Background/Purpose: While the definitive diagnosis of gout and calcium pyrophosphate deposition disease (CPPD) is still based on the identification of characteristic crystals in synovial fluid or tophus, imaging ‒ in particular ultrasound and dual-energy computed tomography (DECT) ‒ is playing an increasingly important role [1]. However, the performance of DECT remains limited by potential artifacts and its minimum spatial resolution (down to ≈250 µm), which lowers its sensitivity and prevents early diagnosis in patients with small crystal deposition volumes. Thanks to its higher spatial resolution (down to ≈100 µm), emerging multi-energy/spectral photon-counting CT (PCCT) has the potential to identify monosodium urate (MSU) and calcium pyrophosphate (CPP) crystal deposition earlier. In this phantom study, we aimed to assess whether point-of-care PCCT can identify smaller deposits of MSU and CPP crystals than DECT.

Methods: We used a soft tissue phantom with synthetic crystal inserts with known concentrations of MSU (200-600 mg/mL) and CPP (50-200 mg/mL). Crystals were suspended in a crystal-free background mimicking synovial tissue attenuation. Concentrations were selected to cover the range of X-ray attenuation characteristics encountered in vivo for MSU and CPP crystal deposits [3, 4]. The phantom was successively scanned using a last-generation DECT operated at 70 and 140 kVp, and a novel point-of-care PCCT system operated at 118 kVp. A single dose level of 3 mGy was used for both CT systems. PCCT images were reconstructed in five non-overlapping energy bins (7-40, 40-50, 50-60, 60-79, and 79-118 keV) at an isotropic voxel size of 0.1×0.1×0.1 mm³. Areas under the curve (AUC) were computed for each crystal suspension pair and with the crystal-free background. We also calculated the minimum pixel sample size (with its corresponding diameter) required to distinguish between two crystal types within a deposit, using the same method as in [2].

Results: Overall, the diagnostic performance of PCCT in identifying MSU and CPP crystal deposits was greater than or equal to that of DECT. When distinguishing between MSU and CPP, diagnostic accuracy was excellent for both PCCT (AUC=0.980, [95%CI=0.960-0.993]) and DECT (AUC=1, [95%CI=1-1]). In contrast, diagnostic accuracy of PCCT (AUC=0.984, [95%CI=0.933-0.994] for MSU; and AUC=0.998, [95%CI=0.994-1.0] for CPP) outperformed that of DECT (AUC=0.650, [95%CI=0.450-0.830] for MSU; and AUC=0.729, [95%CI=0.556-0.903] for CPP) for the identification of both MSU and CPP in a crystal-free background. The minimum pixel sample size required to distinguish between MSU and CPP within a crystal deposit decreased from 3 pixels (0.4 mm diameter) with DECT to 2 pixels (0.2 mm diameter) with PCCT. On the other hand, DECT needed a minimum of 3-4 pixels (0.5 mm diameter) to differentiate between MSU and CPP within a crystal-free soft tissue background, whereas PCCT required minimum sample sizes of 2-3 pixels (0.2 mm diameter) to perform this same diagnostic task.

Conclusion: Point-of-care PCCT can identify smaller volumes of MSU and CPP crystal deposits with comparable accuracy to DECT. Optimization of PCCT protocols should further improve crystal characterization at very high spatial resolution.

« Back to ACR Convergence 2022

ACR Meeting Abstracts - https://acrabstracts.org/abstract/point-of-care-multi-energy-photon-counting-ct-for-earlier-non-invasive-diagnosis-of-gout-and-calcium-pyrophosphate-deposition-disease/