Assessment of methodologies to calculate intraventricular pressure differences in computational models and patients

Intraventricular pressure differences (IVPDs) govern left ventricular (LV) efficient filling and are a significant determinant of LV diastolic function. Our primary aim is to assess the performance of available methods (color M-mode (CMM) and 1D/2D MRI-based methods) to determine IVPDs from intracardiac flow measurements. Performance of three methods to calculate IVPDs was first investigated via an LV computational fluid dynamics (CFD) model. CFD velocity data were derived along a modifiable scan line, mimicking ultrasound/MRI acquisition of 1D (IVPD_CMM/IVPD_{1D MRI}) and 2D (IVPD_{2D MRI}) velocity-based IVPD information. CFD pressure data (IVPD_CFD) was used as a ground truth. Methods were also compared in a small cohort (n = 13) of patients with heart failure with preserved ejection fraction (HFpEF). In silico data showed a better performance of the IVPD_{2D MRI} approach: RMSE values for a well-aligned scan line were 0.2550 mmHg (IVPD_{1D MRI}), 0.0798 mmHg (IVPD_{2D MRI}), and 0.2633 mmHg (IVPD_CMM). In vivo data exhibited moderate correlation between techniques. Considerable differences found may be attributable to different timing of measurements and/or integration path. CFD modeling demonstrated an advantage using 2D velocity information to compute IVPDs, and therefore, a 2D MRI-based method should be favored. However, further studies are needed to support the clinical significance of MRI-based computation of IVPDs over CMM.