Abstract
BACKGROUND: Load-independent indices of right ventricular (RV) dysfunction aid in the prognosis of patients with pulmonary hypertension (PH), but acquisition of these indices remains difficult. Simpler image-based tools could bring these metrics to everyday practice. OBJECTIVES: This study sought to develop a novel, artificial intelligence-based pipeline that estimates load-independent RV functional indices using a pressure-time waveform and stroke volume from clinical right-sided heart catheterization. METHODS: Clinical data and pressure-volume-time data were collected from 76 patients referred for right-sided heart catheterization for known or suspected PH from 3 centers. A computational pipeline was developed to determine the RV pressure-volume loop and extract load-independent RV indices using computer vision image processing and single-beat analysis. Agreement with gold standard single-beat analysis and prognostic value were evaluated. RESULTS: Strong concordance was observed between both methods for end-systolic elastance (Ees: R = 0.96; concordance correlation coefficient [CCC] = 0.58), effective arterial elastance (Ea: R = 0.97; CCC = 0.88), end-diastolic elastance (Eed: R = 0.87; CCC = 0.47), and Ees/Ea ratio (R = 0.93; CCC = 0.71) in both the validation and external cohorts. Prognostic analyses showed that calculated Ea (HR: 2.09; 95% CI: 1.04-4.20) and Ees/Ea (HR: 0.27; 95% CI: 0.08-0.87) were significant predictors of clinical outcomes. Cluster analysis using single-beat indices identified 2 RV subphenotypes with distinct hemodynamic features that were more predictive of poor outcomes than analysis using standard clinical features. CONCLUSIONS: Study investigators have developed a novel computational pipeline that digitizes and generates single-beat estimates of RV-pulmonary arterial coupling from an image of the RV pressure waveform and stroke volume. Its output correlates with single-beat methods and predicts clinical outcomes.