February 14, 2023

QA4IQI – Quality Assessment for Interoperable Quantitative CT Imaging

2nd Call, Tech Trans

Short Summary

QA4IQI project is an infrastructure project that aims to characterize the sensitivity of commonly used texture and shape (radiomics) features extracted from Computed Tomography (CT) images to variations in acquisition parameters. We do this through phantom imaging at different hospitals in Switzerland, statistical analysis and simulation studies.

Goals

The project performs phantom based controlled imaging experiments in four different clinical centers across Switzerland. The experimental data is analyzed with statistical methods and backed up by simulation analyses. The ultimate goals are to identify radiomics features that can be reliably measured from imaging data across Switzerland despite variations in scanning conditions and scanner vendors, and to construct the pipeline that can be repeated for identifying such features.
The results as well as the phantom data will be shared with the public at the end of the project. The phantom will be kept as a basic infrastructure that can be used for other hospital and imaging centers. More important, analysis code and outcomes will be made available for the use of the larger scientific and clinical community in Switzerland. The code will allow users to assess the changes they apply to their imaging protocols on the CT-based radiomics features.

Significance

Medical imaging is a crucial tool for patient care. In its current state, besides some elementary geometric measurements such as volume and diameter, interpretation of medical images is mostly qualitative based on visual inspection. Attributes in the image assessed by visual inspection are not quantitative measurements and thus can be observer dependent, meaning it can show large variation depending on the expert observing the image. Successful completion of this project will remove an important obstacle in front of using a richer set of quantitative visual measurements in CT imaging exams by building tools to identifying robust radiomics-based measurements.

Background

Quantitative image analysis is becoming an integral part of medical imaging for extracting measurements to support clinical tasks. Modern techniques allow extracting numerous measurements, i.e., features, quantifying a diverse set of visual attributes of a given image. However, not all these measurements are robust across images taken with different acquisition parameters and scanners from different vendors. As a result, not all of the measurements can be used in practice. Identifying the measurements that can be used is crucial for adaptation of advanced image-based measurement in clinical practice.

Publications

Flouris, K., Jimenez-del-Toro, O., Aberle, C. et al. Assessing radiomics feature stability with simulated CT acquisitions. Sci Rep 12, 4732 (2022). https://doi.org/10.1038/s41598-022-08301-1
Jimenez-del-Toro, Oscar MD, PhD∗; Aberle, Christoph PhD†; Bach, Michael MD†; Schaer, Roger BSc∗; Obmann, Markus M. MD†; Flouris, Kyriakos PhD‡; Konukoglu, Ender PhD‡; Stieltjes, Bram MD, PhD†; Müller, Henning PhD∗,§; Depeursinge, Adrien PhD∗,∥ The Discriminative Power and Stability of Radiomics Features With Computed Tomography Variations, Investigative Radiology: December 2021 – Volume 56 – Issue 12 – p 820-825 doi: 10.1097/RLI.0000000000000795