Enhancing Daily QC for Photon-Counting CT: Incorporating 40 keV VMI Evaluations with the ACR CT Phantom 📝

Author: Lakshmi Ananthakrishnan, Xinhui Duan, Liqiang Ren, Gary Xu, Kuan Zhang, Yue Zhang 👨‍🔬

Affiliation: Department of Radiology, UT Southwestern Medical Center, Imaging Services, UT Southwestern Medical Center 🌍

Abstract:

Purpose: Daily quality control (QC) programs for diagnostic CT include water CT number (HU), standard deviation (SD), and visual artifact evaluations. At our institution, 70 keV virtual monoenergetic images (VMIs) are used for these evaluations on the clinical photon-counting CT (PCCT). However, 70 keV VMIs are less sensitive to CT number instability and image artifacts than lower keV levels. This study aims to enhance the QC program by incorporating evaluations of 40 keV VMIs.

Methods: An ACR CT phantom was scanned daily on our PCCT as part of an existing QC program. Helical and axial scans were performed at 120 kV with standard collimation of 144 × 0.4 mm and a fixed radiation dose (CTDIvol: 25 mGy) for both scan modes. In addition to daily evaluations of 70 keV VMIs, 40 keV VMIs were also assessed recording water HUs, SDs, and artifacts. Action limits (pass/fail criteria) were established by analyzing HU and SD from 40 keV VMIs collected daily for one month. The interquartile range (IQR = Q3 − Q1) was calculated, and the lower and upper bounds were defined as Q1 − 1.5 × IQR and Q3 + 1.5 × IQR, respectively. Measurements were subsequently performed by technologists and monitored by physicists.

Results: Action limits for 40 keV VMIs were established: Water HU [-30, -6] and SD [5, 8] for helical mode, and Water HU [-22, -6] and SD [5, 8] for axial mode. While water HU and SD on 70 keV VMIs remained within ACR limits with no artifacts, those on 40 keV VMIs exhibited greater variability and correlated with artifact severity.

Conclusion: Measurements from 40 keV VMIs were more sensitive to HU and SD deviations and artifacts compared to 70 keV VMIs. Incorporating 40 keV VMIs enhances QC programs and may help predict potential detector instabilities earlier.

Back to List