Author: Sarah Huo, David Zander, Wei Zhou 👨🔬
Affiliation: Cherry Creek High School, University of Colorado Anschutz Medical Campus, University of Colorado Anschutz Medical Campus, School of Medicine 🌍
Purpose:
Ring artifact is commonly seen in CT, and it is usually caused by detector element failure or miscalibration. Although many ring artifacts are apparent and easy to recognize, those originating from central detector elements can be more subtle and may lead to misdiagnosis, particularly when they overlap with contrast-enhanced small vessels, mimicking occlusion. Identifying the center of rotation (COR) in an axial image can help radiologists detect subtle ring artifacts more effectively, especially when the COR is clearly marked on the images.
Methods:
A webapp was created using python and streamlit for quick implementation. The webapp is hosted on a computer within the intranet for PACS access. Image information (MRN, Accession, Series Number, Instance Number) was input by the users and the suspected DICOM image was query/retrieved from the PACS server. DICOM tags [Data Collection Center (0018,9313), Reconstruction Target Cetner (0018,9318), and Pixel Spacing (0028,0030)] were retrieved from DICOM header and used to calculate the COR position in the image. The identified COR location was marked with a red circle (radius: 5 pixels) on the suspected image and displayed on the website. Users can change the window/level, zoom level, and size of the circle to better view the suspected ring artifact region. A local DICOM image upload option is also enabled.
Results:
Multiple CT images, including retrospective cases with subtle ring artifacts, were successfully tested using the developed web-based tool. Radiologists and physicists can easily bookmark this local website, allowing them to quickly upload images and check the COR location whenever they suspect a potential ring artifact.
Conclusion:
A convenient web-based tool to help identify CT ring artifacts has been successfully developed. Using python and streamlit significantly reduced the complexity of development, and similar approaches could be used to solve other simple but specific clinical issues quickly.