Region of interest cone beam computed tomography (ROI-CBCT)
Chityala, Ravishankar N.
MetadataShow full item record
CT image quality is improving dramatically but it may require increased dose to the patient. Reduction of the imaged volume to a Region of Interest (ROI) could reduce integral dose but can also result in artifacts. We have developed three techniques for recovering accurate reconstructions with minimal artifacts when ROI data are acquired. In the first technique, the sinogram completion , the data outside the ROI is estimated based on the a priori information that every feature in a sinogram traces a sine curve. The estimated data is then used to complete the sinogram outside the ROI to obtain good reconstruction inside an ROI. In the second technique, ROI filter acquisition , region outside an ROI is obtained at a lower dose than region inside by placing an ROI filter in the x-ray beam. The intensity variation between inside and outside the ROI are equalized and then reconstructed to obtain good reconstruction at lower integral dose. In the third technique, Dual resolution, dual detector ROI CT reconstruction , the region inside is obtained at standard dose using a high resolution detector. The region outside the ROI, the full field-of-view (FFOV), was obtained at lower dose using a low resolution detector. The two sets of images were intensity equalized, aligned and reconstructed to obtain good reconstruction inside and outside the ROI. It also provides better spatial resolution inside the ROI compared to ROI filter acquisition based reconstruction while at lower integral dose compared to FFOV reconstruction. Percent differences between ROI and FFOV reconstructions are about 1-2% for dose reduction factors of 3-5 outside the ROI and hence reduced integral dose. To obtain projection images for ROI CBCT using a commercial angiographic system, a Portable Tomographic Platform (PTP) was constructed. The PTP setup ensures acquisition with least setup time and repeatability. The accuracy in the initial setup of the system is more than 0.5 0 and further refinement were obtained by using calibration objects.