Synthesizing 3D Multi-Contrast Brain Tumor MRIs Using Tumor Mask Conditioning

Data scarcity and data imbalance are two major challenges in training deep learning models on medical images, such as brain tumor MRI data. The recent advancements in generative artificial intelligence have opened new possibilities for synthetically generating MRI data, including brain tumor MRI scans. This approach can be a potential solution to mitigate the data scarcity problem and enhance training data availability. This work focused on adapting the 2D latent diffusion models to generate 3D multi-contrast brain tumor MRI data with a tumor mask as the condition. The framework comprises two components: a 3D autoencoder model for perceptual compression and a conditional 3D Diffusion Probabilistic Model (DPM) for generating high-quality and diverse multi-contrast brain tumor MRI samples, guided by a conditional tumor mask. Unlike existing works that focused on generating either 2D multi-contrast or 3D single-contrast MRI samples, our models generate multi-contrast 3D MRI samples. We also integrated a conditional module within the UNet backbone of the DPM to capture the semantic class-dependent data distribution driven by the provided tumor mask to generate MRI brain tumor samples based on a specific brain tumor mask. We trained our models using two brain tumor datasets: The Cancer Genome Atlas (TCGA) public dataset and an internal dataset from the University of Texas Southwestern Medical Center (UTSW). The models were able to generate high-quality 3D multi-contrast brain tumor MRI samples with the tumor location aligned by the input condition mask. The quality of the generated images was evaluated using the Fréchet Inception Distance (FID) score. This work has the potential to mitigate the scarcity of brain tumor data and improve the performance of deep learning models involving brain tumor MRI data.

MRI-Based Deep Learning Method for Classification of IDH Mutation Status

Isocitrate dehydrogenase (IDH) mutation status has emerged as an important prognostic marker in gliomas. This study sought to develop deep learning networks for non-invasive IDH classification using T2w MR images while comparing their performance to a multi-contrast network. Methods: Multi-contrast brain tumor MRI and genomic data were obtained from The Cancer Imaging Archive (TCIA) and The Erasmus Glioma Database (EGD). Two separate 2D networks were developed using nnU-Net, a T2w-image-only network (T2-net) and a multi-contrast network (MC-net). Each network was separately trained using TCIA (227 subjects) or TCIA + EGD data (683 subjects combined). The networks were trained to classify IDH mutation status and implement single-label tumor segmentation simultaneously. The trained networks were tested on over 1100 held-out datasets including 360 cases from UT Southwestern Medical Center, 136 cases from New York University, 175 cases from the University of Wisconsin-Madison, 456 cases from EGD (for the TCIA-trained network), and 495 cases from the University of California, San Francisco public database. A receiver operating characteristic curve (ROC) was drawn to calculate the AUC value to determine classifier performance. Results: T2-net trained on TCIA and TCIA + EGD datasets achieved an overall accuracy of 85.4% and 87.6% with AUCs of 0.86 and 0.89, respectively. MC-net trained on TCIA and TCIA + EGD datasets achieved an overall accuracy of 91.0% and 92.8% with AUCs of 0.94 and 0.96, respectively. We developed reliable, high-performing deep learning algorithms for IDH classification using both a T2-image-only and a multi-contrast approach. The networks were tested on more than 1100 subjects from diverse databases, making this the largest study on image-based IDH classification to date.

Deep-learning-enabled brain hemodynamic mapping using resting-state fMRI

Cerebrovascular disease is a leading cause of death globally. Prevention and early intervention are known to be the most effective forms of its management. Non-invasive imaging methods hold great promises for early stratification, but at present lack the sensitivity for personalized prognosis. Resting-state functional magnetic resonance imaging (rs-fMRI), a powerful tool previously used for mapping neural activity, is available in most hospitals. Here we show that rs-fMRI can be used to map cerebral hemodynamic function and delineate impairment. By exploiting time variations in breathing pattern during rs-fMRI, deep learning enables reproducible mapping of cerebrovascular reactivity (CVR) and bolus arrival time (BAT) of the human brain using resting-state CO2 fluctuations as a natural "contrast media". The deep-learning network is trained with CVR and BAT maps obtained with a reference method of CO2-inhalation MRI, which includes data from young and older healthy subjects and patients with Moyamoya disease and brain tumors. We demonstrate the performance of deep-learning cerebrovascular mapping in the detection of vascular abnormalities, evaluation of revascularization effects, and vascular alterations in normal aging. In addition, cerebrovascular maps obtained with the proposed method exhibit excellent reproducibility in both healthy volunteers and stroke patients. Deep-learning resting-state vascular imaging has the potential to become a useful tool in clinical cerebrovascular imaging.

Corrigendum to: A novel fully automated MRI-based deep-learning method for classification of 1p/19q co-deletion status in brain gliomas

[This corrects the article DOI: 10.1093/noajnl/vdaa066.].

Errata: Brain tumor IDH, 1p/19q, and MGMT molecular classification using MRI-based deep learning: an initial study on the effect of motion and motion correction

[This corrects the article DOI: 10.1117/1.JMI.9.1.016001.].

Federated learning enables big data for rare cancer boundary detection

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

Association of distal hyperintense vessel sign and recurrent stroke in patients with symptomatic intracranial stenosis

OBJECTIVE

The distal hyperintense vessel sign (DHV) on fluid-attenuated inversion recovery magnetic resonance image (MRI) is an imaging biomarker of slow leptomeningeal collateral flow in the presence of large artery stenosis or occlusion reflecting impaired cerebral hemodynamics. In this study, we aim to investigate the significance of the DHV sign in patients with symptomatic ≥ 70% intracranial atherosclerotic stenosis.

METHODS

We retrospectively reviewed patients with ischemic stroke or transient ischemic attack admitted to a single center from January 2010 to December 2017. Patients were included if they had symptomatic ≥ 70% atherosclerotic stenosis of the intracranial internal carotid artery or middle cerebral artery. The presence of the DHV sign was evaluated by blinded neuroradiologist and vascular neurologists. Recurrent ischemic stroke in the vascular territory of symptomatic intracranial artery was defined as new neurological deficits with associated neuroimaging findings during the follow up period.

RESULTS

A total of 109 patients were included in the study, of which 55 had DHV sign. Average duration of follow up was 297 ± 326 days. Four patients were lost during follow up. Patients with the DHV sign had a higher rate of recurrent ischemic stroke (38%), compared to patients without the DHV sign (17%; p=0.018). In multivariate regression analysis, the presence of DHV sign was an independent predictor of recurrent ischemic stroke. A DHV score of ≥ 2 had a 63% sensitivity and 69% specificity for recurrent ischemic stroke.

INTERPRETATION

In patients with severe symptomatic intracranial atherosclerotic stenosis, those with a DHV sign on MRI are at higher risk of recurrent ischemic stroke.

Adverse Radiation Therapy Effects in the Treatment of Head and Neck Tumors

Whether used as a single modality or as part of a combined approach, radiation therapy (RT) plays an essential role in the treatment of several head and neck malignancies. Despite the improvement in radiation delivery techniques, normal structures in the vicinity of the target area remain susceptible to a wide range of adverse effects. Given their high incidence, some of these effects are referred to as expected postradiation changes (eg, mucositis, sialadenitis, and edema), while others are considered true complications, meaning they should not be expected and can even represent life-threatening conditions (eg, radionecrosis, fistulas, and radiation-induced neoplasms). Also, according to their timing of onset, these deleterious effects can be divided into four groups: acute (during RT), subacute (within weeks to months), delayed onset (within months to years), and very delayed onset (after several years).The authors provide a comprehensive review of the most important radiation-induced changes related to distinct head and neck sites, focusing on their typical cross-sectional imaging features and correlating them with the time elapsed after treatment. Radiologists should not only be familiar with these imaging findings but also actively seek essential clinical data at the time of interpretation (including knowledge of the RT dose and time, target site, and manifesting symptoms) to better recognize imaging findings, avoid pitfalls and help guide appropriate management. ^© RSNA, 2022.

Dynamic 13 C MR spectroscopy as an alternative to imaging for assessing cerebral metabolism using hyperpolarized pyruvate in humans

PURPOSE

This study is to investigate time-resolved ¹³ C MR spectroscopy (MRS) as an alternative to imaging for assessing pyruvate metabolism using hyperpolarized (HP) [1-¹³ C]pyruvate in the human brain.

METHODS

Time-resolved ¹³ C spectra were acquired from four axial brain slices of healthy human participants (n = 4) after a bolus injection of HP [1-¹³ C]pyruvate. ¹³ C MRS with low flip-angle excitations and a multichannel ¹³ C/¹ H dual-frequency radiofrequency (RF) coil were exploited for reliable and unperturbed assessment of HP pyruvate metabolism. Slice-wise areas under the curve (AUCs) of ¹³ C-metabolites were measured and kinetic analysis was performed to estimate the production rates of lactate and HCO3- . Linear regression analysis between brain volumes and HP signals was performed. Region-focused pyruvate metabolism was estimated using coil-wise ¹³ C reconstruction. Reproducibility of HP pyruvate exams was presented by performing two consecutive injections with a 45-minutes interval.

RESULTS

[1-¹³ C]Lactate relative to the total ¹³ C signal (tC) was 0.21-0.24 in all slices. [¹³ C] HCO3- /tC was 0.065-0.091. Apparent conversion rate constants from pyruvate to lactate and HCO3- were calculated as 0.014-0.018 s^-1 and 0.0043-0.0056 s^-1 , respectively. Pyruvate/tC and lactate/tC were in moderate linear relationships with fractional gray matter volume within each slice. White matter presented poor linear regression fit with HP signals, and moderate correlations of the fractional cerebrospinal fluid volume with pyruvate/tC and lactate/tC were measured. Measured HP signals were comparable between two consecutive exams with HP [1-¹³ C]pyruvate.

CONCLUSIONS

Dynamic MRS in combination with multichannel RF coils is an affordable and reliable alternative to imaging methods in investigating cerebral metabolism using HP [1-¹³ C]pyruvate.

Non-contrast hemodynamic imaging of Moyamoya disease with MR fingerprinting ASL: A feasibility study

PURPOSE

MR Fingerprinting (MRF) Arterial Spin Labeling (ASL) is a non-contrast technique to estimate multiple brain hemodynamic and structural parameters in a single scan. The purpose of this study is to examine the feasibility and initial utility of MRF-ASL in Moyamoya disease.

METHODS

MRF-ASL, conventional single-delay ASL, Time-of-flight (TOF) MR angiography, and contrast-based dynamic-susceptibility-contrast (DSC) MRI were prospectively collected from a group of Moyamoya patients in North America (N = 21, 4 men and 17 women). Sixteen healthy subjects (7 men and 9 women) also underwent an MRF-ASL scan. Cerebral blood flow (CBF), bolus arrival time (BAT), and tissue T1 were compared between Moyamoya patients and healthy controls. Perfusion parameters from MRF-ASL were compared to those from other MRI sequences. Multi-linear regression was used for comparisons of parameter values between Moyamoya and control groups. Linear mixed-effects models was used when comparing MRF-ASL to PCASL and DSC parameters. Spearman's Rank Correlation Coefficient was calculated when comparing MRF-ASL to and MRA grades. A P value of 0.05 or less was considered significant.

RESULTS

BAT in stenotic internal carotid artery (ICA) territories was prolonged (P < 0.001) in Moyamoya patients, when compared with healthy controls. CBF in stenotic ICA territories of Moyamoya patients was not different from CBF in healthy controls; but in the PCA territories, CBF in Moyamoya patients was higher (P < 0.01) than controls. Quantitative T1 values in the stenotic ICA territories was longer (P < 0.05) than that in controls. Hemodynamic parameters estimated from MRF-ASL were significantly correlated with single-delay ASL and DSC. Longer BAT was associated with more severe intracranial artery stenosis in ICA.

CONCLUSIONS

MRF-ASL is a promising technique to assess perfusion and structural abnormalities in Moyamoya patients.

Brain tumor IDH, 1p/19q, and MGMT molecular classification using MRI-based deep learning: an initial study on the effect of motion and motion correction

Purpose: Deep learning has shown promise for predicting the molecular profiles of gliomas using MR images. Prior to clinical implementation, ensuring robustness to real-world problems, such as patient motion, is crucial. The purpose of this study is to perform a preliminary evaluation on the effects of simulated motion artifact on glioma marker classifier performance and determine if motion correction can restore classification accuracies. Approach: T2w images and molecular information were retrieved from the TCIA and TCGA databases. Simulated motion was added in the k-space domain along the phase encoding direction. Classifier performance for IDH mutation, 1p/19q co-deletion, and MGMT methylation was assessed over the range of 0% to 100% corrupted k-space lines. Rudimentary motion correction networks were trained on the motion-corrupted images. The performance of the three glioma marker classifiers was then evaluated on the motion-corrected images. Results: Glioma marker classifier performance decreased markedly with increasing motion corruption. Applying motion correction effectively restored classification accuracy for even the most motion-corrupted images. For isocitrate dehydrogenase (IDH) classification, 99% accuracy was achieved, exceeding the original performance of the network and representing a new benchmark in non-invasive MRI-based IDH classification. Conclusions: Robust motion correction can facilitate highly accurate deep learning MRI-based molecular marker classification, rivaling invasive tissue-based characterization methods. Motion correction may be able to increase classification accuracy even in the absence of a visible artifact, representing a new strategy for boosting classifier performance.

Imaging characteristics of 4th ventricle subependymoma

PURPOSE

Subependymomas located within the 4th ventricle are rare, and the literature describing imaging characteristics is sparse. Here, we describe the clinical and radiological characteristics of 29 patients with 4th ventricle subependymoma.

METHODS

This is a retrospective multi-center study performed after Institutional Review Board (IRB) approval. Patients diagnosed with suspected 4th ventricle subependymoma were identified. A review of clinical, radiology, and pathology reports along with magnetic resonance imaging (MRI) images was performed.

RESULTS

Twenty-nine patients, including 6 females, were identified. Eighteen patients underwent surgery with histopathological confirmation of subependymoma. The median age at diagnosis was 52 years. Median tumor volume for the operative cohort was 9.87 cm³, while for the non-operative cohort, it was 0.96 cm³. Thirteen patients in the operative group exhibited symptoms at diagnosis. For the total cohort, the majority of subependymomas (n = 22) were isointense on T1, hyperintense (n = 22) on T2, and enhanced (n = 24). All tumors were located just below the body of the 4th ventricle, terminating near the level of the obex. Fourteen cases demonstrated extension of tumor into foramen of Magendie or Luschka.

CONCLUSION

To the best of our knowledge, this is the largest collection of 4th ventricular subependymomas with imaging findings reported to date. All patients in this cohort had tumors originating between the bottom of the body of the 4th ventricle and the obex. This uniform and specific site of origin aids with imaging diagnosis and may infer possible theories of origin.

Alterations in the RB Pathway With Inactivation of RB1 Characterize Glioblastomas With a Primitive Neuronal Component

A primitive neuronal component is a feature of some glioblastomas but defining molecular alterations of this histologic variant remains uncertain. We performed next-generation sequencing of 1500 tumor related genes on tissue from 9 patients with glioblastoma with a primitive component (G/PN) and analyzed 27 similar cases from the Cancer Genome Atlas (TCGA) dataset. Alterations in the RB pathway were identified in all of our patients' tumors and 81% of TCGA tumors with the retinoblastoma tumor suppressor gene (RB1) commonly affected. Although RB1 mutations were observed in some conventional glioblastomas, the allelic fractions of these mutations were significantly higher in tumors with a primitive neuronal component in both our and TCGA cohorts (median, 72% vs 25%, p < 0.001 and 80% vs 40%, p < 0.02, respectively). Further, in 78% of patients in our cohort, RB expression was lost by immunohistochemistry. Our findings indicate that alterations in the RB pathway are common in G/PNs and suggest that inactivation of RB1 may be a driving mechanism for the phenotype.

Preoperative imaging of glioblastoma patients using hyperpolarized 13C pyruvate: Potential role in clinical decision making

Background

Glioblastoma remains incurable despite treatment with surgery, radiation therapy, and cytotoxic chemotherapy, prompting the search for a metabolic pathway unique to glioblastoma cells.¹³C MR spectroscopic imaging with hyperpolarized pyruvate can demonstrate alterations in pyruvate metabolism in these tumors.

Methods

Three patients with diagnostic MRI suggestive of a glioblastoma were scanned at 3 T 1–2 days prior to tumor resection using a ¹³C/¹H dual-frequency RF coil and a ¹³C/¹H-integrated MR protocol, which consists of a series of ¹H MR sequences (T₂ FLAIR, arterial spin labeling and contrast-enhanced [CE] T₁) and ¹³C spectroscopic imaging with hyperpolarized [1-¹³C]pyruvate. Dynamic spiral chemical shift imaging was used for ¹³C data acquisition. Surgical navigation was used to correlate the locations of tissue samples submitted for histology with the changes seen on the diagnostic MR scans and the ¹³C spectroscopic images.

Results

Each tumor was histologically confirmed to be a WHO grade IV glioblastoma with isocitrate dehydrogenase wild type. Total hyperpolarized ¹³C signals detected near the tumor mass reflected altered tissue perfusion near the tumor. For each tumor, a hyperintense [1-¹³C]lactate signal was detected both within CE and T₂-FLAIR regions on the ¹H diagnostic images (P = .008). [¹³C]bicarbonate signal was maintained or decreased in the lesion but the observation was not significant (P = .3).

Conclusions

Prior to surgical resection, ¹³C MR spectroscopic imaging with hyperpolarized pyruvate reveals increased lactate production in regions of histologically confirmed glioblastoma.

Intrasession Reliability of Arterial Spin-Labeled MRI-Measured Noncontrast Perfusion in Glioblastoma at 3 T

Arterial spin-labeled magnetic resonance imaging can provide quantitative perfusion measurements in the brain and can be potentially used to evaluate therapy response assessment in glioblastoma (GBM). The reliability and reproducibility of this method to measure noncontrast perfusion in GBM, however, are lacking. We evaluated the intrasession reliability of brain and tumor perfusion in both healthy volunteers and patients with GBM at 3 T using pseudocontinuous labeling (pCASL) and 3D turbo spin echo (TSE) using Cartesian acquisition with spiral profile reordering (CASPR). Two healthy volunteers at a single time point and 6 newly diagnosed patients with GBM at multiple time points (before, during, and after chemoradiation) underwent scanning (total, 14 sessions). Compared with 3D GraSE, 3D TSE-CASPR generated cerebral blood flow maps with better tumor-to-normal background tissue contrast and reduced image distortions. The intraclass correlation coefficient between the 2 runs of 3D pCASL with TSE-CASPR was consistently high (≥0.90) across all normal-appearing gray matter (NAGM) regions of interest (ROIs), and was particularly high in tumors (0.98 with 95% confidence interval [CI]: 0.97-0.99). The within-subject coefficients of variation were relatively low in all normal-appearing gray matter regions of interest (3.40%-7.12%), and in tumors (4.91%). Noncontrast perfusion measured using 3D pCASL with TSE-CASPR provided robust cerebral blood flow maps in both healthy volunteers and patients with GBM with high intrasession repeatability at 3 T. This approach can be an appropriate noncontrast and noninvasive quantitative perfusion imaging method for longitudinal assessment of therapy response and management of patients with GBM.

A Fully Automated Deep Learning Network for Brain Tumor Segmentation

We developed a fully automated method for brain tumor segmentation using deep learning; 285 brain tumor cases with multiparametric magnetic resonance images from the BraTS2018 data set were used. We designed 3 separate 3D-Dense-UNets to simplify the complex multiclass segmentation problem into individual binary-segmentation problems for each subcomponent. We implemented a 3-fold cross-validation to generalize the network's performance. The mean cross-validation Dice-scores for whole tumor (WT), tumor core (TC), and enhancing tumor (ET) segmentations were 0.92, 0.84, and 0.80, respectively. We then retrained the individual binary-segmentation networks using 265 of the 285 cases, with 20 cases held-out for testing. We also tested the network on 46 cases from the BraTS2017 validation data set, 66 cases from the BraTS2018 validation data set, and 52 cases from an independent clinical data set. The average Dice-scores for WT, TC, and ET were 0.90, 0.84, and 0.80, respectively, on the 20 held-out testing cases. The average Dice-scores for WT, TC, and ET on the BraTS2017 validation data set, the BraTS2018 validation data set, and the clinical data set were as follows: 0.90, 0.80, and 0.78; 0.90, 0.82, and 0.80; and 0.85, 0.80, and 0.77, respectively. A fully automated deep learning method was developed to segment brain tumors into their subcomponents, which achieved high prediction accuracy on the BraTS data set and on the independent clinical data set. This method is promising for implementation into a clinical workflow.

MRI-Based Deep-Learning Method for Determining Glioma MGMT Promoter Methylation Status

BACKGROUND AND PURPOSE

O6-Methylguanine-DNA methyltransferase (MGMT) promoter methylation confers an improved prognosis and treatment response in gliomas. We developed a deep learning network for determining MGMT promoter methylation status using T2 weighted Images (T2WI) only.

MATERIALS AND METHODS

Brain MR imaging and corresponding genomic information were obtained for 247 subjects from The Cancer Imaging Archive and The Cancer Genome Atlas. One hundred sixty-three subjects had a methylated MGMT promoter. A T2WI-only network (MGMT-net) was developed to determine MGMT promoter methylation status and simultaneous single-label tumor segmentation. The network was trained using 3D-dense-UNets. Three-fold cross-validation was performed to generalize the performance of the networks. Dice scores were computed to determine tumor-segmentation accuracy.

RESULTS

The MGMT-net demonstrated a mean cross-validation accuracy of 94.73% across the 3 folds (95.12%, 93.98%, and 95.12%, [SD, 0.66%]) in predicting MGMT methylation status with a sensitivity and specificity of 96.31% [SD, 0.04%] and 91.66% [SD, 2.06%], respectively, and a mean area under the curve of 0.93 [SD, 0.01]. The whole tumor-segmentation mean Dice score was 0.82 [SD, 0.008].

CONCLUSIONS

We demonstrate high classification accuracy in predicting MGMT promoter methylation status using only T2WI. Our network surpasses the sensitivity, specificity, and accuracy of histologic and molecular methods. This result represents an important milestone toward using MR imaging to predict prognosis and treatment response.

Glycine by MR spectroscopy is an imaging biomarker of glioma aggressiveness

BACKGROUND

High-grade gliomas likely remodel the metabolic machinery to meet the increased demands for amino acids and nucleotides during rapid cell proliferation. Glycine, a non-essential amino acid and intermediate of nucleotide biosynthesis, may increase with proliferation. Non-invasive measurement of glycine by magnetic resonance spectroscopy (MRS) was evaluated as an imaging biomarker for assessment of tumor aggressiveness.

METHODS

We measured glycine, 2-hydroxyglutarate (2HG), and other tumor-related metabolites in 35 glioma patients using an MRS sequence tailored for co-detection of glycine and 2HG in gadolinium-enhancing and non-enhancing tumor regions on 3T MRI. Glycine and 2HG concentrations as measured by MRS were correlated with tumor cell proliferation (MIB-1 labeling index), expression of mitochondrial serine hydroxymethyltransferase (SHMT2), and glycine decarboxylase (GLDC) enzymes, and patient overall survival.

RESULTS

Elevated glycine was strongly associated with presence of gadolinium enhancement, indicating more rapidly proliferative disease. Glycine concentration was positively correlated with MIB-1, and levels higher than 2.5 mM showed significant association with shorter patient survival, irrespective of isocitrate dehydrogenase status. Concentration of 2HG did not correlate with MIB-1 index. A high glycine/2HG concentration ratio, >2.5, was strongly associated with shorter survival (P < 0.0001). GLDC and SHMT2 expression were detectable in all tumors with glycine concentration, demonstrating an inverse correlation with GLDC.

CONCLUSIONS

The data suggest that aggressive gliomas reprogram glycine-mediated one-carbon metabolism to meet the biosynthetic demands for rapid cell proliferation. MRS evaluation of glycine provides a non-invasive metabolic imaging biomarker that is predictive of tumor progression and clinical outcome.

KEY POINTS

1. Glycine and 2-hydroxyglutarate in glioma patients are precisely co-detected using MRS at 3T.2. Tumors with elevated glycine proliferate and progress rapidly.3. A high glycine/2HG ratio is predictive of shortened patient survival.

Cerebrovascular Reactivity Mapping Using Resting-State BOLD Functional MRI in Healthy Adults and Patients with Moyamoya Disease

Background Cerebrovascular reserve, the potential capacity of brain tissue to receive more blood flow when needed, is a desirable marker in evaluating ischemic risk. However, current measurement methods require acetazolamide injection or hypercapnia challenge, prompting a clinical need for resting-state (RS) blood oxygen level-dependent (BOLD) functional MRI data to measure cerebrovascular reactivity (CVR). Purpose To optimize and evaluate an RS CVR MRI technique and demonstrate its relationship to neurosurgical treatment. Materials and Methods In this HIPAA-compliant study, RS BOLD functional MRI data collected in 170 healthy controls between December 2008 and September 2010 were retrospectively evaluated to identify the optimal frequency range of temporal filtering on the basis of spatial correlation with the reference standard CVR map obtained with CO₂ inhalation. Next, the optimized RS method was applied in a new, prospective cohort of 50 participants with Moyamoya disease who underwent imaging between June 2014 and August 2019. Finally, CVR values were compared between brain hemispheres with and brain hemispheres without revascularization surgery by using Mann-Whitney U test. Results A total of 170 healthy controls (mean age ± standard deviation, 51 years ± 20; 105 women) and 100 brain hemispheres of 50 participants with Moyamoya disease (mean age, 41 years ± 12; 43 women) were evaluated. RS CVR maps based on a temporal filtering frequency of [0, 0.1164 Hz] yielded the highest spatial correlation (r = 0.74) with the CO₂ inhalation CVR results. In patients with Moyamoya disease, 77 middle cerebral arteries (MCAs) had stenosis. RS CVR in the MCA territory was lower in the group that did not undergo surgery (n = 30) than in the group that underwent surgery (n = 47) (mean, 0.407 relative units [ru] ± 0.208 vs 0.532 ru ± 0.182, respectively; P = .006), which is corroborated with the CO₂ inhalation CVR data (mean, 0.242 ru ± 0.273 vs 0.437 ru ± 0.200; P = .003). Conclusion Cerebrovascular reactivity mapping performed by using resting-state blood oxygen level-dependent functional MRI provided a task-free method to measure cerebrovascular reserve and depicted treatment effect of revascularization surgery in patients with Moyamoya disease comparable to that with the reference standard of CO₂ inhalation MRI. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Wolf and Ware in this issue.

Predicting Parkinson's disease trajectory using clinical and neuroimaging baseline measures

INTRODUCTION

Predictive biomarkers of Parkinson's Disease progression are needed to expedite neuroprotective treatment development and facilitate prognoses for patients. This work uses measures derived from resting-state functional magnetic resonance imaging, including regional homogeneity (ReHo) and fractional amplitude of low frequency fluctuations (fALFF), to predict an individual's current and future severity over up to 4 years and to elucidate the most prognostic brain regions.

METHODS

ReHo and fALFF are measured for 82 Parkinson's Disease subjects and used to train machine learning predictors of baseline clinical and future severity at 1 year, 2 years, and 4 years follow-up as measured by the Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS). Predictive performance is measured with nested cross-validation, validated on an external dataset, and again validated through leave-one-site-out cross-validation. Important predictive features are identified.

RESULTS

The models explain up to 30.4% of the variance in current MDS-UPDRS scores, 55.8% of the variance in year 1 scores, and 47.1% of the variance in year 2 scores (p < 0.0001). For distinguishing high and low-severity individuals at each timepoint (MDS-UPDRS score above or below the median, respectively), the models achieve positive predictive values up to 79% and negative predictive values up to 80%. Higher ReHo and fALFF in several regions, including components of the default motor network, predicted lower severity across current and future timepoints.

CONCLUSION

These results identify an accurate prognostic neuroimaging biomarker which may be used to better inform enrollment in trials of neuroprotective treatments and enable physicians to counsel their patients.

A novel fully automated MRI-based deep-learning method for classification of 1p/19q co-deletion status in brain gliomas

BACKGROUND

One of the most important recent discoveries in brain glioma biology has been the identification of the isocitrate dehydrogenase (IDH) mutation and 1p/19q co-deletion status as markers for therapy and prognosis. 1p/19q co-deletion is the defining genomic marker for oligodendrogliomas and confers a better prognosis and treatment response than gliomas without it. Our group has previously developed a highly accurate deep-learning network for determining IDH mutation status using T2-weighted (T2w) MRI only. The purpose of this study was to develop a similar 1p/19q deep-learning classification network.

METHODS

Multiparametric brain MRI and corresponding genomic information were obtained for 368 subjects from The Cancer Imaging Archive and The Cancer Genome Atlas. 1p/19 co-deletions were present in 130 subjects. Two-hundred and thirty-eight subjects were non-co-deleted. A T2w image-only network (1p/19q-net) was developed to perform 1p/19q co-deletion status classification and simultaneous single-label tumor segmentation using 3D-Dense-UNets. Three-fold cross-validation was performed to generalize the network performance. Receiver operating characteristic analysis was also performed. Dice scores were computed to determine tumor segmentation accuracy.

RESULTS

1p/19q-net demonstrated a mean cross-validation accuracy of 93.46% across the 3 folds (93.4%, 94.35%, and 92.62%, SD = 0.8) in predicting 1p/19q co-deletion status with a sensitivity and specificity of 0.90 ± 0.003 and 0.95 ± 0.01, respectively and a mean area under the curve of 0.95 ± 0.01. The whole tumor segmentation mean Dice score was 0.80 ± 0.007.

CONCLUSION

We demonstrate high 1p/19q co-deletion classification accuracy using only T2w MR images. This represents an important milestone toward using MRI to predict glioma histology, prognosis, and response to treatment.

Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of Glioblastoma

PURPOSE

To determine the influence of preprocessing on the repeatability and redundancy of radiomics features extracted using a popular open-source radiomics software package in a scan-rescan glioblastoma MRI study.

MATERIALS AND METHODS

In this study, a secondary analysis of T2-weighted fluid-attenuated inversion recovery (FLAIR) and T1-weighted postcontrast images from 48 patients (mean age, 56 years [range, 22-77 years]) diagnosed with glioblastoma were included from two prospective studies (ClinicalTrials.gov NCT00662506 [2009-2011] and NCT00756106 [2008-2011]). All patients underwent two baseline scans 2-6 days apart using identical imaging protocols on 3-T MRI systems. No treatment occurred between scan and rescan, and tumors were essentially unchanged visually. Radiomic features were extracted by using PyRadiomics (https://pyradiomics.readthedocs.io/) under varying conditions, including normalization strategies and intensity quantization. Subsequently, intraclass correlation coefficients were determined between feature values of the scan and rescan.

RESULTS

Shape features showed a higher repeatability than intensity (adjusted P < .001) and texture features (adjusted P < .001) for both T2-weighted FLAIR and T1-weighted postcontrast images. Normalization improved the overlap between the region of interest intensity histograms of scan and rescan (adjusted P < .001 for both T2-weighted FLAIR and T1-weighted postcontrast images), except in scans where brain extraction fails. As such, normalization significantly improves the repeatability of intensity features from T2-weighted FLAIR scans (adjusted P = .003 [z score normalization] and adjusted P = .002 [histogram matching]). The use of a relative intensity binning strategy as opposed to default absolute intensity binning reduces correlation between gray-level co-occurrence matrix features after normalization.

CONCLUSION

Both normalization and intensity quantization have an effect on the level of repeatability and redundancy of features, emphasizing the importance of both accurate reporting of methodology in radiomics articles and understanding the limitations of choices made in pipeline design. Supplemental material is available for this article. © RSNA, 2020See also the commentary by Tiwari and Verma in this issue.

Phase II trial of carboplatin and bevacizumab in patients with breast cancer brain metastases

BACKGROUND

We aimed to examine the safety and efficacy of bevacizumab and carboplatin in patients with breast cancer brain metastases.

METHODS

We enrolled patients with breast cancer and > 1 measurable new or progressive brain metastasis. Patients received bevacizumab 15 mg/kg intravenously (IV) on cycle 1 day 1 and carboplatin IV AUC = 5 on cycle 1 day 8. Patients with HER2-positive disease also received trastuzumab. In subsequent cycles, all drugs were administered on day 1 of each cycle. Contrast-enhanced brain MRI was performed at baseline, 24-96 h after the first bevacizumab dose (day + 1), and every 2 cycles. The primary endpoint was objective response rate in the central nervous system (CNS ORR) by composite criteria. Associations between germline VEGF single nucleotide polymorphisms (rs699947, rs2019063, rs1570360, rs833061) and progression-free survival (PFS) and overall survival (OS) were explored, as were associations between early (day + 1) MRI changes and outcomes.

RESULTS

Thirty-eight patients were enrolled (29 HER2-positive, 9 HER2-negative); all were evaluable for response. The CNS ORR was 63% (95% CI, 46-78). Median PFS was 5.62 months and median OS was 14.10 months. As compared with an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0, patients with ECOG PS 1-2 had significantly worse PFS and OS (all P < 0.01). No significant associations between VEGF genotypes or early MRI changes and clinical outcomes were observed.

CONCLUSIONS

The combination of bevacizumab and carboplatin results in a high rate of durable objective response in patients with brain metastases from breast cancer. This regimen warrants further investigation.

TRIAL REGISTRATION

NCT01004172 . Registered 28 October 2009.

Imaging Acute Metabolic Changes in Patients with Mild Traumatic Brain Injury Using Hyperpolarized [1-13C]Pyruvate

Traumatic brain injury (TBI) involves complex secondary injury processes following the primary injury. The secondary injury is often associated with rapid metabolic shifts and impaired brain function immediately after the initial tissue damage. Magnetic resonance spectroscopic imaging (MRSI) coupled with hyperpolarization of ¹³C-labeled substrates provides a unique opportunity to map the metabolic changes in the brain after traumatic injury in real-time without invasive procedures. In this report, we investigated two patients with acute mild TBI (Glasgow coma scale 15) but no anatomical brain injury or hemorrhage. Patients were imaged with hyperpolarized [1-¹³C]pyruvate MRSI 1 or 6 days after head trauma. Both patients showed significantly reduced bicarbonate (HCO₃^–) production, and one showed hyperintense lactate production at the injured sites. This study reports the feasibility of imaging altered metabolism using hyperpolarized pyruvate in patients with TBI, demonstrating the translatability and sensitivity of the technology to cerebral metabolic changes after mild TBI.

•

Clinical translation of hyperpolarized pyruvate to TBI was demonstrated

•

Patients with mild TBI were imaged with hyperpolarized [1-¹³C]pyruvate

•

Altered lactate and HCO₃^– production in the brain nearest the site of trauma

Combining inhomogeneous magnetization transfer and multipoint Dixon acquisition: Potential utility and evaluation

PURPOSE

The recently introduced inhomogeneous magnetization transfer (ihMT) method has predominantly been applied for imaging the central nervous system. Future applications of ihMT, such as in peripheral nerves and muscles, will involve imaging in the vicinity of adipose tissues. This work aims to systematically investigate the partial volume effect of fat on the ihMT signal and to propose an efficient fat-separation method that does not interfere with ihMT measurements.

METHODS

First, the influence of fat on ihMT signal was studied using simulations. Next, the ihMT sequence was combined with a multi-echo Dixon acquisition for fat separation. The sequence was tested in 9 healthy volunteers using a 3T human scanner. The ihMT ratio (ihMTR) values were calculated in regions of interest in the brain and the spinal cord using standard acquisition (no fat saturation), water-only, in-phase, and out-of-phase reconstructions. The values obtained were compared with a standard fat suppression method, spectral presaturation with inversion recovery.

RESULTS

Simulations showed variations in the ihMTR values in the presence of fat, depending on the TEs used. The IhMTR values in the brain and spinal cord derived from the water-only ihMT multi-echo Dixon images were in good agreement with values from the unsuppressed sequence. The ihMT-spectral presaturation with inversion recovery combination resulted in 24%-35% lower ihMTR values compared with the standard non-fat-suppressed acquisition.

CONCLUSION

The presence of fat within a voxel affects the ihMTR calculations. The IhMT multi-echo Dixon method does not compromise the observable ihMT effect and can potentially be used to remove fat influence in ihMT.

Long-Term Physical Exercise and Mindfulness Practice in an Aging Population

Previous studies have shown that physical exercise and mindfulness meditation can both lead to improvement in physical and mental health. However, it is unclear whether these two forms of training share the same underlying mechanisms. We compared two groups of older adults with 10 years of mindfulness meditation (integrative body-mind training, IBMT) or physical exercise (PE) experience to demonstrate their effects on brain, physiology and behavior. Healthy older adults were randomly selected from a large community health project and the groups were compared on measures of quality of life, autonomic activity (heart rate, heart rate variability, skin conductance response, respiratory amplitude/rate), immune function (secretory Immunoglobulin A, sIgA), stress hormone (cortisol) and brain imaging (resting state functional connectivity, structural differences). In comparison with PE, we found significantly higher ratings for the IBMT group on dimensions of life quality. Parasympathetic activity indexed by skin conductance response and high-frequency heart rate variability also showed more favorable outcomes in the IBMT group. However, the PE group showed lower basal heart rate and greater chest respiratory amplitude. Basal sIgA level was significantly higher and cortisol concentration was lower in the IBMT group. Lastly, the IBMT group had stronger brain connectivity between the dorsal anterior cingulate cortex (dACC) and the striatum at resting state, as well as greater volume of gray matter in the striatum. Our results indicate that mindfulness meditation and physical exercise function in part by different mechanisms, with PE increasing physical fitness and IBMT inducing plasticity in the central nervous systems. These findings suggest combining physical and mental training may achieve better health and quality of life results for an aging population.

A novel fully automated MRI-based deep-learning method for classification of IDH mutation status in brain gliomas

BACKGROUND

Isocitrate dehydrogenase (IDH) mutation status has emerged as an important prognostic marker in gliomas. Currently, reliable IDH mutation determination requires invasive surgical procedures. The purpose of this study was to develop a highly accurate, MRI-based, voxelwise deep-learning IDH classification network using T2-weighted (T2w) MR images and compare its performance to a multicontrast network.

METHODS

Multiparametric brain MRI data and corresponding genomic information were obtained for 214 subjects (94 IDH-mutated, 120 IDH wild-type) from The Cancer Imaging Archive and The Cancer Genome Atlas. Two separate networks were developed, including a T2w image-only network (T2-net) and a multicontrast (T2w, fluid attenuated inversion recovery, and T1 postcontrast) network (TS-net) to perform IDH classification and simultaneous single label tumor segmentation. The networks were trained using 3D Dense-UNets. Three-fold cross-validation was performed to generalize the networks' performance. Receiver operating characteristic analysis was also performed. Dice scores were computed to determine tumor segmentation accuracy.

RESULTS

T2-net demonstrated a mean cross-validation accuracy of 97.14% ± 0.04 in predicting IDH mutation status, with a sensitivity of 0.97 ± 0.03, specificity of 0.98 ± 0.01, and an area under the curve (AUC) of 0.98 ± 0.01. TS-net achieved a mean cross-validation accuracy of 97.12% ± 0.09, with a sensitivity of 0.98 ± 0.02, specificity of 0.97 ± 0.001, and an AUC of 0.99 ± 0.01. The mean whole tumor segmentation Dice scores were 0.85 ± 0.009 for T2-net and 0.89 ± 0.006 for TS-net.

CONCLUSION

We demonstrate high IDH classification accuracy using only T2-weighted MR images. This represents an important milestone toward clinical translation.

Spiral T1 Spin-Echo for Routine Postcontrast Brain MRI Exams: A Multicenter Multireader Clinical Evaluation

BACKGROUND AND PURPOSE

Spiral MR imaging has several advantages compared with Cartesian MR imaging that can be leveraged for added clinical value. A multicenter multireader study was designed to compare spiral with standard-of-care Cartesian postcontrast structural brain MR imaging on the basis of relative performance in 10 metrics of image quality, artifact prevalence, and diagnostic benefit.

MATERIALS AND METHODS

Seven clinical sites acquired 88 total subjects. For each subject, sites acquired 2 postcontrast MR imaging scans: a spiral 2D T1 spin-echo, and 1 of 4 routine Cartesian 2D T1 spin-echo/TSE scans (fully sampled spin-echo at 3T, 1.5T, partial Fourier, TSE). The spiral acquisition matched the Cartesian scan for scan time, geometry, and contrast. Nine neuroradiologists independently reviewed each subject, with the matching pair of spiral and Cartesian scans compared side-by-side, and scored on 10 image-quality metrics (5-point Likert scale) focused on intracranial assessment. The Wilcoxon signed rank test evaluated relative performance of spiral versus Cartesian, while the Kruskal-Wallis test assessed interprotocol differences.

RESULTS

Spiral was superior to Cartesian in 7 of 10 metrics (flow artifact mitigation, SNR, GM/WM contrast, image sharpness, lesion conspicuity, preference for diagnosing abnormal enhancement, and overall intracranial image quality), comparable in 1 of 10 metrics (motion artifacts), and inferior in 2 of 10 metrics (susceptibility artifacts, overall extracranial image quality) related to magnetic susceptibility (P < .05). Interprotocol comparison confirmed relatively higher SNR and GM/WM contrast for partial Fourier and TSE protocol groups, respectively (P < .05).

CONCLUSIONS

Spiral 2D T1 spin-echo for routine structural brain MR imaging is feasible in the clinic with conventional scanners and was preferred by neuroradiologists for overall postcontrast intracranial evaluation.

Classification of brain tumor isocitrate dehydrogenase status using MRI and deep learning

Isocitrate dehydrogenase (IDH) mutation status is an important marker in glioma diagnosis and therapy. We propose an automated pipeline for noninvasively predicting IDH status using deep learning and T2-weighted (T2w) magnetic resonance (MR) images with minimal preprocessing (N4 bias correction and normalization to zero mean and unit variance). T2w MR images and genomic data were obtained from The Cancer Imaging Archive dataset for 260 subjects (120 high-grade and 140 low-grade gliomas). A fully automated two-dimensional densely connected model was trained to classify IDH mutation status on 208 subjects and tested on another held-out set of 52 subjects using fivefold cross validation. Data leakage was avoided by ensuring subject separation during the slice-wise randomization. Mean classification accuracy of 90.5% was achieved for each axial slice in predicting the three classes of no tumor, IDH mutated, and IDH wild type. Test accuracy of 83.8% was achieved in predicting IDH mutation status for individual subjects on the test dataset of 52 subjects. We demonstrate a deep learning method to predict IDH mutation status using T2w MRI alone. Radiologic imaging studies using deep learning methods must address data leakage (subject duplication) in the randomization process to avoid upward bias in the reported classification accuracy.

Prevalence of and Risk Factors for Cerebral Microbleeds in Moyamoya Disease and Syndrome in the American Population

BACKGROUND AND PURPOSE

Cerebral microbleeds (CMB) are reported to be frequent in moyamoya disease (MMD) and moyamoya syndrome (MMS) in the Asian population. It is associated with an increased risk of intracerebral hemorrhage. The significance of CMB in MMD/MMS in non-Asian populations has not been well established. Our study aimed to investigate the prevalence of CMB in MMD/MMS in a moymoya cohort with a majority of non-Asians and to identify risk factors for developing a CMB and its predictive value for subsequent vascular events.

METHODS

The moyamoya database was compiled by screening for MMD/MMS among patients admitted to the Zale-Lipshy University Hospital at the University of Texas Southwestern Medical Center. We identified and analyzed data of 67 patients with MMD or MMS. Patients were characterized as CMB+ or CMB- based on MRI findings. In CMB+ patients, the total number and location of CMB were identified. Univariate and multivariate logistic regression were used to identify risk factors for developing CMB and whether CMB are associated with the development of subsequent vascular events.

RESULTS

Out of a total of 67 patients, 11 (16%) had CMB. Males had significantly higher odds of having CMB as compared to females (OR 1.76; 95% CI 1.40-24.3, p = 0.021). The incidence of CMB was also associated with age at diagnosis (mean age of CMB+ patients vs. CMB- patients: 44 vs. 34 years, respectively, p = 0.024), smoking (p = 0.006), and hemorrhagic stroke at presentation (p = 0.034). Logistic regression with multivariate analysis found that gender and age at diagnosis remained statistically significant. New ischemic events occurred in 2 (20%) out of 10 CMB+ patients and 13 (23%) out of 55 CMB- patients, respectively (p = 0.79). While 2 (3%) CMB- patients had a new cerebral hemorrhage during follow-up, none of the CMB+ patients did.

CONCLUSIONS

CMB are less prevalent in MMD/MMS in the USA than in Asia. An older age at diagnosis and male gender were associated with CMB. The presence of CMB was not associated with an increased risk of a subsequent ischemic or hemorrhagic stroke.

Automatic assessment of glioma burden: a deep learning algorithm for fully automated volumetric and bidimensional measurement

BACKGROUND

Longitudinal measurement of glioma burden with MRI is the basis for treatment response assessment. In this study, we developed a deep learning algorithm that automatically segments abnormal fluid attenuated inversion recovery (FLAIR) hyperintensity and contrast-enhancing tumor, quantitating tumor volumes as well as the product of maximum bidimensional diameters according to the Response Assessment in Neuro-Oncology (RANO) criteria (AutoRANO).

METHODS

Two cohorts of patients were used for this study. One consisted of 843 preoperative MRIs from 843 patients with low- or high-grade gliomas from 4 institutions and the second consisted of 713 longitudinal postoperative MRI visits from 54 patients with newly diagnosed glioblastomas (each with 2 pretreatment "baseline" MRIs) from 1 institution.

RESULTS

The automatically generated FLAIR hyperintensity volume, contrast-enhancing tumor volume, and AutoRANO were highly repeatable for the double-baseline visits, with an intraclass correlation coefficient (ICC) of 0.986, 0.991, and 0.977, respectively, on the cohort of postoperative GBM patients. Furthermore, there was high agreement between manually and automatically measured tumor volumes, with ICC values of 0.915, 0.924, and 0.965 for preoperative FLAIR hyperintensity, postoperative FLAIR hyperintensity, and postoperative contrast-enhancing tumor volumes, respectively. Lastly, the ICCs for comparing manually and automatically derived longitudinal changes in tumor burden were 0.917, 0.966, and 0.850 for FLAIR hyperintensity volume, contrast-enhancing tumor volume, and RANO measures, respectively.

CONCLUSIONS

Our automated algorithm demonstrates potential utility for evaluating tumor burden in complex posttreatment settings, although further validation in multicenter clinical trials will be needed prior to widespread implementation.

Robust pCASL perfusion imaging using a 3D Cartesian acquisition with spiral profile reordering (CASPR)

PURPOSE

To improve the robustness of arterial spin-labeled measured perfusion using a novel Cartesian acquisition with spiral profile reordering (CASPR) 3D turbo spin echo (TSE) in the brain and kidneys.

METHODS

The CASPR view ordering followed a pseudo-spiral trajectory on a Cartesian grid, by sampling the center of k-space at the beginning of each echo train of a segmented 3D TSE acquisition. With institutional review board approval and written informed consent, 14 normal subjects (9 brain and 5 kidneys) were scanned with pCASL perfusion imaging using 3D CASPR and compared against 3D linear TSE (brain and kidneys), the established 2D EPI and 3D gradient and spin echo perfusion (brain), and 2D single-shot turbo spin-echo perfusion (kidneys). The SNR and the quantitative perfusion values were compared among different acquisitions.

RESULTS

3D CASPR TSE achieved robust perfusion across all slices compared to 3D linear TSE in the brain and kidneys. Compared to 2D EPI, 3D CASPR TSE showed higher SNR across the brain (P < 0.01), and exhibited good agreement (36.4 ± 4.7 and 36.9 ± 5.3 mL/100 g/min with 2D EPI and 3D CASPR, respectively), and with 3D gradient and spin echo (27.9 ± 7.2 mL/100 g/min). Compared to a single slice 2D single-shot turbo spin-echo acquisition, 3D CASPR TSE achieved robust perfusion across the entire kidneys in similar scan time with comparable quantified perfusion values (154.1 ± 74.6 and 151.7 ± 70.6 mL/100 g/min with 2D single-shot turbo spin-echo and 3D CASPR, respectively).

CONCLUSION

The CASPR view ordering with 3D TSE achieves robust arterial spin-labeled perfusion in the brain and kidneys because of the sampling of the center of k-space at the beginning of each echo train.

Abstract WP168: Cerebrovascular Reactivity Predicts Surgical Decisions in Moyamoya Patients

Introduction: Moyamoya disease (MMD) is characterized by chronic occlusion of the distal intracranial internal carotid arteries and can be treated by revascularization surgery. At present, surgical decisions are primarily based on symptomatology and imaging studies using DSA and SPECT, which are are costly, invasive, time consuming and qualitative. Here we applied a novel iVas-MRI technique that provides quantitative assessment of multiple hemodynamic parameters in a 9-minute scan in MMD patients, and evaluated the ability of iVas parametric maps to predict surgical decisions in such patients.

Methods: Sixteen MMD patients were scanned on 3T MRI. Each patient had at least one hemisphere pending decision regarding surgery at the time of MRI. During iVas-MRI, a concomitant CO2/O2 breathing challenge was performed while BOLD images were continuously collected. BOLD images and end-tidal (Et) CO2 and O2 traces were used to calculate maps of cerebrovascular reactivity (CVR, based on BOLD signal change to EtCO2 change), cerebral blood volume (CBV, based on BOLD signal change to EtO2 change), and bolus arrival time (BAT, based on the time lag between EtCO2/O2 and BOLD signal). Parametric values from the ICA perfusion territory were compared between two groups, surgical and medical, with decision made by the treating neurosurgeons blinded to research data. The area-under-the-ROC-curve (AUC) was calculated to evaluate the performance of the MRI indices in predicting surgical decisions.

Results: Out of the 23 hemispheres under consideration, 9 were diagnosed as surgical and 14 considered medical. CVR, CO2-BAT and O2-BAT values showed significant differences between the two groups (p=0.0002, 0.005 and 0.001, respectively), while CBV values showed no difference (p=0.45). Brain hemispheres that required revascularization surgeries had lower CVR and longer BAT, compared to medically-managed hemispheres. ROC analyses revealed an AUC of 0.89, 0.84 and 0.95 for CVR, O2-BAT, and CO2-BAT respectively.

Conclusion: Our results showed that CVR and BAT had a great accuracy in predicting surgical decisions. Therefore, iVas-MRI may be a cost-effective and reliable method to select between medical and surgical treatments for MMD patients.

Microstructural correlates of 3D steady-state inhomogeneous magnetization transfer (ihMT) in the human brain white matter assessed by myelin water imaging and diffusion tensor imaging

PURPOSE

To compare the recently introduced inhomogeneous magnetization transfer (ihMT) technique with more established MRI techniques including myelin water imaging (MWI) and diffusion tensor imaging (DTI), and to evaluate the microstructural attributes correlating with this new contrast method in the human brain white matter.

METHODS

Eight adult healthy volunteers underwent T₁ -weighted, ihMT, MWI, and DTI imaging on a 3T human scanner. The ihMT ratio (ihMTR), myelin water fraction (MWF), fractional anisotropy (FA), radial diffusivity (RD), axial diffusivity (AD), and mean diffusivity (MD) values were calculated from different white matter tracts. The angle ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>θ</mml:mi></mml:math> ) between the directions of the principal eigenvector, as measured by DTI, and the main magnetic field was calculated for all voxels from various fiber tracts. The ihMTR was correlated with MWF and DTI metrics.

RESULTS

A strong correlation was found between ihMTR and MWF (ρ = 0.77, P < 0.0001). This was followed by moderate to weak correlations between ihMTR and DTI metrics: RD (ρ = -0.30, P < 0.0001), FA (ρ = 0.20, P < 0.0001), MD (ρ = -0.19, P < 0.0001), AD (ρ = 0.02, P < 0.0001). A strong correlation was found between ihMTR and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>θ</mml:mi></mml:math> (ρ = -0.541, P < 0.0001).

CONCLUSION

The strong correlation with myelin water imaging and its low coefficient of variation suggest that ihMT has the potential to become a new structural imaging marker of myelin. The substantial orientational dependence of ihMT should be taken into account when evaluating and quantitatively interpreting ihMT results.

Measurement of glycine in healthy and tumorous brain by triple-refocusing MRS at 3 T in vivo

Glycine (Gly) has been implicated in several neurological disorders, including malignant brain tumors. The precise measurement of Gly is challenging largely as a result of the spectral overlap with myo-inositol (mI). We report a new triple-refocusing sequence for the reliable co-detection of Gly and mI at 3 T and for the evaluation of Gly in healthy and tumorous brain. The sequence parameters were optimized with density-matrix simulations and phantom validation. With a total TE of 134 ms, the sequence gave complete suppression of the mI signal between 3.5 and 3.6 ppm and, consequently, well-defined Gly (3.55 ppm) and mI (3.64 ppm) peaks. In vivo ¹ H magnetic resonance spectroscopy (MRS) data were acquired from the gray matter (GM)-dominant medial occipital and white matter (WM)-dominant left parietal regions in six healthy subjects, and analyzed with LCModel using in-house-calculated basis spectra. Tissue segmentation was performed to obtain the GM and WM contents within the MRS voxels. Metabolites were quantified with reference to GM-rich medial occipital total creatine at 8 mM. The Gly and mI concentrations were estimated to be 0.63 ± 0.05 and 8.6 ± 0.6 mM for the medial occipital and 0.34 ± 0.05 and 5.3 ± 0.8 mM for the left parietal regions, respectively. From linear regression of the metabolite estimates versus fractional GM content, the concentration ratios between pure GM and pure WM were estimated to be 2.6 and 2.1 for Gly and mI, respectively. Clinical application of the optimized sequence was performed in four subjects with brain tumor. The Gly levels in tumors were higher than those of healthy brain. Gly elevation was more extensive in a post-contrast enhancing region than in a non-enhancing region. The data indicate that the optimized triple-refocusing sequence may provide reliable co-detection of Gly and mI, and alterations of Gly in brain tumors can be precisely evaluated.

Echo-planar spectroscopic imaging with dual-readout alternated gradients (DRAG-EPSI) at 7 T: Application for 2-hydroxyglutarate imaging in glioma patients

PURPOSE

To develop echo-planar spectroscopic imaging (EPSI) with large spectral width and accomplish high-resolution imaging of 2-hydroxyglutarate (2HG) at 7 T.

METHODS

We designed a new EPSI readout scheme at 7 T. Data were recorded with dual-readout alternated gradients and combined according to the gradient polarity. Following validation of its performance in phantoms, the new readout scheme, together with previously reported 2HG-optimized magnetic resonance spectroscopy (point-resolved spectroscopy echo time of 78 ms), was used for time-efficient and high-resolution imaging of 2HG and other metabolites in five glioma patients before treatment. Unsuppressed water, acquired with EPSI, was used as reference for multichannel combination, eddy-current compensation, and metabolite quantification. Spectral fitting was conducted with the LCModel using in-house calculated basis sets.

RESULTS

Using a readout gradient strength of 9.5 mT/m and slew rate of 90 mT/m/ms, dual-readout alternated gradients EPSI permitted 1638-Hz spectral width with 6 × 6 mm² in-plane resolution at 7 T. Phantom data indicated that dual-readout alternated gradients EPSI provides proper metabolite signals and induces much less frequency drifts than conventional EPSI. For a spatial resolution of 0.5 mL, 2HG was detected in tumors with precision (Cramer-Rao lower bound < 10%). The 2HG was estimated to be 2.3 to 3.3 mM in tumors of three patients with biopsy-proven isocitrate dehydrogenase (IDH) mutant gliomas. The 2HG was undetectable in an IDH wild-type glioblastoma. For a radiographically suggested glioma, the estimated 2HG of 2.3 ± 0.2 mM (Cramer-Rao lower bound < 10%) indicated that the lesion may be an IDH mutant glioma.

CONCLUSIONS

The data indicated that the dual-readout alternated gradients EPSI can provide reliable high-resolution imaging of 2HG in glioma patients at 7 T in vivo. Magn Reson Med 79:1851-1861, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Early changes in glioblastoma metabolism measured by MR spectroscopic imaging during combination of anti-angiogenic cediranib and chemoradiation therapy are associated with survival

Precise assessment of treatment response in glioblastoma during combined anti-angiogenic and chemoradiation remains a challenge. In particular, early detection of treatment response by standard anatomical imaging is confounded by pseudo-response or pseudo-progression. Metabolic changes may be more specific for tumor physiology and less confounded by changes in blood-brain barrier permeability. We hypothesize that metabolic changes probed by magnetic resonance spectroscopic imaging can stratify patient response early during combination therapy. We performed a prospective longitudinal imaging study in newly diagnosed glioblastoma patients enrolled in a phase II clinical trial of the pan-vascular endothelial growth factor receptor inhibitor cediranib in combination with standard fractionated radiation and temozolomide (chemoradiation). Forty patients were imaged weekly during therapy with an imaging protocol that included magnetic resonance spectroscopic imaging, perfusion magnetic resonance imaging, and anatomical magnetic resonance imaging. Data were analyzed using receiver operator characteristics, Cox proportional hazards model, and Kaplan-Meier survival plots. We observed that the ratio of total choline to healthy creatine after 1 month of treatment was significantly associated with overall survival, and provided as single parameter: (1) the largest area under curve (0.859) in receiver operator characteristics, (2) the highest hazard ratio (HR = 85.85, P = 0.006) in Cox proportional hazards model, (3) the largest separation (P = 0.004) in Kaplan-Meier survival plots. An inverse correlation was observed between total choline/healthy creatine and cerebral blood flow, but no significant relation to tumor volumetrics was identified. Our results suggest that in vivo metabolic biomarkers obtained by magnetic resonance spectroscopic imaging may be an early indicator of response to anti-angiogenic therapy combined with standard chemoradiation in newly diagnosed glioblastoma.

Multiparametric estimation of brain hemodynamics with MR fingerprinting ASL

PURPOSE

Assessment of brain hemodynamics without exogenous contrast agents is of increasing importance in clinical applications. This study aims to develop an MR perfusion technique that can provide noncontrast and multiparametric estimation of hemodynamic markers.

METHODS

We devised an arterial spin labeling (ASL) method based on the principle of MR fingerprinting (MRF), referred to as MRF-ASL. By taking advantage of the rich information contained in MRF sequence, up to seven hemodynamic parameters can be estimated concomitantly. Feasibility demonstration, flip angle optimization, comparison with Look-Locker ASL, reproducibility test, sensitivity to hypercapnia challenge, and initial clinical application in an intracranial steno-occlusive process, Moyamoya disease, were performed to evaluate this technique.

RESULTS

Magnetic resonance fingerprinting ASL provided estimation of up to seven parameters, including B1+, tissue T₁ , cerebral blood flow (CBF), tissue bolus arrival time (BAT), pass-through arterial BAT, pass-through blood volume, and pass-through blood travel time. Coefficients of variation of the estimated parameters ranged from 0.2 to 9.6%. Hypercapnia resulted in an increase in CBF by 57.7%, and a decrease in BAT by 13.7 and 24.8% in tissue and vessels, respectively. Patients with Moyamoya disease showed diminished CBF and lengthened BAT that could not be detected with regular ASL.

CONCLUSION

Magnetic resonance fingerprinting ASL is a promising technique for noncontrast, multiparametric perfusion assessment. Magn Reson Med 78:1812-1823, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Detection of 2-hydroxyglutarate in brain tumors by triple-refocusing MR spectroscopy at 3T in vivo

PURPOSE

To test the efficacy of triple-refocusing MR spectroscopy (MRS) for improved detection of 2-hydroxyglutarate (2HG) in brain tumors at 3T in vivo.

METHODS

The triple-refocusing sequence parameters were tailored at 3T, with density-matrix simulations and phantom validation, for enhancing the 2HG 2.25-ppm signal selectivity with respect to the adjacent resonances of glutamate (Glu), glutamine (Gln), and gamma-aminobutyric acid (GABA). In vivo MRS data were acquired from 15 glioma patients and analyzed with LCModel using calculated basis spectra. Metabolites were quantified with reference to water.

RESULTS

A triple-refocusing sequence (echo time = 137 ms) was obtained for 2HG detection. The 2HG 2.25-ppm signal was large and narrow while the Glu and Gln signals between 2.2 and 2.3 ppm were minimal. The optimized triple refocusing offered improved separation of 2HG from Glu, Gln and GABA when compared with published MRS methods. 2HG was detected in all 15 patients, the estimated 2HG concentrations ranging from 2.4 to 15.0 mM, with Cramer-Rao lower bounds of 2%-11%. The 2HG estimates did not show significant correlation with total choline.

CONCLUSION

The optimized triple refocusing provides excellent 2HG signal discrimination from adjacent resonances and may confer reliable in vivo measurement of 2HG at relatively low concentrations. Magn Reson Med 78:40-48, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

In vivo detection of 2-hydroxyglutarate in brain tumors by optimized point-resolved spectroscopy (PRESS) at 7T

PURPOSE

To test the efficacy of 7T MRS for in vivo detection of 2-hydroxyglutarate (2HG) in brain tumors.

METHODS

The subecho times of point-resolved spectroscopy (PRESS) were optimized at 7T with density-matrix simulations and phantom validation to improve the 2HG signal selectivity with respect to the neighboring resonances of γ-aminobutyric acid (GABA), glutamate (Glu), and glutamine (Gln). MRS data were acquired from 12 subjects with gliomas in vivo and analyzed with LCModel using calculated basis spectra. Metabolite levels were quantified using unsuppressed short echo time (TE) water as a reference.

RESULTS

The PRESS TE was optimized as TE = 78 ms (TE₁ = 58 ms and TE₂ = 20 ms), at which the 2HG 2.25 ppm resonance appeared as a temporally maximum inverted narrow peak and the GABA, Glu, and Gln resonances between 2.2 and 2.5 ppm were all positive peaks. The PRESS TE = 78 ms method offered improved discrimination of 2HG from Glu, Gln, and GABA when compared with short-TE MRS. 2HG was detected in all patients enrolled in the study, the estimated 2HG concentrations ranging from 1.0 to 6.2 mM, with percentage standard deviation of 2%-7%.

CONCLUSION

Data indicate that the optimized MRS provides good selectivity of 2HG from other metabolite signals and may confer reliable in vivo detection of 2HG at relatively low concentrations. Magn Reson Med 77:936-944, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Basic MR relaxation mechanisms and contrast agent design

The diagnostic capabilities of magnetic resonance imaging (MRI) have undergone continuous and substantial evolution by virtue of hardware and software innovations and the development and implementation of exogenous contrast media. Thirty years since the first MRI contrast agent was approved for clinical use, a reliance on MR contrast media persists, largely to improve image quality with higher contrast resolution and to provide additional functional characterization of normal and abnormal tissues. Further development of MR contrast media is an important component in the quest for continued augmentation of diagnostic capabilities. In this review we detail the many important considerations when pursuing the design and use of MR contrast media. We offer a perspective on the importance of chemical stability, particularly kinetic stability, and how this influences one's thinking about the safety of metal-ligand-based contrast agents. We discuss the mechanisms involved in MR relaxation in the context of probe design strategies. A brief description of currently available contrast agents is accompanied by an in-depth discussion that highlights promising MRI contrast agents in the development of future clinical and research applications. Our intention is to give a diverse audience an improved understanding of the factors involved in developing new types of safe and highly efficient MR contrast agents and, at the same time, provide an appreciation of the insights into physiology and disease that newer types of responsive agents can provide.

A generic support vector machine model for preoperative glioma survival associations

PURPOSE

To develop a generic support vector machine (SVM) model by using magnetic resonance (MR) imaging-based blood volume distribution data for preoperative glioma survival associations and to prospectively evaluate the diagnostic effectiveness of this model in autonomous patient data.

MATERIALS AND METHODS

Institutional and regional medical ethics committees approved the study, and all patients signed a consent form. Two hundred thirty-five preoperative adult patients from two institutions with a subsequent histologically confirmed diagnosis of glioma after surgery were included retrospectively. An SVM learning technique was applied to MR imaging-based whole-tumor relative cerebral blood volume (rCBV) histograms. SVM models with the highest diagnostic accuracy for 6-month and 1-, 2-, and 3-year survival associations were trained on 101 patients from the first institution. With Cox survival analysis, the diagnostic effectiveness of the SVM models was tested on independent data from 134 patients at the second institution.

RESULTS

were adjusted for known survival predictors, including patient age, tumor size, neurologic status, and postsurgery treatment, and were compared with survival associations from an expert reader.

RESULTS

Compared with total qualitative assessment by an expert reader, the whole-tumor rCBV-based SVM model was the strongest parameter associated with 6-month and 1-, 2-, and 3-year survival in the independent patient data (area under the receiver operating characteristic curve, 0.794-0.851; hazard ratio, 5.4-21.2).

DISCUSSION

Machine learning by means of SVM in combination with whole-tumor rCBV histogram analysis can be used to identify early patient survival in aggressive gliomas. The SVM model returned higher diagnostic accuracy values than an expert reader, and the model appears to be insensitive to patient, observer, and institutional variations.

Proton T2 measurement and quantification of lactate in brain tumors by MRS at 3 Tesla in vivo

PURPOSE

To evaluate the T2 relaxation time of lactate (Lac) in brain tumors and the correlation of the T2 and concentration with tumor grades.

METHODS

Eight pairs of the subecho time sets of point-resolved spectroscopy were selected between 58 and 268 ms, with numerical and phantom analyses, for Lac T2 measurement. In vivo spectra were acquired from 24 subjects with gliomas (13 low grade and 11 high grade) and analyzed with LCModel using numerically-calculated basis spectra. The metabolite T2 relaxation time was obtained from monoexponential fitting of the multi-echo time (TE) signal estimates versus TE. The metabolite concentration was estimated from the zero-TE extrapolation of the T2 fits.

RESULTS

The Lac T2 was estimated to be approximately 240 ms, without a significant difference between low and high grade tumors. The Lac concentration was estimated to be 4.1 ± 3.4 and 7.0 ± 4.7 mM for low and high grades respectively, but the difference was not significant.

CONCLUSION

The Lac T2 was similar among gliomas regardless of their tumor grades. This suggests that the T2 value from this study may be applicable to obtain the T2 relaxation-free estimates of Lac in a subset of brain tumors.

Low incidence of pseudoprogression by imaging in newly diagnosed glioblastoma patients treated with cediranib in combination with chemoradiation

BACKGROUND

Chemoradiation (CRT) can significantly modify the radiographic appearance of malignant gliomas, especially within the immediate post-CRT period. Pseudoprogression (PsP) is an increasingly recognized phenomenon in this setting, and is thought to be secondary to increased permeability as a byproduct of the complex process of radiation-induced tissue injury, possibly enhanced by temozolomide. We sought to determine whether the addition of a vascular endothelial growth factor (VEGF) signaling inhibitor (cediranib) to conventional CRT had an impact on the frequency of PsP, by comparing two groups of patients with newly diagnosed glioblastoma before, during, and after CRT.

METHODS

All patients underwent serial magnetic resonance imaging as part of institutional review board-approved clinical studies. Eleven patients in the control group received only chemoradiation, whereas 29 patients in the study group received chemoradiation and cediranib until disease progression or toxicity. Response assessment was defined according to Response Assessment in Neuro-Oncology criteria, and patients with enlarging lesions were classified into true tumor progressions (TTP) or PsP, based on serial radiographic follow-up.

RESULTS

Two patients in the study group (7%) showed signs of apparent early tumor progression, and both were subsequently classified as TTP. Six patients in the control group (54%) showed signs of apparent early tumor progression, and three were subsequently classified as TTP and three as PsP. The frequency of PsP was significantly higher in the control group.

CONCLUSION

Administration of a VEGF inhibitor during and after CRT modifies the expression of PsP by imaging.