Sustainable four-dimensional cardiac CT: Minimizing data storage via reduced matrix size with optimized...

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Radiological Reconstruction Techniques

Background:

• Increasing volumes of 4D cardiac CT data pose challenges for transcatheter aortic valve repair (TAVR) planning, particularly concerning data storage.
• Accurate measurement of aortic valve anatomy is critical for successful TAVR device selection and procedural outcomes.

Purpose of the Study:

• To evaluate a modified 256x256 CT reconstruction method designed to reduce data storage requirements for TAVR planning.
• To assess the accuracy of aortic valve measurements and the reliability of device selection using the modified reconstruction method compared to standard methods.

Main Methods:

• Retrospective analysis of 75 TAVR-CT scans using dual-source energy-integrating detector CT.
• Comparison of virtual basal ring (BR) areas derived from reference 512x512, standard 256x256, and modified 256x256 reconstructions.
• Assessment of measurement bias, reproducibility (ICC), limits of agreement, and valve selection consistency by CT and interventional radiology technologists.

Main Results:

• The modified 256x256 method showed significantly smaller mean bias (-2.5 mm²) compared to the reference 512x512 images than the standard 256x256 method (4.1 mm²).
• Excellent reproducibility of BR measurements was observed (ICC = 0.996-0.997) with comparable limits of agreement between technologist teams.
• Valve selection consistency was high at 95% (Cohen's Kappa = 0.92).

Conclusions:

• The modified 256x256 reconstruction method effectively preserves the accuracy of basal ring measurements crucial for TAVR planning.
• This optimized reconstruction technique offers a viable solution for reducing CT data storage while maintaining diagnostic quality and ensuring reliable TAVR device selection.