SUMMARY: Neuroradiologists play a key role in brain tumor diagnosis and management. Staying current with the latest classification systems and diagnostic markers is important to provide optimal patient care. Publication of the 2016 World Health Organization Classification of Tumors of the Central Nervous System introduced a paradigm shift in the diagnosis of CNS neoplasms. For the first time, both histologic features and genetic alterations were incorporated into the diagnostic framework, classifying and grading brain tumors. The newly published 2021 World Health Organization Classification of Tumors of the Central Nervous System, May 2021, 5th edition, has added even more molecular features and updated pathologic diagnoses. We present, summarize, and illustrate the most salient aspects of the new 5th edition. We have selected the key “must know” topics for practicing neuroradiologists.

SUMMARY: Monoclonal antibodies are emerging disease-modifying therapies for Alzheimer disease that require brain MR imaging for eligibility assessment as well as for monitoring for amyloid-related imaging abnormalities. Amyloid-related imaging abnormalities result from treatment-related loss of vascular integrity and may occur in 2 forms. Amyloid-related imaging abnormalities with edema or effusion are transient, treatment-induced edema or sulcal effusion, identified on T2-FLAIR. Amyloid-related imaging abnormalities with hemorrhage are treatment-induced microhemorrhages or superficial siderosis identified on T2* gradient recalled-echo. As monoclonal antibodies become more widely available, treatment screening and monitoring brain MR imaging examinations may greatly increase neuroradiology practice volumes. Radiologists must become familiar with the imaging appearance of amyloid-related imaging abnormalities, how to select an appropriate imaging protocol, and report findings in clinical practice. On the basis of clinical trial literature and expert experience from clinical trial imaging, we summarize imaging findings of amyloid-related imaging abnormalities, describe potential interpretation pitfalls, and provide recommendations for a standardized imaging protocol and an amyloid-related imaging abnormalities reporting template. Standardized imaging and reporting of these findings are important because an amyloid-related imaging abnormalities severity score, derived from the imaging findings, is used along with clinical status to determine patient management and eligibility for continued monoclonal antibody dosing.

BACKGROUND AND PURPOSE: Distinct molecular subgroups of pediatric medulloblastoma confer important differences in prognosis and therapy. Currently, tissue sampling is the only method to obtain information for classification. Our goal was to develop and validate radiomic and machine learning approaches for predicting molecular subgroups of pediatric medulloblastoma. MATERIALS AND METHODS: In this multi-institutional retrospective study, we evaluated MR imaging datasets of 109 pediatric patients with medulloblastoma from 3 children9s hospitals from January 2001 to January 2014. A computational framework was developed to extract MR imaging–based radiomic features from tumor segmentations, and we tested 2 predictive models: a double 10-fold cross-validation using a combined dataset consisting of all 3 patient cohorts and a 3-dataset cross-validation, in which training was performed on 2 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 to evaluate model performance. RESULTS: Of 590 MR imaging–derived radiomic features, including intensity-based histograms, tumor edge-sharpness, Gabor features, and local area integral invariant features, extracted from imaging-derived tumor segmentations, tumor edge-sharpness was most useful for predicting sonic hedgehog and group 4 tumors. Receiver operating characteristic analysis revealed superior performance of the double 10-fold cross-validation model for predicting sonic hedgehog, group 3, and group 4 tumors when using combined T1- and T2-weighted images (area under the curve = 0.79, 0.70, and 0.83, respectively). With the independent 3-dataset cross-validation strategy, select radiomic features were predictive of sonic hedgehog (area under the curve = 0.70–0.73) and group 4 (area under the curve = 0.76–0.80) medulloblastoma. CONCLUSIONS: This study provides proof-of-concept results for the application of radiomic and machine learning approaches to a multi-institutional dataset for the prediction of medulloblastoma subgroups.

We report MRI findings from 2 pediatric clinical trials of diffuse intrinsic brainstem glioma (BSG) incorporating concurrent radiation therapy (RT) with molecularly targeted agents (gefitinib and tipifarnib). We determined associations of MRI variables with progression-free survival and overall survival and investigated effects of treatment on these variables. MRI (including diffusion and perfusion) was done before treatment, every 8 weeks (first year), every 12 weeks (thereafter), and at the end of treatment or disease progression. Reduced tumor volume (P < .0001) and tumor diffusion values (P <.0001) were apparent on the first post-RT/drug studies. Decreases in tumor volume correlated with pre-RT volume (P < .0001) and pre-RT diffusion values (P < .0001); larger decreases were noted for tumors with higher volumes and diffusion values. Patients with larger pre-RT tumors had longer progression-free survival (P < .0001). Patients with ≥25% decrease in tumor volume and diffusion values after RT had longer progression-free survival (P = .028) and overall survival (P = .0009). Enhancement at baseline and over time was significantly associated with shorter survival. Tumor diffusion values with baseline enhancement were significantly lower than those without (P = .0002). RT of BSG is associated with decreased tumor volume and intralesional diffusion values; patients with ≥25% decrease in values post-RT had relatively longer survival intervals, apparently providing an early imaging-based surrogate for relative outcomes. Patients with larger tumors and greater decreases in tumor volume and diffusion values had longer survival intervals. Tumor enhancement was associated with shorter survival, lower tumor diffusion values (increased cellularity), and a smaller drop in diffusion values after RT (P = .006). These associations justify continued investigation in other large clinical trials of brainstem glioma patients.

Diffuse intrinsic pontine glioma (DIPG) is associated with poor survival regardless of therapy. We used volumetric apparent diffusion coefficient (ADC) histogram metrics to determine associations with progression-free survival (PFS) and overall survival (OS) at baseline and after radiation therapy (RT). Baseline and post-RT quantitative ADC histograms were generated from fluid-attenuated inversion recovery (FLAIR) images and enhancement regions of interest. Metrics assessed included number of peaks (ie, unimodal or bimodal), mean and median ADC, standard deviation, mode, skewness, and kurtosis. Based on FLAIR images, the majority of tumors had unimodal peaks with significantly shorter average survival. Pre-RT FLAIR mean, mode, and median values were significantly associated with decreased risk of progression; higher pre-RT ADC values had longer PFS on average. Pre-RT FLAIR skewness and standard deviation were significantly associated with increased risk of progression; higher pre-RT FLAIR skewness and standard deviation had shorter PFS. Nonenhancing tumors at baseline showed higher ADC FLAIR mean values, lower kurtosis, and higher PFS. For enhancing tumors at baseline, bimodal enhancement histograms had much worse PFS and OS than unimodal cases and significantly lower mean peak values. Enhancement in tumors only after RT led to significantly shorter PFS and OS than in patients with baseline or no baseline enhancement. ADC histogram metrics in DIPG demonstrate significant correlations between diffusion metrics and survival, with lower diffusion values (increased cellularity), increased skewness, and enhancement associated with shorter survival, requiring future investigations in large DIPG clinical trials.

BACKGROUND AND PURPOSE: Posterior fossa tumors are the most common pediatric brain tumors. MR imaging is key to tumor detection, diagnosis, and therapy guidance. We sought to develop an MR imaging–based deep learning model for posterior fossa tumor detection and tumor pathology classification. MATERIALS AND METHODS: The study cohort comprised 617 children (median age, 92 months; 56% males) from 5 pediatric institutions with posterior fossa tumors: diffuse midline glioma of the pons (n = 122), medulloblastoma (n = 272), pilocytic astrocytoma (n = 135), and ependymoma (n = 88). There were 199 controls. Tumor histology served as ground truth except for diffuse midline glioma of the pons, which was primarily diagnosed by MR imaging. A modified ResNeXt-50-32x4d architecture served as the backbone for a multitask classifier model, using T2-weighted MRIs as input to detect the presence of tumor and predict tumor class. Deep learning model performance was compared against that of 4 radiologists. RESULTS: Model tumor detection accuracy exceeded an AUROC of 0.99 and was similar to that of 4 radiologists. Model tumor classification accuracy was 92% with an F₁ score of 0.80. The model was most accurate at predicting diffuse midline glioma of the pons, followed by pilocytic astrocytoma and medulloblastoma. Ependymoma prediction was the least accurate. Tumor type classification accuracy and F₁ score were higher than those of 2 of the 4 radiologists. CONCLUSIONS: We present a multi-institutional deep learning model for pediatric posterior fossa tumor detection and classification with the potential to augment and improve the accuracy of radiologic diagnosis.