Search

Cardiovascular imaging studies provide a multitude of structural and functional data to better understand disease mechanisms. While pooling data across studies enables more powerful and broader applications, performing quantitative comparisons across datasets with varying acquisition or analysis methods is problematic due to inherent measurement biases specific to each protocol. We show how dynamic time warping and partial least squares regression can be applied to effectively map between left ventricular geometries derived from different imaging modalities and analysis protocols to account for such differences. To demonstrate this method, paired real-time 3D echocardiography (3DE) and cardiac magnetic resonance (CMR) sequences from 138 subjects were used to construct a mapping function between the two modalities to correct for biases in left ventricular clinical cardiac indices, as well as regional shape. Leave-one-out cross-validation revealed a significant reduction in mean bias, narrower limits of agreement, and higher intraclass correlation coefficients for all functional indices between CMR and 3DE geometries after spatiotemporal mapping. Meanwhile, average root mean squared errors between surface coordinates of 3DE and CMR geometries across the cardiac cycle decreased from 7 ± 1 to 4 ± 1 mm for the total study population. Our generalised method for mapping between time-varying cardiac geometries obtained using different acquisition and analysis protocols enables the pooling of data between modalities and the potential for smaller studies to leverage large population databases for quantitative comparisons., (© 2023. The Author(s).)

One of the main limitations in using inverse methods for non-invasively imaging cardiac electrical activity in a clinical setting is the difficulty in readily obtaining high-quality data sets to reconstruct accurately a patient-specific geometric model of the heart and torso. This issue was addressed by investigation into the feasibility of using a pseudo-3D ultrasound system and a hand-held laser scanner to reconstruct such a model. This information was collected in under 20 min prior to a catheter ablation or pacemaker study in the electrophysiology laboratory. Using the models created from these data, different activation field maps were computed using several different inverse methods. These were independently validated by comparison of the earliest site of activation with the physical location of the pacing electrodes, as determined from orthogonal fluoroscopy images. With an estimated average geometric error of approximately 8 mm, it was also possible to reconstruct the site of initial activation to within 17.3 mm and obtain a quantitatively realistic activation sequence. The study demonstrates that it is possible rapidly to construct a geometric model that can then be used non-invasively to reconstruct an activation field map of the heart.

Three-dimensional transesophageal echocardiography offers promise for improved understanding of mitral leaflet pathology, but it has not been validated quantitatively, nor has the minimum number of imaging planes for satisfactory reconstruction been determined with a rotational scanning geometry. This study assessed its accuracy in vitro by comparing, on a 1 x 1-mm grid, the surfaces of mitral leaflets derived from 5-degree rotational ultrasonic scans with those derived from laser scans of casts of the atrial side of the leaflets. Overall, the ultrasonically derived surface had a mean absolute deviation of 0.65 +/- 0.12 mm from the laser-derived surface. Using only alternate imaging planes (10-degree increments) made no significant difference in the overall distribution of deviations (P =.56), although the distributions on some individual specimens differed markedly. We conclude that 5-degree rotational scanning in vitro can reconstruct the mitral valve leaflets with sufficient accuracy and detail to render clinically important features.

Accurate measurement of left-ventricular (LV) volume and function are important to monitor disease progression and assess prognosis in patients with heart disease. Existing methods of three-dimensional (3-D) imaging of the heart using ultrasound have shown the potential of this modality, but each suffers from inherent restrictions which limit its applicability to the full range of clinical situations. We have developed a technique for image acquisition using a magnetic-field system to track the 3-D echocardiographic imaging planes and 3-D image analysis software including the piecewise smooth subdivision method for surface reconstruction. The technique offers several advantages over existing methods of 3-D echocardiography. The results of validation using in vitro LV's show that the technique allows accurate measurement of LV volume and anatomically accurate 3-D reconstruction of LV shape and is, therefore, suitable for analysis of regional as well as global function.

Three-dimensional (3D) reconstruction from a single esophageal scanning position requires a stable relationship between the probe and the heart. The purpose of this study was to examine the movement of a transesophageal echocardiographic probe during 3D image acquisition. A new dual-axis multiplane probe was used that includes a miniature (6 x 6 x 9 mm) magnetic sensor in the tip. The sensor identifies the probe's 3D position and 3D orientation in space with respect to the location of a magnetic field generator placed beneath the subject. In vivo 3D scanning was performed in five anesthetized, ventilated dogs, with positional determinations acquired every 66 msec. Probe movement was estimated by computing the deviations of each x, y, and z position and orientation determination, compared with the average values during each 3D scan or cardiac cycle. Ten 3D scans were analyzed, involving 263 cardiac cycles and 2328 determinations. The range and SD of the translational movement of the transducer were 2.3 and 0.8 mm, 1.7 and 0.5 mm, and 2.4 and 0.7 mm in x, y, and z directions, respectively, during 3D scanning. Translational movement was more dominant than was rotational movement. Misregistration of three-dimensional reconstructions may be due to subtle probe movement. The ability to monitor probe movement may be helpful in optimizing 3D data sets.