Search

OBJECTIVE Resection of brain metastases (BMs) may be associated with increased risk of leptomeningeal disease (LMD). This study examined rates and predictors of LMD, including imaging subtypes, in patients who underwent resection of a BM followed by postoperative radiation. METHODS A retrospective, single-center study was conducted examining overall LMD, classic LMD (cLMD), and nodular LMD (nLMD) risk. Logistic regression, Cox proportional hazards, and random forest analyses were performed to identify risk factors associated with LMD. RESULTS Of the 217 patients in the cohort, 47 (21.7%) developed postoperative LMD, with 19 cases (8.8%) of cLMD and 28 cases (12.9%) of nLMD. Six-, 12-, and 24-month LMD-free survival rates were 92.3%, 85.6%, and 71.4%, respectively. Patients with cLMD had worse survival outcomes from the date of LMD diagnosis compared with nLMD (median 2.4 vs 6.9 months, p = 0.02, log-rank test). Cox proportional hazards analysis identified cerebellar/insular/occipital location (hazard ratio [HR] 3.25, 95% confidence interval [CI] 1.73–6.11, p = 0.0003), absence of extracranial disease (HR 2.49, 95% CI 1.27–4.88, p = 0.008), and ventricle contact (HR 2.82, 95% CI 1.5–5.3, p = 0.001) to be associated with postoperative LMD. A predictive model using random forest analysis with an area under the receiver operating characteristic curve of 0.87 in a test cohort identified tumor location, systemic disease status, and tumor volume as the most important factors associated with LMD. CONCLUSIONS Tumor location, absence of extracranial disease at the time of surgery, ventricle contact, and increased tumor volume were associated with LMD. Further work is needed to determine whether escalating therapies in patients at risk of LMD prevents disease dissemination.

Meningiomas are the most common primary intracranial tumors. There are no effective medical therapies for meningioma patients, and new treatments have been encumbered by limited understanding of meningioma biology. Here, we use DNA methylation profiling on 565 meningiomas integrated with genetic, transcriptomic, biochemical, proteomic and single-cell approaches to show meningiomas are composed of three DNA methylation groups with distinct clinical outcomes, biological drivers and therapeutic vulnerabilities. Merlin-intact meningiomas (34%) have the best outcomes and are distinguished by NF2/Merlin regulation of susceptibility to cytotoxic therapy. Immune-enriched meningiomas (38%) have intermediate outcomes and are distinguished by immune infiltration, HLA expression and lymphatic vessels. Hypermitotic meningiomas (28%) have the worst outcomes and are distinguished by convergent genetic and epigenetic mechanisms driving the cell cycle and resistance to cytotoxic therapy. To translate these findings into clinical practice, we show cytostatic cell cycle inhibitors attenuate meningioma growth in cell culture, organoids, xenografts and patients.

Objectives To evaluate whether combining fast acquisitions with deep-learning reconstruction can provide diagnostically useful images and quantitative assessment comparable to standard-of-care acquisitions for lumbar spine MRI. Methods Eighteen patients were imaged with both standard protocol and fast protocol using reduced signal averages, each protocol including sagittal fat-suppressed T2-weighted, sagittal T1-weighted, and axial T2-weighted 2D fast spin-echo sequences. Fast-acquisition data was additionally reconstructed using vendor-supplied deep-learning reconstruction with three different noise reduction factors. For qualitative analysis, standard images as well as fast images with and without deep-learning reconstruction were graded by three radiologists on five different categories. For quantitative analysis, convolutional neural networks were applied to sagittal T1-weighted images to segment intervertebral discs and vertebral bodies, and disc heights and vertebral body volumes were derived. Results Based on noninferiority testing on qualitative scores, fast images without deep-learning reconstruction were inferior to standard images for most categories. However, deep-learning reconstruction improved the average scores, and noninferiority was observed over 24 out of 45 comparisons (all with sagittal T2-weighted images while 4/5 comparisons with sagittal T1-weighted and axial T2-weighted images). Interobserver variability increased with 50 and 75% noise reduction factors. Deep-learning reconstructed fast images with 50% and 75% noise reduction factors had comparable disc heights and vertebral body volumes to standard images (r2 ≥ 0.86 for disc heights and r2 ≥ 0.98 for vertebral body volumes). Conclusions This study demonstrated that deep-learning-reconstructed fast-acquisition images have the potential to provide noninferior image quality and comparable quantitative assessment to standard clinical images.

SummaryCell state evolution underlies tumor development and response to therapy1, but mechanisms specifying cancer cell states and intratumor heterogeneity are incompletely understood. Schwannomas are the most common tumors of the peripheral nervous system and are treated with surgery and ionizing radiation2–5. Schwannomas can oscillate in size for many years after radiotherapy6,7, suggesting treatment may reprogram schwannoma cells or the tumor microenvironment. Here we show epigenetic reprogramming shapes the cellular landscape of schwannomas. We find schwannomas are comprised of 2 molecular groups distinguished by reactivation of neural crest development pathways or misactivation of nerve injury mechanisms that specify cancer cell states and the architecture of the tumor immune microenvironment. Schwannoma molecular groups can arise independently, but ionizing radiation is sufficient for epigenetic reprogramming of neural crest to immune-enriched schwannoma by remodeling chromatin accessibility, gene expression, and metabolism to drive schwannoma cell state evolution and immune cell infiltration. To define functional genomic mechanisms underlying epigenetic reprograming of schwannomas, we develop a technique for simultaneous interrogation of chromatin accessibility and gene expression coupled with genetic and therapeutic perturbations in single-nuclei. Our results elucidate a framework for understanding epigenetic drivers of cancer evolution and establish a paradigm of epigenetic reprograming of cancer in response to radiotherapy.

Supplemental material is available for this article. Keywords: Informatics, MR Diffusion Tensor Imaging, MR Perfusion, MR Imaging, Neuro-Oncology, CNS, Brain/Brain Stem, Oncology, Radiogenomics, Radiology-Pathology Integration © RSNA, 2022.

Neural networks were trained for segmentation and longitudinal assessment of posttreatment diffuse glioma. A retrospective cohort (from January 2018 to December 2019) of 298 patients with diffuse glioma (mean age, 52 years ± 14 [SD]; 177 men; 152 patients with glioblastoma, 72 patients with astrocytoma, and 74 patients with oligodendroglioma) who underwent two consecutive multimodal MRI examinations were randomly selected into training (n = 198) and testing (n = 100) samples. A posttreatment tumor segmentation three-dimensional nnU-Net convolutional neural network with multichannel inputs (T1, T2, and T1 postcontrast and fluid-attenuated inversion recovery [FLAIR]) was trained to segment three multiclass tissue types (peritumoral edematous, infiltrated, or treatment-changed tissue [ED]; active tumor or enhancing tissue [AT]; and necrotic core). Separate longitudinal change nnU-Nets were trained on registered and subtracted FLAIR and T1 postlongitudinal images to localize and better quantify and classify changes in ED and AT. Segmentation Dice scores, volume similarities, and 95th percentile Hausdorff distances ranged from 0.72 to 0.89, 0.90 to 0.96, and 2.5 to 3.6 mm, respectively. Accuracy rates of the posttreatment tumor segmentation and longitudinal change networks being able to classify longitudinal changes in ED and AT as increased, decreased, or unchanged were 76%-79% and 90%-91%, respectively. The accuracy levels of the longitudinal change networks were not significantly different from those of three neuroradiologists (accuracy, 90%-92%; κ, 0.58-0.63; P > .05). The results of this study support the potential clinical value of artificial intelligence-based automated longitudinal assessment of posttreatment diffuse glioma. Keywords: MR Imaging, Neuro-Oncology, Neural Networks, CNS, Brain/Brain Stem, Segmentation, Quantification, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2022.

PurposeTo evaluate whether an AI-based method could be used routinely as part of patient care to assist in detecting non-enhancing glioma progression.Materials and MethodsA 3D U-Net trained (n=481) and validated (n=121) to segment post-surgical lower grade gliomas was used to measure tumor volumes over time and assess progression in a clinical test set. Eight prospective and eight retrospective patients (total 72 exams) who were suspected of progression during their routine outpatient imaging were clinically assessed. Gold standards for progression were derived from clinical reports a posteriori using visual read, and radiologists were blinded to the AI decision at time of reporting.ResultsProgression assessments were presented to radiologists via an easy-to-use, interactive, and interpretable environment in under 10 minutes. Combining prospective and retrospective cases, a final sensitivity of 0.72 and specificity of 0.75 was achieved at progression detection.ConclusionsAutomated detection of glioma progression would provide valuable decision support for routine use.

Quantitative susceptibility mapping (QSM) is a promising tool for investigating iron dysregulation in neurodegenerative diseases, including Huntington's disease (HD). A diverse range of methods have been proposed to generate accurate and robust QSM images. In this study, we evaluated the performance of different dipole inversion algorithms for iron-sensitive susceptibility imaging at 7T on healthy subjects of large age range and patients with HD. We compared an iterative least-squares-based method (iLSQR), iterative methods that use regularization, single-step approaches, and deep learning-based techniques. Their performance was evaluated by comparing: (1) deviations from a multiple-orientation QSM reference; (2) visual appearance of QSM maps and the presence of artifacts; (3) susceptibility in subcortical brain regions with age; (4) regional brain susceptibility with published postmortem brain iron quantification; and (5) susceptibility in HD-affected basal ganglia regions between HD subjects and healthy controls. We found that single-step QSM methods with either total variation or total generalized variation constraints (SSTV/SSTGV) and the single-step deep learning method iQSM generally provided the best performance in terms of correlation with iron deposition and were better at differentiating between healthy controls and premanifest HD individuals, while deep learning QSM methods trained with multiple-orientation susceptibility data created QSM maps that were most similar to the multiple orientation reference and with the best visual scores.

Many studies in patients with brain tumors evaluating innovative PET tracers have been published in recent years, and the initial results are promising. Here, the Response Assessment in Neuro-Oncology (RANO) PET working group provides an overview of the literature on novel investigational PET tracers for brain tumor patients. Furthermore, newer indications of more established PET tracers for the evaluation of glucose metabolism, amino acid transport, hypoxia, cell proliferation, and others are also discussed. Based on the preliminary findings, these novel investigational PET tracers should be further evaluated considering their promising potential. In particular, novel PET probes for imaging of translocator protein and somatostatin receptor overexpression as well as for immune system reactions appear to be of additional clinical value for tumor delineation and therapy monitoring. Progress in developing these radiotracers may contribute to improving brain tumor diagnostics and advancing clinical translational research.

Background Glioblastoma is the most common primary brain malignancy, yet treatment options are limited, and prognosis remains guarded. Individualized tumor genetic assessment has become important for accurate prognosis and for guiding emerging targeted therapies. However, challenges remain for widespread tumor genetic testing due to costs and the need for tissue sampling. The aim of this study is to evaluate a novel artificial intelligence method for predicting clinically relevant genetic biomarkers from preoperative brain MRI in patients with glioblastoma. Methods We retrospectively analyzed preoperative MRI data from 400 patients with glioblastoma, IDH-wildtype or WHO grade 4 astrocytoma, IDH mutant who underwent resection and genetic testing. Nine genetic biomarkers were assessed: hotspot mutations of IDH1 or TERT promoter, pathogenic mutations of TP53, PTEN, ATRX, or CDKN2A/B, MGMT promoter methylation, EGFR amplification, and combined aneuploidy of chromosomes 7 and 10. Models were developed to predict biomarker status from MRI data using radiomics features, convolutional neural network (CNN) features, and a combination of both. Results Combined model performance was good for IDH1 and TERT promoter hotspot mutations, pathogenic mutations of ATRX and CDKN2A/B, and combined aneuploidy of chromosomes 7 and 10, with receiver operating characteristic area under the curve (ROC AUC) >0.85 and was fair for all other tested biomarkers with ROC AUC >0.7. Combined model performance was statistically superior to individual radiomics and CNN feature models for prediction chromosome 7 and 10 aneuploidy, MGMT promoter methylation, and PTEN mutation. Conclusions Combining radiomics and CNN features from preoperative MRI yields improved noninvasive genetic biomarker prediction performance in patients with WHO grade 4 diffuse astrocytic gliomas.

Real-time metabolic conversion of intravenously-injected hyperpolarized [1-Dynamic HP-Both denoising methods improved metabolite SNR and regional signal coverage. In patient data, the average increase in peak dynamic metabolite SNR was 2-fold using TRI and 4-5 folds using GL-HOSVD denoising compared to acquired data. Denoising reduced kPost-processing denoising methods significantly improved the SNR of dynamic HP

To investigate multi-resolution hyperpolarized (HP)HP [1-High resolution 7.5 × 7.5 mmAcquiring HP

Background Genomic profiling studies of diffuse gliomas have led to new improved classification schemes that better predict patient outcomes compared to conventional histomorphology alone. One example is the recognition that patients with IDH-wildtype diffuse astrocytic gliomas demonstrating lower-grade histologic features but genomic and/or epigenomic profile characteristic of glioblastoma typically have poor outcomes similar to patients with histologically diagnosed glioblastoma. Here we sought to determine the clinical impact of prospective genomic profiling for these IDH-wildtype diffuse astrocytic gliomas lacking high-grade histologic features but with molecular profile of glioblastoma. Methods Clinical management and outcomes were analyzed for 38 consecutive adult patients with IDH-wildtype diffuse astrocytic gliomas lacking necrosis or microvascular proliferation on histologic examination that were genomically profiled on a prospective clinical basis revealing criteria for an integrated diagnosis of “diffuse astrocytic glioma, IDH-wildtype, with molecular features of glioblastoma, WHO grade IV” per cIMPACT-NOW criteria. Results We identified that this diagnosis consists of two divergent clinical scenarios based on integration of radiologic, histologic, and genomic features that we term “early/evolving” and “undersampled” glioblastoma, IDH-wildtype. We found that prospective genomically guided identification of early/evolving and undersampled IDH-wildtype glioblastoma resulted in more aggressive patient management and improved clinical outcomes compared to a biologically matched historical control patient cohort receiving standard-of-care therapy based on histomorphologic diagnosis alone. Conclusions These results support routine use of genomic and/or epigenomic profiling to accurately classify glial neoplasms, as these assays not only improve diagnostic classification but critically lead to more appropriate patient management that can improve clinical outcomes.

Non-invasive imaging is a key clinical tool for detection and treatment monitoring of infections. Existing clinical imaging techniques are frequently unable to distinguish infection from tumors or sterile inflammation. This challenge is well-illustrated by prosthetic joint infections that often complicate joint replacements. D-methyl-614.5 ± 100.2 MBq of D-D-D

Purpose Brain herniation into arachnoid granulations (BHAG) of the dural venous sinuses is a recently described finding of uncertain etiology. The purpose of this study was to investigate the prevalence of BHAG in a cohort of patients with pulsatile tinnitus (PT) and to clarify the physiologic and clinical implications of these lesions. Methods The imaging and charts of consecutive PT patients were retrospectively reviewed. All patients were examined with MRI including pre- and post-contrast T1- and T2-weighted sequences. Images were reviewed separately by three blinded neuroradiologists to identify the presence of BHAG. Their location, signal intensity, size, presence of arachnoid granulation, and associated dural venous sinus stenosis were documented. Clinical records were further reviewed for idiopathic intracranial hypertension, history of prior lumbar puncture, and opening pressure. Results Two hundred sixty-two consecutive PT patients over a 4-year period met inclusion criteria. PT patients with BHAG were significantly more likely to have idiopathic intracranial hypertension than PT patients without BHAG (OR 4.2, CI 1.5–12, p = 0.006). Sixteen out of 262 (6%) patients were found to have 18 BHAG. Eleven out of 16 (69%) patients had unilateral temporal or occipital lobe herniations located in the transverse sinus or the transverse-sigmoid junction. Three out of 16 (19%) patients had unilateral cerebellar herniations and 2/16 (13%) patients had bilateral BHAG. Conclusion In patients with PT, BHAG is a prevalent MRI finding that is strongly associated with the clinical diagnosis of IIH. The pathogenesis of BHAG remains uncertain, but recognition should prompt comprehensive evaluation for IIH.

BackgroundReal-time metabolic conversion of intravenously-injected hyperpolarized [1-13C]pyruvate to [1-13C]lactate and [13C]bicarbonate in the brain can be measured using dynamic hyperpolarized carbon-13 (HP-13C) MRI. However, voxel-wise evaluation of metabolism in patients with glioma is challenged by the limited signal-to-noise ratio (SNR) of downstream 13C metabolites, especially within lesions. The purpose of this study was to evaluate the ability of higher-order singular value decomposition (HOSVD) denoising methods to enhance dynamic HP [1-13C]pyruvate MRI data acquired from patients with glioma.MethodsDynamic HP-13C MRI were acquired from 14 patients with glioma. The effects of two HOSVD denoising techniques, tensor rank truncation-image enhancement (TRI) and global-local HOSVD (GL-HOSVD), on the SNR and kinetic modeling were analyzed in [1-13C]lactate data with simulated noise that matched the levels of [13C]bicarbonate signals. Both methods were then evaluated in patient data based on their ability to improve [1-13C]pyruvate, [1-13C]lactate and [13C]bicarbonate SNR. The effects of denoising on voxel-wise kinetic modeling of kPL and kPB was also evaluated. The number of voxels with reliable kinetic modeling of pyruvate-to-lactate (kPL) and pyruvate-to-bicarbonate (kPB) conversion rates within regions of interest (ROIs) before and after denoising was then compared.ResultsBoth denoising methods improved metabolite SNR and regional signal coverage. In patient data, the average increase in peak dynamic metabolite SNR was 2-fold using TRI and 4-5 folds using GL-HOSVD denoising compared to acquired data. Denoising reduced kPL modeling errors from a native average of 23% to 16% (TRI) and 15% (GL-HOSVD); and kPB error from 42% to 34% (TRI) and 37% (GL-HOSVD) (values were averaged voxelwise over all datasets). In contrast-enhancing lesions, the average number of voxels demonstrating within-tolerance kPL modeling error relative to the total voxels increased from 48% in the original data to 84% (TRI) and 90% (GL-HOSVD), while the number of voxels showing within-tolerance kPB modeling error increased from 0% to 15% (TRI) and 8% (GL-HOSVD).ConclusionPost-processing denoising methods significantly improved the SNR of dynamic HP-13C imaging data, resulting in a greater number of voxels satisfying minimum SNR criteria and maximum kinetic modeling errors in tumor lesions. This enhancement can aid in the voxel-wise analysis of HP-13C data and thereby improve monitoring of metabolic changes in patients with glioma following treatment.

Background Monitoring lower-grade gliomas (LrGGs) for disease progression is made difficult by the limits of anatomical MRI to distinguish treatment related tissue changes from tumor progression. MR spectroscopic imaging (MRSI) offers additional metabolic information that can help address these challenges. The goal of this study was to compare longitudinal changes in multiparametric MRI, including diffusion weighted imaging, perfusion imaging, and 3D MRSI, for LrGG patients who progressed at the final time-point and those who remained clinically stable. Methods Forty-one patients with LrGG who were clinically stable were longitudinally assessed for progression. Changes in anatomical, diffusion, perfusion and MRSI data were acquired and compared between patients who remained clinically stable and those who progressed. Results Thirty-one patients remained stable, and 10 patients progressed. Over the study period, progressed patients had a significantly greater increase in normalized choline, choline-to-N-acetylaspartic acid index (CNI), normalized creatine, and creatine-to-N-acetylaspartic acid index (CRNI), than stable patients. CRNI was significantly associated with progression status and WHO type. Progressed astrocytoma patients had greater increases in CRNI than stable astrocytoma patients. Conclusions LrGG patients in surveillance with tumors that progressed had significantly increasing choline and creatine metabolite signals on MRSI, with a trend of increasing T2 FLAIR volumes, compared to LrGG patients who remained stable. These data show that MRSI can be used in conjunction with anatomical imaging studies to gain a clearer picture of LrGG progression, especially in the setting of clinical ambiguity.

SUMMARYMeningiomas arising from the meningothelial central nervous system lining are the most common primary intracranial tumors, and a significant cause of neurologic morbidity and mortality1. There are no effective medical therapies for meningioma patients2,3, and new treatments have been encumbered by limited understanding of meningioma biology. DNA methylation profiling provides robust classification of central nervous system tumors4, and can elucidate targets for molecular therapy5. Here we use DNA methylation profiling on 565 meningiomas integrated with genetic, transcriptomic, biochemical, and single-cell approaches to show meningiomas are comprised of 3 epigenetic groups with distinct clinical outcomes and biological features informing new treatments for meningioma patients. Merlin-intact meningiomas (group A, 34%) have the best outcomes and are distinguished by a novel apoptotic tumor suppressor function ofNF2/Merlin. Immune-enriched meningiomas (group B, 38%) have intermediate outcomes and are distinguished by immune cell infiltration,HLAexpression, and lymphatic vessels. Hypermitotic meningiomas (group C, 28%) have the worst outcomes and are distinguished by convergent genetic mechanisms misactivating the cell cycle. Consistently, we find cell cycle inhibitors block meningioma growth in cell culture, organoids, xenografts, and patients. Our results establish a framework for understanding meningioma biology, and provide preclinical rationale for new therapies to treat meningioma patients.

Meningiomas are the most common primary tumors of the central nervous system and as such they are often encountered at neuroimaging. Fortunately, meningiomas are readily diagnosed with anatomic computed tomography and magnetic resonance imaging. While conventional imaging is the mainstay for initial diagnosis and delineating tumor for treatment planning and posttreatment follow-up, the last couple of decades have given rise to advanced physiologic and metabolic imaging techniques that serve as powerful tools in the management of meningioma. These modern approaches are allowing imaging to expand its utility to include extraction of biologic and potentially prognostic information that will ultimately improve care for meningioma patients.

Despite widespread adoption of electronic health records (EHRs), most hospitals are not ready to implement data science research in the clinical pipelines. Here, we develop MEDomics, a continuously learning infrastructure through which multimodal health data are systematically organized and data quality is assessed with the goal of applying artificial intelligence for individual prognosis. Using this framework, currently composed of thousands of individuals with cancer and millions of data points over a decade of data recording, we demonstrate prognostic utility of this framework in oncology. As proof of concept, we report an analysis using this infrastructure, which identified the Framingham risk score to be robustly associated with mortality among individuals with early-stage and advanced-stage cancer, a potentially actionable finding from a real-world cohort of individuals with cancer. Finally, we show how natural language processing (NLP) of medical notes could be used to continuously update estimates of prognosis as a given individual's disease course unfolds.

PURPOSE: To ameliorate tradeoffs between a fixed spatial resolution and signal-to-noise ratio (SNR) for hyperpolarized (13)C MRI. METHODS: In MRI, SNR is proportional to voxel volume but retrospective downsampling or voxel averaging only improves SNR by the square root of voxel size. This can be exploited with a metabolite-selective imaging approach that independently encodes each compound, yielding high-resolution images for the injected substrate and coarser resolution images for downstream metabolites, while maintaining adequate SNR for each. To assess the efficacy of this approach, hyperpolarized [1-(13)C]pyruvate data were acquired in healthy Sprague-Dawley rats (n = 4) and in two healthy human subjects. RESULTS: Compared with a constant resolution acquisition, variable-resolution data sets showed improved detectability of metabolites in pre-clinical renal studies with a 3.5-fold, 8.7-fold, and 6.0-fold increase in SNR for lactate, alanine, and bicarbonate data, respectively. Variable-resolution data sets from healthy human subjects showed cardiac structure and neuro-vasculature in the higher resolution pyruvate images (6.0 × 6.0 mm(2) for cardiac and 7.5 × 7.5 mm(2) for brain) that would otherwise be missed due to partial-volume effects and illustrates the level of detail that can be achieved with hyperpolarized substrates in a clinical setting. CONCLUSION: We developed a variable-resolution strategy for hyperpolarized (13)C MRI using metabolite-selective imaging and demonstrated that it mitigates tradeoffs between a fixed spatial resolution and SNR for hyperpolarized substrates, providing both high resolution pyruvate and coarse resolution metabolite data sets in a single exam. This technique shows promise to improve future studies by maximizing metabolite SNR while minimizing partial-volume effects from the injected substrate.