Premium
SU‐E‐I‐46: Comparison Study of Reconstruction Algorithms for Prototype Digital Breast Tomosynthesis Using Various Type of Breast Phantoms
Author(s) -
Kim Y,
Kim H
Publication year - 2015
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 - 10.1118/1.4924043
Subject(s) - microcalcification , imaging phantom , tomosynthesis , algorithm , digital breast tomosynthesis , image quality , mammography , iterative reconstruction , reconstruction algorithm , noise (video) , computer science , artificial intelligence , mathematics , nuclear medicine , medicine , breast cancer , image (mathematics) , cancer
Purpose: Qualitative evaluations for digital breast tomosynthesis were previously performed on a phantom with unrealistic model with heterogeneous background and noise which is not representative of real breast. The purpose of this study is to compare reconstruction algorithms for digital breast tomosynthesis using various type of breast phantoms Methods: DBT was performed on a prototype unit which is optimized for very low exposures with rapid readout. Three algorithms, a BP algorithm, a FBP algorithm, an iterative EM algorithm were compared in this study. To compare algorithms, three type of breast phantoms (uniform background phantom, heterogeneous background phantom and anthropomorphic breast phantom) were evaluated. The in‐plane image quality was evaluated by the signal difference to noise ratio (SDNR) and out‐of‐plane artifact was evaluated by artifact spread function (ASF). Results: The results showed that the SDNR of mass with the BP algorithm is slightly higher than that with the EM algorithm, while the FBP algorithm results in much lower SDNR due to the high fluctuations of background noise. In case of microcalcification, the FBP algorithm results in high SDNR. The EM algorithm shows better ASF with regard to full width half maximum. The FBP algorithm provides the best conspicuity for larger calcification by enhancing their contrast and sharpness compared to other algorithms. On the other hand the contrast and the quality were poor for soft tissue and EM algorithm demonstrated high conspicuity of both of microcalcification and mass intensity. Conclusion: In this study, three algorithms (BP, FBP, and EM) were compared. Higher conspicuity of microcalcification with FBP and better ASF and SDNR results with EM were demonstrated. Future work for algorithms considering the type of the breast and the acquisition techniques (e.g., angular range, dose distribution) should be considered with actual patient or patient‐like phantoms to increase the potential of practical applications. This research was financially supported by the Ministry of Trade, Industry & Energy(MOTIE), Korea Institute for Advancement of Technology(KIAT) and Gangwon Institute for Regional Program Evaluation(GWIRPE) through the Economic and Regional Cooperation Industry