Image processing as a potential tool for CT dose optimization. (September 2016)

• Record Type: Journal Article
• Authors: Amorim, J.; Mendes, B.; Ribau, E.; Gouvêa, M.; Sarmento, S.
• Is Part Of: Physica medica. Volume 32(2016)Supplement 3
• Page Start: 316
• Publication Date: 2016-09
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2016.07.194
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Introduction: Automatic tube current modulation (ATCM) is an important tool for dose optimization in computed tomography (CT). Optimization of ATCM parameters is usually done by trial and error. In a 2009 study by IAEA, image processing (noise addition) was used to simulate lower dose acquisitions. Purpose: The aim of this work is to assess the potential of image processing as a routine auxiliary tool for CT dose optimization. Materials and methods: In this first stage, only CT exams without iterative reconstruction (IR) were considered. A process to add noise to DICOM® images was developed, to simulate acquisitions with lower tube current (I). This was tested using phantom images, acquired with different I values, and then applied to real exams. CT images in the liver region were selected from 13 archived exams, and dose reductions of 15% and 33% were simulated. Two experienced Radiologists evaluated the original and virtual images in terms of diagnostic quality, noise and artifacts. Results: Standard deviation (SD) in virtual images of phantoms agreed well with the expected values, within the range of variation observed for acquisitions with the same parameters. Diagnostic quality scores were slightly lower for the virtual images, but the results obtained highlight the subjectivity of image quality evaluation. Some virtual images were classified as having less artifacts than the original ones, as a result of reduced sampling. Conclusion: This study was well received by radiologists, and implemented using commonly available software. The challenge is to develop a robust method of image quality evaluation. Disclosure: The authors have no relevant financial or non-financial relationships to disclose.