Premium
SU‐E‐E‐04: Automated Analysis of Quality Assurance Phantom Scan Image in Functional Lung CT Study by Using Simulated CT Imaging Technique
Author(s) -
Heo C,
Yang Z,
Kim J
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 - 10.1118/1.4814072
Subject(s) - imaging phantom , quality assurance , image quality , image registration , nuclear medicine , medical imaging , computer science , artificial intelligence , computer vision , biomedical engineering , medicine , image (mathematics) , external quality assessment , pathology
Purpose: Quality assurance is of crucial importance in functional lung studies and yet requires time consuming manual measurements on phantom scan images. We present an automated method for analyzing CT scan images of quality assurance phantom. Methods: COPDgene™ quality assurance phantom was scanned in 4 commercial scanners (Discovery750, GE, Sensation16, Siemens, Ingenuity, Philips, Aquillion One, Toshiba) with 45 different combinations of mAs, reconstruction kernels and display FOV. In addition, simulated CT images were generated by using geometric and attenuation properties of design sheet of phantom and applying projection and filtered backprojection algorithms. In order to mimic real CT images, filter kernels were derived from measured MTFs of the CT scanners. Rigid image registration technique was used to find best matching translation and rotation parameters between simulated and real CT scan images, followed by measurements of FWHM and ROI statistic for the known positions and sizes of artificial airways and simulated lung foam materials. Results: Image registration was performed successfully in all CT scan images, yielding rotation and translation parameters within accuracy range less than 0.8 mm. FWHM measurements were extracted for 14 holes and 6 tubes ranging 2mm ∼ 10mm in diameter. Discrepancies between measured FWHM and ground truth value ranged from 0.02 ∼ 1.5mm. Conclusion: Proposed method could successfully recognize locations and types of objects in the quality assurance phantom, and thereby carrying out measurements required for the quality assurance procedure. This technique has the potential to improve quality of functional lung CT studies.