Investigation of Trimodality Image Registration for Guiding Cardiovascular Diseases Treatment Procedure
This study introduced a registration system that combines preoperative cardiac computed tomography (CT) and Magnetic Resonance Imaging (MRI) volume data with 2D ultrasound (US) images of cardiovascular diseases, specifically on aortic valves. By integrating these three different imaging modalitie...
Saved in:
| Summary: | This study introduced a registration system that combines preoperative cardiac
computed tomography (CT) and Magnetic Resonance Imaging (MRI) volume data with
2D ultrasound (US) images of cardiovascular diseases, specifically on aortic valves. By
integrating these three different imaging modalities (US-CT-MRI), the registration
process enhances the accuracy of diagnosing aortic valve disorders and provides
surgical guidance during the placement of the transcatheter aortic valve. The
registration process involved acquiring short-axis "Mercedes Benz" images of the aortic
valve from 20 patients who underwent the procedure. It encompasses temporal
synchronization and spatial registration as the key components of the image registration
framework. Temporal synchronization involves identifying frames in the CT and MRI
volume that align with the cardiac phase of the US time-series data. Spatial registration
utilizes an intensity-based normalized mutual information method in combination with
a pattern search optimization algorithm to generate interpolated cardiac CT and MRI
images that align with the US image. The accuracy of the trimodal registration method
was evaluated using the Dice similarity coefficient, which yielded values of 0.92 (±
0.05) and 0.92 (± 0.04) when compared to US-CT and US-MRI, respectively, in the
short-axis "Mercedes Benz" views. Additionally, the Hausdorff distance, a measure of
dissimilarity, was found to be 1.49 (± 0.20) and 1.49 (± 0.19) mm for both pairings.
These results demonstrate a level of precision comparable to manual registration by an
expert. The registration process also enabled accurate measurements of the aortic
annulus diameter, showing no significant differences between the manually registered
MRI scans and automatically registered CT scans. These findings highlight the
excellent accuracy of the proposed technique in improving image-guided systems for
aortic valve surgical guidance, particularly in Transcatheter Aortic Valve Implantation
(TAVI) and Transcatheter Aortic Valve Replacement (TAVR) procedures. |
|---|