American Journal of Neuroradiology (2021).

Nominated for 2021 Lucien Levy Best Research Article. AJNR Blog Annoucement

Abstract:

Background and Purpose. T2-FLAIR mismatch sign is a validated imaging sign of IDH-mutant 1p/19q non-codeleted gliomas. It is identified by radiologists through visual inspection of pre-operative MRI scans, and has been shown to identify IDH-mutant 1p/19q non-codeleted gliomas with high positive predictive value. We have developed an approach to quantify the T2-FLAIR mismatch signature, and use it to predict molecular status of lower-grade gliomas (LGG).

Materials and Methods. We used multi-parametric MRI scans and segmentation labels of 108 pre-operative LGG tumors from The Cancer Imaging Archive. Clinical information and T2-FLAIR mismatch sign labels were obtained from supplementary material of relevant publications. We adopted an objective analytical approach to estimate this sign through a geographically weighted regression (GWR), and use the residuals for each case to construct a probability density function (serving as residual signature). These functions are then analyzed using an appropriate statistical framework.

Results. We observe statistically significant (p-value = 0.05) differences between the averages of residual signatures for IDH-mutated 1p/19q non-codeleted class of tumors versus other categories. Our classifier predicts these cases with area under the curve (AUC) of 0.98, high specificity and sensitivity. It also predicts T2-FLAIR mismatch sign within these cases with an AUC of 0.93.

Conclusions. Based on this retrospective study, we show that GWR-based residual signatures are highly informative of T2-FLAIR mismatch sign, and can identify IDH mutation and 1p/19q codeletion status with high predictive power. The utility of proposed quantification of T2-FLAIR mismatch sign can be potentially validated through a prospective multi-institutional study.