Background and Purpose: Our goal was to develop and validate radiomics and machine learning approaches for predicting molecular subgroups of pediatric medulloblastoma (MB). Materials and Methods: In this multi-institutional retrospective study, we evaluated MRI datasets of 109 pediatric MB patients from three children’s hospitals from January 2001 to January 2014. A computational framework was developed to extract MRI-based radiomic features from tumor segmentations, and two predictive models were tested: a double 10-fold cross validation using a combined dataset consisting of all three patient cohorts and a three-dataset cross-validation, in which training was performed on two cohorts and testing was performed on the third independent cohort. We used the Wilcoxon rank sum test for feature selection with assessment of area under the receiver operating characteristic curve (AUC) to evaluate model performance. Results: Of 590 MRI-derived radiomic features, including intensity-based histograms, tumor edge sharpness, Gabor features, and Local Area Integral Invariant (LAII) features, extracted from imaging-derived tumor segmentations, tumor edge sharpness was most significant for predicting sonic hedgehog (SHH) and group 4 tumors. ROC analysis revealed superior performance of the double 10-fold cross validation model for predicting SHH, group 3, and group 4 tumors when using combined T1- and T2-weighted images (AUC=0.79, 0.70, and 0.83, respectively). Using the independent three-dataset cross validation strategy, select radiomics features were predictive of SHH (AUC=0.70–0.73) and group 4 (AUC=0.76–0.80) MB. Conclusion: This study provides proof-of-concept results for the application of radiomics and machine learning approaches to a multi-institutional dataset for the prediction of MB subgroups.

Imaging plays a vital role by providing the information necessary for AVM management. Here, we discuss the background, natural history, clinical presentation, and imaging of AVMs. In addition, we explain advances in techniques for imaging AVMs.

Background and Purpose Magnetic susceptibility measured with quantitative susceptibility mapping (QSM) has been proposed as a biomarker for demyelination and inflammation in MS patients, but investigations have mostly been on white matter (WM) lesions. A detailed characterization of cortical lesions has yet to be performed. The purpose of this study was to evaluate magnetic susceptibility in both cortical and WM lesions in MS using QSM. Materials and Methods Fourteen patients with MS were scanned at a 7T MR scanner with T1, T2 and T2*-weighted sequences. The T2*-weighted sequence was used to perform QSM and generate tissue susceptibility maps. The susceptibility contrast of a lesion was quantified as the relative susceptibility between the lesion and its adjacent normal-appearing parenchyma. The susceptibility difference between cortical and WM lesions was assessed using the t-test. Results The mean relative susceptibility was significantly negative for cortical lesions (p<10−7) but positive for WM lesions (p<10−22). A similar pattern was also observed in the cortical (p=0.054) and WM portions (p=0.043) of mixed lesions. Conclusion The negative susceptibility in cortical lesions suggests that iron loss dominates the susceptibility contrast in cortical lesions. The opposite susceptibility contrast between cortical and WM lesions may reflect both their structural (degree of myelination) and pathological differences (degree of inflammation), where the latter may lead to a faster release of iron in cortical lesions.

No abstract