A Comparative Analysis of Single-Energy Computed Tomography versus Dual-Energy Computed Tomography to Detect Crystal Deposits in Gout and Calcium Crystal Deposition Diseases

Purpose or Learning Objective: Dual-energy computed tomography is known as a characterization tool with lower sensitivity than single-energy computed tomography to assess high-contrast spatial resolution tasks. We compared the performance of single-energy versus double-energy computed tomography for the detection of monosodium urate, calcium pyrophosphate, and hydroxyapatite crystal deposition, using both human observers and mathematical model observers.

Methods or Background: We used dedicated computed tomography phantoms containing 0.5-, 1-, 1.5-, and 2-mm-diameter synthetic crystals at known concentrations of 150, 300, 450, and 600 mg/cm³ for monosodium urate, and 50, 100, 150, and 200 mg/cm³ for calcium pyrophosphate and hydroxyapatite, in a soft tissue–mimicking background. The phantoms were scanned using single-energy computed tomography at 100 kVp and double-energy computed tomography at 40, 70, and 140 keV. Noise and spatial resolution were integrated into a model observer, with or without adjustment for internal noise (non-prewhitening observer with eye filter and internal noise or non-prewhitening with eye filter, respectively). Three human observers assessed the detectability of crystal deposits within the phantom. The area under the curve was calculated to assess crystal detection performance.

Results or Findings: For all crystal types, detectability increased with increasing crystal concentration and deposit diameter for both single-energy computed tomography and double-energy computed tomography, except for monosodium urate at low concentration (150 mg/cm³) and 140 keV. For 0.5-mm-diameter monosodium urate deposits, double-energy computed tomography at 40 keV demonstrated superior detectability compared with single-energy computed tomography at concentrations of 150, 300, 450, and 600 mg/cm³, yielding area under the curve values of 0.60, 0.70, 0.73, and 0.82, respectively. Single-energy computed tomography exhibited lower performance, with area under the curve values of 0.51, 0.65, 0.65, and 0.79, respectively. For 1-mm-diameter calcium pyrophosphate deposits, double-energy computed tomography at 40 keV achieved higher detectability than single-energy computed tomography at 50, 100, 150, and 200 mg/cm³, yielding area under the curve values of 0.93, 1, 1, and 1, respectively. Single-energy computed tomography showed slightly lower performance, with area under the curve values of 0.80, 0.99, 1, and 1, respectively. The same trend was observed for hydroxyapatite deposits, with higher area under the curve values. Results from human observers were lower than those obtained with an non-prewhitening with eye filter model observer, but higher than those calculated with an non-prewhitening with eye filter and internal noise model observer.

Conclusion: Double-energy computed tomography is not only a characterization tool because at 40 keV it outperformed single-energy computed tomography in detecting small low-concentration monosodium urate and calcium crystal deposits. Mathematical model observers must be fine-tuned with data from multiple human observers.

Publication History

Article published online:
02 June 2025