Accuracy of dual-energy computed tomography (CT), ultrasound, cone-beam CT, and spectral photon-counting CT for detecting calcium crystal deposition in the osteoarthritic hand: a cross-sectional diagnostic test study

Objective: To evaluate the diagnostic performance of dual-energy computed tomography (DECT), ultrasound (US), cone-beam computed tomography (CBCT), and multi-energy spectral photon-counting computed tomography (SPCCT) for detecting calcium pyrophosphate (CPP) and hydroxyapatite crystal deposition. Method: We prospectively enrolled patients scheduled for finger-joint surgery due to osteoarthritis. Preoperative cross-sectional assessment included visual analogue scale pain, and in vivo index tests with DECT, US, and CBCT; the postoperative ex vivo index test was SPCCT of excised joint material. Reference tests comprised compensated polarized light microscopy of synovial fluid and histology with Alizarin Red S staining of excised joint material. Tests for crystals were scored on a binary basis. We calculated sensitivity, specificity, and likelihood ratios (LRs); a test with positive LR > 10 or negative LR < 0.1 was considered ‘very good’. Results: We included 12 participants, nine of whom had at least one positive reference test for calcium crystals. None of the index tests met our predefined cut-offs for a very good test. The best sensitivity and specificity were 0.29 [95% confidence interval (CI) 0.04–0.71] and 0.67 (0.09–0.99) for DECT, 0.89 (0.52–1.00) and 1.00 (0.16–1.00) for US, 0.33 (0.07–0.70) and 0.67 (0.09–0.99) for CBCT, and 1.00 (0.66–1.00) and 0.00 (0.00–0.98) for SPCCT. Wide CIs reflected the small sample size. Participants with CPP in the synovial fluid had more pain than those without. Conclusion: SPCCT and US had high sensitivity for calcium crystal detection, although no test achieved the strict definition of a very good test. Trial registration: ClinicalTrials.gov (NCT04585113).