Evaluation of Femoral Head Avascular Necrosis With Virtual Noncalcium Dual-Energy Computed Tomography

ObjectiveOur aim was to investigate the effectiveness of the dual-energy computed tomography (DECT) virtual noncalcium (VNCa) technique in avascular necrosis (AVN) for detecting bone marrow edema (BME) and staging.MethodsThis prospective study included adult patients diagnosed with unilateral or bilateral femoral head AVN between January 2023 and December 2023, who had magnetic resonance imaging (MRI) and DECT. Two participants were excluded from the study due to undergoing surgical procedures during the period between the scans. Two reviewers, blinded to MRI images and clinical data, visually examined color-coded VNCa pictures to assess BME using a binary classification (0 = normal bone marrow, 1 = BME). Same 2 reviewers also used color-coded and nonmapped images to stage AVN in accordance to the "Association for Research on Osseous Circulation" (ARCO) staging system. Interobserver agreements for the visual evaluation and staging were calculated with kappa coefficient. Following a visual assessment of BME and the staging of AVN, same 2 reviewers conducted CT density measurements on regions of BME regions utilizing DECT noncalcium images. An independent third investigator (reference standard) utilized MRI, x-ray, and clinical data to confirm the definitive diagnosis and staging of AVN. A P value less than 0.05 was considered statistically significant.ResultsFifty patients (28 men, 22 women, mean age: 44.2 +/- 13.1 years, range: 25-75 years) were included in the final analysis. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the VNCa technique in detecting BME were 96.0%, 94.4%, 97.9%, 89.4%, and 95.6%, respectively, for reviewer 1; and 96.0%, 88.9%, 96.0%, 88.9%, and 94.1%, respectively, for reviewer 2. Interobserver agreement was almost perfect (kappa = 0.84). Both reviewer 1 and reviewer 2 accurately classified 92.7% of the AVNs. The density measurements showed a statistically significant difference (P = 0.001) between the edema regions and the normal marrow regions. No statistically significant difference was observed in the density measurements of edema regions at different stages (P = 0.13).ConclusionsDECT VNCa technique exhibits excellent performance in detecting BME in hip AVN cases, as well as accurately determining the stage of AVN.

Journal

JOURNAL OF COMPUTER ASSISTED TOMOGRAPHY

Year

2025

Volume

49

Issue

1

Pages

133-139

ISSN

0363-8715

eISSN

1532-3145

DOI

10.1097/RCT.0000000000001655

WoS UID

WOS:001399827800002