Application of an automated analysis framework for pulsed-wave Doppler cardiac ultrasound measurements to generate reference data in adult zebrafish

Abstract

High-frequency cardiac ultrasound is the only well-established method to characterize in vivo cardiovascular function in adult zebrafish noninvasively. Pulsed-wave Doppler imaging allows measurements of blood flow velocities at well-defined anatomical positions, but the measurements and results obtained using this technique need to be analyzed carefully, taking into account the substantial baseline variability within one recording and the possibility for operator bias. To address these issues and to increase throughput by limiting hands-on analysis time, we have developed a fully automated processing pipeline. This framework enables the fast, unbiased analysis of all cardiac cycles in a zebrafish pulsed-wave Doppler recording of both atrioventricular valve flow as well as aortic valve flow without operator-dependent inputs. Applying this automated pipeline to a large number of recordings from wild-type zebrafish shows a strong agreement between the automated results and manual annotations performed by an experienced operator. The reference data obtained from this analysis showed that the early wave peak during ventricular inflow is lower for female compared with male zebrafish. We also found that the peaks of the ventricular inflow and outflow waves as well as the peaks of the regurgitation waves are all correlated positively with body surface area. In general, the presented reference data, as well as the automated Doppler measurement processing tools developed and validated in this study will facilitate future (high-throughput) cardiovascular phenotyping studies in adult zebrafish ultimately leading to a more comprehensive understanding of human (genetic) cardiovascular diseases.