Prediction of meningioma subtypes using qualitative assessment of fractional anisotropy maps.

AIM: In this study, we explore the role of FA maps in predicting the histopathological subtypes of meningioma using a qualitative ordinal scale and quantitative histogram features.

MATERIAL &

METHODS

Retrospective analysis of grey-scale FA maps of 96 cases of meningioma was done by two observers blinded to the histopathological diagnosis. An ordinal scale of 1-4 was used to grade the degree of FA in each lesion. Histogram features were calculated from the region of interest and statistical analysis was carried out for discriminating between the above-mentioned qualitative gradings.

RESULTS

Out of 98 cases, there were 9 meningothelial/15 transitional/14 chordoid/8 angiomatous/15 microcystic/17 fibroblastic/18 atypical meningiomas. Interobserver reliability had intraclass correlation coefficient of 0.92. The intergroup comparison revealed significantly low FA grade in microcystic/angiomatous/transitional meningioma compared to meningothelial/chordoid/ fibroblastic/ atypical meningioma with transitional being a relatively heterogenous subgroup compared to the former two. While all fibroblastic meningiomas showed high FA grade, it is relatively non-specific since several grade 3 and 4 meningiomas were also seen among meningothelial/chordoid/ atypical subtypes. All quantitative median FA values were found to be significantly correlated with qualitative FA grades as assigned by the interpreting radiologist (Spearman's Rho for mean is 0.502 (P < 0.001), for 75th percentile is 0.505 (P < 0.001). However, the strength of correlation for all metrics was moderate and positive.

CONCLUSION

Qualitative grading of FA maps is useful in predicting the meningioma subtype. Low FA is characteristically seen in microcystic and angiomatous variants. High FA within tumour although a consistent feature of Fibroblastic variant, is nonspecific.