Search

Purpose Prognostically favorable IDH-mutant gliomas are known to produce oncometabolite D-2-hydroxyglutarate (2HG). In this study, we investigated metabolite-based features of patients with grade 2 and 3 glioma using 2HG-specific in vivo MR spectroscopy, to determine their relationship with image-guided tissue pathology and predictive role in progression-free survival (PFS). Methods Forty-five patients received pre-operative MRIs that included 3-D spectroscopy optimized for 2HG detection. Spectral data were reconstructed and quantified to compare metabolite levels according to molecular pathology (IDH1R132H, 1p/19q, and p53); glioma grade; histological subtype; and T2 lesion versus normal-appearing white matter (NAWM) ROIs. Levels of 2HG were correlated with other metabolites and pathological parameters (cellularity, MIB-1) from image-guided tissue samples using Pearson’s correlation test. Metabolites predictive of PFS were evaluated with Cox proportional hazards models. Results Quantifiable levels of 2HG in 39/42 (93%) IDH+ and 1/3 (33%) IDH– patients indicated a 91.1% apparent detection accuracy. Myo-inositol/total choline (tCho) showed reduced values in astrocytic (1p/19q-wildtype), p53-mutant, and grade 3 (vs. 2) IDH-mutant gliomas (p p p p p p = 0.002), total NAA (R = − 0.61; p = 0.002) and cellularity (R = 0.37; p = 0.04) but not MIB-1. Increasing levels of 2HG/tCr (p = 0.0007, HR 5.594) and thresholding (≥ 0.905, median value; p = 0.02) predicted adverse PFS. Conclusion In vivo 2HG detection can reasonably be achieved on clinical scanners and increased levels may signal adverse PFS.

PURPOSE In patients with diffuse low-grade glioma (LGG), the extent of surgical tumor resection (EOR) has a controversial role, in part because a randomized clinical trial with different levels of EOR is not feasible. METHODS In a 20-year retrospective cohort of 392 patients with IDH-mutant grade 2 glioma, we analyzed the combined effects of volumetric EOR and molecular and clinical factors on overall survival (OS) and progression-free survival by recursive partitioning analysis. The OS results were validated in two external cohorts (n = 365). Propensity score analysis of the combined cohorts (n = 757) was used to mimic a randomized clinical trial with varying levels of EOR. RESULTS Recursive partitioning analysis identified three survival risk groups. Median OS was shortest in two subsets of patients with astrocytoma: those with postoperative tumor volume (TV) > 4.6 mL and those with preoperative TV > 43.1 mL and postoperative TV ≤ 4.6 mL. Intermediate OS was seen in patients with astrocytoma who had chemotherapy with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL in addition to oligodendroglioma patients with either preoperative TV > 43.1 mL and residual TV ≤ 4.6 mL or postoperative residual volume > 4.6 mL. Longest OS was seen in astrocytoma patients with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL who received no chemotherapy and oligodendroglioma patients with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL. EOR ≥ 75% improved survival outcomes, as shown by propensity score analysis. CONCLUSION Across both subtypes of LGG, EOR beginning at 75% improves OS while beginning at 80% improves progression-free survival. Nonetheless, maximal resection with preservation of neurological function remains the treatment goal. Our findings have implications for surgical strategies for LGGs, particularly oligodendroglioma.

Background The mechanistic basis for neurocognitive deficits in central nervous system (CNS) lymphoma and other brain tumors is incompletely understood. We tested the hypothesis that tumor metabolism impairs neurotransmitter pathways and neurocognitive function. Methods We performed serial cerebrospinal fluid (CSF) metabolomic analyses using liquid chromatography-electrospray tandem mass spectrometry to evaluate changes in the tumor microenvironment in 14 patients with recurrent CNS lymphoma, focusing on 18 metabolites involved in neurotransmission and bioenergetics. These were paired with serial mini-mental state examination (MMSE) and MRI studies for tumor volumetric analyses. Patients were analyzed in the setting of the phase I trial of lenalidomide/rituximab. Associations were assessed by Pearson and Spearman correlation coefficient. Generalized estimating equation (GEE) models were also established, adjusting for within-subject repeated measures. Results Of 18 metabolites, elevated CSF lactate correlated most strongly with lower MMSE score (P < 8E-8, ρ = −0.67). High lactate was associated with lower gamma-aminobutyric acid (GABA), higher glutamate/GABA ratio, and dopamine. Conversely, high succinate correlated with higher MMSE scores. Serial analysis demonstrated a reproducible, time-dependent, reciprocal correlation between changes in lactate and GABA concentrations. While high lactate and low GABA correlated with tumor contrast-enhancing volume, they correlated more significantly with lower MMSE scores than tumor volumes. Conclusions We provide evidence that lactate production and Warburg metabolism may impact neurotransmitter dysregulation and neurocognition in CNS lymphomas. We identify novel metabolomic biomarkers that may be applied in future studies of neurocognition in CNS lymphomas. Elucidation of mechanistic interactions between lymphoma metabolism, neurotransmitter imbalance, and neurocognition may promote interventions that preserve cognitive function.

INTRODUCTION Hyperpolarized carbon-13 (HP-13C) MR enables rapid dynamic imaging of metabolic pathways in the human brain using non-toxic, non-radioactive metabolites as tracers. This study presents our unique experience on the benefit of using [1-13C]pyruvate and [2-13C]pyruvate MR for evaluating patients with glioma. METHODS 132 scans (71 using an integrated 13C /1H coil) including steady-state 1H-MRSI (~10 min) and dynamic HP-13C imaging (60 sec) following injection of HP [1-13C]pyruvate (N=125) or [2-13C]pyruvate (N=7) were acquired from 46 patients with glioma (18F/28M; 15 IDH-mutant, 27 IDH-wildtype, 4 IDH-status-unknown). Maps of temporally-summed 13C-metabolite signals, ratios, and kinetic rate constants were calculated for contrast-enhancing, nonenhancing, and normal-appearing-white-matter (NAWM) regions and compared to steady-state metabolic metrics. RESULTS The only adverse event, in a single patient, was a burning sensation after the injection that resolved after saline flush. The mean time-to-injection of HP probes was 58.6±14.0 sec. Signal-to-noise ratios of [1-13C]lactate and [13C]bicarbonate within the NAWM from the HP [1-13C]pyruvate data were 53±39 and 13±6, respectively. The SD/mean of repeated injections (N=3) for lactate/pyruvate and pyruvate-to-lactate conversion rates were 3.8±3.1% and 6.5±3.1% in the NAWM, respectively. Patients with progressive GBM had significantly higher lactate/pyruvate and lower bicarbonate/lactate (p< 0.05) in contrast- and non-enhancing lesions compared to NAWM. Significantly elevated lactate/pyruvate and reduced bicarbonate/lactate (p< 0.01) were found in contrast-enhancing compared to nonenhancing regions, whereas choline/NAA and steady-state 1H-lactate levels were similar. HP [2-13C]pyruvate data showed reduced glutamate/pyruvate and pyruvate-to-glutamate conversion rates in T2 lesions compared to contralateral normal-appearing brain in IDH-mutant gliomas, consistent with known metabolic reprogramming. CONCLUSION This study demonstrates the potential benefit of dynamic HP-13C MRI for evaluating patients with glioma, which provides unique and spatially distinct contrast compared to steady-state metabolic imaging. Ongoing studies will further characterize dynamic metabolism in specific glioma subtypes and provide biomarkers for evaluating responses to treatment.

Although physiologic (diffusion-weighted and perfusion-weighted) MRI has shown promise in identifying regions of recurrent tumor (rTumor) in patients with glioblastoma suspected of progression, distinguishing treatment-induced effects (TxE) from rTumor on anatomical MRI remains a challenge. Whereas prior larger-scale machine learning (ML)-based studies mostly utilize anatomical imaging alone and/or perform lesion-level predictions, this study aimed to develop a non-invasive, radiopathomic tool for regional probabilistic mapping of TxE using 208 tissue-samples (55 pathologically-confirmed TxE, 153 recurrent glioblastoma) acquired from 107 patients with known spatial coordinates on pre-surgical MRI. We tested the hypothesis that applying a deep-learning (DL) model that included physiological MRI can: 1) more accurately identify areas of TxE that mimic rTumor on anatomical MRI and 2) better generalize to an independent test set than ML-models or a DL-model that uses anatomical MRI alone. An 80/20 split for training/validation was used after 1/3 of the patients were withheld for testing. Oversampling of TxE samples was employed to address class imbalance and an equal proportion of TxE samples was maintained across all datasets. Three ML-models, their ensemble, and a deep 4D-convolutional-neural-network were trained based on normalized anatomical (post-contrast T1, T2-FLAIR), diffusion-weighted (ADC, FA), and DSC perfusion-weighted (PeakHeight, %recovery) images cropped to 10mm-cubic patches centered on the coordinates from where tissue was obtained. Although Random Forest and voting-ensembled ML-models using all imaging and the anatomical DL-model had the best validation performance (AUC=0.81-0.82), these models did not generalize (test AUC=0.58-0.59). The DL-model including physiologic images had slightly lower validation AUC (0.78) but the best overall test AUC (0.795), indicating superior generalizability. Elevated blood volume (nPeakHeight) was the most important feature. Our DL-model’s interpretability was also demonstrated by disrupting class separation after shuffling voxels within each input patch. These results suggest that using deep-learning with physiologic MRI can improve intratumoral classification of TxE from rTumor.

INTRODUCTION Mutations in isocitrate dehydrogenase (IDH) have been investigated as a prognostic biomarker in glioma. The presence of the IDH mutation (IDHm) is associated with 2-hydroxyglutarate (2HG) production and inhibition of glutamate synthesis (McBrayer, Cell 2018). Hyperpolarized carbon-13 (HP-13C) MRI enables dynamic measurements of in-vivo metabolism using a [2-13C]pyruvate labeled probe that undergoes conversion to [2-13C]lactate and [5-13C]glutamate. Here, we present HP [2-13C]pyruvate data from healthy volunteers and patients with IDHm diffuse glioma. Due to its intrinsic low signal-to-noise ratio (SNR), we demonstrate the ability of post-processing denoising to improve its utility and aid in detection of metabolic changes associated with IDHm. METHODS Dynamic HP 13C data were acquired following intravenous injection of [2-13C]pyruvate from five healthy volunteers and one patient with IDHm grade III astrocytoma. A novel multi-resolution frequency specific multislice EPI sequence was used to obtain [2-13C]pyruvate, [5-13C]glutamate, and downfield and upfield [2-13C]lactate signals (3s temporal resolution, pyruvate/lactate/glutamate spatial resolutions = 0.75x0.75cm2/ 2.25x2.25cm2/ 2.25x2.25cm2, 5 slices 3cm thick). Following phase correction, patch-based tensor decomposition denoising was applied to metabolite images. Metabolite differences between normal-appearing white matter (NAWM) and T2 lesion were examined for the patient data. RESULTS HP [2-13C]pyruvate imaging is able to simultaneously probe glycolytic ([2-13C]lactate) and oxidative ([5-13C]glutamate) metabolism. Denoised pyruvate/lactate/glutamate signals achieved a 4-9/3-6/3-7 fold increase in SNR. T2 lesion exhibited decreased glutamate-to-pyruvate and glutamate-to-lactate AUC ratios versus contralateral NAWM (p< 0.018, p < 1.5e-5), consistent with IDH mutant status. CONCLUSION We successfully demonstrated the feasibility of applying variable resolution HP [2-13C]pyruvate metabolic imaging to detect IDHm specific metabolism. This technique addresses a major hurdle in HP 13C MRI by improving SNR while permitting robust metabolism quantification. Future studies will optimize methods for acquiring and processing data to evaluate further data acquired from IDHm glioma patients. Supported by NIH T32 CA151022, P01 CA118816, and NICO.

BackgroundDiagnostic classification of diffuse gliomas now requires an assessment of molecular features, often including IDH-mutation and 1p19q-codeletion status. Because genetic testing requires an invasive process, an alternative noninvasive approach is attractive, particularly if resection is not recommended. The goal of this study was to evaluate the effects of training strategy and incorporation of biologically relevant images on predicting genetic subtypes with deep learning.MethodsOur dataset consisted of 384 patients with newly diagnosed gliomas who underwent preoperative MRI with standard anatomical and diffusion-weighted imaging, and 147 patients from an external cohort with anatomical imaging. Using tissue samples acquired during surgery, each glioma was classified into IDH-wildtype (IDHwt), IDH-mutant/1p19q-noncodeleted (IDHmut-intact), and IDH-mutant/1p19q-codeleted (IDHmut-codel) subgroups. After optimizing training parameters, top performing convolutional neural network (CNN) classifiers were trained, validated, and tested using combinations of anatomical and diffusion MRI with either a 3-class or tiered structure. Generalization to an external cohort was assessed using anatomical imaging models.ResultsThe best model used a 3-class CNN containing diffusion-weighted imaging as an input, achieving 85.7% (95% CI: [77.1, 100]) overall test accuracy and correctly classifying 95.2%, 88.9%, 60.0% of the IDHwt, IDHmut-intact, and IDHmut-codel tumors. In general, 3-class models outperformed tiered approaches by 13.5%-17.5%, and models that included diffusion-weighted imaging were 5%-8.8% more accurate than those that used only anatomical imaging.ConclusionTraining a classifier to predict both IDH-mutation and 1p19q-codeletion status outperformed a tiered structure that first predicted IDH-mutation, then 1p19q-codeletion. Including apparent diffusion coefficient (ADC), a surrogate marker of cellularity, more accurately captured differences between subgroups.

OBJECTIVES: Treatment-induced lesions represent a great challenge in neuro-oncology. The aims of this study were (i) to characterize treatment induced lesions in glioblastoma patients treated with chemoradiotherapy and heat-shock protein (HSP) vaccine and (ii) to evaluate the diagnostic accuracy of diffusion weighted imaging (DWI) for differentiation between treatment-induced lesions and tumor progression. METHODS: Twenty-seven patients with newly diagnosed glioblastoma treated with heat-shock protein vaccine and chemoradiotherapy were included. Serial magnetic resonance imaging (MRI) evaluation was performed to detect treatment-induced lesions and assess their growth. Quantitative analysis of the apparent diffusion coefficient (ADC) was performed to discriminate treatment-induced lesions from tumor progression. Mann Whitney U-test and receiver operating characteristic (ROC) curves were used for analysis. RESULTS: Thirty-three percent of patients developed treatment-induced lesions. Five treatment-related lesions appeared between end of radiotherapy and the first vaccine administration; 4 lesions within the first 4 months from vaccine initiation and 1 at 3.5 years. Three patients with pathology proven treatment-induced lesions showed a biphasic growth pattern progressed shortly after. Apparent diffusion coefficient (ADC) ratio between the peripheral enhancing rim and central necrosis showed an accuracy of 0.84 (95%CI: 0.63–1) for differentiation between progression and treatment-induced lesions. CONCLUSION: Our findings do not support the iRANO recommendation of a 6-month time window in which progressive disease should not be declared after immunotherapy initiation. A biphasic growth pattern of pathologically proven treatment-induced lesions was associated with a dismal prognosis. The presence of lower ADC values in the central necrotic portion of the lesions compared to the enhancing rim shows high specificity for detection of treatment-induced lesions.

INTRODUCTION Noninvasive, radiopathomic mapping of tumor aggressiveness can benefit patients with glioma by guiding the selection of tissue samples for diagnosis, increasing extent of resection, and non-invasively characterizing residual tumor burden for subsequent treatment. Although prior studies have demonstrated the utility of metabolic metrics quantified from 1H-MR Spectroscopy (MRS) in probing tumor pathology, this study evaluated the benefit of using the entire 1D-spectrum and deep learning for radiopathomic mapping of intratumoral cellularity, proliferation (ki-67), and a new tumor aggressiveness index (TAI) defined as log((n(ki−67)+n(cellularity))*tumor-score). METHODS Multi-voxel 1H-MRS was acquired on 281 patients newly diagnosed with a glioma (47% IDH-wildtype) immediately before surgical resection. After reconstructing individual spectra at the locations where tissue samples were obtained during surgery and normalizing by NAA in contralateral normal-appearing-white-matter, 607 spectra with corresponding histopathology were deemed of sufficient quality for analysis. A 1D convolutional-neural-network with bidirectional long- and short-term memory deep-learning model using the entire spectrum (0.6-3.6ppm) was compared to mixed-effects regression (with choline-to-NAA index[CNI]) and Random Forest (with CNI+normalized peak heights) models for predicting ki-67, cellularity, and TAI. Results & DISCUSSION Using deep-learning on the entire spectrum resulted in 10.3%-22.1% lower mean absolute error (MAE) and 0.32-0.37 higher R2 values compared to using CNI alone or a random forest model with multiple metabolic metrics. MAE values for all 3 deep-learning models were 26-44% < 1 standard deviation of the ground truth, demonstrating reasonable prediction accuracy within the test data set. Although the lowest MAE (0.16) and highest R2 (0.41) was attained when predicting TAI with deep-learning, the prediction of cellularity resulted in the lowest %MAE. Colormaps of predicted pathology identified regions of heightened aggressiveness surrounding tissue samples with most abnormal pathological features that sometimes extended beyond the non-enhancing lesion. Current work is evaluating the clinical utility of our deep-learning model and predicted maps of aggressiveness.

Treatment failure in glioblastoma, the most common and lethal primary brain tumor in adults, is often attributed to intratumoral heterogeneity as it fosters tumor evolution and selection of resistant clones. Sources of intratumoral heterogeneity may include genomic alterations, such as single-nucleotide and structural variants, and epigenomic alterations, such as changes in chromatin structure and transcriptional regulation. The relative extent and functional significance of these contributors to intratumoral heterogeneity in glioblastoma have yet to be elucidated. In collaboration with neurosurgeons and neuro-imaging experts, we have established a novel approach towards characterizing intratumoral heterogeneity in three-dimensional (3D) space. For patients undergoing tumor resection, we utilize 3D surgical neuronavigation to safely acquire ~10 samples representing maximal anatomical diversity. Samples are mapped by 3D spatial coordinates and integrated with patient MRI scans for 360º visualization of sample location in context of the brain. We have now conducted whole-exome sequencing (Exome-Seq), assay for transposase-accessible chromatin (ATAC-Seq), and RNA-sequencing (RNA-Seq) for 83 spatially mapped samples obtained from 8 patients with primary IDH-WT glioblastoma. Integrative data analysis provides unprecedented insight into sources of intratumoral heterogeneity in glioblastoma and their 3D spatial patterning. We find that tumor cells show aberrant transcription factor activity, activation of fetal brain programs, and variable chromatin accessibility at CTCF-binding loop anchors indicating plasticity in higher-order chromatin structure. Chromosome conformation capture analysis by Hi-C extends these findings and reveals intratumoral differences in long-range chromatin interactions due to structural variants and enhancer hijacking. Further, we use deconvolution to identify microenvironmental contributors to epigenomic intratumoral heterogeneity including neural, glial, and immune populations. We define chromatin signatures associated with microenvironmental cell types and states, revealing their 3D spatial patterning, and validate these findings by single-cell ATAC-Seq. Our work thus establishes both tumor and microenvironmental sources of intratumoral heterogeneity in glioblastoma, revealing their chromatin programs and 3D spatial patterning. As a resource for further investigation, we have developed an interactive data sharing platform that enables visualization of brain tumor intratumoral heterogeneity in 360 degrees. Citation Format: Radhika Mathur, Qixuan Wang, Patrick Schupp, Ana Nikolic, Takafumi Yamaguchi, Stephanie Hilz, Chibo Hong, Ivan Smirnov, Marisa LaFontaine, Joanna Phillips, Susan Chang, Yan Li, Janine Lupo, Paul Boutros, Marco Gallo, Michael Oldham, Feng Yue, Joseph Costello. 3D spatial sampling and integrated omics reveal sources and patterning of intratumoral heterogeneity in glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3621.

INTRODUCTION Mutations in isocitrate dehydrogenase (IDH1/2) genes are both diagnostic and prognostic biomarkers for lower grade gliomas, which can be noninvasively evaluated via MRS detection of 2-hydroxyglutarate (2HG). A high percentage of lower grade gliomas harbor IDH mutations that have a major impact on tumor biology. METHODS Forty IDH-mutant lower grade glioma patients were recruited and scanned using 2HG-optimized 97ms PRESS spectroscopy prior to surgical resection, whereupon image-guided tissue samples were acquired and analyzed via histopathology. Spectra were processed and quantified using LCModel. Wilcoxon ranksum tests were used to compare the difference in metabolic parameters between patient subgroups. Cox proportional hazard models were used to evaluate the influence of metabolic parameters on progression free survival (PFS). Pearson and Kendall Tau rank correlations were used for metabolites and histopathology parameters. RESULTS Levels of 2HG were quantifiable in 37/40 (92.5%) IDH+ patients, yielding a detection rate of 92.5%. A significant difference in myoinositol/total choline was found between astrocytoma and oligodendroglioma in the newly diagnosed grade II population (N=10/13, p=0.042) and also in the overall population (N=25/15, p=0.022). 2HG/creatine was the only significant prognostic indicator on PFS in newly diagnosed patients (p=0.036, HR=4.435; N=33), who had a median PFS of 1472 days [95%CI 994 1950 days] (15 censored). Significantly decreased glutamate/creatine and increased (2HG+GABA)/creatine (p< 0.0001) were found in the T2 lesion compared to those in the normal appearing white matter. At the location where the biopsy samples were taken, there was a significant relationship between 2HG/creatine and glycine/creatine (0.93±0.01; mean±SD Tau; p=0.0005) as well as 2HG/creatine and glutathione/creatine (0.96±0.02; mean±SD Tau; p=0.001) but no significance between the metabolites and pathology parameters were found. CONCLUSIONS The non-invasive detection of 2HG can provide clinically relevant information for patient management and our study demonstrated both the feasibility of 2HG detection and its relationship with PFS.

BACKGROUND The overall prognostic significance of maximal surgical resection in patients with diffuse low-grade glioma has been well established. Nonetheless, prior studies omit the combined importance of molecular subclass, patient characteristics, and chemoradiation. Similar to findings recently published in newly diagnosed glioblastoma, incorporation of these interactive factors may redefine the relative benefit of cytoreductive surgery. METHODS We examine the interactive effects of volumetric extent of resection with molecular and clinical factors to develop a new roadmap for cytoreductive surgery. Based on a 20-year retrospective cohort of 556 patients with WHO II diffuse low-grade glioma treated with surgery at UCSF 444 had complete records for survival modeling and recursive partitioning (RPA) to investigate multivariate relationships of overall and progression free survival. RESULTS Regardless of molecular subtype, patients with tumor volume under 55cm3 and postoperative volume of residual under 1.9cm3 experience the longest OS (median OS: not reached). Patients with volume of residual over 1.9cm3 experience a OS similar to that of patients with large (over 55cm3) oligodendrogliomas (median OS: not reached). Patients faring worst have large (over 55cm3) astrocytic gliomas (median OS: 84.8 months). Patients not treated with chemotherapy and either ATRX wild-type tumors or ATRX-mutant tumors with small (under 1cm3) volume of residual have the longest PFS together with chemotherapy treated patients who receive either no radiation or radiation for p53-mutant tumors under 30cm3 (median PFS 119 months). Patients with the shortest PFS are under 32-years with larger volume of residual (>1cm3), who receive no chemotherapy for ATRX-mutant tumors together with patients who receive both chemoradiation for larger (>30cm3) p53 mutant tumors (median PFS 30.8 months). CONCLUSION This is the first study to combine extent of resection with molecular and clinical information which paves the way for rethinking surgical strategies for individual patients with newly diagnosed low-grade gliomas.

Although combined spin- and gradient-echo (SAGE) dynamic susceptibility-contrast (DSC) MRI can provide perfusion quantification that is sensitive to both macrovessels and microvessels while correcting for T1 -shortening effects, spatial coverage is often limited in order to maintain a high temporal resolution for DSC quantification. In this work, we combined a SAGE echo-planar imaging (EPI) sequence with simultaneous multi-slice (SMS) excitation and blipped controlled aliasing in parallel imaging (blipped CAIPI) at 3 T to achieve both high temporal resolution and whole brain coverage. Two protocols using this sequence with multi-band (MB) acceleration factors of 2 and 3 were evaluated in 20 patients with treated gliomas to determine the optimal scan parameters for clinical use. ΔR2 *(t) and ΔR2 (t) curves were derived to calculate dynamic signal-to-noise ratio (dSNR), ΔR2 *- and ΔR2 -based relative cerebral blood volume (rCBV), and mean vessel diameter (mVD) for each voxel. The resulting SAGE DSC images acquired using MB acceleration of 3 versus 2 appeared visually similar in terms of image distortion and contrast. The difference in the mean dSNR from normal-appearing white matter (NAWM) and that in the mean dSNR between NAWM and normal-appearing gray matter were not statistically significant between the two protocols. ΔR2 *- and ΔR2 -rCBV maps and mVD maps provided unique contrast and spatial heterogeneity within tumors.

BackgroundDifferentiating treatment-induced injury from recurrent high-grade glioma is an ongoing challenge in neuro-oncology, in part due to lesion heterogeneity. This study aimed to determine whether different MR features were relevant for distinguishing recurrent tumor from the effects of treatment in contrast-enhancing lesions (CEL) and non-enhancing lesions (NEL).MethodsThis prospective study analyzed 291 tissue samples (222 recurrent tumor, 69 treatment-effect) with known coordinates on imaging from 139 patients who underwent preoperative 3T MRI and surgery for a suspected recurrence. 8 MR parameter values were tested from perfusion-weighted, diffusion-weighted, and MR spectroscopic imaging at each tissue sample location for association with histopathological outcome using generalized estimating equation models for CEL and NEL tissue samples. Individual cutoff values were evaluated using receiver operating characteristic curve analysis with 5-fold cross-validation.ResultsIn tissue samples obtained from CEL, elevated relative cerebral blood volume (rCBV) was associated with the presence of recurrent tumor pathology (P < 0.03), while increases in normalized choline (nCho) and choline-to-NAA index (CNI) were associated with the presence of recurrent tumor pathology in NEL tissue samples (P < 0.008). A mean CNI cutoff value of 2.7 had the highest performance, resulting in mean sensitivity and specificity of 0.61 and 0.81 for distinguishing treatment-effect from recurrent tumor within the NEL.ConclusionAlthough our results support prior work that underscores the utility of rCBV in distinguishing the effects of treatment from recurrent tumor within the contrast enhancing lesion, we found that metabolic parameters may be better at differentiating recurrent tumor from treatment-related changes in the NEL of high-grade gliomas.

Although current advances for automated glioma lesion segmentation and volumetric measurements using deep learning have yielded high performance on newly-diagnosed patients, response assessment in neuro-oncology still relies on manually-drawn, cross-sectional areas of the tumor because these models do not generalize to patients in the post-treatment setting, where they are most needed in the clinic. Surgical resections, adjuvant treatment, or disease progression can alter the characteristics of these lesions on T2-weighted imaging, causing measures of segmentation accuracy, typically measured by Dice coefficients of overlap (DCs), to drop by ~15%. To improve the generalizability of T2-lesion segmentation to patients with glioma post-treatment, we evaluated the effects of: 1) training with different proportions of newly-diagnosed and treated gliomas, 2) applying transfer learning from pre- to post-treatment domains, and 3) incorporating a loss term that spatially weights the lesion boundaries with greater emphasis in training. Using 425 patients (208 newly-diagnosed, 217 post-Tx, with 25 treated patients withheld as a test set) and a top-performing model previously trained on newly-diagnosed gliomas, we found that DCs increased by 10% (to 0.84) then plateaued after including ~25% of post-treatment patients in training. Transfer learning (pre-training on newly-diagnosed and finetuning with post-treatment data) significantly improved Hausdorf distances (HDs), a measure more sensitive to changes at the lesion boundaries, by 17% after including 26% post-treatment images in training, while DCs remained similar. Although modifying our loss functions with boundary-weighted penalizations resulted in comparable DCs to using standard DC loss, HD measures were further reduced by 26%, suggesting that HDs may be a more sensitive metric to subtle changes in segmentation accuracy than DCs. Current work is evaluating their utility in providing accurate volumes for real-time response assessment in the clinic using workflows that have recently been deployed on our clinical PACs system.

INTRODUCTION The goal of this study was to characterize progressive and pseudoprogressive GBM using multi-parametric hyperpolarized (HP)-13C / 1H MRI. METHODS Dynamic HP-13C MRI was acquired from 13 patients with progressive GBM [patients (scans): 2(3) IDH-mutant; 11(13) IDH-wildtype] and 2 IDH-wildtype patients (3 scans) demonstrating pseudo-progression following intravenous injection of HP [1-13C]pyruvate. Frequency-selective echo-planar imaging (3s temporal resolution, 3.38 cm3 spatial resolution) captured [1-13C]pyruvate metabolism to [1-13C]lactate and 13C-bicarbonate in the brain. Dynamic 13C data were kinetically modeled to obtain the pyruvate-to-lactate conversion rate constant k PL and temporally summed to calculate 13C-metabolite percentiles and ratios (linearly interpolated 2x in-plane). 1H imaging included T2, post-Gd T1, perfusion (nCBV, %recovery), diffusion (ADC), and lactate-edited spectroscopy (CNI, choline-to-NAA index; 1H-lactate). The normal-appearing white matter (NAWM), non-enhancing lesion (NEL), and contrast-enhancing lesion (CEL) were segmented from 1H images. 13C-resolution masks were iteratively applied on a voxel-wise basis to evaluate 1H imaging parameters within each ROI and multi-parametric data were collectively evaluated using a mixed effects model in R. RESULTS Progressive IDH-mutant GBM compared to wildtype counterparts displayed increased perfusion %recovery (p < 0.001) and k PL (p < 0.01), together with reduced 1H-lactate (p < 0.001) and pyruvate percentile (p < 0.01), in the T2 lesion. Among IDH-wildtype progressive GBM, the CEL was distinguished from NEL/NAWM by increased nCBV (p < 0.05/0.001), 1H-lactate (p < 0.05/0.001); and decreased bicarbonate / lactate (p < 0.05/0.001). The CEL and NEL were collectively distinguished from NAWM by elevated CNI (p < 0.001/0.001), ADC (p < 0.05/0.001), pyruvate percentile (p < 0.001/0.001), lactate percentile (p < 0.001/0.001), and relative lactate / pyruvate (p < 0.001/0.05). Psuedo-progressive IDH-wildtype GBM displayed lower k PL (T2 Lesion; p < 0.01) and nCBV (CEL; p < 0.01) compared to progressive GBM. CONCLUSION HP-13C parameters can potentially augment proton imaging and demonstrated Warburg-associated metabolic alterations.

Importance Per the World Health Organization 2016 integrative classification, newly diagnosed glioblastomas are separated into isocitrate dehydrogenase gene 1 or 2 (IDH)–wild-type andIDH-mutant subtypes, with median patient survival of 1.2 and 3.6 years, respectively. Although maximal resection of contrast-enhanced (CE) tumor is associated with longer survival, the prognostic importance of maximal resection within molecular subgroups and the potential importance of resection of non–contrast-enhanced (NCE) disease is poorly understood. Objective To assess the association of resection of CE and NCE tumors in conjunction with molecular and clinical information to develop a new road map for cytoreductive surgery. Design, Setting, and Participants This retrospective, multicenter cohort study included a development cohort from the University of California, San Francisco (761 patients diagnosed from January 1, 1997, through December 31, 2017, with 9.6 years of follow-up) and validation cohorts from the Mayo Clinic (107 patients diagnosed from January 1, 2004, through December 31, 2014, with 5.7 years of follow-up) and the Ohio Brain Tumor Study (99 patients with data collected from January 1, 2008, through December 31, 2011, with a median follow-up of 10.9 months). Image accessors were blinded to patient groupings. Eligible patients underwent surgical resection for newly diagnosed glioblastoma and had available survival, molecular, and clinical data and preoperative and postoperative magnetic resonance images. Data were analyzed from November 15, 2018, to March 15, 2019. Main Outcomes and Measures Overall survival. Results Among the 761 patients included in the development cohort (468 [61.5%] men; median age, 60 [interquartile range, 51.6-67.7] years), younger patients withIDH–wild-type tumors and aggressive resection of CE and NCE tumors had survival similar to that of patients withIDH-mutant tumors (median overall survival [OS], 37.3 [95% CI, 31.6-70.7] months). Younger patients withIDH–wild-type tumors and reduction of CE tumor but residual NCE tumors fared worse (median OS, 16.5 [95% CI, 14.7-18.3] months). Older patients withIDH–wild-type tumors benefited from reduction of CE tumor (median OS, 12.4 [95% CI, 11.4-14.0] months). The results were validated in the 2 external cohorts. The association between aggressive CE and NCE in patients withIDH–wild-type tumors was not attenuated by the methylation status of the promoter region of the DNA repair enzyme O6-methylguanine-DNA methyltransferase. Conclusions and Relevance This study confirms an association between maximal resection of CE tumor and OS in patients with glioblastoma across all subgroups. In addition, maximal resection of NCE tumor was associated with longer OS in younger patients, regardless ofIDHstatus, and among patients withIDH–wild-type glioblastoma regardless of the methylation status of the promoter region of the DNA repair enzyme O6-methylguanine-DNA methyltransferase. These conclusions may help reassess surgical strategies for individual patients with newly diagnosed glioblastoma.

Graphical abstract, Highlights • Serial HP 13C MRI was evaluated for data consistency and abnormal metabolism. • Metabolism of [1-13C]pyruvate to lactate and bicarbonate was kinetically modeled. • Conversion rates within NAWM were consistent in healthy volunteer and patient scans • Progressed tumor lesions showed higher relative conversion rates to [1-13C]lactate. • Globally elevated rate constants were observed with anti-angiogenic treatment., Background Hyperpolarized carbon-13 (HP-13C) MRI is a non-invasive imaging technique for probing brain metabolism, which may improve clinical cancer surveillance. This work aimed to characterize the consistency of serial HP-13C imaging in patients undergoing treatment for brain tumors and determine whether there is evidence of aberrant metabolism in the tumor lesion compared to normal-appearing tissue. Methods Serial dynamic HP [1-13C]pyruvate MRI was performed on 3 healthy volunteers (6 total examinations) and 5 patients (21 total examinations) with diffuse infiltrating glioma during their course of treatment, using a frequency-selective echo-planar imaging (EPI) sequence. HP-13C imaging at routine clinical timepoints overlapped treatment, including radiotherapy (RT), temozolomide (TMZ) chemotherapy, and anti-angiogenic/investigational agents. Apparent rate constants for [1-13C]pyruvate conversion to [1-13C]lactate (kPL) and [13C]bicarbonate (kPB) were simultaneously quantified based on an inputless kinetic model within normal-appearing white matter (NAWM) and anatomic lesions defined from 1H MRI. The inter/intra-subject consistency of kPL-NAWM and kPB-NAWM was measured in terms of the coefficient of variation (CV). Results When excluding scans following anti-angiogenic therapy, patient values of kPL-NAWM and kPB-NAWM were 0.020 s−1 ± 23.8% and 0.0058 s−1 ± 27.7% (mean ± CV) across 17 HP-13C MRIs, with intra-patient serial kPL-NAWM/kPB-NAWM CVs ranging 6.8–16.6%/10.6–40.7%. In 4/5 patients, these values (0.018 s−1 ± 13.4% and 0.0058 s−1 ± 24.4%; n = 13) were more similar to those from healthy volunteers (0.018 s−1 ± 5.0% and 0.0043 s−1 ± 12.6%; n = 6) (mean ± CV). The anti-angiogenic agent bevacizumab was associated with global elevations in apparent rate constants, with maximum kPL-NAWM in 2 patients reaching 0.047 ± 0.001 and 0.047 ± 0.003 s−1 (±model error). In 3 patients with progressive disease, anatomic lesions showed elevated kPL relative to kPL-NAWM of 0.024 ± 0.001 s−1 (±model error) in the absence of gadolinium enhancement, and 0.032 ± 0.008, 0.040 ± 0.003 and 0.041 ± 0.009 s−1 with gadolinium enhancement. The lesion kPB in patients was reduced to unquantifiable values compared to kPB-NAWM. Conclusion Serial measures of HP [1-13C]pyruvate metabolism displayed consistency in the NAWM of healthy volunteers and patients. Both kPL and kPB were globally elevated following bevacizumab treatment, while progressive disease demonstrated elevated kPL in gadolinium-enhancing and non-enhancing lesions. Larger prospective studies with homogeneous patient populations are planned to evaluate metabolic changes following treatment.

INTRODUCTION Hyperpolarized (HP) 13C MRI is a novel metabolic imaging that allows for real time in vivo probing of pathway-specific metabolic processes relevant to gliomas. The first human brain application reports were published recently. In this study, we characterized longitudinal dynamic and static brain metabolism changes from serial HP 13C and 1H metabolic imaging in a patient with multiply recurrent glioma. METHODS Eight dynamic HP 13C imaging, 1H MRI and MRSI examinations were acquired over a period of 15 months in a patient with multiply recurrent glioma that underwent malignant transformation. The patient was imaged after initial tumor debulking, after chemoradiotherapy, at first recurrence, after re-irradiation and anti-angiogentic therapy, and at second recurrence. Ratios of pyruvate-to-lactate and pyruvate-to-bicarbonate were calculated from HP 13C imaging and choline-to-NAA index (CNI) was derived from 1H MRSI. RESULTS/ DISCUSSION Successful therapies are usually associated with a drop in pyruvate-to-lactate conversion, mediated by different mechanisms from various treatments. We identified three key findings over repeat HP 13C MRIs: 1) a marked response to therapy with a reduction in pyruvate-to-lactate conversion corresponding to associated imaging response to therapy at the primary site of disease. Following chemoradiotherapy, the volume of T2 hyperintensity decreased from 28.21cc to 15.29 cc, and the volume of metabolic abnormality (CNI >3) decreased from 27.84 cc to 10.36 cc. 2) increasing pyruvate-to-lactate conversion in a region of recurrence before morphologic signal abnormality, and 3) a reduction in pyruvate-to-lactate conversion in a recurrent lesion following anti-angiogenic therapy. CONCLUSION Here we show that one can reliably image glioma patients using HP 13C MRI alongside standard of care morphologic and physiologic MRI including diffusion, perfusion, and 1H MRSI. Serial HP 13C MRI can be performed in the neuro-oncologic clinical setting and pyruvate-to-lactate metabolism may be useful to monitor treatment response.

INTRODUCTION Multi-contrast-weighted (MCW) MRI allows derivation of T1-weighted, T2-weighted, and T2-FLAIR images and their parametric maps. Under current practice, these high-resolution 3D MR images are acquired separately with a total scan time is approximately 30 minutes. The goal of this study was to develop a novel sequence and processing strategy for MCW MRI and lesion detection of gliomas with 1mm isotropic resolution in 6 minutes, which can significantly improve clinical imaging workflow. The proposed whole brain, single-scan method enables the generation of segmented tumor regions that correspond the hyperintense lesion on T2-weighted or T2-FLAIR images (T2L). METHODS Our strategy involves using a highly accelerated 3D MRI sequence to achieve whole brain, MCW images that can be jointly segmented with modified k-means clustering. Six patients with enhancing gliomas were scanned with this sequence before the injection of gadolinium, 4 of which also received a repeat scan post-gadolinium. Our automatic segmentation of the T2L and manual definition of the contrast-enhancing lesion (CEL) was compared against manually segmented lesions drawn from conventional anatomical imaging. RESULTS Whole-brain MCW MRI was achieved with 1mm isotropic resolution in 6 mins. Grey-matter, white-matter, CSF, and the T2L could be automatically segmented with our clustering method. The resulting automatically-segmented T2L volume was not significantly different between our new sequence and clinically defined regions (average Dice Coefficient=0.81(+/-0.05); p=0.95). CONCLUSIONS Our novel, 6-minute scan and processing strategy was able to generate MCW anatomical images and automatically segment grey-matter, white-matter, and T2L volumes that were similar to manual segmentation of conventional anatomical imaging. We are currently extending the automatic segmentation to the CEL. This approach has the potential to improve clinical workflow by allowing time for therapy-specific advanced imaging and providing a noninvasive tool for monitoring tumor evolution. Validation is currently being performed in a larger cohort.

Treatment failure in glioblastoma is often attributed to intratumoral heterogeneity (ITH), which fosters tumor evolution and selection of therapy-resistant clones. While genomic alterations are known contributors to ITH, emerging studies highlight functional roles for epigenomic ITH which integrates differentiation status, stochastic events, and microenvironmental inputs. Here, we have established a novel platform for integrative characterization of genomic and epigenomic ITH of glioblastoma in three-dimensional (3-D) space. In collaboration with neurosurgeons and biomedical imaging experts, we utilize 3-D surgical neuro-navigation to safely acquire ~10 tumor samples per patient representing maximal anatomical diversity. We conduct whole-exome sequencing, RNA sequencing, and assay for transposase-accessible chromatin using sequencing (ATAC-Seq) on each sample. The spatial location of each sample is mapped by its 3-D coordinates, allowing 360-degree visualization of genomic and epigenomic ITH for each patient. We demonstrate this approach on 8 patients with primary IDH-WT glioblastoma (83 spatially mapped samples), providing unprecedented insight into their spatial organization at the genomic and epigenomic levels. We link genetically defined tumor subclones to patterns of open chromatin and gene regulation, revealing underlying transcription factor binding at active promoters and enhancers. We also identify ITH in whole-genome doubling and focal oncogene amplification events in multiple patients, which we then link with epigenomic ITH. Further, to study microenvironmental inputs and their contribution to epigenomic ITH, we conduct deconvolution of RNA sequencing and ATAC-Seq data by analyzing feature co-variation. We resolve the 3-D spatial organization of immune, neural, and other nontumor cell types present in glioblastoma, characterizing their functional states and interactions with tumor cells. This work provides the most comprehensive spatial characterization of genomic and epigenomic ITH to date in glioblastoma. As a resource for further investigation, we have developed an interactive data sharing platform – The 3D Glioma Atlas – that enables 360-degree visualization of both genomic and epigenomic ITH.

BACKGROUND The prognostic importance of maximal resection of contrast enhancing and non-contrast enhancing disease has been established. Nonetheless, glioblastomas exist within the framework of complex neural circuitry serving cognition, movement, and behavior consequential leading to neurological impairments. The prognostic importance of neurological impairments on survival remains poorly understood. METHODS This is a retrospective, single cohort study from UCSF including 316 eligible patients diagnosed over 20 years with 9.6 years of follow-up. All patients underwent surgical resection for newly diagnosed glioblastoma for whom survival, molecular, preoperative and postoperative MRI images, and clinical data were available. All patients had chemoradiation treated IDH-wild-type glioblastoma with available preoperative and 1-month post-surgical resection neurological outcomes. We employed survival models and recursive partitioning (RPA) to investigate multivariate relationships of overall survival (OS). RESULTS Preoperative neurological impairments were present in 75.6% (n= 239) and new post resection impairments were identified in 37.3% (n=117). Univariate analysis confirmed that new postoperative cognitive impairment [HR 7.91, 95% CI 2.47-25.33] and hemiplegia [HR 3.38, 95% CI 0.83-13.67] (not hemiparesis) impact OS. Risk stratified grouping by RPA demonstrated that gross total resection of contrast enhancing tumor in patients with no new postoperative neurological impairments confers the longest OS (median OS 27.1 months 95%CI 21.5-33.7). Patients with any residual tumor volume after surgery but no new neurological deficits experience a similar survival to younger patients (under 65) with 1 or more new postoperative neurological deficits (median OS 16.6 months 95%CI 15.2-19.2). Shortest OS is identified in patients with any volume of residual tumor plus 1 or more new postoperative neurological deficits and age over 65 (median OS 11.4 months, 95%CI 9.3-13.5). CONCLUSIONS This study confirms that new postoperative neurological impairments impact overall survival in patients with chemoradiation treated IDH-wild-type glioblastoma.

Recurrent high grade gliomas (rHGG) are difficult to diagnose accurately as they are often confounded with treatment-induced effects (TxE) because on conventional T1w and T2w anatomical MR imaging these two phenomena appear identical. This could result in removing a patient from an effective second-line therapy, confound the results of a clinical trial of a new therapeutic, and expose a patient to unnecessary surgical intervention. Advanced MR imaging modalities have shown promise in distinguishing between TxE and rHGG, but prior studies are unable to account for heterogeneity within the same lesion and often lack pathological confirmation. The goal of this study was to identify imaging parameters that could spatially map directly to pathology to account for heterogeneity within lesions. 1–8 tissue samples were collected upon surgical resection of suspected rHGG. A total of 484 samples were collected from 183 patients; 332 of these (163 patients) were labeled as either pathologically confirmed treatment effect or having no tumor cells (tumor score=0) within the imaging abnormality (TxE; 80/51 samples/patients) or rHGG (tumor score=2–3; 252/118 samples/patients) by a pathologist. Preoperative 3T MRI scans included T1w-Gd and T2w FLAIR; diffusion tensor imaging (b=1000s/mm2); 3D MRSI; and dynamic susceptibility-contrast perfusion MRI. Univariate models were created using generalized estimating equations to evaluate each variable’s ability to distinguish TxE from rHGG tissue samples. Normalized choline (median=1.1 TxE & 1.2 rHGG), choline-to-NAA index (CNI, median=2.4 TxE & 3.5 rHGG), and normalized cerebral blood volume (nCBV, median=1.2 TxE & 1.4 rHGG) were significantly associated with biopsy-level classification (p=0.034, 0.015, 0.017). These results are being consolidated into a multiparametric algorithm for predicting TxE or rHGG, which could be of clinical importance in managing patients with recurrent rHGG. Downstream analysis will create a spatial map of rHGG to help guide future surgical sampling.

Although the overall prognostic significance of maximal surgical resection of contrast-enhancing tumor in glioblastoma patients is well established, prior studies have not evaluated the combined importance of resection, molecular markers, patient characteristics, and chemoradiation. Incorporation of these factors may redefine the relative benefit of cytoreductive surgery and establish differing thresholds for extent of resection in varying clinical presentations. In the first study of its kind, we examine the interactive effects of volumetric extent of resection with molecular and clinical factors to develop a new roadmap for cytoreductive surgery. Based on a 20-year retrospective cohort of 850 glioblastoma patients who had initial surgery at UCSF, we employed survival models and recursive partitioning (RPA) to investigate multivariate relationships of overall survival (OS), both in the entire cohort as well as a subset diagnosed since 2005 (Stupp-era) with IDH1 mutation status available (n=470). For the entire cohort and the Stupp-era subset, the RPAs elucidate the combinatorial consequence of treatment, age, IDH1 status (in the subset), and resection of both enhancing and non-enhancing tumor. In the Stupp-era, temozolomide-treated patients that are IDH-wildtype and >65 clearly benefit from a reduction of the enhancing tumor (median OS: 10.1 vs 15.8 months). IDH-wildtype, temozolomide-treated patients under 65 benefit from reduction of both enhancing and non-enhancing tumor with a median survival similar to that of IDH-mutant, temozolomide-treated patients (combined median OS: 33.7 months). The patients faring worst are those that did not receive temozolomide that are >65 and/or have ≥0.3 cm(3) residual enhancing tumor (median OS: 4 months). These risk models outperform all published prognostic models. This is the first study to combine resection of contrast-enhancing and non-enhancing tumor in conjunction with molecular and clinical information in a large single-institution study, and paves the way for rethinking surgical strategies for individual patients with newly diagnosed glioblastoma.

INTRODUCTION: The safety and efficacy of a heat shock protein peptide complex- 96 vaccine (HSPPC-96, Prophage) has been previously studied in phase II single-arm trials for the treatment of newly diagnosed and recurrent glioblastoma (GBM). These studies demonstrated modest improvements in survival compared with historical standards. PTEN loss has been recently associated with immunoresistance in GBM patients, mediated in part by B7-H1. PTEN status has not shown clear prognostic value in GBM patients treated with standard of care therapies. The aim of this study is to evaluate the prognostic significance of PTEN status in newly diagnosed GBM patients treated with autologous HSP vaccine and standard chemoradiation. METHODS: Our institutional cohort of patients enrolled in a single arm, phase II study of adult GBM patients treated with autologous HSP vaccine and standard chemoradiation (n=27) was analyzed. Differences in overall survival (OS) by PTEN status were evaluated via Kaplan-Meier curves and Log-rank test. RESULTS: Median overall survival (n=27) was 26 months. 23 patients had PTEN status available. PTEN loss was found in 16 patients (69.6%) whereas retained PTEN was present in 7 patients (30.4%). Median OS was 59 months (95% CI, 0–120 months) in patients with retained PTEN and 23 months (95% CI, 15–30 months) in patients with PTEN loss. The difference in OS was statistically significant (p=0.037). CONCLUSION: Retained PTEN expression was associated with extended survival in GBM patients treated with HSP vaccine. This finding suggests that PTEN loss may be associated with resistance to vaccine treatment and emphasizes the need for subgroup analysis in further immunotherapy studies.

Proton magnetic resonance spectroscopic imaging (1H MRSI) is a powerful noninvasive method for assessing the spatial extent and properties of abnormal metabolism in patients with glioma. Hyperpolarized 13C metabolic imaging is a new molecular imaging modality that can be used to assess real-time changes in metabolism and has been shown to be safe and feasible in patients with glioma. In this study, both 3D 1H lactate-edited MRSI and hyperpolarized [1-13C]pyruvate imaging data were obtained and compared in thirteen patients with glioma to assess steady state versus dynamic brain metabolism. FLAIR images from the 1H examination were registered to the FSE images from the 13C examination for each subject. The corresponding transformation matrix was then applied to spectra and metabolite maps from the excitation volume of the 1H MRSI data. The volumes of the anatomic and metabolic lesions varied between patients, and using a multi-slice frequency specific EPI sequence allowed improved coverage of the lesion of interest compared to a 2D dynamic EPSI sequence. The choline-to-NAA index values and levels of steady state lactate peak heights from the 1H MRSI data were elevated in the lesion, while 13C bicarbonate levels were reduced and 13C lactate levels were similar or lower compared with normal appearing brain. The initial results indicated that there was a negative association between estimates of hyperpolarized 13C lactate/pyruvate and steady state normalized lactate peak heights, and a positive association between hyperpolarized 13C lactate/pyruvate and Cho/NAA in the T2 lesion. For one patient who received repeated examinations within 2 months and was assessed as having stable disease after treatment, the 13C lactate-to-pyruvate ratio in the T2 lesion was 0.65 vs. 0.35 and 0.43 vs. 0.43 in normal appearing brain. Future studies will evaluate a larger patient population to see whether these relationships hold up.

BACKGROUND Previous studies of solid tumors have been restricted in their ability to map how heterogeneous cell populations evolved within the tumor in three-dimensional (3D) space due to insufficient sampling, typically one sample per tumor, and a lack of knowledge of where within the tumor the sample was obtained. Knowledge of the extensivity of heterogeneity and how it is spatially distributed is crucial for assessing whether a therapeutic target is truly tumor-wide, and for exploring how mutations relate to heterogeneity in the local microenvironment. METHODS We developed a novel platform to integrate and visualize in 3D multi-omics data generated from each of 8–10 spatially mapped samples per tumor. Together, the 171 samples collected using this approach from 16 adult diffuse glioma at diagnosis and recurrence form a novel resource – the 3D Glioma Atlas. RESULTS By maximally sampling the tumor geography without excluding samples based on low cancer cell fraction (CCF), we identify a subpopulation of glioblastoma with pervasively lower CCF likely excluded by other atlases, such as the TCGA, that used stringent CCF cutoffs. Exome sequencing of 3D-mapped samples from lower-grade tumors revealed that clonal expansions are typically spatially segregated, implying minimal tumor-wide intermixing of genetically heterogenous cells. Heterogeneity is less spatially segregated for faster-growing high-grade tumors, suggesting that cell populations expand in these tumors differently. Recurrent low-grade tumors have greater intratumoral mutational heterogeneity than newly diagnosed tumors, though this did not translate into greater dissimilarity in gene expression profiles for recurrent tumors, suggesting minimal functional impact of this additional mutational diversity on gene expression. CONCLUSIONS The delineation of spatial patterns of heterogeneity that our work provides enables more informed interpretation of biopsies and greater insight into the factors shaping intratumoral variation of gene expression patterns. Ongoing work is exploring the spatial patterning of amplification events and the tumor microenvironment.

Treatment-related changes often mimic or mask tumor on standard anatomic imaging, making it difficult to monitor disease recurrence. Hyperpolarized (HP) carbon-13 MR imaging allows for real-time non-invasive measurement of metabolism, which may improve patient surveillance. Here, we focused on characterizing serial HP scans in patients undergoing treatment compared to healthy controls. Serial dynamic HP C-13 MRI scans were performed on 5 patients with recurrent glioma (22 total) and 3 healthy controls (4 total) using an echo-planar imaging sequence (2.88-8cm3 spatial resolution, 3s temporal resolution, 60s), following injection of 0.43mL/kg of 250mM HP [1-13C]pyruvate. Apparent rate constants were modeled for enzymatic conversion of pyruvate-to-lactate (kPL) via cytosolic lactate dehydrogenase and pyruvate-to-bicarbonate (kPB) via mitochondrial pyruvate dehydrogenase and carbonic anhydrase. Regions of interest included normal-appearing white matter (NAWM) and T2-hyperintense lesions (T2L), which were segmented from H-1 MR images and then aligned to the HP data. Carbon voxels containing >30% of NAWM or T2L were included in the analysis. Healthy controls demonstrated consistent kPL and kPB values over 4 scans in NAWM with SD/Mean of 5% and 12%, respectively. Compared to the median kPL-NAWM of 0.022s-1 in controls, the 5 patients had median serial kPL-NAWM values of 0.023, 0.023, 0.023, 0.029, and 0.015s-1, and mean serial ratios of kPL between T2L and NAWM (kPL-T2L/kPL-NAWM) of 1.22, 1.27, 1.05, 1.32, and 1.37s-1, indicating higher values in putative tumor. Median kPB-NAWM in controls was 0.004s-1 and ranged in patients 0.003-0.006s-1. Two patients with >4 serial scans, showed consistent kPL-NAWM over standard-of-care treatment and elevated kPL-T2L within new lesions, but up to 85% increase in kPL-NAWM with bevacizumab, which may be attributed to reduced BBB permeability. Stable patients generally demonstrated consistent kPL-T2L values that were lower compared to progressive patients. Future studies will include multi-parametric 1H imaging analysis in a larger patient population.

Three-dimensional proton magnetic resonance spectroscopic imaging (MRSI) is a powerful non-invasive tool for characterizing spatial variations in metabolic profiles for patients with glioma. Metabolic parameters obtained using this technique have been shown to predict treatment response, disease progression, and transformation to a more malignant phenotype. The availability of ultra-high-field MR systems has the potential to improve the characterization of metabolites. The purpose of this study was to compare the metabolite profiles acquired with conventional long echo time (TE) MRSI at 3T with those obtained with short TE MRSI at 3T and 7T in patients with glioma. The data acquisition parameters were optimized separately for each echo time and field strength to obtain volumetric coverage within clinically feasible data acquisition times of 5-10 min. While a higher field strength did provide better detection of metabolites with overlapping peaks, spatial coverage was reduced and the use of inversion recovery to reduce lipid precluded the detection of lipid in regions of necrosis. For serial evaluation of large, heterogeneous lesions, the use of 3T short TE MRSI may thus be preferred. Despite the limited number of metabolites that it is able to detect, the use of 3T long TE MRSI gives the best contrast in choline/N-acetyl aspartate between normal appearing brain and tumor and also allows the separate detection of lactate and lipid. It may therefore be preferred for serial evaluation of patients with high-grade glioma and for detection of malignant transformation in patients with low-grade glioma.

INTRODUCTION The purpose of this work was to investigate the trajectory of tumor growth and multiparametric MRI characteristics in grade II and III gliomas leading up to the onset of new contrast enhancement or the expansion of FLAIR signal abnormality. METHODS Fifty-one grade II and III glioma patients (39 grade II, 12 grade III) underwent serial MR imaging at standard of care timepoints. Multiparametric MR imaging was performed and maps of the apparent diffusion coefficient (ADC) and percentiles (10th, 50th, 90th) from histograms of normalized signal intensities were calculated. Regions of Interest (ROIs) were defined for the entire T2/FLAIR lesion, and when present, the contrast-enhancing (CE) lesion. Serial imaging parameters were acquired for each of these ROIs. Linear rates of change were calculated for CE volume, T2/FLAIR volume, normalized T2 FSE intensity within the FLAIR volume, normalized T2/FLAIR intensity within the FLAIR volume, and normalized median, 10th percentile and 90th% ADC values within the FLAIR volume. RESULTS Thirty-nine patients progressed while on study (7 by CE growth, 32 by T2/FLAIR growth), and 12 patients remained stable, according to RANO criteria. In progressed patients, T2/FLAIR volume and T2 FSE intensity increased over time to progression. The rate of T2/FLAIR volume growth was significantly greater in progressed than stable patients (mean = .42, -.07 mL/week, respectively). The rate of normalized T2/FLAIR signal intensity change was also significantly different in progressed than stable patients (mean = -.002, .0035 mL/week). There were no significant interactions with grade or molecular subgroup by WHO criteria. Additionally, 44% (17/39) of progressed patients by RANO criteria were characterized as having stable disease by radiologic assessment. CONCLUSION Tumor volumes remain the gold standard for response assessment in lower grade gliomas and there exists a need for standardized measurement as visual assessment for determination of tumor status is inconsistent.

The goal of this project is to identify serial advanced imaging markers that can reveal tumor progression in grade II and III gliomas prior to the to the onset of new contrast enhancement or the expansion of FLAIR abnormality. Nineteen lower grade glioma patients were identified with tumor progression during their participation in research on serial advanced MR imaging and spectroscopy at standard of care timepoints (2–6 months). The average number of scans acquired prior to progression was 7 (range = 12–3). The 3T MRI examination included T1-weighted IRSPGR, T2-weighted FSE, T2 FLAIR, and 6 directional diffusion weighted imaging. Maps of the apparent diffusion coefficient (ADC) and fractional anisotropy (FA) were calculated, and percentiles (10(th), 50(th), 90(th)) from histograms of normalized intensities. Serial images were aligned to the baseline image. Regions of Interest (ROIs) were defined for the entire T2 FLAIR lesion, and when present, the contrast-enhancing (CE) T1 lesion at each timepoint. In addition, we created a Progression ROI consisting of the region that was normal on the baseline T2 FLAIR but abnormal at the time of clinical progression, and a Baseline ROI consisting of the region that was abnormal at baseline and progression. Serial imaging parameters were acquired for each of these ROIs at every timepoint. We determined the imaging parameters with significant changes within the ROIs leading up to the time of progression using repeated-measures analyses of variance. Leading up to the time of progression, there were significant declines in median and 10% normalized ADC and 10% normalized FA within the Baseline T2 FLAIR abnormality ROI. Median normalized ADC also declined significantly in the Progression ROI, indicating changes in diffusion parameters prior to the onset of T2 FLAIR abnormality and the clinical determination of progression.

As a non-invasive method for probing in vivo metabolism, hyperpolarized carbon-13 metabolic imaging may enhance the diagnostic assessment of response-to-treatment in patients with glioma. Having demonstrated the feasibility of applying this technique to human glioma (Park et al. 2017), the current study seeks to evaluate the reproducibility of carbon data by measuring serial metabolic changes in white matter. In total, 4 radiologically stable patients (2 male, 2 female) who were treated with radiation and TMZ for heterogeneous diagnoses (GBM, grade III oligo, gliosarcoma, secondary GBM) received either 2 (n=1) or 3 (n=3) serial exams. Following injection of hyperpolarized [1-(13)C]pyruvate, dynamic echo-planar imaging was acquired with a 3s temporal resolution and 2–8 cm(3) spatial resolution. Dynamic data were summed together and the ratio of metabolically-produced lactate relative to pyruvate substrate (lac/pyr) was evaluated in segmented normal-appearing white matter (NAWM, >50% voxel). Median and SD of lac/pyr values in NAWM are reported for serial scans, along with %change per interval: P1-male [0.32 ± 0.14, 0.28 ± 0.11 (-12.5%), 0.33 ± 0.10 (+17.9%)]; P2-male [0.32 ± 0.10, 0.28 ± 0.10 (-12.5%)]; P3-female [0.37 ± 0.09, 0.45 ± 0.14 (+21.6%), 0.59 ± 0.19 (+31.1%)]; and P4-female [0.49 ± 0.14, 0.43 ± 0.12 (-12.2%), 0.44 ± 0.10 (2.3%)]. These results show an average absolute change of 15.7% in lac/pyr within NAWM over the entire population, with considerable overlap of the variance. Patient P3, who was treated with adjuvant agent pembrolizumab, also notably showed a serial increase in lac/pyr. When separating the population according to gender, females demonstrated a higher average lac/pyr value of 0.46, compared to 0.31 in males. In summary, our initial experience with serial hyperpolarized carbon data indicates that the range of lactate-to-pyruvate ratios is relatively consistent across patient scans, but may differ according to gender and adjuvant treatment. Future studies will evaluate changes in the T2 lesion relative to NAWM over the course of treatment.

Background: The etiologic basis for neurocognitive and neuropsychological deficits in cancer patients and cancer survivors is poorly understood. Because of ongoing improvements in therapy and advances in survival, the problem of cancer-associated cognitive dysfunction is increasingly significant. While there is agreement that T2-weighted imaging abnormalities on MRI correlate with neurocognitive deficits, there has been limited insight into the neurochemical abnormalities associated with cognitive dysfunction in brain tumor patients. We are testing the hypothesis that tumor metabolism directly impairs neurotransmitter pathways and cognitive function, independent of anatomic extent of the cancer. Methods: Our approach has been to focus on a type of brain tumor, CNS lymphoma, in which we are able to simultaneously monitor cognitive function, as assayed by repeat Mini-Mental Status Examinations (MMSE), tumor volume, as assessed by MRI, and dynamic changes in the tumor metabolic microenvironment, as characterized by quantitative measurement of tumor-associated metabolites in correlation with neurotransmitters that we hypothesize to be linked to normal cognitive function. Our initial study was to evaluate 14 subjects with CNS lymphoma that were treated on a phase I trial of the immunomodulatory agent lenalidomide. Volumetric analysis of CNS lymphomas was performed using Smartbrush Software (Brainlab) on pre-and post-therapy MRI's conducted at baseline and at monthly restaging. Metabolomic analysis of cerebrospinal fluid (CSF), using GC/MS, was conducted at baseline and at timepoints within 1 week of corresponding MRI. MMSE tests were conducted in all subjects at baseline and at corresponding monthly restaging examinations. Results: Of 20 CSF metabolites analyzed, including 12 neurotransmitters, elevated CSF lactate correlated most strongly with impaired neurocognitive function as measured by MMSE score. (P=2.5e-6; rho= -0.66). Patients with high lactate had lower relative CSF concentration of the inhibitory neurotransmitter GABA, and higher concentrations of the excitotoxic glutamate. Notably, we determined that CSF lactate concentration more significantly correlated with lower MMSE score than size of the brain tumor, as quantified by volumetric analysis of tumor T2 hyperintensity and lesional contrast-enhancing volume. Conclusions: To our knowledge, this is the first data linking cancer metabolism, neurotransmitter dysregulation, and neurocognitive deficits in a brain tumor patient population. We anticipate that elucidation of the mechanistic basis between tumor lactate metabolism, neurotransmitter imbalance, and neurocognitive deficits will provide potential opportunities for pharmacologic intervention to preserve neurologic function and potentially minimize cognitive and neuropsychological deficits in cancer patients. Citation Format: Lakshmi Subbaraj, Huimin Geng, Jigyasa Sharma, Marisa LaFontaine, James L. Rubenstein. Tumor metabolism and cognitive dysfunction in CNS lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4975.

The increasing interest in enhancing the RANO criteria by using quantitative assessments of changes in lesion size and image intensities has highlighted the need for rapid, easy-to-use tools that provide DICOM compatible outputs for evaluation of patients with glioma. To evaluate the performance of the SmartBrush software (Brainlab AG), which provides computer-assisted definitions of regions of interest (ROIs), a cohort of 20 patients with glioma (equal number having high and low grade and treated and un-treated) were scanned using a 3T whole-body MR system prior to surgical resection. The T2-weighted FLAIR, pre- and post-contrast T1-weighted gradient echo DICOM images were pushed from the scanner to an offline workstation where analysis of lesion volumes was performed using SmartBrush. Volumes of the T2Ls ranged from 7.9 to 110.2cm3 and the volumes of the CELs was 0.1 to 28.5 cm3 with 19/20 of the subjects having CELs and all 20 having T2Ls. the computer-assisted analysis was performed rapidly and efficiently, with the mean time for defining both lesions per subject was 5.77 (range 3.5 to 7.5) minutes. Prior analysis of ROIS with the SLICER package (www.slicer.org) took approximately 30 minutes/subject. SmartBrush provides lesion volumes and cross-sectional diameter as a PDF report, which can be stored in DICOM. The ROIs were also saved as DICOM objects and transferred to other packages for performing histogram analysis from ADC or other functional parameter maps. Ongoing studies that will be reported in this presentation are performing a similar analysis with multiple users in order to compare the relative intra- and inter-operator variations in terms of both the speed of analysis and the ROIs that are identified. Acknowledgements: The authors would like to acknowledge Rowena Thomson and Natalie Wright from Brainlab for helping to set up this study.