A Tool for Rapid Comparative Analysis of 3D Echocardiography-Generated Volume–Time Curves

Background: Volume–time curves (VTC) of the left ventricle can be generated by three-dimensional echocardiography (3DE). They contain functional information like ejection fraction (EF), the relation of systole and diastole, and velocity changes over time (dV/dt). The question of how to establish a normalized VTC from a large cohort and how to enable a statistical comparison with pathological VTC remains. We describe a method to visually compare a patient’s VTC to a normalized VTC based on a large cohort of healthy subjects. In addition, we propose distance metrics to statistically describe differences.

Methods: 3DE datasets of the left ventricle of 371 healthy subjects (age range 1 day–18 years) were used to compute normalized VTC and their first derivatives, normalized dV/dt curves. We then compared the VTC generated by 3DE of a toddler on a left ventricular assist device during a weaning attempt with different beats per minute (bpm) of the pump to the normalized curves. Here, two distance metrics were used to define the degree of deviation from the median of the healthy cohort expressed in z-scores. The area difference to the median was used for comparing VTC, and dynamic time warping distance was used for comparing dV/dt curves. Furthermore, we applied this approach to a subset of 20 healthy controls (age range 1 day–26 months) as a validation group.

Results: The VTC and dV/dt curves of the patient under decreasing bpm on the pump are plotted against the normal curves. This allows a fast visual perception revealing a decreasing EF, decreasing ejection velocities, and loss of E- and A-wave with prolongation of systole and paradoxal movement under reduction of bpm. Moreover, the distance metrics revealed increasing deviation with decreasing pumping frequency for the VTC (z = −0.19 to −3.22) and the dV/dt curve (z = 1.75–5.99). The healthy controls were, except for some outliers, within the range of norm values.

Conclusion: We were able to generate normalized VTC and dV/dt curves for the left ventricle from 3DE datasets based on a large cohort of healthy subjects. We established a tool that allows plotting of patient’s VTC and fast visual perception in comparison to the normalized curves. Moreover, the distance metric allows a statistical assessment of the degree of shape abnormality of a new VTC. All this can serve as a tool for longitudinal follow-up of left ventricular function, including systolic and diastolic aspects, by acquisition of one 3DE dataset.

Publikationsverlauf

Artikel online veröffentlicht:
11. Februar 2025