Dual-energy CT virtual non-calcium imaging for differentiating pathologic from non-pathologic femoral fractures: a study of inexperienced non-radiologist readers.

Abstract

Objective: To assess the effect of dual-energy CT (DECT) virtual non-calcium (VNCa) imaging on the ability of inexperienced non-radiologist readers to differentiate pathologic from non-pathologic femoral fractures.

Materials and methods: In this retrospective diagnostic accuracy study, 45 eligible emergency department patients with femoral fractures who underwent DECT between January 2022 and September 2024 were included, comprising 20 patients with pathologic fractures and 25 with non-pathologic fractures. The reference standard for fracture classification was established using surgical records or histopathologic results. Two inexperienced non-radiologist readers independently evaluated conventional CT alone and conventional CT combined with color-coded DECT VNCa images in separate reading sessions. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were compared between the two sessions. AUCs were compared using the paired DeLong test.

Results: The addition of DECT VNCa significantly improved sensitivity for detecting pathologic fractures in both readers. For Reader 1, sensitivity increased from 65 to 95% (p = 0.031), and accuracy increased from 77.8% to 95.6%. For Reader 2, sensitivity increased from 50 to 85% (p = 0.016), and accuracy increased from 68.9% to 91.1%. Specificity remained high without significant differences between sessions. AUC increased from 0.765 (95% CI, 0.617-0.913) to 0.955 (95% CI, 0.883-1.0) for Reader 1 and from 0.670 (95% CI, 0.506-0.834) to 0.905 (95% CI, 0.802-1.0) for Reader 2 (p = 0.0135 and < 0.01, respectively).

Conclusion: Adding color-coded DECT VNCa to conventional CT improves the diagnostic performance of inexperienced non-radiologist readers in detecting pathologic femoral fractures.