Premium
WE‐G‐134‐04: Four‐Dimensional Dual Cone‐Beam CT (4D‐DCBCT): Preliminary Experimental Results
Author(s) -
Li H,
Vergalasova I,
Giles W,
Bowsher J,
Yin F
Publication year - 2013
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 - 10.1118/1.4815673
Subject(s) - imaging phantom , cone beam computed tomography , detector , flat panel detector , projection (relational algebra) , beam (structure) , nuclear medicine , computer science , optics , physics , computed tomography , algorithm , medicine , radiology
Purpose: One of the major challenges for clinical implementation of 4D‐CBCT is long scan time. This study aims to develop a 4D‐DCBCT technique to improve the efficiency of 4D imaging (for motion management in radiation therapy). Methods: A bench‐top DCBCT system, which consists of two orthogonal 40x30cm flat panel detectors and two conventional x‐ray tubes with two individual high‐voltage generators, sharing the same rotational axis, was used to develop the technique. The x‐ray source to detector distance was 150 cm and x‐ray source to rotational axis distance was 100 cm for both subsystems. The dual CBCT system utilized 110° of projection data from one detector and 90° from the other, as opposed to a single CBCT utilizing 200° of projection data per each detector. Motion phantom studies were conducted to validate the efficiencies by comparing 4D images generated from 4D‐DCBCT and 4D‐CBCT. First, a simple sinusoidal profile was used to confirm the scan time reduction. Next, both irregular sinusoidal and patient‐derived profiles were used to investigate the advantage of temporally correlated orthogonal projections due to a reduced scan time. Normalized mutual information (NMI) between 4D‐DCBCT and 4D‐CBCT was used for quantitative evaluation. Results: For the simple sinusoidal profile, the average NMI for ten phases between two single 4D‐CBCTs was 0.336, indicating the maximum NMI that can be achieved for this study. The average NMIs between 4D‐DCBCT and each single 4D‐CBCT were 0.331 and 0.320. For both irregular sinusoidal and patient‐derived profiles, 4D‐DCBCT generated phase images with less motion blurring when compared with single 4D‐CBCT. Conclusion: The 4D‐DCBCT provides an efficient 4D imaging technique for motion management. The scan time is approximately reduced by a factor of two. The temporally correlated orthogonal projections improved the image blur across 4D phase images.