Maximum intensity projection (MIP) imaging using slice-stacking MRI

Purpose: To evaluate the feasibility of acquiring maximum intensity projection (MIP) images using a novel slice-stacking MRI (SS-MRI) technique. Methods: The proposed technique employed a steady state acquisition sequence to image multiple axial slices. At each axial slice, the scan is repeated throughout one respiratory cycle. Four objects (small, medium, and large triangles, and a cylinder) moving with a patient breathing trajectory were imaged repeatedly for six times using the slice-stacking MRI and 4D-CT. MIPSS-MRI and MIP4D-CT were reconstructed. The internal target volume (ITV) was segmented for each object on the six scans and compared between MIPSS-MRI and MIP4D-CT. The medium triangle was also imaged with various motion patterns using slice-stacking MRI, 4D-CT, and sagittal cine-MRI. The corresponding MIP images were reconstructed and volume/area measurements were performed and compared between different imaging methods. Three healthy volunteers underwent the slice-stacking MRI and sagittal cine-MRI scans. A region of interest (ROI) was selected and contoured for each subject in both MIPSS-MRI and MIPcine-MRI. The area of the selected ROI was computed and compared. Results: Volume comparison between MIPSS-MRI and MIP4D-CT showed statistically insignificant (p 0.05 in all cases) difference in the mean ITVs for all four objects. For the study of the medium triangle with multiple motion patterns, there was a good agreement in the measured ITVs between MIPSS-MRI and MIP4D-CT (p=0.46, correlation coefficient=0.91), with a mean difference of 1.4%±4.4%. The area measurements between MIPSS-MRI and MIPcine-MRI also showed good agreement (p=0.47, correlation coefficient=0.97), with a mean difference of 0.2%±2.9%. For the healthy volunteer study, the average difference in the area of selected ROI was -2.5%±2.5% between MIPSS-MRI and MIPcine-MRI. Conclusions: These preliminary results showed good agreement in volume/area measurements between the slice-stacking MRI technique and 4D-CT/cine-MRI, indicating that it is feasible to use this technique for MIP imaging.