Brain tumor classification of virtual NMR voxels based on realistic blood vessel-induced spin dephasing using support vector machines.

Remodeling of tissue microvasculature commonly promotes neoplastic growth; however, there is no imaging modality in oncology yet that noninvasively quantifies microvascular changes in clinical routine. Although blood capillaries cannot be resolved in typical magnetic resonance imaging (MRI) measurements, their geometry and distribution influence the integral nuclear magnetic resonance (NMR) signal from each macroscopic MRI voxel. We have numerically simulated the expected transverse relaxation in NMR voxels with different dimensions based on the realistic microvasculature in healthy and tumor-bearing mouse brains (U87 and GL261 glioblastoma). The 3D capillary structure in entire, undissected brains was acquired using light sheet fluorescence microscopy to produce large datasets of the highly resolved cerebrovasculature. Using this data, we trained support vector machines to classify virtual NMR voxels with different dimensions based on the simulated spin dephasing accountable to field inhomogeneities caused by the underlying vasculature. In prediction tests with previously blinded virtual voxels from healthy brain tissue and GL261 tumors, stable classification accuracies above 95% were reached. Our results indicate that high classification accuracies can be stably attained with achievable training set sizes and that larger MRI voxels facilitated increasingly successful classifications, even with small training datasets. We were able to prove that, theoretically, the transverse relaxation process can be harnessed to learn endogenous contrasts for single voxel tissue type classifications on tailored MRI acquisitions. If translatable to experimental MRI, this may augment diagnostic imaging in oncology with automated voxel-by-voxel signal interpretation to detect vascular pathologies.

Structural Nerve Remodeling at 3-T MR Neurography Differs between Painful and Painless Diabetic Polyneuropathy in Type 1 or 2 Diabetes.

Background The pathophysiologic mechanisms underlying painful symptoms in diabetic polyneuropathy (DPN) are poorly understood. They may be associated with MRI characteristics, which have not yet been investigated. Purpose To investigate correlations between nerve structure, load and spatial distribution of nerve lesions, and pain in patients with DPN. Materials and Methods In this prospective single-center cross-sectional study, participants with type 1 or 2 diabetes volunteered between June 2015 and March 2018. Participants underwent 3-T MR neurography of the sciatic nerve with a T2-weighed fat-suppressed sequence, which was preceded by clinical and electrophysiologic tests. For group comparisons, analysis of variance or the Kruskal-Wallis test was performed depending on Gaussian or non-Gaussian distribution of data. Spearman correlation coefficients were calculated for correlation analysis. Results A total of 131 participants (mean age, 62 years ± 11 [standard deviation]; 82 men) with either type 1 (n = 45) or type 2 (n = 86) diabetes were evaluated with painful (n = 64), painless (n = 37), or no (n = 30) DPN. Participants who had painful diabetic neuropathy had a higher percentage of nerve lesions in the full nerve volume (15.2% ± 1.6) than did participants with nonpainful DPN (10.4% ± 1.7, P = .03) or no DPN (8.3% ± 1.7; P < .001). The amount and extension of T2-weighted hyperintense nerve lesions correlated positively with the neuropathy disability score (r = 0.37; 95% confidence interval [CI]: 0.21, 0.52; r = 0.37; 95% CI: 0.20, 0.52, respectively) and the neuropathy symptom score (r = 0.41; 95% CI: 0.25, 0.55; r = 0.34; 95% CI: 0.17, 0.49, respectively). Negative correlations were found for the tibial nerve conduction velocity (r = -0.23; 95% CI: -0.44, -0.01; r = -0.37; 95% CI: -0.55, -0.15, respectively). The cross-sectional area of the nerve was positively correlated with the neuropathy disability score (r = 0.23; 95% CI: 0.03, 0.36). Negative correlations were found for the tibial nerve conduction velocity (r = -0.24; 95% CI: -0.45, -0.01). Conclusion The amount and extension of T2-weighted hyperintense fascicular nerve lesions were greater in patients with painful diabetic neuropathy than in those with painless diabetic neuropathy. These results suggest that proximal fascicular damage is associated with the evolution of painful sensory symptoms in diabetic polyneuropathy. © RSNA, 2019 Online supplemental material is available for this article.

Gibbs point field model quantifies disorder in microvasculature of U87-glioblastoma.

Microvascular proliferation in glioblastoma multiforme is a biological key mechanism to facilitate tumor growth and infiltration and a main target for treatment interventions. The vascular architecture can be obtained by Single Plane Illumination Microscopy (SPIM) to evaluate vascular heterogeneity in tumorous tissue. We make use of the Gibbs point field model to quantify the order of regularity in capillary distributions found in the U87 glioblastoma model in a murine model and to compare tumorous and healthy brain tissue. A single model parameter Γ was assigned that is linked to tissue-specific vascular topology through Monte-Carlo simulations. Distributions of the model parameter Γ differ significantly between glioblastoma tissue with mean 〈ΓG〉=2.1±0.4, as compared to healthy brain tissue with mean 〈ΓH〉=4.9±0.4, suggesting that the average Γ-value allows for tissue differentiation. These results may be used for diagnostic magnetic resonance imaging, where it has been shown recently that Γ is linked to tissue-inherent relaxation parameters.

A PRDX1-p38α heterodimer amplifies MET-driven invasion of IDH-wildtype and IDH-mutant gliomas.

The Peroxiredoxin 1 (PRDX1) gene maps to chromosome arm 1p and is hemizygously deleted and epigenetically silenced in isocitrate dehydrogenase 1 or 2 (IDH)-mutant and 1p/19q-codeleted oligodendroglial tumors. In contrast, IDH-wildtype astrocytic gliomas including glioblastomas mostly lack epigenetic silencing and express PRDX1 protein. In our study, we investigated how PRDX1 contributes to the infiltrative growth of IDH-wildtype gliomas. Focusing on p38α-dependent pathways, we analyzed clinical data from 133 patients of the NOA-04 trial cohort to look for differences in the gene expression profiles of gliomas with wildtype or mutant IDH. Biochemical interaction studies as well as in vitro and ex vivo migration studies were used to establish a biological role of PRDX1 in maintaining pathway activity. Whole-brain high-resolution ultramicroscopy and survival analyses of pre-clinical mouse models for IDH-wildtype gliomas were then used for in vivo confirmation. Based on clinical data, we found that the absence of PRDX1 is associated with changes in the expression of MET/HGF signaling components. PRDX1 forms a heterodimer with p38α mitogen-activated protein kinase 14 (MAPK14), stabilizing phospho-p38α in glioma cells. This process amplifies hepatocyte growth factor (HGF)-mediated signaling and stimulates actin cytoskeleton dynamics that promote glioma cell migration. Whole-brain high-resolution ultramicroscopy confirms these findings, indicating that PRDX1 promotes glioma brain invasion in vivo. Finally, reduced expression of PRDX1 increased survival in mouse glioma models. Thus, our preclinical findings suggest that PRDX1 expression levels may serve as a molecular marker for patients who could benefit from targeted inhibition of MET/HGF signaling.

Vessel radius mapping in an extended model of transverse relaxation.

OBJECTIVES: Spin dephasing of the local magnetization in blood vessel networks can be described in the static dephasing regime (where diffusion effects may be ignored) by the established model of Yablonskiy and Haacke. However, for small capillary radii, diffusion phenomena for spin-bearing particles are not negligible. MATERIAL AND METHODS: In this work, we include diffusion effects for a set of randomly distributed capillaries and provide analytical expressions for the transverse relaxation times T2* and T2 in the strong collision approximation and the Gaussian approximation that relate MR signal properties with microstructural parameters such as the mean local capillary radius. RESULTS: Theoretical results are numerically validated with random walk simulations and are used to calculate capillary radius distribution maps for glioblastoma mouse brains at 9.4 T. For representative tumor regions, the capillary maps reveal a relative increase of mean radius for tumor tissue towards healthy brain tissue of [Formula: see text] (p < 0.001). CONCLUSION: The presented method may be used to quantify angiogenesis or the effects of antiangiogenic therapy in tumors whose growth is associated with significant microvascular changes.

Advanced Computational Methods to Evaluate Vascular Heterogeneity in Tumor Tissue Based on Single Plane Illumination Microscopy.

During tumor growth, the complex composition of vasculature is prone to dynamic changes due to mechanic and biochemical challenges. Perivascular invasion of tumor cells to co-opt existing vasculature, but also formation of de-novo vasculature and other effects on the vascular network, may lead to altered geometric vessel properties as well as changes in vascular network topology, which is defined by vascular multifurcations and connections between vessel segments. The intricate organization and heterogeneity of the vascular network can be analyzed with advanced computational methods to uncover vascular network signatures that may allow differentiating between pathological and physiological vessel regions. Herein, we present a protocol to evaluate vascular heterogeneity in whole vascular networks, using morphological and topological measures. The protocol was developed for single plane illumination microscopy images of mice brain vasculature but can be applied to any vascular network.

Large-scale characterization of the microvascular geometry in development and disease by tissue clearing and quantitative ultramicroscopy.

Three-dimensional assessment of optically cleared, entire organs and organisms has recently become possible by tissue clearing and selective plane illumination microscopy ("ultramicroscopy"). Resulting datasets can be highly complex, encompass over a thousand images with millions of objects and data of several gigabytes per acquisition. This constitutes a major challenge for quantitative analysis. We have developed post-processing tools to quantify millions of microvessels and their distribution in three-dimensional datasets from ultramicroscopy and demonstrate the capabilities of our pipeline within entire mouse brains and embryos. Using our developed acquisition, segmentation, and analysis platform, we quantify physiological vascular networks in development and the healthy brain. We compare various geometric vessel parameters (e.g. vessel density, radius, tortuosity) in the embryonic spinal cord and brain as well as in different brain regions (basal ganglia, corpus callosum, cortex). White matter tract structures (corpus callosum, spinal cord) showed lower microvascular branch densities and longer vessel branch length compared to grey matter (cortex, basal ganglia). Furthermore, we assess tumor neoangiogenesis in a mouse glioma model to compare tumor core and tumor border. The developed methodology allows rapid quantification of three-dimensional datasets by semi-automated segmentation of fluorescently labeled objects with conventional computer hardware. Our approach can aid preclinical investigations and paves the way towards "quantitative ultramicroscopy".

Diffusion Tensor Imaging of the Sciatic Nerve as a Surrogate Marker for Nerve Functionality of the Upper and Lower Limb in Patients With Diabetes and Prediabetes.

BACKGROUND: Nerve damage in diabetic neuropathy (DN) is assumed to begin in the distal legs with a subsequent progression to hands and arms at later stages. In contrast, recent studies have found that lower limb nerve lesions in DN predominate at the proximal sciatic nerve and that, in the upper limb, nerve functions can be impaired at early stages of DN. MATERIALS AND METHODS: In this prospective, single-center cross-sectional study, participants underwent diffusion-weighted 3 Tesla magnetic resonance neurography in order to calculate the sciatic nerve's fractional anisotropy (FA), a surrogate parameter for structural nerve integrity. Results were correlated with clinical and electrophysiological assessments of the lower limb and an examination of hand function derived from the Purdue Pegboard Test. RESULTS: Overall, 71 patients with diabetes, 11 patients with prediabetes and 25 age-matched control subjects took part in this study. In patients with diabetes, the sciatic nerve's FA showed positive correlations with tibial and peroneal nerve conduction velocities (r = 0.62; p < 0.001 and r = 0.56; p < 0.001, respectively), and tibial and peroneal nerve compound motor action potentials (r = 0.62; p < 0.001 and r = 0.63; p < 0.001, respectively). Moreover, the sciatic nerve's FA was correlated with the Pegboard Test results in patients with diabetes (r = 0.52; p < 0.001), prediabetes (r = 0.76; p < 0.001) and in controls (r = 0.79; p = 0.007). CONCLUSION: This study is the first to show that the sciatic nerve's FA is a surrogate marker for functional and electrophysiological parameters of both upper and lower limbs in patients with diabetes and prediabetes, suggesting that nerve damage in these patients is not restricted to the level of the symptomatic limbs but rather affects the entire peripheral nervous system.

Troponin T Parallels Structural Nerve Damage in Type 2 Diabetes: A Cross-sectional Study Using Magnetic Resonance Neurography.

Clinical studies have suggested that changes in peripheral nerve microcirculation may contribute to nerve damage in diabetic polyneuropathy (DN). High-sensitivity troponin T (hsTNT) assays have been recently shown to provide predictive values for both cardiac and peripheral microangiopathy in type 2 diabetes (T2D). This study investigated the association of sciatic nerve structural damage in 3 Tesla (3T) magnetic resonance neurography (MRN) with hsTNT and N-terminal pro-brain natriuretic peptide serum levels in patients with T2D. MRN at 3T was performed in 51 patients with T2D (23 without DN, 28 with DN) and 10 control subjects without diabetes. The sciatic nerve's fractional anisotropy (FA), a marker of structural nerve integrity, was correlated with clinical, electrophysiological, and serological data. In patients with T2D, hsTNT showed a negative correlation with the sciatic nerve's FA (r = -0.52, P < 0.001), with a closer correlation in DN patients (r = -0.66, P < 0.001). hsTNT further correlated positively with the neuropathy disability score (r = 0.39, P = 0.005). Negative correlations were found with sural nerve conduction velocities (NCVs) (r = -0.65, P < 0.001) and tibial NCVs (r = -0.44, P = 0.002) and amplitudes (r = -0.53, P < 0.001). This study is the first to show that hsTNT is a potential indicator for structural nerve damage in T2D. Our results indirectly support the hypothesis that microangiopathy contributes to structural nerve damage in T2D.

Diabetic Polyneuropathy Is Associated With Pathomorphological Changes in Human Dorsal Root Ganglia: A Study Using 3T MR Neurography.

Diabetic neuropathy (DPN) is one of the most severe and yet most poorly understood complications of diabetes mellitus. In vivo imaging of dorsal root ganglia (DRG), a key structure for the understanding of DPN, has been restricted to animal studies. These have shown a correlation of decreased DRG volume with neuropathic symptom severity. Our objective was to investigate correlations of DRG morphology and signal characteristics at 3 Tesla (3T) magnetic resonance neurography (MRN) with clinical and serological data in diabetic patients with and without DPN. In this cross-sectional study, participants underwent 3T MRN of both L5 DRG using an isotropic 3D T2-weighted, fat-suppressed sequence with subsequent segmentation of DRG volume and analysis of normalized signal properties. Overall, 55 diabetes patients (66 ± 9 years; 32 men; 30 with DPN) took part in this study. DRG volume was smaller in patients with severe DPN when compared to patients with mild or moderate DPN (134.7 ± 21.86 vs 170.1 ± 49.22; p = 0.040). In DPN patients, DRG volume was negatively correlated with the neuropathy disability score (r = -0.43; 95%CI = -0.66 to -0.14; p = 0.02), a measure of neuropathy severity. DRG volume showed negative correlations with triglycerides (r = -0.40; 95%CI = -0.57 to -0.19; p = 0.006), and LDL cholesterol (r = -0.33; 95%CI = -0.51 to -0.11; p = 0.04). There was a strong positive correlation of normalized MR signal intensity (SI) with the neuropathy symptom score in the subgroup of patients with painful DPN (r = 0.80; 95%CI = 0.46 to 0.93; p = 0.005). DRG SI was positively correlated with HbA1c levels (r = 0.30; 95%CI = 0.09 to 0.50; p = 0.03) and the triglyceride/HDL ratio (r = 0.40; 95%CI = 0.19 to 0.57; p = 0.007). In this first in vivo study, we found DRG morphological degeneration and signal increase in correlation with neuropathy severity. This elucidates the potential importance of MR-based DRG assessments in studying structural and functional changes in DPN.

Glioblastoma multiforme restructures the topological connectivity of cerebrovascular networks.

Glioblastoma multiforme alters healthy tissue vasculature by inducing angiogenesis and vascular remodeling. To fully comprehend the structural and functional properties of the resulting vascular network, it needs to be studied collectively by considering both geometric and topological properties. Utilizing Single Plane Illumination Microscopy (SPIM), the detailed capillary structure in entire healthy and tumor-bearing mouse brains could be resolved in three dimensions. At the scale of the smallest capillaries, the entire vascular systems of bulk U87- and GL261-glioblastoma xenografts, their respective cores, and healthy brain hemispheres were modeled as complex networks and quantified with fundamental topological measures. All individual vessel segments were further quantified geometrically and modular clusters were uncovered and characterized as meta-networks, facilitating an analysis of large-scale connectivity. An inclusive comparison of large tissue sections revealed that geometric properties of individual vessels were altered in glioblastoma in a relatively subtle way, with high intra- and inter-tumor heterogeneity, compared to the impact on the vessel connectivity. A network topology analysis revealed a clear decomposition of large modular structures and hierarchical network organization, while preserving most fundamental topological classifications, in both tumor models with distinct growth patterns. These results augment our understanding of cerebrovascular networks and offer a topological assessment of glioma-induced vascular remodeling. The findings may help understand the emergence of hypoxia and necrosis, and prove valuable for therapeutic interventions such as radiation or antiangiogenic therapy.

Dual-contrast pCASL using simultaneous gradient-echo/spin-echo multiband EPI.

A 2D gradient-echo EPI is commonly employed for arterial spin labeling (ASL) readout to achieve fast whole brain coverage measurements. However, such a readout suffers from susceptibility artifacts induced by magnetic field inhomogeneities. To reduce these susceptibility effects, single-shot spin-echo EPI was proposed to be used for acquisitions in continuous ASL (CASL). To minimize functional and physiological variations, a gradient-echo (GE)/spin-echo (SE) dual-echo EPI readout of the CASL sequence is needed for a comparison between GE- and SE-based determination of cerebral blood flow (CBF). In this study, we employed a simultaneous GE/SE multiband EPI as the readout of a pseudo-CASL (pCASL) sequence. Motor cortex activations derived from a finger-tapping task and functional networks from resting state fMRI were compared for both GE and SE contrasts. Direct comparison of SE and GE contrasts revealed that GE ASL provides an improved sensitivity of functional activity in finger-tapping and in resting-state imaging. SE ASL, on the other hand, suffered less from susceptibility artifacts induced by magnetic field inhomogeneities and pulsatile flow artifacts.

Vessel architecture imaging using multiband gradient-echo/spin-echo EPI.

OBJECTIVES: To apply the MB (multiband) excitation and blipped-CAIPI (blipped-controlled aliasing in parallel imaging) techniques in a spin and gradient-echo (SAGE) EPI sequence to improve the slice coverage for vessel architecture imaging (VAI). MATERIALS AND METHODS: Both MB excitation and blipped-CAIPI with in-plane parallel imaging were incorporated into a gradient-echo (GE)/spin-echo (SE) EPI sequence for simultaneous tracking of the dynamic MR signal changes in both GE and SE contrasts after the injection of contrast agent. MB and singleband (SB) excitation were compared using a 20-channel head coil at 3 Tesla, and high-resolution MB VAI could be performed in 32 glioma patients. RESULTS: Whole-brain covered high resolution VAI can be achieved after applying multiband excitation with a factor of 2 and in-plane parallel imaging with a factor of 3. The quality of the images resulting from MB acceleration was comparable to those from the SB method: images were reconstructed without any loss of spatial resolution or severe distortions. In addition, MB and SB signal-to-noise ratios (SNR) were similar. A relative low g-factor induced from the MB acceleration method was achieved after using a blipped-CAIPI technique (1.35 for GE and 1.33 for SE imaging). Performing quantitative VAI, we found that, among all VAI parametric maps, microvessel type indicator (MTI), distance map (I) and vascular-induced bolus peak-time shift (VIPS) were highly correlated. Likewise, VAI parametric maps of slope, slope length and short axis were highly correlated. CONCLUSIONS: Multiband accelerated SAGE successfully doubles the number of readout slices in the same measurement time when compared to conventional readout sequences. The corresponding VAI parametric maps provide insights into the complexity and heterogeneity of vascular changes in glioma.

Association of Serum Cholesterol Levels With Peripheral Nerve Damage in Patients With Type 2 Diabetes.

Importance: Lowering serum cholesterol levels is a well-established treatment for dyslipidemia in patients with type 2 diabetes (T2D). However, nerve lesions in patients with T2D increase with lower serum cholesterol levels, suggesting that lowering serum cholesterol levels is associated with diabetic polyneuropathy (DPN) in patients with T2D. Objective: To investigate whether there is an association between serum cholesterol levels and peripheral nerve lesions in patients with T2D with and without DPN. Design, Setting, and Participants: This single-center, cross-sectional, prospective cohort study was performed from June 1, 2015, to March 31, 2018. Observers were blinded to clinical data. A total of 256 participants were approached, of whom 156 were excluded. A total of 100 participants consented to undergo magnetic resonance neurography of the right leg at the Department of Neuroradiology and clinical, serologic, and electrophysiologic assessment at the Department of Endocrinology, Heidelberg University Hospital, Heidelberg, Germany. Exposures: Quantification of the nerve's diameter and lipid equivalent lesion (LEL) load with a subsequent analysis of all acquired clinical and serologic data with use of 3.0-T magnetic resonance neurography of the right leg with 3-dimensional reconstruction of the sciatic nerve. Main Outcomes and Measures: The primary outcome was lesion load and extension. Secondary outcomes were clinical, serologic, and electrophysiologic findings. Results: A total of 100 participants with T2D (mean [SD] age, 64.6 [0.9] years; 68 [68.0%] male) participated in the study. The LEL load correlated positively with the nerve's mean cross-sectional area (r = 0.44; P < .001) and the maximum length of a lesion (r = 0.71; P < .001). The LEL load was negatively associated with total serum cholesterol level (r = -0.41; P < .001), high-density lipoprotein cholesterol level (r = -0.30; P = .006), low-density lipoprotein cholesterol level (r = -0.33; P = .003), nerve conduction velocities of the tibial (r = -0.33; P = .01) and peroneal (r = -0.51; P < .001) nerves, and nerve conduction amplitudes of the tibial (r = -0.31; P = .02) and peroneal (r = -0.28; P = .03) nerves. Conclusions and Relevance: The findings suggest that lowering serum cholesterol levels in patients with T2D and DPN is associated with a higher amount of nerve lesions and declining nerve conduction velocities and amplitudes. These findings may be relevant to emerging therapies that promote an aggressive lowering of serum cholesterol levels in patients with T2D.

Correlated MRI and Ultramicroscopy (MR-UM) of Brain Tumors Reveals Vast Heterogeneity of Tumor Infiltration and Neoangiogenesis in Preclinical Models and Human Disease.

Diffuse tumor infiltration into the adjacent parenchyma is an effective dissemination mechanism of brain tumors. We have previously developed correlated high field magnetic resonance imaging and ultramicroscopy (MR-UM) to study neonangiogenesis in a glioma model. In the present study we used MR-UM to investigate tumor infiltration and neoangiogenesis in a translational approach. We compare infiltration and neoangiogenesis patterns in four brain tumor models and the human disease: whereas the U87MG glioma model resembles brain metastases with an encapsulated growth and extensive neoangiogenesis, S24 experimental gliomas mimic IDH1 wildtype glioblastomas, exhibiting infiltration into the adjacent parenchyma and along white matter tracts to the contralateral hemisphere. MR-UM resolves tumor infiltration and neoangiogenesis longitudinally based on the expression of fluorescent proteins, intravital dyes or endogenous contrasts. Our study demonstrates the huge morphological diversity of brain tumor models regarding their infiltrative and neoangiogenic capacities and further establishes MR-UM as a platform for translational neuroimaging.