Novel Centroid Method for Robust Evaluation of Return Time of Reflected Waves in the Systemic Arterial Network

Abstract

A diastolic-to-systolic shift in the return time (RT) of backward waves to central arteries is expected with ageing. However, current methods of estimating RT-inflection point, zero crossing, and foot-depend on a single waveform feature and produce systolic RT throughout life. We propose a novel centroid method that accounts for the entire backward pressure waveform. We assess the accuracy of the various methods against a ground truth RT (GTRT) and their sensitivity to diastolic/systolic RT. Linear wave tracking was implemented in a one-dimensional systemic arterial tree model and GTRT was calculated as the amplitude-weighted mean RT of backward waves at the ascending aorta. The sensitivity of the methods to diastolic/systolic RT was also assessed in ten sheep. A balloon catheter in the descending thoracic aorta generated a backward-running pulse that arrived at the ascending aorta at different times during diastole or systole, allowing the 'bulk' RT of the backward-running wave ensemble to be manipulated. Using a virtual cohort of 1200 patients, the centroid RT was closest to GTRT compared to the zero crossing, inflection point, and foot methods; mean differences (limits of agreement) were -8 (-47, 30), vs -42 (-136, 52), -78 (-305, 149), and -197 (-379,-15) ms, respectively. Furthermore, only the centroid method was sensitive to both diastolic and systolic RT; other methods were only sensitive to systolic RT. The centroid method had the highest accuracy and robustness in estimating RT. This can provide insight into the diastolic-to-systolic shift in RT of backward waves with ageing.