Improving the clinical understanding of hypertrophic cardiomyopathy by combining patient data, machine learning and computer simulations: A case study

Figures

• 

  Figure 1. Approach and clinical impact of the case-study. Using signal processing, mathematical modeling and clustering, we identified four different phenotypes in hypertrophic cardiomyopathy based on the ECG. This had a clinical impact by providing improved patient risk stratification. High performance simulations then investigated the potential mechanisms underlying these phenotypes. This provided new options for individual patient management and different therapeutic approaches.

• 

  Figure 2. Methodological approach to the identification of four HCM phenotypes (from [13]).

• 

  Figure 3. Four phenotypes were identified in HCM, exhibiting differences in ECG morphology (A), distribution of hypertrophy (B) and arrhythmic risk (C) (from [13]).

• 

  Figure 4. Effect of the anatomy on the ECG. Groups 1B, 3 and 1A exhibit similar ECG morphologies despite differences in extent and distribution of hypertrophy. Anatomy alone does not explain Group 3 and 1A specific abnormalities. Taken from [14].

• 

  Figure 5. Effect of conduction changes in various regions of the myocardium on the QRS complex (from [14]).

• 

  Figure 6. Effect of activation blocks in various regions of the myocardium on the QRS complex (from [14]).

• 

  Figure 7. Abnormal Purkinje-myocardium coupling led to a patchy activation (A, B) with areas of late activation (C), and translated into deep S waves in lead V6, explaining Group 3 abnormalities. Taken from [14].

• 

  Figure 8. HCM ionic remodeling in various regions of the myocardium. Ionic remodeling in the region of hypertrophy (A) explains the inverted T waves with normal QRS observed in Group 1A patients [14].