Conventional MR-based Preoperative Nomograms for Prediction of IDH/1p19q Subtype in Low-Grade Glioma.

RATIONALE AND

OBJECTIVES

To develop nomogram models incorporating MR and clinical features for preoperative prediction of isocitrate dehydrogenase (IDH)/1p19q subtype in patients with lower-grade gliomas (LGG).

MATERIALS AND

METHODS

We classified LGG (149 patients) into three categories: (1) IDH mutation and 1p/19q codeletion, (2) IDH mutation and no 1p/19q codeletion, and (3) wild-type IDH. The correlation between clinical and MR features and IDH/1p19q subtype was analyzed.

RESULTS

(1) Multivariate analysis showed that hemorrhage (yes vs no odds ratio [OR]: 12.775), enhancing margin (poorly vs well defined OR: 17.87), and SVZ (SVZ+ vs SVZ- OR: 0.304 were associated with a higher incidence of IDHmut-codel status (All p < 0.05). (2) Multivariate analysis showed that age (≥40 years vs <40 years OR: 0.139), hemorrhage (yes vs no OR: 0.095), enhancing margin (poorly vs well defined OR: 0.275), volume (>60 cm3 vs ≤60 cm3 OR: 5.111), and the shortest distance from the tumor centroid to the edge of the lateral ventricles (CS) (>30 mm vs ≤30 mm OR: 3.766) were associated with a higher incidence of IDHmut-noncodel status. (3) Multivariate analysis showed age (≥40 years vs <40 years OR: 17.311), tumor site (other vs frontal lobe OR: 4.696), volume (>60 cm3 vs ≤60 cm3 OR: 0.188), CS (>30 mm vs ≤30 mm OR: 0.285), necrosis (yes vs no OR: 0.193), and proportion CE tumor (>5% vs ≤5% OR: 5.253) were associated with a higher incidence of IDHwt status. Three nomogram models showed good discrimination (all area under the curve > 0.8) and calibration.

CONCLUSION

Clinical and MR features may therefore be used to facilitate the preoperative prediction of LGG IDH/1p19q subtype.