Click here to

  1. Home
  2. Meeting Program
  3. All Sessions
  4. Abstracts
Session: Imaging: CT Artifacts, Reconstruction, and Imaging Performance Assessment [Return to Session]

Development of An Automated Analysis Software for Daily CT QA

A Hart1*, A Watkins2, M Bazalova-Carter1, C Uribe3,4, (1) University of Victoria, Victoria, BC, CA, (2) Simon Fraser University, Burnaby, BC, CA, (3) University of British Columbia, Vancouver, BC, CA, (4) BC Cancer, Vancouver, BC, CA

Presentations

TU-IePD-TRACK 1-1 (Tuesday, 7/27/2021) 3:00 PM - 3:30 PM [Eastern Time (GMT-4)]

Purpose: Positron emission tomography/computed tomography (PET/CT) imaging is used for a wide range of applications from diagnosis, to staging and radiotherapy treatment planning. Daily quality assurance (QA) testing is an essential duty of both physicists and technologists in order to verify that the x-ray CT images produced by PET/CT scanners meet performance standards. As part of an effort to bring scanners into compliance with standards in an efficient manner, a software package was developed to automate the analysis of x-ray CT images of the GE VCT QA phantom. This work aimed at developing an automatic solution that would analyze the results of the daily CT QA scans. Our tool reduces the burden on technologists and contributes to a more robust QA procedure.

Methods: The software was developed using Python (v3.8.2) to aid in the daily CT QA procedure for GE DMI and D690 PET/CT scanners. The developed software automates each analysis step by automatically placing ROIs and performing calculations for the following tests: laser alignment, contrast scale, spatial resolution, noise and uniformity, slice thickness, and low contrast detectability.

Results: The software has been implemented clinically in three centres. The software can automatically log data through an API connection. All analysis tests conducted using the automatic software agreed with the manual analysis and results have been within manufacturer specification. For example, the differences in measured image noise (ΔSD = 0.16HU) and slice thickness (Δx = 0.01mm) were not statistically significant (p>0.05).

Conclusion: The automated analysis software developed in this work requires fewer manual hours from the technologists and allows the physicists overseeing the QA program to easily audit the performance of scanners located in several centres from a centralized location. This software is also compatible with standalone CT systems that use the GE VCT QA phantom for daily tests.

ePosters

    Keywords

    Quality Assurance, CT, PET

    Taxonomy

    IM- CT: Quality Control and Image Quality Assessment

    Contact Email

    Share:

    Additional Links