Abstract Background A non-invasive clinical method for construction of the systolic part of the left ventricular (LV) pressure curve was recently introduced. There is, however, no clinical method available to construct the LV diastolic pressure curve. Purpose The aim of this study was construction of patient-specific LV diastolic pressure curves from echocardiographic data and have initial validation against simultaneously acquired pressure curves by high-fidelity LV pressure catheters (micromanometers). Methods The study included 108 patients referred for diagnostic left heart catheterization due to suspected coronary artery disease. 81 patients were used as a training cohort. The construction of a diastolic pressure curve is based on a reference unitless pressure trace and seven key pressure-time diastolic events estimated in each patient. The reference diastolic pressure trace was generated from a group of 8 patients with LV pressures recorded by micromanometer. Each individual pressure trace from the reference cohort was first annotated with the occurrence of the key diastolic events: mitral valve (MV) opening, minimum LV pressure, peak of early-diastolic mitral flow velocity (E) and start and peak of atrial contraction-induced velocity (A), and MV closure. Then the temporal intervals between these events were normalized before calculating an averaged pressure curve (Figure 1, panels A-C). A patient-specific LV diastolic pressure curve was constructed by adjusting the standard curve according to estimated LV pressure at the key diastolic events (shown in Figure 1D). The estimation of the pressure-time instants is based on the LV volume temporal transient, used to estimate filling flow rate (Q=dV/dt) and flow acceleration and deceleration (dQ/dt=d2V/dt2), as well as a set of biomarkers including left atrial reservoir strain, peak systolic LV pressure, and body mass index. LV volume temporal traces were generated by combining global 2D volumes with global longitudinal strain. The ability to generate patient-specific diastolic LV pressure curves was validated in a second cohort of 7 patients studied with micromanometer-tipped catheters. Echocardiography and LV pressure were recorded from the same series of beats and recordings were done during breath-hold at end-expiration. The agreement was accessed qualitatively and considering the mean diastolic LV pressure. Results There was good agreement between measured and estimated LV diastolic pressure curves, qualitatively and quantitatively, via mean diastolic PLV: Bias 0.2 mmHg, limits of agreement <3.2 mmHg (Figure 1D-F). Conclusions The results suggest that the LV diastolic pressure curve can be estimated by echocardiography. The method needs validation in larger populations.Non-invasive diastolic LV pressure curve
Read full abstract