction, especially in suspected PVC-induced cardiomyopathy. However, many patients fail to normalize LVEF despite successful ablation, and current tools do not reliably distinguish true PVC-induced cardiomyopathy from underlying cardiomyopathy exacerbated by PVCs. OBJECTIVE To develop and externally validate a machine learning (ML) model using routinely available clinical, echocardiographic, and electrocardiographic variables to predict LVEF recovery after PVC ablation. METHODS In this retrospective multicenter study, 256 patients with LVEF <50% undergoing successful PVC ablation at three international referral centers were included. Predictors were selected using the Boruta algorithm, and five ML models were trained. Performance was assessed with 10-fold cross-validation and ROC curve analysis. The best-performing model underwent calibration and threshold analysis and was externally validated in an independent cohort from three additional centers. RESULTS The Random Forest model showed the best performance, with an AUC of 0.88 (95% CI 0.79-0.98) in the internal test set and good calibration (Hosmer-Lemeshow p=0.562). External validation confirmed consistent discrimination (AUC 0.83, 95% CI 0.72-0.95). Key predictors included baseline PVC burden, QRS duration in sinus rhythm, and preprocedural LVEF. CONCLUSION This ML-based tool, built on widely available variables, accurately estimates the probability of LVEF recovery after PVC ablation and may support clinical decision-making and patient counselling.