Related Experiment Video
Updated: Jul 17, 2026
07:13
Multimodal Cross-Device and Marker-Free Co-Registration of Preclinical Imaging Modalities
Published on: October 27, 2023
Extending Regression Without Truth to Integrate Ground-Truth Measurements for Evaluating Quantitative Imaging Methods
Yan Liu1, Abhinav K Jha1,2
1Department of Biomedical Engineering, Washington University in St. Louis, St. Louis, MO, USA.
Summary
Evaluating quantitative imaging (QI) methods is challenging without gold standards. This study introduces a novel approach combining patient data and ground-truth datasets to improve QI method evaluation, especially with limited patient samples.
Area of Science:
- Medical Imaging
- Biostatistics
- Quantitative Imaging Analysis
Background:
- Objective evaluation of quantitative imaging (QI) methods is crucial but often limited by the absence of gold standards in patient data.
- Existing regression-without-truth (RWT) techniques require extensive patient samples, which are not always feasible, particularly for rare diseases or novel imaging procedures.
Purpose of the Study:
- To develop and validate a novel approach for the objective evaluation of QI methods using limited patient samples.
- To integrate information from patient data lacking ground truth with datasets possessing known ground truth.
Main Methods:
- Proposed an approach that combines patient data without ground truth and datasets with known ground truth.
- Validated the approach using numerical studies to assess its performance in ranking QI methods.
Main Results:
- The proposed approach demonstrated improved performance in ranking QI methods compared to traditional RWT techniques.
- Numerical studies confirmed the effectiveness of integrating both data types for QI method evaluation.
Conclusions:
- The developed approach shows significant potential for evaluating QI methods when patient data is limited.
- Further validation using clinically realistic simulations and clinical data is warranted to confirm its broader applicability.
