Kim, Jiwon

BACKGROUND: Defining qualitative grades of echocardiographic metrics of right heart chamber size and function is critical for screening, clinical assessment, and measurement of therapeutic response in individuals with pulmonary vascular disease (PVD). In a population enriched for PVD, we sought to establish qualitative grades and prognostic value of right heart chamber size and function. METHODS: We investigated 1053 study participants in PVDOMICS (Redefining Pulmonary Hypertension Through the PVD Phenomics Program) to determine clinical and echocardiographic differences associated with increasing pulmonary vascular resistance severity. Right heart chamber size and function were qualitatively assessed using a percentile-based approach above the median values to create a clinical grading system for right heart adaptation. The relationship between echocardiographic categories and all-cause mortality was examined using survival analyses adjusted for potential confounders. RESULTS: A stepwise increase in adverse right heart remodeling was observed with a concomitant decrease in functional parameters by pulmonary vascular resistance strata ( P <0.001 for all). Mild, moderate, and severe categories of right heart chamber size and dysfunction were defined using a percentile-based approach across the spectrum of PVD. During a median follow-up of 2.07 years (interquartile range, 1.23–3.01 years), 130 participants died (11.4%). Progressive pulmonary vascular resistance increase and 2DE evidence of right heart dysfunction inclusive of fractional area change, and right ventricular global longitudinal strain were independently associated with increased all-cause mortality risk in multivariate analysis adjusted for age, disease duration, and male sex. CONCLUSIONS: In this well-characterized sample of adults with diverse etiologies and varying PVD severity, we define categories of abnormal right heart chamber size and function. Further, we demonstrate a stepwise relationship between these categories of abnormal morphology and function and all-cause mortality. Defining grades of right ventricular dysfunction in individuals with known PVD has important clinical implications for monitoring disease progression and response to therapies.

Abstract Purpose Echocardiography (echo) is widely used for right ventricular (RV) assessment. Current techniques for RV evaluation require additional imaging and manual analysis; machine learning (ML) approaches have the potential to provide efficient, fully automated quantification of RV function. Methods An automated ML model was developed to track the tricuspid annulus on echo using a convolutional neural network approach. The model was trained using 7791 image frames, and automated linear and circumferential indices quantifying annular displacement were generated. Automated indices were compared to an independent reference of cardiac magnetic resonance (CMR) defined RV dysfunction (RVEF < 50%). Results A total of 101 patients prospectively underwent echo and CMR: Fully automated annular tracking was uniformly successful; analyses entailed minimal processing time (<1 second for all) and no user editing. Findings demonstrate all automated annular shortening indices to be lower among patients with CMR‐quantified RV dysfunction (all P < .001). Magnitude of ML annular displacement decreased stepwise in relation to population‐based tertiles of TAPSE, with similar results when ML analyses were localized to the septal or lateral annulus (all P ≤ .001). Automated segmentation techniques provided good diagnostic performance (AUC 0.69–0.73) in relation to CMR reference and compared to conventional RV indices (TAPSE and S ′) with high negative predictive value (NPV 84%–87% vs 83%–88%). Reproducibility was higher for ML algorithm as compared to manual segmentation with zero inter‐ and intra‐observer variability and ICC 1.0 (manual ICC: 0.87–0.91). Conclusions This study provides an initial validation of a deep learning system for RV assessment using automated tracking of the tricuspid annulus.