Diagnostic performance and reader confidence of MRI-derived CT-like images in patients with femoral head necrosis
摘要
Femoral head necrosis (FHN) requires accurate imaging for staging and treatment planning. While MRI is sensitive, conventional sequences may miss subchondral fractures. A CT-like MRI sequence combines CT-like structural detail with soft-tissue contrast. We evaluated its performance for Association Research Circulation Osseous (ARCO) staging compared with MRI and radiography. In this retrospective study, 21 patients (33 hips) with confirmed FHN underwent radiography and MRI, including T1- and T2-weighted sequences and a T1-weighted gradient-echo CT-like sequence. Two musculoskeletal imaging specialists assessed ARCO stage, reader confidence, and image quality. Interrater reliability was evaluated using the intraclass correlation coefficient, intermodality agreement with weighted Cohen’s κ, and modality comparisons with Wilcoxon signed-rank tests. CT-like MRI demonstrated higher ARCO staging than radiography (p = 0.004), T1-weighted (p = 0.031), and T2-weighted MRI (p = 0.046). Upstaging compared with conventional modalities occurred in 6 of 33 hips (18.2%). Reader confidence was highest for CT-like MRI (all p ≤ 0.003), and image quality was superior compared with T1- and T2-weighted MRI (both p < 0.001). Agreement in ARCO staging was excellent with T1-weighted and T2-weighted images and substantial with radiography. CT-like MRI improved image quality and diagnostic confidence, enabling more accurate FHN assessment and detection of subchondral collapse missed by conventional MRI.
Relevance statementCT-like MRI improves detection of early femoral head collapse, enabling more accurate staging and treatment planning.
Key PointsFemoral head necrosis requires precise imaging to accurately stage the disease. Conventional MRI may miss early subchondral femoral head fractures. CT-like MRI enhances visualization of bone microstructure, diagnostic confidence and image quality. Improved staging with CT-like MRI may influence clinical management.