Premium
SU‐E‐J‐05: Validation of an Iterative Tomosynthesis Algorithm for Low Dose on Board Cone Beam CT Patient Localization
Author(s) -
Izaguirre EW,
Price SG,
Yang D,
Rangaraj D
Publication year - 2012
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4734837
Subject(s) - tomosynthesis , imaging phantom , image quality , iterative reconstruction , cone beam computed tomography , algorithm , medical imaging , computer science , projection (relational algebra) , image resolution , computer vision , reconstruction algorithm , artificial intelligence , nuclear medicine , medicine , mammography , radiology , image (mathematics) , computed tomography , cancer , breast cancer
Purpose: Cone beam CT (CBCT) is a well established technique to localize patients using bone and soft tissue anatomy. Current protocols are limited to one weekly CBCT due to the considerable imaging dose delivered to the patient. The purpose of this project is to develop and validate a low dose CBCT algorithm to reduce dose and imaging time of current 3D imaging localization procedures using a novel iterative tomosynthesis algorithm to allow daily CBCT for patient positioning and target localization. Methods: The algorithm is based on the combination of a tomosynthesis filtered back propagation (TFBP) acquisition geometry algorithm and a maximum likelihood expectation maximization (MLEM) iterative reconstruction. Circular or arc acquisition trajectory, projection number, and angular projection position are optimized according to the anatomical treatment site and region of interest. The TFBP method provides the first 3D image estimate, and the MLEM improves its quality. In this study, we focused on head and neck treatment localization imaging. Results: We studied the performance of our tomosynthesis algorithm imaging resolution on an anthropomorphic head and neck phantom to determine image quality as a function of dose reduction techniques. Reconstructed anatomy shows that a 1/8 dose reduction provides similar image quality and resolution as current CBCT protocols. Seven iterations show an optimal compromise between image quality and reconstruction time. Tomosynthesis images provide digitally reconstructed radiographs with similar resolution and contrast as full CBCT. We verified that the iterative process eliminates phantom images originated by the acquired sparse angular data projections. Conclusions: We developed and validated an iterative algorithm for low dose cone beam CT based on circular or arc tomosynthesis geometries and iterative reconstruction techniques. The algorithm combines the strengths of both techniques to provide a novel low dose method to image patient anatomy for patient positioning and target localization.