Respiratory motion can influence the efficacy and safety of radiation therapy in the thorax and abdomen. Free-breathing 4D-MRI is a promising alternative to 4D-CT for motion management in MRI-guided linac treatments, providing excellent soft-tissue contrast with no ionizing radiation. High-quality MR images free of motion artifacts can distinguish lesions from normal tissue. MRI techniques, however, currently necessitate multiple scans with lengthy scan times.
Researchers are working on a technique that will allow them to use a single MRI scan for motion detection, motion-resolved 4D-MRI, and motion-integrated 3D-MRI reconstruction. It is possible with a self-navigated MRI technique and deep learning-based image reconstruction in less than a minute.
The three-stage technique starts with CAPTURE, a self-navigated respiratory motion detection sequence that is a variant of the stack-of-stars MRI sequence. Despite differences in respiratory patterns between subjects and individual scans, CAPTURE can detect respiratory motion. The corresponding frequency spectra identified the individual frequency components.
The researchers created 4D-MRIs using three reconstruction techniques: multi-coil non-uniform inverse fast Fourier transform (MCNUFFT), compressed sensing, and deep learning-based Phase2Phase (P2P).
On a low-field MRI-guided linac, a rapid single MRI scan combined with CAPTURE, P2P, and MOTIF can generate high-quality 4D-MR images for lesion motion range determination and 3D-MR images for lesion delineation.