Distinguishing recurrent brain tumor from treatment effects, including late time-to-onset radiation necrosis (RN), presents an on-going challenge in post-treatment imaging of neuro-oncology patients. Experiments were performed in a novel mouse model that recapitulates the relevant clinical histologic features of recurrent glioblastoma growing in a RN environment, the mixed tumor/RN model. The goal of this work was to apply single-voxel deuterium (2H) magnetic resonance spectroscopy (MRS), in concert with administration of deuterated glucose, to determine if the metabolic signature of aerobic glycolysis (Warburg effect: glucose → lactate in the presence of O2), a distinguishing characteristic of proliferating tumor, provides a quantitative readout of the tumor fraction (percent) in a mixed tumor/RN lesion. 2H MRS employed the SPin-ECho full-Intensity Acquired Localized (SPECIAL) MRS pulse sequence and outer volume suppression at 11.74 T. For each subject, a single 2H MRS voxel was placed over the mixed lesion as defined by contrast enhanced (CE) 1H T1-weighted MRI. Following intravenous administration of [6,6-2H2]glucose (Glc), 2H MRS monitored the glycolytic conversion to [3,3-2H2]lactate (Lac) and glutamate + glutamine (Glu + Gln = Glx). Based on previous work, the tumor fraction of the mixed lesion was quantified as the ratio of tumor volume, defined by 1H magnetization transfer experiments, vs. the total mixed-lesion volume. Metabolite 2H MR spectral-amplitude values were converted to metabolite concentrations using the natural-abundance semi-heavy water (1HO2H) resonance as an internal concentration standard. The 2H MR-determined [Lac] / [Glx] ratio was strongly linearly correlated with tumor fraction in the mixed lesion (n = 9), Pearson’s r = 0.87, and 77

PURPOSE:Tail-vein catheterization and subsequent in-magnet infusion is a common route of administration of deuterium (2 H)-labeled substrates in small-animal deuterium (D) MR studies. With mice, because of the tail vein's small diameter, this procedure is challenging. It requires considerable personnel training and practice, is prone to failure, and may preclude serial studies. Motivated by the need for an alternative, the time courses for common small-molecule deuterated substrates and downstream metabolites in brain following subcutaneous infusion were determined in mice and are presented herein. METHODS:Three 2 H-labeled substrates-[6,6-2 H2 ]glucose, [2 H3 ]acetate, and [3,4,4,4-2 H4 ]beta-hydroxybutyrate-and 2 H2 O were administered to mice in-magnet via subcutaneous catheter. Brain time courses of the substrates and downstream metabolites (and semi-heavy water) were determined via single-voxel DMRS. RESULTS:Subcutaneous catheter placement and substrate administration was readily accomplished with limited personnel training. Substrates reached pseudo-steady state in brain within ∼30-40 min of bolus infusion. Time constants characterizing the appearance in brain of deuterated substrates or semi-heavy water following 2 H2 O administration were similar (∼15 min). CONCLUSION:Administration of deuterated substrates via subcutaneous catheter for in vivo DMRS experiments with mice is robust, requires limited personnel training, and enables substantial dosing. It is suitable for metabolic studies where pseudo-steady state substrate administration/accumulation is sufficient. It is particularly advantageous for serial longitudinal studies over an extended period because it avoids inevitable damage to the tail vein following multiple catheterizations.

Employing mouse models of radiation necrosis (one that displays the histologic hallmarks of the clinical condition) and glioblastoma (GL261), deuterium ( 2 H) MR spectroscopy, in concert with infusion of 2 H-labelled glucose, was employed to ascertain whether the 2 H MRS signatures could differentiate the two lesions, an unmet clinical need. The 2 H MR metabolic profiles of the two lesions were markedly different. In the tumor, the Warburg effect (aerobic glycolysis, fermentation) converted glucose nearly exclusively to lactate. In radiation necrosis, oxidative phosphorylation (respiration) dominated, with glucose converting to TCA cycle intermediates glutamate and glutamine. Thus, 2 H MR distinguishes glioblastoma vs. radiation necrosis.

Purpose To assess changes in intracellular diffusion as a mechanism for the reduction in water ADC that accompanies brain injury. Using NAA as a marker of neuronal cytoplasmic diffusion, NAA diffusion was measured before and after global ischemia (immediately postmortem) in the female Sprague–Dawley rat. Methods Diffusion‐weighted PRESS spectra, with diffusion encoding in a single direction, were acquired from large voxels of rat brain gray matter in vivo and postischemia employing either pairs of pulsed half‐sine–shaped gradients (in vivo and postischemia, b max = 19 ms/μm 2 ) or sinusoidal oscillating gradients (in vivo only) with frequencies of 99.2–250 Hz. A 2D randomly oriented cylinder (neurite) model gave estimates of longitudinal and transverse diffusivities ( D L and D T , respectively). In this model, D L represents the “free” diffusivity of NAA, whereas D T reflects highly restricted diffusion. Using oscillating gradients, the frequency dependence of D T [ D T (ω)] gave estimates of the cylinder (axon/dendrite) radius. Results A 10% decrease in D L,NAA followed global ischemia, dropping from 0.391 ± 0.012 μm 2 /ms to 0.350 ± 0.009 μm 2 /ms. Modeling D T,NAA (ω) provided an estimate of the neurite radius of 1.0 ± 0.6 μm. Conclusion Whereas the increase in apparent intraneuronal viscosity suggested by changes in D L,NAA may contribute to the overall reduction in water ADC associated with brain injury, it is not sufficient to be the sole explanation. Estimates of neurite radius based on D T (ω) were consistent with literature values.

NMR in BiomedicineVolume 36, Issue 1 e4841 EDITORIAL Metabolic activity diffusion imaging (MADI): A new paradigm Jeffrey J. Neil, Jeffrey J. Neil Washington University School of Medicine, St. Louis, Missouri, USASearch for more papers by this authorJoseph J. H. Ackerman, Corresponding Author Joseph J. H. Ackerman [email protected] Washington University School of Medicine, St. Louis, Missouri, USA Correspondence Joseph J. H. Ackerman, Washington University School of Medicine, 660 South Euclid, St. Louis, MO 63110, USA. Email: [email protected]Search for more papers by this author Jeffrey J. Neil, Jeffrey J. Neil Washington University School of Medicine, St. Louis, Missouri, USASearch for more papers by this authorJoseph J. H. Ackerman, Corresponding Author Joseph J. H. Ackerman [email protected] Washington University School of Medicine, St. Louis, Missouri, USA Correspondence Joseph J. H. Ackerman, Washington University School of Medicine, 660 South Euclid, St. Louis, MO 63110, USA. Email: [email protected]Search for more papers by this author First published: 02 November 2022 https://doi.org/10.1002/nbm.4841Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL No abstract is available for this article. Volume36, Issue1January 2023e4841 RelatedInformation

Purpose Distinguishing radiation necrosis (RN) from recurrent tumor remains a vexing clinical problem with important health-care consequences for neuro-oncology patients. Here, mouse models of pure tumor, pure RN, and admixed RN/tumor are employed to evaluate hydrogen (1H) and deuterium (2H) magnetic resonance methods for distinguishing RN vs. tumor. Furthermore, proof-of-principle, range-finding deuterium (2H) metabolic magnetic resonance is employed to assess glycolytic signatures distinguishing RN vs. tumor. Materials and Methods A pipeline of common quantitative 1H MRI contrasts, including an improved magnetization transfer ratio (MTR) sequence, and 2H magnetic resonance spectroscopy (MRS) following administration of 2H-labeled glucose, was applied to C57BL/6 mouse models of the following: (i) late time-to-onset RN, occurring 4–5 weeks post focal 50-Gy (50% isodose) Gamma Knife irradiation to the left cerebral hemisphere, (ii) glioblastoma, growing ~18–24 days post implantation of 50,000 mouse GL261 tumor cells into the left cerebral hemisphere, and (iii) mixed model, with GL261 tumor growing within a region of radiation necrosis (1H MRI only). Control C57BL/6 mice were also examined by 2H metabolic magnetic resonance. Results Differences in quantitative 1H MRI parametric values of R1, R2, ADC, and MTR comparing pure tumor vs. pure RN were all highly statistically significant. Differences in these parameter values and DCEAUC for tumor vs. RN in the mixed model (tumor growing in an RN background) are also all significant, demonstrating that these contrasts—in particular, MTR—can effectively distinguish tumor vs. RN. Additionally, quantitative 2H MRS showed a highly statistically significant dominance of aerobic glycolysis (glucose ➔ lactate; fermentation, Warburg effect) in the tumor vs. oxidative respiration (glucose ➔ TCA cycle) in the RN and control brain. Conclusions These findings, employing a pipeline of quantitative 1H MRI contrasts and 2H MRS following administration of 2H-labeled glucose, suggest a pathway for substantially improving the discrimination of tumor vs. RN in the clinic.

PurposeClinical evidence suggests radiation induces changes in the brain microenvironment that affect subsequent response to treatment. This study investigates the effect of previous radiation, delivered six weeks prior to orthotopic tumor implantation, on subsequent tumor growth and therapeutic response to anti-PD-L1 therapy in an intracranial mouse model, termed the Radiation Induced Immunosuppressive Microenvironment (RI2M) model.Method and MaterialsC57Bl/6 mice received focal (hemispheric) single-fraction, 30-Gy radiation using the Leksell GammaKnife® Perfexion™, a dose that does not produce frank/gross radiation necrosis. Non-irradiated GL261 glioblastoma tumor cells were implanted six weeks later into the irradiated hemisphere. Lesion volume was measured longitudinally by in vivo MRI. In a separate experiment, tumors were implanted into either previously irradiated (30 Gy) or non-irradiated mouse brain, mice were treated with anti-PD-L1 antibody, and Kaplan-Meier survival curves were constructed. Mouse brains were assessed by conventional hematoxylin and eosin (H&E) staining, IBA-1 staining, which detects activated microglia and macrophages, and fluorescence-activated cell sorting (FACS) analysis.ResultsTumors in previously irradiated brain display aggressive, invasive growth, characterized by viable tumor and large regions of hemorrhage and necrosis. Mice challenged intracranially with GL261 six weeks after prior intracranial irradiation are unresponsive to anti-PD-L1 therapy. K-M curves demonstrate a statistically significant difference in survival for tumor-bearing mice treated with anti-PD-L1 antibody between RI2M vs. non-irradiated mice. The most prominent immunologic change in the post-irradiated brain parenchyma is an increased frequency of activated microglia.ConclusionsThe RI2M model focuses on the persisting (weeks-to-months) impact of radiation applied to normal, control-state brain on the growth characteristics and immunotherapy response of subsequently implanted tumor. GL261 tumors growing in the RI2M grew markedly more aggressively, with tumor cells admixed with regions of hemorrhage and necrosis, and showed a dramatic loss of response to anti-PD-L1 therapy compared to tumors in non-irradiated brain. IHC and FACS analyses demonstrate increased frequency of activated microglia, which correlates with loss of sensitivity to checkpoint immunotherapy. Given that standard-of-care for primary brain tumor following resection includes concurrent radiation and chemotherapy, these striking observations strongly motivate detailed assessment of the late effects of the RI2M on tumor growth and therapeutic efficacy.

Preclinical imaging is critical in the development of translational strategies to detect diseases and monitor response to therapy. The National Cancer Institute Co-Clinical Imaging Resource Program was launched, in part, to develop best practices in preclinical imaging. In this context, the objective of this work was to develop a 1-hour, multiparametric magnetic resonance image-acquisition pipeline with triple-negative breast cancer patient-derived xenografts (PDXs). The 1-hour, image-acquisition pipeline includes T1- and T2-weighted scans, quantitative T1, T2, and apparent diffusion coefficient (ADC) parameter maps, and dynamic contrast-enhanced (DCE) time-course images. Quality-control measures used phantoms. The triple-negative breast cancer PDXs used for this study averaged 174 ± 73 μL in volume, with region of interest–averaged T1, T2, and ADC values of 1.9 ± 0.2 seconds, 62 ± 3 milliseconds, and 0.71 ± 0.06 μm2/ms (mean ± SD), respectively. Specific focus was on assessing the within-subject test–retest coefficient-of-variation (CVWS) for each of the magnetic resonance imaging metrics. Determination of PDX volume via manually drawn regions of interest is highly robust, with ∼1% CVWS. Determination of T2 is also robust with a ∼3% CVWS. Measurements of T1 and ADC are less robust with CVWS values in the 6%–11% range. Preliminary DCE test–retest time-course determinations, as quantified by area under the curve and Ktrans from 2-compartment exchange (extended Tofts) modeling, suggest that DCE is the least robust protocol, with ∼30%–40% CVWS.

Purpose: There is mounting evidence that, in addition to angiogenesis, hypoxiainduced inflammation via the hypoxia-inducible factor 1 alpha (HIF-1 alpha) eCXC chemokine receptor 4 (CXCR4) pathway may contribute to the pathogenesis of late-onset, irradiation-induced necrosis. This study investigates the mitigative efficacy of an HIF-1 alpha inhibitor, topotecan, and a CXCR4 antagonist, AMD3100, on the development of radiation necrosis (RN) in an intracranial mouse model. Methods and Materials: Mice received a single-fraction, 50-Gy dose of hemispheric irradiation from the Leksell Gamma Knife Perfexion and were then treated with either topotecan, an HIF-1 alpha inhibitor, from 1 to 12 weeks after irradiation, or AMD3100, a CXCR4 antagonist, from 4 to 12 weeks after irradiation. The onset and progression of RN were monitored longitudinally via noninvasive, in vivo magnetic resonance imaging (MRI) from 4 to 12 weeks after irradiation. Conventional hematoxylin-eosin staining and immunohistochemistry staining were performed to evaluate the treatment response. Results: The progression of brain RN was significantly mitigated for mice treated with either topotecan or AMD3100 compared with control animals. MRI-derived lesion volumes were significantly smaller for both of the treated groups, and histologic findings correlated well with the MRI data. By hematoxylin-eosin staining, both treated groups demonstrated reduced irradiation-induced tissue damage compared with controls. Furthermore, immunohistochemistry results revealed that expression levels of vascular endothelial growth factor, CXC chemokine ligand 12, CD68, CD3, and tumor necrosis factor a in the lesion area were significantly lower in treated (topotecan or AMD3100) brains versus control brains, while ionized calcium-binding adapter molecule 1 (Iba1) and HIF-1 alpha expression was similar, though somewhat reduced. CXCR4 expression was reduced only in topotecan-treated mice, while interleukin 6 expression was unaffected by either topotecan or AMD3100. Conclusions: By reducing inflammation, both topotecan and AMD3100 can, independently, mitigate the development of RN in the mouse brain. When combined with first-line, antiangiogenic treatment, anti-inflammation therapy may provide an adjuvant therapeutic strategy for clinical, postirradiation management of tumors, with additional benefits in the mitigation of RN development. (C) 2017 Elsevier Inc. All rights reserved.

Recently, a number of magnetic resonance imaging protocols have been reported that seek to exploit the effect of dissolved oxygen (O2, paramagnetic) on the longitudinal 1H relaxation of tissue water, thus providing image contrast related to tissue oxygen content. However, tissue water relaxation is dependent on a number of mechanisms and this raises the issue of how best to model the relaxation data. This problem, the model selection problem, occurs in many branches of science and is optimally addressed by Bayesian probability theory. High signal-to-noise, densely sampled, longitudinal 1H relaxation data were acquired from rat brain in vivo and from a cross-linked bovine serum albumin (xBSA) phantom, a sample that recapitulates the relaxation characteristics of tissue water in vivo. Bayesian-based model selection was applied to a cohort of five competing relaxation models: (1) monoexponential, (2) stretched-exponential, (3) biexponential, (4) Gaussian (normal) R 1-distribution, and (5) gamma R 1-distribution. Bayesian joint analysis of multiple replicate datasets revealed that water relaxation of both the xBSA phantom and in vivo rat brain was best described by a biexponential model, while xBSA relaxation datasets truncated to remove evidence of the fast relaxation component were best modeled as a stretched exponential. In all cases, estimated model parameters were compared to the commonly used monoexponential model. Reducing the sampling density of the relaxation data and adding Gaussian-distributed noise served to simulate cases in which the data are acquisition-time or signal-to-noise restricted, respectively. As expected, reducing either the number of data points or the signal-to-noise increases the uncertainty in estimated parameters and, ultimately, reduces support for more complex relaxation models.

Primary and metastatic CNS malignancies remain incurable despite aggressive therapy with surgery and radiation. Immunotherapy has shown promise in many solid and hematologic malignancies, however, results have been disappointing in recurrent primary CNS tumors. Additionally, while upfront immune checkpoint blockade therapy (ICB) has shown equal efficacy intracranially and extracranially, these responses are abrogated in previously irradiated brain lesions. Together, we hypothesize that radiation modulates the brain microenvironment to permit tumor growth and suppress anti-tumor immunity. As such, we developed a mouse model whereby orthotopic transplantation of murine gliomas into previously irradiated normal brain results in a more aggressive tumor phenotype. Moreover, the ICB-sensitive glioma cell line, GL261, is resistant to ICB therapy when implanted into a previously irradiated brain microenvironment. Immunophenotyping revealed a decreased ratio of CD8:CD4 T cells within tumor-infiltrating lymphocytes isolated from previously irradiated mice though the relative frequency of neoantigen-specific CD8 T cells was slightly increased and no difference in PD-1 expression was observed. Alternatively, the frequency of microglia and tumor-infiltrating CD11b+ Gr-1+ myeloid-derived suppressor cells (MDSC) was increased following irradiation suggesting a potential role for these myeloid cells in the immunosuppressive effects noted. Consistently, the administration of a CSF-1R inhibitor, which has been shown to reduce the number of microglia and block MDSCs, partially resensitizes GL261 cells to ICB therapy. In summary, we have developed a model that recapitulates the late effects of radiation on immunotherapy-resistance in CNS tumors. Preliminary results suggest that these late radiation effects are mediated through an increased myeloid population, and that inhibition of these cell subsets via the CSF-1R pathway can partially restore efficacy of ICB therapy. Furthermore, this model may provide further insight into additional therapeutic strategies that can be used to overcome these mechanisms of resistance induced by radiation therapy in the CNS.

Objective: The infant temperament behavioral inhibition is a potent risk factor for development of an anxiety disorder. It is difficult to predict risk for behavioral inhibition at birth, however, and the neural underpinnings are poorly understood. The authors hypothesized that neonatal functional connectivity of the ventral attention network is related to behavioral inhibition at age 2 years beyond sociodemographic and familial factors. This hypothesis is supported by the ventral attention network's role in attention to novelty, a key feature of behavioral inhibition. Method: Using a longitudinal design (N=45), the authors measured functional connectivity using MRI in neonates and behavioral inhibition at age 2 using the Infant-Toddler Social and Emotional Assessment. Whole-brain connectivity maps were computed for regions from the ventral attention, default mode, and salience networks. Regression analyses related these maps to behavioral inhibition at age 2, covarying for sex, social risk, and motion during scanning. Results: Decreased neonatal functional connectivity of three connections was associated with increased behavioral inhibition at age 2. One connection (between the right ventrolateral prefrontal cortex and the right temporal-parietal junction) included the ventral attention network seed, and two connections (between the medial prefrontal cortex and both the right superior parietal lobule and the left lateral occipital cortex) included the default mode network seed. Conclusions: Neonatal functional connectivity of the ventral attention and default mode networks is associated with behavioral inhibition at age 2. These results inform the developmental neurobiology of behavioral inhibition and anxiety disorders and may aid in early risk assessment and intervention.

PurposeTo determine the intracellular water preexchange lifetime, τi, the “average residence time” of water, in the intracellular milieu of neurons and astrocytes. The preexchange lifetime is important for modeling a variety of MR data sets, including relaxation, diffusion‐sensitive, and dynamic contrast–enhanced data sets.MethodsHerein, τi in neurons and astrocytes is determined in a microbead‐adherent, cultured cell system. In concert with thin‐slice selection, rapid flow of extracellular media suppresses extracellular signal, allowing determination of the transcytolemmal‐exchange‐dominated, intracellular T1. With this knowledge, and that of the intracellular T1 in the absence of exchange, τi can be derived.ResultsUnder normal culture conditions, τi for neurons is 0.75 ± 0.05 s versus 0.57 ± 0.03 s for astrocytes. Both neuronal and astrocytic τis decrease within 30 min after the onset of oxygen‐glucose deprivation, with the astrocytic τi showing a substantially greater decrease than the neuronal τi.ConclusionsGiven an approximate intra‐ to extracellular volume ratio of 4:1 in the brain, these data imply that, under normal physiological conditions, an MR experimental characteristic time of less than 0.012 s is required for a nonexchanging, two‐compartment (intra‐ and extracellular) model to be valid for MR studies. This characteristic time shortens significantly (i.e., 0.004 s) under injury conditions. Magn Reson Med 79:1616–1627, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Purpose: Glioblastoma (GBM) remains incurable, despite state-of-the-art treatment involving surgical resection, chemotherapy, and radiation. GBM invariably recurs as a highly invasive and aggressive phenotype, with the majority of recurrences within the radiation therapy treatment field. Although a large body of literature reporting on primary GBM exists, comprehensive studies of how prior irradiation alters recurrent tumor growth are lacking. An animal model that replicates the delayed effects of radiation therapy on the brain microenvironment, and its impact on the development of recurrent GBM, would be a significant advance. Methods and Materials: Cohorts of mice received a single fraction of 0, 20, 30, or 40 Gy Gamma Knife irradiation. Naive, nonirradiated mouse GBM tumor cells were implanted into the ipsilateral hemisphere 6 weeks postirradiation. Tumor growth was measured by magnetic resonance imaging, and animal survival was assessed by monitoring weight loss. Magnetic resonance imaging results were supported by hemotoxylin and eosin histology. Results: Tumorous lesions generated from orthotopic implantation of nonirradiated mouse GBM tumor cells into irradiated mouse brain grew far more aggressively and invasively than implantation of these same cells into nonirradiated brain. Lesions in irradiated brain tissue were significantly larger, more necrotic, and more vascular than those in control animals with increased invasiveness of tumor cells in the periphery, consistent with the histologic features commonly observed in recurrent high-grade tumors in patients. Conclusions: Irradiation of normal brain primes the targeted cellular microenvironment for aggressive tumor growth when naive (not previously irradiated) cancer cells are subsequently introduced. The resultant growth pattern is similar to the highly aggressive pattern of tumor regrowth observed clinically after therapeutic radiation therapy. The mouse model offers an avenue for determining the cellular and molecular basis for the aggressiveness of recurrent GBM. (C) 2018 Published by Elsevier Inc.

The apparent diffusion coefficient (ADC) is analyzed for the case of oscillating diffusion-sensitizing gradients in the high-frequency regime. We provide a concise derivation of the analytical expression for the ADC for an arbitrary number of gradient oscillations N and initial phase phi. It is demonstrated that an ultimate goal - to determine the surface-to-volume ratio (S/V) from MR measurements by using oscillating gradients - can be achieved with cosine-type gradients (phi = 0) for an arbitrary N. However, to determine S/V employing gradients with phi not equal 0 (including the sine-type gradients) and arbitrary N additionally requires prior knowledge of the time-dependent diffusion coefficient D(t). The latter is rarely known a priori but can be estimated under certain limiting conditions: (i) in the short time regime, when the total diffusion time of the measurements, t, is smaller than the characteristic diffusion time of the microstructural system of interest, an analytical expression for D(t) is available (Mitra's expression) and this allows S/V to be determined in the short time regime with sine-type gradients; (ii) in the important case of purely restricted diffusion, D(t) -> 0 at sufficiently long time, the signal becomes independent of phi and behaves as for the cosine-type gradients, thus, allowing determination of S/V. (C) 2018 Elsevier Inc. All rights reserved.

Bayesian probability theory provides optimal parameter estimates and robust model selection from a family of competing data models. However, widespread adoption of the Bayesian approach to the analysis of magnetic resonance and other data types has been hindered by its perceived complexity and heavy computational burden. This manuscript describes the Bayesian Analysis Toolbox, a computationally efficient, robust, and highly optimized suite of data modeling software packages based upon the precepts of Bayesian probability theory. The Toolbox is downloadable at no cost for non-commercial applications from http://bayesiananalysis. wustl. edu. The Toolbox extends Bayesian-based data analysis to a variety of real-world data analysis problems commonly encountered in spectroscopy and imaging, with a focus on magnetic resonance-derived data, making the power of this approach available to the non-expert user.

4,254,778 3/1981 Clow et al. ......................... 128/653 4,363,793 l2/1982 Blau et al. ............................... 424/1 4,447,600 5/1984 Ogura et al. .......................... 536/23 4,586,511 5/1986 Clark ............... ... 128/653 4,663,443 5/1987 Shibayama et al. . ... 536/4.1 4,716,225 12/1987 Ledley et al........ ... 536/22 4,775,522 10/1987 Clark ....................................... 424/9 4,889,125 6/1987 Doddrell et al. ... ... 128/653 4,889,126 6/1987 Doddrell et al. ... ... 128/653 4,893,627 1/1990 Kehayias et al. ................... 128/654

Anti-vascular endothelial growth factor (anti-VEGF) antibodies are a promising new treatment for late time-to-onset radiation-induced necrosis (RN). We sought to evaluate and validate the response to anti-VEGF antibody in a mouse model of RN. Mice were irradiated with the Leksell Gamma Knife Perfexion™ and then treated with anti-VEGF antibody, beginning at post-irradiation (PIR) week 8. RN progression was monitored via anatomic and diffusion MRI from weeks 4–12 PIR. Standard histology, using haematoxylin and eosin (H&E), and immunohistochemistry staining were used to validate the response to treatment. After treatment, both post-contrast T1-weighted and T2-weighted image-derived lesion volumes decreased (P < 0.001), while the lesion volumes for the control group increased. The abnormally high apparent diffusion coefficient (ADC) for RN also returned to the ADC range for normal brain following treatment (P < 0.001). However, typical RN pathology was still present histologically. Large areas of focal calcification were observed in ~50% of treated mouse brains. Additionally, VEGF and hypoxia-inducible factor 1-alpha (HIF-1α) were continually upregulated in both the anti-VEGF and control groups. Despite improvements observed radiographically following anti-VEGF treatment, lesions were not completely resolved histologically. The subsequent calcification and the continued upregulation of VEGF and HIF-1α merit further preclinical/clinical investigation.

This study aims to develop a constrained local arterial input function (cL-AIF) to improve quantitative analysis of dynamic contrast-enhanced (DCE)-magnetic resonance imaging (MRI) data by accounting for the contrast-agent bolus amplitude error in the voxel-specific AIF. Bayesian probability theory-based parameter estimation and model selection were used to compare tracer kinetic modeling employing either the measured remote-AIF (R-AIF, i.e., the traditional approach) or an inferred cL-AIF against both in silico DCE-MRI data and clinical, cervical cancer DCE-MRI data. When the data model included the cL-AIF, tracer kinetic parameters were correctly estimated from in silico data under contrast-to-noise conditions typical of clinical DCE-MRI experiments. Considering the clinical cervical cancer data, Bayesian model selection was performed for all tumor voxels of the 16 patients (35,602 voxels in total). Among those voxels, a tracer kinetic model that employed the voxel-specific cL-AIF was preferred (i.e., had a higher posterior probability) in 80 % of the voxels compared to the direct use of a single R-AIF. Maps of spatial variation in voxel-specific AIF bolus amplitude and arrival time for heterogeneous tissues, such as cervical cancer, are accessible with the cL-AIF approach. The cL-AIF method, which estimates unique local-AIF amplitude and arrival time for each voxel within the tissue of interest, provides better modeling of DCE-MRI data than the use of a single, measured R-AIF. The Bayesian-based data analysis described herein affords estimates of uncertainties for each model parameter, via posterior probability density functions, and voxel-wise comparison across methods/models, via model selection in data modeling.