POS1129 OPTIMIZATION OF DUAL ENERGY COMPUTED TOMOGRAPHY POST-PROCESSING TO REDUCE LOWER LIMB ARTIFACTS IN GOUT

International audienceBackground: Dual energy computed tomography (DECT) is highly accurate for the diagnosis of gout. However, many artifacts have been described (1,2) such as nail bed, skin, beam hardening, sub-millimeter and vascular artifacts). Their presence can lead to a risk of over-diagnosis (false positives). Objectives: Main objective of this case-control study was to determine the optimal DECT settings post-treatment parameters (ratio and attenuation coefficient (HU)) in order to reduce the frequency of lower limb artifacts in patients with suspected gout of the lower limbs. Methods: Seventy-seven patients hospitalized for suspected gout arthritis (feet/ankles and/or knees) who received a DECT imaging were consecutively included (final diagnosis of 43 gout and 34 other rheumatic disorders). Different post-treatment settings were evaluated from the Syngovia software: an R1 (standard) setting with a ratio at 1.36 and minimum attenuation at 150 HU; an R2 setting with a ratio at 1.28 and minimum attenuation at 170 HU and an R3 setting with a ratio at 1.28 and minimum attenuation at 120 HU. The frequency of each artifact according to the 3 settings was determined. Diagnostic accuracy of R1 and R2 settings has been calculated. Correlations between artefacts and patient’s clinical characteristics were obtained by performing a Spearman test. Results: The R2 setting (170 HU, ratio=1.28) significantly reduced the presence of knee and foot/ankle artifacts compared to the standard R1 setting (85% and 94% decrease in beam hardening and clumpy artifacts in the ankle and foot, respectively (p < 0.001); a decrease of 71%, 60% and 88% respectively of meniscal beam hardening, beam hardening and submillimeter artifacts in the knee (p < 0.001). The use of R3 setting lead to a significant increase of some artifacts (clumpy artifacts, skin artifacts, beam hardening and nail beds.). In addition, our results found a positive correlation between the presence of deposits of knee menisci beam hardening chondrocalcinosis. Body mass index was also positively correlated with the presence of knee beam hardening artifact. Compared to standard setting, the use of R2 settings decreased sensitivity (0.79 [95CI: 0.65;0.88] versus 0.90 [95CI: 0.78;0.96] and increased specificity (0.86 [95CI: 0.71;0.93) versus 0.63 [95CI: 0.47;0.77] (p