Quantify the Magnitudes of Cardiac Motion Artifacts In Respiratory 4DCTs and Estimate the Cardiac Motion Magnitude
Abstract
Purpose
Managing cardiac and respiratory motion is critical for stereotactic arrhythmia radiotherapy (STAR), thoracic, and breast treatments. Breath-hold cardiac 4DCT (c4DCT) is commonly required for assessing cardiac motion, while respiratory 4DCT (r4DCT) is used for assessing respiratory motion. However, c4DCT is often unavailable for clinical patients. The purpose of this study is to present a new method to estimate cardiac motion directly using r4DCT without c4DCT.
Methods
We recently developed a novel image processing algorithm gCGF, for computing respiratory motion in r4DCT and iteratively filtering out coexisting but unmanaged cardiac motion artifacts such as axial slice misalignments. In this study, we extended our gCGF algorithm to compute the magnitudes of the filtered cardiac motion artifacts in r4DCT, and 2) to build linear regression models to map artifact magnitudes to known cardiac motion magnitudes measured in c4DCT of the same patients. The cardiac motion information (95th percentile and mean magnitudes) of 18 patients was used to train and test the regression models with a 70%/30% split. The distances between the fitted curves and the validation points were evaluated.
Results
The average 95th percentile and mean cardiac motion magnitudes were 4.3 mm and 1.9 mm, respectively, measured from the heart surface in each c4DCT phase to the average heart shape. The fitted linear regression models yielded mean prediction errors of 0.67 mm and 0.23 mm, or 19.2% and 12.8%, for the 95th percentile and mean magnitude metrics, respectively.
Conclusion
This study demonstrated that the fitted linear regression models would be useful to estimate the cardiac motion magnitudes with accuracies < 0.7 mm using the unmanaged cardiac motion artifacts measured in a patient’s respiratory 4DCTs. The estimation results will be useful for managing cardiac motion for radiation treatments on or near the heart.