Diagnostic accuracy of dual-layer spectral CT for osteolytic vertebral metastases

Abstract

OBJECTIVES: To evaluate whether dual-layer spectral computed tomography, compared with conventional CT, improves diagnostic accuracy for osteolytic vertebral metastases. Furthermore, to investigate the influence of dual-layer CT on the subjective visibility of metastases. MATERIALS AND METHODS: In this single-center retrospective study, consecutive patients with an untreated primary tumor who underwent dual-layer CT and either MRI or PET-CT as reference standard within 14 days were included. Two independent observers, blinded to the reference, performed two scorings. First, the conventional CT was scored and the results were recorded. Subsequently, Calcium suppression, monoenergetic (monoE40 and monoE200), and Z-effective reconstructions were added. Subjective visibility was compared to conventional CT using a 5-point Likert scale. Diagnostic accuracy measures were calculated with 95% confidence intervals. Sensitivity and specificity were compared using the McNemar's test. RESULTS: Fourteen patients (63 ± 8 years; 64.3% female) and 189 vertebrae were included, with 46 vertebrae showing 58 metastases with a mean diameter of 18 mm (range 5-53 mm). For conventional CT, the sensitivity, specificity, and diagnostic accuracy for observer A and B were, respectively, 57% and 57%, 96% and 90%, 85% and 81%. The diagnostic performance did not improve when using the dual-layer CT reconstructions in addition to conventional CT (p ≥ 0.13). MonoE40 improved the subjective visibility of metastases. Interobserver agreement was moderate for conventional CT (ĸ:0.48), and dual-layer CT reconstructions (ĸ:0.41-0.51). CONCLUSION: Dual-layer CT reconstructions did not improve diagnostic accuracy for osteolytic vertebral metastases compared with conventional CT, although subjective visibility was improved on low monoenergetic reconstructions.