Approaching Higher Dimension Imaging Data Using Cluster-Based Hierarchical Modeling in Patients with Heart Failure Preserved Ejection Fraction

Figures

• 

  Figure 1. Flow chart of patients in the study. Patients were identified using Stanford Translational Research Integrated Database Environment (STRIDE). Manual curation of the chart was performed to confirm a diagnosis of heart failure for hospitalization. Echocardiographic images of each patient were reviewed to exclude other etiologies and comprehensive echocardiographic assessment was performed. If either one of the echocardiographic parameters of LVLS, LAV, LAS, RVLS, RAS or RVSP was not measurable due to the poor 2D image quality or the quality of tricuspid regurgitation signal. LAV, left atrial volume; LAS, left atrial strain; LVLS, left ventricular longitudinal strain; RAS, right atrial strain; RVLS, right ventricular longitudinal strain; RVSP, right ventricular systolic pressure.

• 

  Figure 2. The prevalence of cardiac impairment in patients with HFPEF. The distribution of LV mass index (A), LVLS (B) and RVSP (C). The panel D presents the venn diagram demonstrating the overlap between LV hypertrophy (threshold of LV mass index 115 g/m2 for male and 95 g/m2 for female), impaired LVLS (threshold of LVLS −16%), and pulmonary hypertension (threshold RVSP of 40 mmHg) features. LVLS, left ventricular longitudinal strain; RVSP, right ventricular systolic pressure.

• 

  Figure 3. Cluster Dendrogram of clinical, laboratory and echocardiographic echocardiographic features. Dynamic tree cut algorithm detected six clusters of closely associated features shown in blue, brown, green, red, yellow, and turquoise. AF; atrial fibrillation, AFL; atrial flutter, BMI; body mass index, BUN; blood urea nitrogen, Cr; creatinine, DBP; diastolic blood pressure, eGFR; estimate glomerular filtration rate, Hb; hemoglobin, HR; heart rate, IVSd; diastolic interventricular septum, LAEF; left atrial emptying fraction, LAS; left atrial strain, LAVI; left atrial volume index, LVDd; diastolic left ventricular dimension, LVEF; left ventricular ejection fraction, LVLS; left ventricular longitudinal strain, LVMI; left ventricular mass index, MCV; mean corpuscular volume, Na; sodium, NLR; neutrophil-to-lymphocyte ratio, NT-proBNP; N-terminal pro B-type natriuretic peptide, Plt; platelet, PWd; diastolic posterior wall, RAAI; right atrial area index, RAEF; right atrial emptying fraction, RAP; right atrial pressure, RAS; right atrial strain, RAVI; right atrial volume index, RDW; red cell distribution width, RVFAC; right ventricular fractional area change, RVLS; right ventricular…

• 

  Figure 4. Network analysis of different clinical, laboratory and echocardiographic features. The nodes are colored based on the clusters in Fig. 3. The thickness of the edges reflects the topological overlap between two nodes. Abbreviation; same as Fig. 3.

• 

  Figure 5. The GWTG-HF risk score and its complementarity to RVSP and NT-proBNP. (A) Distribution of GWTG-HF risk score. (B) Chi-square comparison between the models with the GWTG-HF risk score alone, adding NT-proBNP, and further adding RVSP.