Edema-Enhanced CT Derived from Quantitative Dual-Energy CT Reduces Experience-Related Variability in Sacral Fragility Fracture Diagnosis.

Abstract

Sacral fractures are difficult to diagnose. In dual-energy computed tomography (DECT) water/Ca images, edema at fracture sites appears as locally increased water-equivalent density. This study developed an edema-enhanced CT (EECT) technique based on DECT and evaluated its diagnostic performance for sacral fractures. Fifty Japanese patients with suspected sacral fractures (mean age, 82.4 years; 41 women) underwent single-energy CT (SECT), DECT, and magnetic resonance imaging (MRI). On sagittal DECT, the water-equivalent density at the S1 vertebral midline was defined as the standard value (SV); in 16 regions of interest per patient, the relative standard value (RSV) was calculated as the ratio relative to the SV and classified according to MRI findings. Receiver operating characteristic curve analysis determined the optimal RSV cutoff for fracture detection, and EECT images highlighted regions exceeding this threshold. Thirty patients were then independently assessed twice (>3-month interval) by orthopaedic surgeons with <3 years (group A) or >10 years (group B) of experience using SECT (S group), DECT (D group), and EECT (E group) to evaluate fracture presence and localization. Sensitivity, specificity, inter- and intraobserver reliability, and diagnostic accuracy were compared among groups using Cochran Q, McNemar, and Fisher exact tests. The optimal RSV cutoff was 1.000. For both fracture detection and localization, EECT showed the highest sensitivity, specificity, and diagnostic accuracy, particularly for fracture site identification. Inter- and intraobserver reliability were also the highest with EECT, and the performance gap between less experienced and experienced observers seen with SECT was not observed when using EECT. EECT showed better performance than DECT and SECT in sacral fracture detection and localization, providing high accuracy and reliability regardless of observer experience, and it may be useful for sacral fractures. Diagnostic Level IV. See Instructions for Authors for a complete description of levels of evidence.