Search

BackgroundAccurate and repeatable measurement of high-grade glioma (HGG) enhancing (Enh.) and T2/FLAIR hyperintensity/edema (Ed.) is required for monitoring treatment response. 3D measurements can be used to inform the modified Response Assessment in Neuro-oncology criteria. We aim to develop an HGG volumetric measurement and visualization AI algorithm that is generalizable and repeatable.MethodsA single 3D-Convoluted Neural Network, NS-HGlio, to analyze HGG on MRIs using 5-fold cross validation was developed using retrospective (557 MRIs), multicentre (38 sites) and multivendor (32 scanners) dataset divided into training (70%), validation (20%), and testing (10%). Six neuroradiologists created the ground truth (GT). Additional Internal validation (IV, three institutions) using 70 MRIs, and External validation (EV, single institution) using 40 MRIs through measuring the Dice Similarity Coefficient (DSC) of Enh., Ed. ,and Enh. + Ed. (WholeLesion/WL) tumor tissue and repeatability testing on 14 subjects from the TCIA MGH-QIN-GBM dataset using volume correlations between timepoints were performed.ResultsIV Preoperative median DSC Enh. 0.89 (SD 0.11), Ed. 0.88 (0.28), WL 0.88 (0.11). EV Preoperative median DSC Enh. 0.82 (0.09), Ed. 0.83 (0.11), WL 0.86 (0.06). IV Postoperative median DSC Enh. 0.77 (SD 0.20), Ed 0.78. (SD 0.09), WL 0.78 (SD 0.11). EV Postoperative median DSC Enh. 0.75 (0.21), Ed 0.74 (0.12), WL 0.79 (0.07). Repeatability testing; Intraclass Correlation Coefficient of 0.95 Enh. and 0.92 Ed.ConclusionNS-HGlio is accurate, repeatable, and generalizable. The output can be used for visualization, documentation, treatment response monitoring, radiation planning, intra-operative targeting, and estimation of Residual Tumor Volume among others.

Abstract Background The intravenous co-infusion of labradimil, a metabolically stable bradykinin B2 receptor agonist, has been shown to temporarily enhance the transvascular delivery of small chemotherapy drugs, such as carboplatin, across the blood-brain tumor barrier. It has been thought that the primary mechanism by which labradimil does so is by acting selectively on tumor microvasculature to increase the local transvascular flow rate across the blood-brain tumor barrier. This mechanism of action does not explain why, in the clinical setting, carboplatin dosing based on patient renal function over-estimates the carboplatin dose required for target carboplatin exposure. In this study we investigated the systemic actions of labradimil, as well as other bradykinin B2 receptor agonists with a range of metabolic stabilities, in context of the local actions of the respective B2 receptor agonists on the blood-brain tumor barrier of rodent malignant gliomas. Methods Using dynamic contrast-enhanced MRI, the pharmacokinetics of gadolinium-diethyltriaminepentaacetic acid (Gd-DTPA), a small MRI contrast agent, were imaged in rodents bearing orthotopic RG-2 malignant gliomas. Baseline blood and brain tumor tissue pharmacokinetics were imaged with the 1st bolus of Gd-DTPA over the first hour, and then re-imaged with a 2nd bolus of Gd-DTPA over the second hour, during which normal saline or a bradykinin B2 receptor agonist was infused intravenously for 15 minutes. Changes in mean arterial blood pressure were recorded. Imaging data was analyzed using both qualitative and quantitative methods. Results The decrease in systemic blood pressure correlated with the known metabolic stability of the bradykinin B2 receptor agonist infused. Metabolically stable bradykinin B2 agonists, methionine-lysine-bradykinin and labradimil, had differential effects on the transvascular flow rate of Gd-DTPA across the blood-brain tumor barrier. Both methionine-lysine-bradykinin and labradimil increased the blood half-life of Gd-DTPA sufficiently enough to increase significantly the tumor tissue Gd-DTPA area under the time-concentration curve. Conclusion Metabolically stable bradykinin B2 receptor agonists, methionine-lysine-bradykinin and labradimil, enhance the transvascular delivery of small chemotherapy drugs across the BBTB of malignant gliomas by increasing the blood half-life of the co-infused drug. The selectivity of the increase in drug delivery into the malignant glioma tissue, but not into normal brain tissue or skeletal muscle tissue, is due to the inherent porous nature of the BBTB of malignant glioma microvasculature.

Abstract Background Effective transvascular delivery of nanoparticle-based chemotherapeutics across the blood-brain tumor barrier of malignant gliomas remains a challenge. This is due to our limited understanding of nanoparticle properties in relation to the physiologic size of pores within the blood-brain tumor barrier. Polyamidoamine dendrimers are particularly small multigenerational nanoparticles with uniform sizes within each generation. Dendrimer sizes increase by only 1 to 2 nm with each successive generation. Using functionalized polyamidoamine dendrimer generations 1 through 8, we investigated how nanoparticle size influences particle accumulation within malignant glioma cells. Methods Magnetic resonance and fluorescence imaging probes were conjugated to the dendrimer terminal amines. Functionalized dendrimers were administered intravenously to rodents with orthotopically grown malignant gliomas. Transvascular transport and accumulation of the nanoparticles in brain tumor tissue was measured in vivo with dynamic contrast-enhanced magnetic resonance imaging. Localization of the nanoparticles within glioma cells was confirmed ex vivo with fluorescence imaging. Results We found that the intravenously administered functionalized dendrimers less than approximately 11.7 to 11.9 nm in diameter were able to traverse pores of the blood-brain tumor barrier of RG-2 malignant gliomas, while larger ones could not. Of the permeable functionalized dendrimer generations, those that possessed long blood half-lives could accumulate within glioma cells. Conclusion The therapeutically relevant upper limit of blood-brain tumor barrier pore size is approximately 11.7 to 11.9 nm. Therefore, effective transvascular drug delivery into malignant glioma cells can be accomplished by using nanoparticles that are smaller than 11.7 to 11.9 nm in diameter and possess long blood half-lives.

Background: Glioblastoma multiforme (GBM) is a malignant transformation of glial tissue, which presents as intradural, intraaxial lesions with heterogenous contrast enhancement and mass effect. Intratumoral hemorrhage is a common finding in GBM although it is frequently asymptomatic. Massive, symptomatic, intratumoral hemorrhage is uncommon and poses a diagnostic challenge. Case Description: Here we discuss a case of GBM, which initially presented as massive, symptomatic intracerebral hemorrhage with underlying mass. Due to size of the hemorrhage and poor neurological status the patient was taken to the operating room for evacuation of this hematoma. On pathology, the mass was found to be epithelioid glioblastoma. Conclusion: Identification and diagnosis of GBM is generally straightforward. In certain circumstances, the presentation of GBM can vary from the routine. The above case demonstrates how pitfalls in diagnosis can be avoided in order to initiate appropriate therapy. [ABSTRACT FROM AUTHOR]

Gadolinium chelates, which are currently approved for clinical MRI use, provide relaxivities well below their theoretical limit, and they also lack tissue specificity. Recently, the geometrical confinement of Gd3+-based contrast agents (CAs) within porous structures has been proposed as a novel, alternative strategy to improve relaxivity without chemical modification of the CA. Here, we have characterized and optimized the performance of MRI nanoconstructs obtained by loading [Gd(DTPA)(H2O)]2− (Magnevist®) into the pores of injectable mesoporous silicon particles. Nanoconstructs with three different pore sizes were studied, and at 60 MHz, they exhibited longitudinal relaxivities of ~24 m m−1 s−1 for 5-10 nm pores and ~10 m m−1 s−1 for 30 - 40 nm pores. No enhancement in relaxivity was observed for larger pores sizes. Using an outer-sphere compound, [GdTTHA]3−, and mathematical modeling, it was demonstrated that the relaxivity enhancement is due to the increase in rotational correlation times (CA adsorbed on the pore walls) and diffusion correlation times (reduced mobility of the water molecules), as the pore sizes decreases. It was also observed that extensive CA adsorption on the outer surface of the silicon particles negates the advantages offered by nanoscale confinement. Upon incubation with HeLa cells, the nanoconstructs did not demonstrate significant cytotoxicity for up to 3 days post incubation, at different particle/cell ratios. In addition, the nanoconstructs showed complete degradation after 24 h of continuous agitation in phosphate-buffered saline. These data support and confirm the hypothesis that the geometrical confinement of Gd3+-chelate compounds into porous structures offers MRI nanoconstructs with enhanced relaxivity (up to 6 times for [Gd(DTPA)(H2O)]2−, and 4 times for [GdTTHA]3−) and, potentially, improved stability, reduced toxicity and tissue specificity. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

The translation of AI-generated brain metastases (BM) segmentation into clinical practice relies heavily on diverse, high-quality annotated medical imaging datasets. The BraTS-METS 2023 challenge has gained momentum for testing and benchmarking algorithms using rigorously annotated internationally compiled real-world datasets. This study presents the results of the segmentation challenge and characterizes the challenging cases that impacted the performance of the winning algorithms. Untreated brain metastases on standard anatomic MRI sequences (T1, T2, FLAIR, T1PG) from eight contributed international datasets were annotated in stepwise method: published UNET algorithms, student, neuroradiologist, final approver neuroradiologist. Segmentations were ranked based on lesion-wise Dice and Hausdorff distance (HD95) scores. False positives (FP) and false negatives (FN) were rigorously penalized, receiving a score of 0 for Dice and a fixed penalty of 374 for HD95. Eight datasets comprising 1303 studies were annotated, with 402 studies (3076 lesions) released on Synapse as publicly available datasets to challenge competitors. Additionally, 31 studies (139 lesions) were held out for validation, and 59 studies (218 lesions) were used for testing. Segmentation accuracy was measured as rank across subjects, with the winning team achieving a LesionWise mean score of 7.9. Common errors among the leading teams included false negatives for small lesions and misregistration of masks in space.The BraTS-METS 2023 challenge successfully curated well-annotated, diverse datasets and identified common errors, facilitating the translation of BM segmentation across varied clinical environments and providing personalized volumetric reports to patients undergoing BM treatment.

Benefits of listening to music with emotional attachment while recovering from a cerebral ischemic event have been reported. To develop a better understanding of the effects of music listening on the human brain, an algorithm for the graph-theoretical analysis of functional magnetic resonance imaging (fMRI) data was developed. From BOLD data of two paradigms (block-design, first piece: music without emotional attachment, additional visual guidance by a moving cursor in the score sheet; second piece: music with emotional attachment), network graphs were constructed with correlations between signal time courses as edge weights. Functional subunits in these graphs were identified with the MCODE clustering algorithm and mapped back into anatomical space using AFNI. Emotional centers including the right amygdala and bilateral insula were activated by the second piece (emotional attachment) but not by the first piece. Network clustering analysis revealed two separate networks of small-world property corresponding to task-oriented and resting state conditions, respectively. Functional subunits with highest interactions were bilateral precuneus for the first piece and left middle frontal gyrus and right amygdala, bilateral insula, left middle temporal gyrus for the second piece. Our results indicate that fMRI in connection with graph theoretical network analysis is capable of identifying and differentiating functional subunits in the human brain when listening to music with and without emotional attachment.

Graph-theoretical analysis algorithms have been used for identifying subnetworks in the human brain during the Default Mode State. Here, these methods are expanded to determine the interaction of the sensory and the motor subnetworks during the performance of an approach-avoidance paradigm utilizing the correlation strength between the signal intensity time courses as measure of synchrony. From functional magnetic resonance imaging (fMRI) data of 9 healthy volunteers, two signal time courses, one from the primary visual cortex (sensory input) and one from the motor cortex (motor output) were identified and a correlation difference map was calculated. Graph networks were created from this map and visualized with spring-embedded layouts and 3D layouts in the original anatomical space. Functional clusters in these networks were identified with the MCODE clustering algorithm. Interactions between the sensory sub-network and the motor sub-network were quantified through the interaction strengths of these clusters. The percentages of interactions involving the visual cortex ranged from 85 % to 18 % and the motor cortex ranged from 40 % to 9 %. Other regions with high interactions were: frontal cortex (19 ± 18 %), insula (17 ± 22 %), cuneus (16 ± 15 %), supplementary motor area (SMA, 11 ± 18 %) and subcortical regions (11 ± 10 %). Interactions between motor cortex, SMA and visual cortex accounted for 12 %, between visual cortex and cuneus for 8 % and between motor cortex, SMA and cuneus for 6 % of all interactions. These quantitative findings are supported by the visual impressions from the 2D and 3D network layouts.

The integrity of functional brain networks inpatients (n=12) diagnosed with traumatic brain injury (TBI) was compared to age-matched subjects (n=12) with orthopedic injury (OI) during a working memory task. A graph-theoretical analysis algorithm was developed and integrated into the AFNI software. Functional networks with correlations between time courses as edge-weights were automatically created and their integrity was quantified by determining the statistical significance of the following network parameters: diameter, density, clustering coefficient, average path length, two largest eigenvalues, spectral density, and minimum eccentricity. Network graphs using a spring-embedded layout (Cytoscape) and a 3D layout integrated into the anatomical space (Paraview) were created. Functional images were composed by color-coding the degree of each voxel (network node) and transformed into Talairach space. Using the AFNI Talairach atlas, degrees of distinct brain regions were quantified. Reduced averaged BOLD responses were found for the TBI group with a higher network integrity potentially as a compensatory mechanism. Regions of high functional connectivity varied in between groups with largest differences in the cerebellum, the temporal lobes and deep brain structures including the lentiform nucleus, caudate and thalamus.