Accuracy of 3-T MRI using susceptibility-weighted imaging to detect meniscal tears of the knee.

Abstract

To evaluate the diagnostic performance of susceptibility-weighted imaging (SWI) at 3-T magnetic resonance imaging for the assessment of meniscal tears. Ninety-four patients with tears in the medial meniscus (31) or lateral meniscus (64) imaged with conventional magnetic resonance imaging and SWI followed by knee arthroscopy within 1 month were selected. The diagnostic values of SWI for the detection of meniscal tears were evaluated using arthroscopy as the reference standard. The sensitivity, specificity and accuracy between spin-echo T1-weighted imaging (T1WI) and fat-suppressed proton density-weighted imaging (FS-PDWI) were compared. The diagnosis consistency with two radiologists was also compared. Receiver operating characteristic curve analyses were performed for each individual sequence to estimate their diagnostic performance in meniscal tear. Analyses from 31 patients of medial meniscus tears showed that SWI achieved comparable performance with T1WI and FS-PDWI with respect to sensitivity (96.8 vs. 93.5 and 89.2%), specificity (66.7 vs. 66.7 and 66.7%) and accuracy (91.9 vs. 89.2 and 93.5%). In 64 patients of lateral meniscus tears, SWI was found to be a superior method over T1WI and FS-PDWI with regard to sensitivity (98.4 vs. 92.2 and 95.3%), specificity (100 vs. 100 and 100%) and accuracy (98.5 vs. 92.5 and 95.5%). Upon combination of these patients, SWI is similar or superior to T1WI and FS-PDWI with sensitivity (97.9 vs. 92.6 and 94.7%), specificity (77.8 vs. 77.8 and 77.8%) and accuracy (96.2 vs. 89.2 and 93.3%). SWI exhibited similar or better results with respect to sensitivity (97.9 vs. 92.6 and 94.7%), specificity (77.8 vs. 77.8 and 77.8%) and accuracy (96.2 vs. 89.2 and 93.3%) over T1WI and FS-PDWI. These data suggest that SWI can be used for the diagnosis of meniscal tears. The sensitivity, accuracy and negative predictive value were same as those of T1WI and FS-PDWI according to the present study. Diagnostic study, Level II.