ABSTRACT
PURPOSE: The aim of this study was to evaluate three-dimensional (3D) turbo spin-echo (TSE) magnetic resonance imaging (MRI) for semiquantitative assessment of knee OA. MATERIALS AND METHOD: Twenty subjects fulfilling the American College of Rheumatology clinical criteria of knee OA underwent both two-dimensional (2D) and 3D MRIs on the same day. The 2D MRI protocol included triplanar fat-suppressed (FS) intermediate-weighted (Iw) TSE. For the 3D TSE technique, a sagittal FS Iw sequence was acquired and triplanar reformations were constructed. 2D and 3D MRIs were read separately by two radiologists using the Whole-Organ Magnetic Resonance Imaging Score (WORMS) system. Agreement was determined using weighted kappa statistics and percentage of overall agreement. The diagnostic performance of WORMS readings using 3D TSE MRI to detect the presence or absence of features was assessed using readings from 2D TSE images as a reference. RESULTS: Agreement for the scored features ranged between 0.62 (osteophytes (OS)) and 0.94 (meniscal extrusion). The sensitivity of WORMS readings using the 3D TSE technique ranged between 80% (periarticular cysts) and 100% (several features), the specificity ranged between 62.3% (OS) and 100% (several features), and accuracy ranged between 77.2% (OS) and 99.3% (subchondral cysts). CONCLUSIONS: Semiquantitative assessment of knee OA can be reliably performed using 3D TSE MRI, showing substantial to almost perfect agreement and high accuracy when compared to routine 2D TSE MRI. 3D TSE MRI also takes less time, which is important for large OA studies.
