Detection and characterization of crystal suspensions using single-source dual-energy computed tomography: a phantom model of crystal arthropathies

Abstract

Objectives: The aim of this study was to perform phantom measurements to prove the feasibility of single-source dual-energy computed tomography (DECT) of the extremities using a volume scan mode. In addition, we, for the first time, wanted to determine which concentrations of monosodium urate (MSU) in gout and calcium pyrophosphate (CP) in pseudogout are needed to detect or distinguish these soft tissue depositions with DECT.

Materials and methods: We created a hand-shaped plastic phantom assembled with a descending order of concentrations of MSU (6.25%-50%) and CP (1.56%-50%) with similar attenuation in conventional computed tomographic (CT) images. Dual-energy imaging was done on a standard 320-row CT scanner with acquisition of 2 volumes: one at 80 and the other at 135 kV. Using linear regression analysis, dual-energy gradients were calculated for MSU and CP. Thereafter, we selected a specific region of interest on the dual-energy graph to color-code MSU and CP on the images. Three blinded readers scored 10 scans of the randomly equipped phantom, corresponding to 60 samples, to determine the sensitivity and specificity of this technique. Receiver operating characteristics analysis was done to determine the diagnostic power.

Results: We found a dual-energy gradient for MSU of 1.020 ± 0.006 and for CP of 0.673 ± 0.001. Assessment of the randomized phantom scans indicates reliable detection of MSU at concentrations of 12.5 % or higher and that of CP at 6.25 % or higher, corresponding to deposits with mean Hounsfield unit values of 59.8 for MSU and 101.1 for CP. The sensitivity for MSU ranged from 83.3% to 97.3% at 15/90 mA (135/80 kV) and from 86.7% to 97.3% at 100/570 mA. Specificity was 96.7% to 100% in 15/90 mA and 100% in 100/570 mA of scans. However, there was inferior sensitivity for CP owing to lower concentrations. In the receiver operating characteristics analysis, the area under the curve for MSU ranged from 0.867 to 0.947 at 15/90 mA and from 0.867 to 0.919 at 100/570 mA and that for CP from 0.659 to 0.745 and from 0.718 to 0.750, respectively.

Conclusions: This phantom study shows that single-source DECT allows detection and characterization of crystal deposits when present in soft tissue at relatively low concentrations. Further studies in patients have to prove its benefits in diagnostic imaging and treatment monitoring as well as its significance compared with dual-source CT systems.