Min Chen1, Nele Herregods2, Jacob L Jaremko3, Philippe Carron4,5, Dirk Elewaut4,5, Filip Van den Bosch4,5, Lennart Jans2. 1. Department of Radiology, Ghent University Hospital, Corneel Heymanslaan 10, 9000, Ghent, Belgium. Min.Chen@Ugent.be. 2. Department of Radiology, Ghent University Hospital, Corneel Heymanslaan 10, 9000, Ghent, Belgium. 3. Department of Radiology & Diagnostic Imaging, University of Alberta Hospital, 8440-112 Street, Edmonton, Alberta, T6G 2B7, Canada. 4. Department of Rheumatology, Ghent University Hospital, Corneel Heymanslaan 10, 9000, Ghent, Belgium. 5. VIB Inflammation Research Center, Unit for Molecular Immunology and Inflammation, Ghent University, Technologiepark 927, 9052, Ghent, Belgium.
Abstract
OBJECTIVES: To evaluate the feasibility and diagnostic accuracy of dual-energy computed tomography (DECT) for the detection of bone marrow edema (BME) in patients suspected for sacroiliitis. METHODS: Patients aged 18-55 years with clinical suspicion for sacroiliitis were enrolled. All patients underwent DECT and 3.0 T MRI of the sacroiliac joints on the same day. Virtual non-calcium (VNCa) images were calculated from DECT images for demonstration of BME. VNCa images were scored by two readers independently using a binary system (0 = normal bone marrow, 1 = BME). Diagnostic performance was assessed with fluid-sensitive MRI as the reference standard. ROIs were placed on VNCa images, and CT numbers were displayed. Cutoff values for BME detection were determined based on ROC curves. RESULTS: Forty patients (16 men, 24 women, mean age 37.1 years ± 9.6 years) were included. Overall inter-reader agreement for visual image reading of BME on VNCa images was good (κ = 0.70). The sensitivity and specificity of BME detection by DECT were 65.4% and 94.2% on the quadrant level and 81.3% and 91.7% on the patient level. ROC analyses revealed AUCs of 0.90 and 0.87 for CT numbers in the ilium and sacrum, respectively. Cutoff values of - 44.4 HU (for iliac quadrants) and - 40.8 HU (for sacral quadrants) yielded sensitivities of 76.9% and 76.7% and specificities of 91.5% and 87.5%, respectively. CONCLUSIONS: Inflammatory sacroiliac BME can be detected by VNCa images calculated from DECT, with a good interobserver agreement, moderate sensitivity, and high specificity. KEY POINTS: • Virtual non-calcium images calculated from dual-energy CT can detect sacroiliac bone marrow edema in patients suspected for sacroiliitis. • Dual-energy CT has a high specificity in bone marrow edema detection. • Virtual non-calcium images for bone marrow edema in patients with a large amount of red bone marrow or obvious sclerosis near the articular surface should be interpreted with caution.
OBJECTIVES: To evaluate the feasibility and diagnostic accuracy of dual-energy computed tomography (DECT) for the detection of bone marrow edema (BME) in patients suspected for sacroiliitis. METHODS:Patients aged 18-55 years with clinical suspicion for sacroiliitis were enrolled. All patients underwent DECT and 3.0 T MRI of the sacroiliac joints on the same day. Virtual non-calcium (VNCa) images were calculated from DECT images for demonstration of BME. VNCa images were scored by two readers independently using a binary system (0 = normal bone marrow, 1 = BME). Diagnostic performance was assessed with fluid-sensitive MRI as the reference standard. ROIs were placed on VNCa images, and CT numbers were displayed. Cutoff values for BME detection were determined based on ROC curves. RESULTS: Forty patients (16 men, 24 women, mean age 37.1 years ± 9.6 years) were included. Overall inter-reader agreement for visual image reading of BME on VNCa images was good (κ = 0.70). The sensitivity and specificity of BME detection by DECT were 65.4% and 94.2% on the quadrant level and 81.3% and 91.7% on the patient level. ROC analyses revealed AUCs of 0.90 and 0.87 for CT numbers in the ilium and sacrum, respectively. Cutoff values of - 44.4 HU (for iliac quadrants) and - 40.8 HU (for sacral quadrants) yielded sensitivities of 76.9% and 76.7% and specificities of 91.5% and 87.5%, respectively. CONCLUSIONS: Inflammatory sacroiliac BME can be detected by VNCa images calculated from DECT, with a good interobserver agreement, moderate sensitivity, and high specificity. KEY POINTS: • Virtual non-calcium images calculated from dual-energy CT can detect sacroiliac bone marrow edema in patients suspected for sacroiliitis. • Dual-energy CT has a high specificity in bone marrow edema detection. • Virtual non-calcium images for bone marrow edema in patients with a large amount of red bone marrow or obvious sclerosis near the articular surface should be interpreted with caution.
Authors: Tommaso D'Angelo; Moritz H Albrecht; Danilo Caudo; Silvio Mazziotti; Thomas J Vogl; Julian L Wichmann; Simon Martin; Ibrahim Yel; Giorgio Ascenti; Vitali Koch; Giuseppe Cicero; Alfredo Blandino; Christian Booz Journal: Eur Radiol Exp Date: 2021-09-03
Authors: Marco Cavallaro; Tommaso D'Angelo; Moritz H Albrecht; Ibrahim Yel; Simon S Martin; Julian L Wichmann; Lukas Lenga; Silvio Mazziotti; Alfredo Blandino; Giorgio Ascenti; Marcello Longo; Thomas J Vogl; Christian Booz Journal: Eur Radiol Date: 2021-07-02 Impact factor: 5.315