Index by author

MR imaging data and molecular information were retrospectively obtained from The Cancer Imaging Archives for 259 patients with either low- or high-grade gliomas. A convolutional neural network was trained to classify IDH1 mutation status, 1p/19q codeletion, and MGMT promotor methylation status. Classification had high accuracy: IDH1 mutation status, 94%; 1p/19q codeletion, 92%; and MGMT promotor methylation status, 83%. The authors conclude that this shows the feasibility of a deep-learning CNN approach for the accurate classification of individual genetic mutations of both low- and high-grade gliomas and that the relevant MR imaging features acquired from an added dimensionality-reduction technique are concordant with existing literature, showing that neural networks are capable of learning key imaging components without prior feature selection or human directed training.

This study aimed to evaluate morphologic patterns and the delineation of deeply seated collateral networks using ultra-high-field MRA in comparison with conventional DSA in 15 patients. Sequences acquired at 7T were TOF-MRA with 0.22 X 0.22 X 0.41 mm³ resolution and MPRAGE with 0.7 X 0.7 X 0.7 mm³ resolution. The relevant deeply seated collateral networks were classified into 2 categories and 6 pathways. A total of 100 collateral networks were detected on DSA; 106, on TOF-MRA; and 73, on MPRAGE. Delineation of deeply seated collateral networks was comparable between TOF-MRA and DSA. The authors demonstrate excellent delineation of 6 distinct deeply seated collateral network pathways in Moyamoya angiopathy.