A new software for patient dose extraction and assessment from CT DICOM images.

Document Type : Original Paper

Authors

1 Department of Physics, Faculty of Science, University of Damascus, Syria

2 Protection and Safety Department, Atomic Energy Commission of SYRIA, Damascus, SYRIA

Abstract

Introduction: Radiation dose monitoring is an important objective of radiation safety and quality assurance program. This study aims to contribute in establishing a local Diagnostic Reference Levels for Computer Tomography, by developing an automated software, based on MATLAB environment, to extract and analyze patient dose identifiers from CT DICOM image files. In addition, an estimation of effective dose and statistical studies were implemented.

Materials and Methods: A random sample of 1466 patients’ CT DICOM image files were collected from a 64-slice Siemens’ Somatom Perspective CT scanner. The proposed GUI extracts the volumetric CT dose index (CTDIvol), the dose length product (DLP) for each phase of scan session in order to calculate the patient radiation effective dose (E). A graphical layout presenting statistical values was also produced, with filtering capabilities according to study date, patient sex, and CT protocol type.

Results: The GUI performance was verified according to the manually proceeded results. The extraction speed and accuracy of the radiation dose values were satisfactory, as compared to the approaches presented in literatures such as optical character recognition (OCR) technology, and the direct extraction from the metadata of CT image files.

Conclusion: The proposed GUI performs the extraction of CT patient dose metrics CTDIvol, DLP with a satisfactory speed and accuracy. The obtained results could be shown in numerical and graphical formats, and it could be used for radiation dose monitoring and DRLs establishing purposes with multiple filtering capacities.

Keywords

Main Subjects

Iranian Journal of Medical Physics

Articles in Press, Accepted Manuscript
Available Online from 17 July 2022

Files

History

  • Receive Date: 09 December 2021
  • Revise Date: 10 July 2022
  • Accept Date: 17 July 2022
  • First Publish Date: 17 July 2022

Share

How to cite

Statistics