Automated Web-based Software for CT Quality Control Testing of Low-contrast Detectability using Model Observers

Abstract

The Channelized Hotelling observer (CHO) is well correlated with human observer performance in many CT detection/classification tasks but has not been widely adopted in routine CT quality control and performance evaluation, mainly because of the lack of an easily available, efficient, and validated software tool. We developed a highly automated solution - CT image quality evaluation and Protocol Optimization (CTPro), a web-based software platform that includes CHO and other traditional image quality assessment tools such as modulation transfer function and noise power spectrum. This tool can allow easy access to the CHO for both the research and clinical community and enable efficient, accurate image quality evaluation without the need of installing additional software. Its application was demonstrated by comparing the low-contrast detectability on a clinical photon-counting-detector (PCD)-CT with a traditional energy-integrating-detector (EID)-CT, which showed UHR-T3D had 6.2% higher d' than EID-CT with IR (p = 0.047) and 4.1% lower d' without IR (p = 0.122).

Keywords: channelized Hotelling observer (CHO); image quality assessment; low-contrast detectability; protocol optimization; web-based software platform.

Figure 1.

Overall design of the CTPro web software platform.

Figure 2.

The role of frontend website, backend, and database for software modules of image quality metric and database query, where the users select either images from our CTPro database or upload their own DICOM images, query the database, run image quality metric calculation function, and visualize the results on the webpage.

Figure 3.

User interface of the CHO calculation module and database on the CTPro website. (a) The user interface of the CHO module displays the images and key DICOM information and calculates the index of detectability (d’) and area under the ROC curve (AUC) for the 6-, 5-, and 4-mm low-contrast objects of the ACR phantom. (b) Image data query from CTPro database. The images were selected with Siemens Alpha PCD-CT, ACR phantom, CTDIvol at 6 mGy, 144x0.4 mm collimation (SR scan mode), T1 for spectral information (T3D reconstruction mode), and with an IR reconstruction (QIR, Br44-3 kernel).

Figure 4.

Examples of (a) MTF module and (b) NPS module.

Figure 5.

(a) Example images of the ACR low-contrast module from the standard resolution (SR, 144x0.4 mm collimation) and ultra-high-resolution (UHR, 120x0.2 mm) scans modes combined with threshold-low (T3D) and virtual-monoenergetic-image at 70 keV reconstruction modes on PCD-CT at 12 mGy with IR (Br44-3) reconstructions, along with the EID-CT images at matched dose and reconstruction methods. Ensemble average image of EID-CT with IR and PCD-CT UHR-T3D with IR are displayed with the signal (red) and background ROIs (blue and yellow). (b) d’ for the four scan-reconstruction modes on PCD-CT (SR-T3D, SR-70keV, UHR-T3D, and UHR-70keV) and EID-CT with both FBP and IR reconstructions. The d’ distribution came from 10 independent d’ measurements with repositioning of the ACR phantom.