Related Experiment Video
Updated: Jul 11, 2025
In vivo Macrophage Imaging Using MR Targeted Contrast Agent for Longitudinal Evaluation of Septic Arthritis
Published on: October 20, 2013
Deep Learned Segmentations of Inflammation for Novel ⁹⁹mTc-maraciclatide Imaging of Rheumatoid Arthritis
Robert Cobb1, Gary J R Cook2,3, Andrew J Reader1
1Department of Biomedical Engineering, School of Biomedical Engineering and Imaging Sciences, King's College London, London WC2R 2LS, UK.
Abstract:
Rheumatoid arthritis (RA) is an autoimmune disease that causes joint pain, stiffness, and erosion. Power Doppler ultrasound and MRI are imaging modalities used in detecting and monitoring the disease, but they have limitations. ⁹⁹mTc-maraciclatide gamma camera imaging is a novel technique that can detect joint inflammation at all sites in a single examination and has been shown to correlate with power Doppler ultrasound. In this work, we investigate if machine learning models can be used to automatically segment regions of normal, low, and highly inflamed tissue from 192 ⁹⁹mTc-maraciclatide scans of the hands and wrists from 48 patients. Two models were trained: a thresholding model that learns lower and upper threshold values and a neural-network-based nnU-Net model that uses a convolutional neural network (CNN). The nnU-Net model showed 0.94 ± 0.01, 0.51 ± 0.14, and 0.76 ± 0.16 modified Dice scores for segmenting the normal, low, and highly inflamed tissue, respectively, when compared to clinical segmented labels. This outperforms the thresholding model, which achieved modified Dice scores of 0.92 ± 0.01, 0.14 ± 0.07, and 0.35 ± 0.21, respectively. This is an important first step in developing artificial intelligence (AI) tools to assist clinicians' workflow in the use of this new radiopharmaceutical.
Insights
Machine learning models can automatically segment rheumatoid arthritis (RA) inflammation using ⁹⁹ᵐTc-maraciclatide scans. The nnU-Net model shows superior performance in identifying normal, low, and highly inflamed joint tissues.
Area of Science:
- Nuclear Medicine
- Radiopharmaceuticals
- Artificial Intelligence in Medicine
Background:
- Rheumatoid arthritis (RA) is an autoimmune disease causing joint pain and erosion.
- Current imaging techniques like ultrasound and MRI have limitations in detecting RA.
- ⁹⁹ᵐTc-maraciclatide gamma camera imaging offers a novel approach for comprehensive joint inflammation detection.
Purpose of the Study:
- To investigate the efficacy of machine learning models for automated segmentation of RA-affected tissues.
- To compare the performance of a thresholding model and a nnU-Net model in segmenting normal, low, and highly inflamed tissues.
Main Methods:
- Training two machine learning models: a thresholding model and a nnU-Net model (CNN-based).
- Utilizing 192 ⁹⁹ᵐTc-maraciclatide scans of hands and wrists from 48 RA patients.
- Evaluating segmentation performance using modified Dice scores against clinical labels.
Main Results:
- The nnU-Net model achieved superior modified Dice scores: 0.94 for normal, 0.51 for low, and 0.76 for high inflammation.
- The thresholding model achieved lower scores: 0.92 for normal, 0.14 for low, and 0.35 for high inflammation.
- nnU-Net significantly outperformed the thresholding model in segmenting inflamed tissues.
Conclusions:
- Machine learning, particularly the nnU-Net model, shows significant potential for automated analysis of ⁹⁹ᵐTc-maraciclatide scans.
- This represents a crucial step towards developing AI tools to aid clinical workflows in RA management.
- Automated segmentation can enhance the utility of novel radiopharmaceuticals in diagnosing and monitoring rheumatoid arthritis.
