Advanced multiparametric MRI-based scoring for isocitrate dehydrogenase mutation prediction of gliomas.

Abstract

To propose an advanced multiparametric magnetic resonance imaging (MRI)-based scoring system and evaluate its diagnostic accuracy with respect to the isocitrate dehydrogenase (IDH) mutation status of gliomas. This prospective observational study included 50 consecutive patients with suspected gliomas, enrolled for pre-operative MRI. The exclusion criteria were previous surgery, biopsy, or chemo/radiotherapy and contraindications to the gadolinium-based contrasts or MRI acquisition. A standardized brain-MRI protocol using a 3-Tesla machine and 16-channel head coil consisted of pre-contrast axial-T2WI, FLAIR, DTI, 3D-ASL perfusion, SWI, 3D-T1WI, and post-contrast axial-DSC perfusion followed by 3D-T1WI and MR spectroscopy. ROIs were drawn from the tumoral centre, periphery, and peritumoral oedema (3 ROIs for each) followed by normalization using the ROIs over the contralateral normal white matter. The cut-off values for the statistically significant (p <0.05) continuous variables were derived by drawing receiver operating characteristic (ROC) curves. A 7-point "glioma-score" was derived from the 3 categorical (T2/FLAIR-mismatch, contrast enhancement, and intratumoral susceptibility signals) and 4 continuous ROI-based variables (ADC, FA, ASL-CBF, and DSC-CBV). The predictability of IDH mutant status using the multiparametric advanced MRI-based glioma score was statistically significant (sensitivity = 69.23%, specificity = 95.65%, PPV = 94.74%, NPV = 73.33%). A glioma score of more than 4.5 out of 7 predicted the IDH-mutation status with higher specificity and sensitivity compared to each of the individual imaging variables. The advanced multiparametric MRI-based glioma score can predict the IDH-mutation status with high statistical significance.