A rigid motion correction method for helical computed tomography (CT)

• Published in: Physics in medicine & biology 2015-03, Vol.60 (5), p.2047-2073
• Main Authors: Kim, J-H, Nuyts, J, Kyme, A, Kuncic, Z, Fulton, R
• Format: Article
• Language: English
• Subjects: Adult; Algorithms; Brain - diagnostic imaging; computed tomography; Computer Simulation; Head - diagnostic imaging; Head Movements; Humans; Image Interpretation, Computer-Assisted - methods; image reconstruction; Male; motion compensation; motion estimation; Pattern Recognition, Automated - methods; Phantoms, Imaging; Quality Control; Radiographic Image Enhancement - methods; reconstruction algorithms; Tomography, Spiral Computed - methods
• ISSN: 0031-9155; 1361-6560
• DOI: 10.1088/0031-9155/60/5/2047

We propose a method to compensate for six degree-of-freedom rigid motion in helical CT of the head. The method is demonstrated in simulations and in helical scans performed on a 16-slice CT scanner. Scans of a Hoffman brain phantom were acquired while an optical motion tracking system recorded the motion of the bed and the phantom. Motion correction was performed by restoring projection consistency using data from the motion tracking system, and reconstructing with an iterative fully 3D algorithm. Motion correction accuracy was evaluated by comparing reconstructed images with a stationary reference scan. We also investigated the effects on accuracy of tracker sampling rate, measurement jitter, interpolation of tracker measurements, and the synchronization of motion data and CT projections. After optimization of these aspects, motion corrected images corresponded remarkably closely to images of the stationary phantom with correlation and similarity coefficients both above 0.9. We performed a simulation study using volunteer head motion and found similarly that our method is capable of compensating effectively for realistic human head movements. To the best of our knowledge, this is the first practical demonstration of generalized rigid motion correction in helical CT. Its clinical value, which we have yet to explore, may be significant. For example it could reduce the necessity for repeat scans and resource-intensive anesthetic and sedation procedures in patient groups prone to motion, such as young children. It is not only applicable to dedicated CT imaging, but also to hybrid PET/CT and SPECT/CT, where it could also ensure an accurate CT image for lesion localization and attenuation correction of the functional image data.