Qualitative and Quantitative Diagnosis of Meniscal Tears Using SWI Compared with T2mapping at 3-Tesla MRI
Jun Zhao1, Wei Chen1, Jian Wang1, Shuai Li2, and Wen-Jing Hou1

1Radiology, Southwest Hospital, Third Military Medical University, Chongqing, China, People's Republic of, 2MR Collaborations NE Asia, Siemens Healthcare, Beijing, China, People's Republic of

Synopsis

In the past reports, invariably irregularity and high-signal-intensity changes of the free edge of meniscus may lead to a false-positive MR imaging, in addition, MR imaging of the knee invariably missed small meniscal tears, tears and abnormalities of the meniscal free edge, and at times large, unstable tears, result in false-negative. In recent decades, new MR image of water-tissues and collagen-rich tissues, including cartilage, menisci and tendon, has undergone significant progress, which are biological MR image techniques for the characterization tissues .This study was to compare the diagnostic performance of SWI (Susceptibility Weighted Imaging) in the evaluation of meniscal tears at 3T MR with those of a T2 Mapping sequence, using phase value and T2 value as the quantitative parameters. The phase value was a good predictor to diagnose meniscal tears.

Purpose

The purpose of this study was to compare the diagnostic performance of SWI sequence in the evaluation of meniscal tears at 3T MR with those of a T2 Mapping sequence, quantitatively and qualitatively, with arthroscopy as gold standard.

Methods

Thirty-four patients suspected meniscal injury and would undergo arthroscopic surgery were selected. The MRI examination was performed at 3.0T in these patients before surgery by using a conventional protocol with the addition of a sagittal T2 Mapping sequence and a sagittal SWI sequence. One hundred and thirty-six sections of meniscus including anterior and posterior horns of medial and lateral meniscus were evaluated by two radiologists. The sensitivities, specificities and accuracy of T2 Mapping and SWI were analyzed as compared to the golden standard of arthroscopy. The agreement of each radiologist using T2 Mapping and SWI was calculated respectively using κ test. The interobserver agreements of T2 Mapping and SWI were also calculated. Diagnostic performance of T2 Mapping and SWI in the evaluation of meniscal tears was analyzed qualitatively by using McNemar test and quantitatively by using the receiver operating characteristic curve (ROC curve). A dual-normal parameters model was used to compare the difference of sensitivity between T2 Mapping and SWI. Phase value of SWI and T2 value were measured in all sections of meniscus, and the differences between torn and no torn sections of them were analyzed by using Mann-Whitney tests. Spearman’s rank correlation coefficients were calculated between phase values and arthroscopy, T2 value and arthroscopy, respectively.

Results

T2 mapping and SWI had sensitivities of 90% and 87.8%, respectively; specificities of 92.3% and 95.6%, respectively; and accuracy of 91.5% and 93%, respectively for reader 1. For reader 2, T2 mapping and SWI had sensitivities of 84.4% and 87.8%, respectively; specificities of 89% and 92.3%, respectively; and accuracy of 87.5% and 90.8%, respectively. The interobserver agreement had a Cohen κ of 0.97 for T2 mapping and a Cohen of κ 0.89 for SWI, The intraobserver agreements had Cohen κ values of 0.85 and 0.82 for T2 mapping, 0.89 and 0.82 for SWI, respectively. There was no significant difference in detection of meniscal tears on T2 mapping compared with SWI for both readers. There were significantly differences between torn and no torn meniscus of phase value and T2 value (P<0.05). Both phase value and T2 value demonstrated a greater ability to distinguish normal and meniscal tear using receiver operating characteristic (ROC) analysis. The areas under the ROC curves for phase value and T2 value were respectively 0.95 and 0.88, which were significantly different.Significant negative correlations between phase value and arthroscopy(R=-0.77, P<0.01), as well as T2 value and arthroscopy(R=-0.64, P<0.01) were observed.

Discussion

The most important discovery of this study was that SWI was demonstrate to be a quantitative predictor for diagnosing meniscal tear, with the phase value had better sensitivity than that of T2 value, and increased the diagnostic efficiency in meniscal tear, quantitatively. the qualitative results revealed that SWI and T2 mapping had high sensitivities, specificities in the diagnosis of meniscal tear by both two radiologists which was in accord with recent literatures compared with current technologies permitting sensitivities and specificities. The intraobserver agreement of both radiologists was perfect for SWI and T2 mapping, as well as the interobserver agreement for the both sequences, because all of the κ values were greater than 0.8.The diagnostic performance of SWI was the same as that of T2 mapping. The quantitative results revealed that the sensitivity of phase value was higher than that of T2 value, both of them had the ability to distinguish meniscal tears from the normal and neighboring tissues, with negative correlation between them and arthroscopy.

Conclusion

The sequence of SWI can be used into diagnosis of meniscus tear. Both phase value and T2 value can distinguish meniscal tears, and sensitivity of phase value is higher than that of T2 value.

Acknowledgements

No acknowledgement found.

References

No reference found.

Figures

ROC curve of phase value and T2 value to diagnose meniscal tear

A 45-year-old male with medial meniscal tear.a-d sagital T2、T2mapping、SWI and Phase images



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2299