Objective: To examine the accuracy of dual-energy CT (DECT) vs ultrasound or their combination for the diagnosis of gout.
Methods: Using prospectively collected data from an outpatient rheumatology clinic at a tertiary-care hospital, we examined the diagnostic accuracy of either modality alone or their combination, by anatomical site (feet/ankles and/or knees), for the diagnosis of gout. We used two standards: (i) demonstration of monosodium urate crystals in synovial fluid (gold), and (ii) modified (excluding DECT and ultrasound) 2015 ACR-EULAR gout classification criteria (silver).
Results: Of the 147 patients who provided data, 48 (33%) had synovial fluid analysis performed (38 were monosodium urate-crystal positive) and mean symptom duration was 9.2 years. One hundred (68%) patients met the silver standard. Compared with the gold standard, diagnostic accuracy statistics for feet/ankles DECT, feet/ankles ultrasound, knees DECT and knees ultrasound were, respectively: sensitivity: 87%, 84%, 91% and 58%; specificity: 100%, 60%, 87% and 80%; positive predictive value: 100%, 89%, 97% and 92%; negative predictive value: 67%, 50%, 70% and 33%; area under the receiver operating characteristic curve: 0.93, 0.72, 0.89 and 0.66. Combining feet/ankles DECT with ultrasound or knees DECT with ultrasound led to a numerically higher sensitivity compared with DECT alone, but overall accuracy was lower. Similarly, combining imaging knees to feet/ankles also yielded a numerically higher sensitivity and negative predictive values compared with feet/ankles DECT alone, without differences in overall accuracy. Findings were replicated compared with the silver standard, but with lower numbers.
Conclusions: Feet/ankles or knees DECT alone had the best overall accuracy for gout diagnosis. The DECT-US combination or multiple joint imaging offered no additional increase in overall diagnostic accuracy.
Keywords: accuracy; diagnosis; dual-energy computed tomography; gout; predictive value; sensitivity; specificity; ultrasound.
Published by Oxford University Press on behalf of the British Society for Rheumatology 2021. This work is written by US Government employees and is in the public domain in the US.