Room: Exhibit Hall | Forum 2
Purpose: To evaluate the feasibility of pseudo-real-time 4DMRI for improved tumor tracking for radiation therapy in chest, abdominal, and pelvic applications.
Methods: A 4D TRAK-XD sequence was employed on a Philips 1.5T Ingenia MR scanner with the following imaging parameters: Coronal 3D T1W SPGR with FOV=400/360/300mm along FH/LR/AP directions; acquisition resolution = 2/2.5/2.5mm; TR/TE=1.9/0.5ms; FA=12Â°; compressed SENSE factor=12; and with 1% central and 33.3% peripheral k-space view sharing. With these parameters a temporal resolution of 0.4s and footprint of 1.2s per 3D volume was achieved. Approximately ten 3D â€œphasesâ€? were generated within a typical respiratory cycle. A real-time (0.3s/slice) coronal 2D SPGR sequence with comparable spatial resolution and contrast was also tested for comparison.
Results: Pseudo-real-time 4DMRI datasets were generated on the magnet without additional off-line processing. Motion images to track a target within any 2D plane could be easily generated based on the 4D dataset. Motion tracking and image quality were evaluated qualitatively by comparing the reference real-time 2D images with the coronal 2D MPR images reconstructed from 4DMRI datasets. Although the degree of motion blurring between end-exhale and end-inhale on the reconstructed 2D images was more severe than the real-time 2D images, the image quality of the most important end-ex/inhale phases was comparable.
Conclusion: Traditional CINE-based 4DMRI is fundamentally flawed due to its incapability to track non-periodic motion. This pseudo-real-time 3D acquisition approach avoids combining data from different time periods. It therefore can track any motion, including irregular breathing and unpredictable peristalsis. This method has the potential to replace traditional 4DMRI methods for target tracking in radiotherapy applications. Further development in real-time 3D image reconstruction is warranted to make this approach a clinical reality.