Enhancing diagnosis of gout with deep learning in dual-energy computed tomography: a retrospective analysis of crystal and artefact differentiation
Authors' conclusion
Our study demonstrates the effectiveness of the DL algorithm in differentiating between green colour coding indicative of crystal deposition and clumpy artefacts in DECT scans.
Study context
- Population
- 47 patients with gout and 27 gout-free controls undergoing dual-energy computed tomography (DECT); 18,704 regions of interest (ROIs) from green foci18,704 ROIs
- Comparison
- Deep learning (DL) (convolutional neural network (CNN) per lesion, support vector machine (SVM) per patient) for differentiating green colour coding indicative of tophi from clumpy artefacts in DECT
- Outcome
- Diagnostic accuracy (AUC, sensitivity, specificity, PPV, NPV) in differentiating green colour coding indicative of tophi from clumpy artefacts
Methodological detailsretrospective, deep learning, per-lesion CNN analysis, per-patient SVM analysis
- Retrospective
- Deep learning
- Per-lesion CNN analysis
- Per-patient SVM analysis
Performance
As reported in the abstract
Diagnostic performance of DL models
| Test or model | Sensitivity | Specificity | Accuracy |
|---|---|---|---|
| CNN small ROIs | 81.5% | 96.1% | 88.5% |
| CNN medium ROIs | 82.7% | 96.1% | 88.6% |
| CNN large ROIs | 91.8% | 86.9% | 91.0% |
| SVM per-patient | 87.2% | 100% | 91.8% |
Sensitivity
CNN small ROIs81.5%
CNN medium ROIs82.7%
CNN large ROIs91.8%
SVM per-patient87.2%
Specificity
CNN small ROIs96.1%
CNN medium ROIs96.1%
CNN large ROIs86.9%
SVM per-patient100%
Accuracy
CNN small ROIs88.5%
CNN medium ROIs88.6%
CNN large ROIs91.0%
SVM per-patient91.8%
Original publication
Enhancing diagnosis of gout with deep learning in dual-energy computed tomography: a retrospective analysis of crystal and artefact differentiation. Rheumatology (Oxford). Published 20 November 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →