A supervised deep neural network approach with standardized targets for enhanced accuracy of IVIM parameter estimation from multi-SNR images.

Extraction of intravoxel incoherent motion (IVIM) parameters from noisy diffusion-weighted (DW) images using a biexponential fitting model is computationally challenging, and the reliability of the estimated perfusion-related quantities represents a limitation of this technique. Artificial intelligence can overcome the current limitations and be a suitable solution to advance use of this technique in both preclinical and clinical settings. The purpose of this work was to develop a deep neural network (DNN) approach, trained on numerical simulated phantoms with different signal to noise ratios (SNRs), to improve IVIM parameter estimation. The proposed approach is based on a supervised fully connected DNN having 3 hidden layers, 18 inputs and 3 targets with standardized values. 14 × 103 simulated DW images, based on a Shepp-Logan phantom, were randomly generated with varying SNRs (ranging from 10 to 100). 7 × 103 images (1000 for each SNR) were used for training. Performance accuracy was assessed in simulated images and the proposed approach was compared with the state-of-the-art Bayesian approach and other DNN algorithms. The DNN approach was also evaluated in vivo on a high-field MRI preclinical scanner. Our DNN approach showed an overall improvement in accuracy when compared with the Bayesian approach and other DNN methods in most of the simulated conditions. The in vivo results demonstrated the feasibility of the proposed approach in real settings and generated quantitative results comparable to those obtained using the Bayesian and unsupervised approaches, especially for D and f, and with lower variability in homogeneous regions. The DNN architecture proposed in this work outlines two innovative features as compared with other studies: (1) the use of standardized targets to improve the estimation of parameters, and (2) the implementation of a single DNN to enhance the IVIM fitting at different SNRs, providing a valuable alternative tool to compute IVIM parameters in conditions of high background noise.

Correlation and Agreement of Yttrium-90 Positron Emission Tomography/Computed Tomography with Ex Vivo Radioembolization Microsphere Deposition in the Rabbit VX2 Liver Tumor Model.

PURPOSE: To demonstrate a stronger correlation and agreement of yttrium-90 (90Y) positron emission tomography (PET)/computed tomography (CT) measurements with explant liver tumor dosing compared with the standard model (SM) for radioembolization. MATERIALS AND METHODS: Hepatic VX2 tumors were implanted into New Zealand white rabbits, with growth confirmed by 7 T magnetic resonance imaging. Seventeen VX2 rabbits provided 33 analyzed tumors. Treatment volumes were calculated from manually drawn volumes of interest (VOI) with three-dimensional surface renderings. Radioembolization was performed with glass 90Y microspheres. PET/CT imaging was completed with scatter and attenuation correction. Three-dimensional ellipsoid VOI were drawn to encompass tumors on fused images. Tumors and livers were then explanted for inductively coupled plasma (ICP)-optical emission spectroscopy (OES) analysis of microsphere content. 90Y PET/CT and SM measurements were compared with reference standard ICP-OES measurements of tumor dosing with Pearson correlation and Bland-Altman analyses for agreement testing with and without adjustment for tumor necrosis. RESULTS: The median infused activity was 33.3 MBq (range, 5.9-152.9). Tumor dose was significantly correlated with 90Y PET/CT measurements (r = 0.903, P < .001) and SM estimates (r = 0.607, P < .001). Bland-Altman analyses showed that the SM tended to underestimate the tumor dosing by a mean of -8.5 Gy (CI, -26.3-9.3), and the degree of underestimation increased to a mean of -18.3 Gy (CI, -38.5-1.9) after the adjustment for tumor necrosis. CONCLUSIONS: 90Y PET/CT estimates were strongly correlated and had better agreement with reference measurements of tumor dosing than SM estimates.

MerTK Cleavage on Resident Cardiac Macrophages Compromises Repair After Myocardial Ischemia Reperfusion Injury.

RATIONALE: Clinical benefits of reperfusion after myocardial infarction are offset by maladaptive innate immune cell function, and therapeutic interventions are lacking. OBJECTIVE: We sought to test the significance of phagocytic clearance by resident and recruited phagocytes after myocardial ischemia reperfusion. METHODS AND RESULTS: In humans, we discovered that clinical reperfusion after myocardial infarction led to significant elevation of the soluble form of MerTK (myeloid-epithelial-reproductive tyrosine kinase; ie, soluble MER), a critical biomarker of compromised phagocytosis by innate macrophages. In reperfused mice, macrophage Mertk deficiency led to decreased cardiac wound debridement, increased infarct size, and depressed cardiac function, newly implicating MerTK in cardiac repair after myocardial ischemia reperfusion. More notably, Mertk(CR) mice, which are resistant to cleavage, showed significantly reduced infarct sizes and improved systolic function. In contrast to other cardiac phagocyte subsets, resident cardiac MHCIILOCCR2- (major histocompatibility complex II/C-C motif chemokine receptor type 2) macrophages expressed higher levels of MerTK and, when exposed to apoptotic cells, secreted proreparative cytokines, including transforming growth factor-ß. Mertk deficiency compromised the accumulation of MHCIILO phagocytes, and this was rescued in Mertk(CR) mice. Interestingly, blockade of CCR2-dependent monocyte infiltration into the heart reduced soluble MER levels post-ischemia reperfusion. CONCLUSIONS: Our data implicate monocyte-induced MerTK cleavage on proreparative MHCIILO cardiac macrophages as a novel contributor and therapeutic target of reperfusion injury.

18F-FDG PET Biomarkers Help Detect Early Metabolic Response to Irreversible Electroporation and Predict Therapeutic Outcomes in a Rat Liver Tumor Model.

Purpose To test the hypothesis that biomarkers of fluorine 18 (18F) fluorodeoxyglucose (FDG) positron emission tomography (PET) can be used for the early detection of therapeutic response to irreversible electroporation (IRE) of liver tumor in a rodent liver tumor model. Materials and Methods The institutional animal care and use committee approved this study. Rats were inoculated with McA-RH7777 liver tumor cells in the left median and left lateral lobes. Tumors were allowed to grow for 7 days to reach a size typically at least 5 mm in longest diameter, as verified with magnetic resonance (MR) imaging. IRE electrodes were inserted, and eight 100-µsec, 2000-V pulses were applied to ablate the tumor tissue in the left median lobe. Tumor in the left lateral lobe served as a control in each animal. PET/computed tomography (CT) and MR imaging measurements were performed at baseline and 3 days after IRE for each animal. Additional MR imaging measurements were obtained 14 days after IRE. After 14-day follow-up MR imaging, rats were euthanized and tumors harvested for hematoxylin-eosin, CD34, and caspase-3 staining. Change in the maximum standardized uptake value (ΔSUVmax) was calculated 3 days after IRE. The maximum lesion diameter change (ΔDmax) was measured 14 days after IRE by using axial T2-weighted imaging. ΔSUVmax and ΔDmax were compared. The apoptosis index was calculated by using caspase-3-stained slices of apoptotic tumor cells. Pearson correlation coefficients were calculated to assess the relationship between ΔSUVmax at 3 days and ΔDmax (or apoptosis index) at 14 days after IRE treatment. Results ΔSUVmax, ΔDmax, and apoptosis index significantly differed between treated and untreated tumors (P < .001 for all). In treated tumors, there was a strong correlation between ΔSUVmax 3 days after IRE and ΔDmax 14 days after IRE (R = 0.66, P = .01) and between ΔSUVmax 3 days after IRE and apoptosis index 14 days after IRE (R = 0.57, P = .04). Conclusion 18F-FDG PET imaging biomarkers can be used for the early detection of therapeutic response to IRE treatment of liver tumors in a rodent model. © RSNA, 2017.

Pretrial functional connectivity differentiates behavioral outcomes during trace eyeblink conditioning in the rabbit.

Fluctuations in neural activity can produce states that facilitate and accelerate task-related performance. Acquisition of trace eyeblink conditioning (tEBC) in the rabbit is enhanced when trials are contingent on optimal pretrial activity in the hippocampus. Other regions which are essential for whisker-signaled tEBC, such as the cerebellar interpositus nucleus (IPN), somatosensory and prelimbic cortices, may also show optimal connectivity prior to successful performance. Functional magnetic resonance imaging (fMRI) was acquired in nine rabbits during tEBC on the first and tenth days of initial training and once again after a 30-d, training-free hiatus. Data acquired during the intertrial interval was parsed depending on whether or not a conditioned response (CR) occurred on the upcoming trial and seed-based functional connectivity was calculated among the IPN, hippocampus, somatosensory, and prelimbic cortices. Functional connectivity between the left somatosensory cortex and right IPN, regions critical for establishing and producing CRs evoked by right vibrissae vibration and right corneal airpuff, was significantly negative prior to successful, CR trials as compared with unsuccessful, non-CR trials. Differences were not observed for any of the other possible combinations of connectivity. Our results demonstrate that specific pretrial functional connectivity exists within the rabbit brain and differentiates between upcoming behavioral response outcomes. Online analysis of network fluctuations has the potential to be used as the basis for therapeutic interventions to facilitate learning and memory.

Intrinsic connectivity of neural networks in the awake rabbit.

The way in which the brain is functionally connected into different networks has emerged as an important research topic in order to understand normal neural processing and signaling. Since some experimental manipulations are difficult or unethical to perform in humans, animal models are better suited to investigate this topic. Rabbits are a species that can undergo MRI scanning in an awake and conscious state with minimal preparation and habituation. In this study, we characterized the intrinsic functional networks of the resting New Zealand White rabbit brain using BOLD fMRI data. Group independent component analysis revealed seven networks similar to those previously found in humans, non-human primates and/or rodents including the hippocampus, default mode, cerebellum, thalamus, and visual, somatosensory, and parietal cortices. For the first time, the intrinsic functional networks of the resting rabbit brain have been elucidated demonstrating the rabbit's applicability as a translational animal model. Without the confounding effects of anesthetics or sedatives, future experiments may employ rabbits to understand changes in neural connectivity and brain functioning as a result of experimental manipulation (e.g., temporary or permanent network disruption, learning-related changes, and drug administration).

Activity-induced manganese-dependent MRI (AIM-MRI) and functional MRI in awake rabbits during somatosensory stimulation.

Activity-induced manganese-dependent MRI (AIM-MRI) is a powerful tool to track system-wide neural activity using high resolution, quantitative T1-weighted MRI in animal models and has significant advantages for investigating neural activity over other modalities including BOLD fMRI. With AIM-MRI, Mn(2+) ions enter neurons via voltage-gated calcium channels preferentially active during the time of experimental exposure. A broad range of AIM-MRI studies using different species studying different phenomena have been performed, but few of these studies provide a systematic evaluation of the factors influencing the detection of Mn(2+) such as dosage and the temporal characteristics of Mn(2+) uptake. We identified an optimal dose of Mn(2+) (25 mg/kg, s.c.) in order to characterize the time-course of Mn(2+) accumulation in active neural regions in the rabbit. T1-weighted MRI and functional MRI were collected 0-3, 6-9, and 24-27 h post-Mn(2+) injection while the vibrissae on the right side were vibrated. Significant BOLD activation in the left somatosensory (SS) cortex and left ventral posteromedial (VPM) thalamic nucleus was detected during whisker vibration. T1-weighted signal intensities were extracted from these regions, their corresponding contralateral regions and the visual cortex (to serve as controls). A significant elevation in T1-weighted signal intensity in the left SS cortex (relative to right) was evident 6-9 and 24-27 h post-Mn(2+) injection while the left VPM thalamus showed a significant enhancement (relative to the right) only during the 24-27 h session. Visual cortex showed no hemispheric difference at any timepoint. Our results suggest that studies employing AIM-MRI would benefit by conducting experimental manipulations 6-24 h after subcutaneous MnCl2 injections to optimize the concentration of contrast agent in the regions active during the exposure.

Novel method for functional brain imaging in awake minimally restrained rats.

Functional magnetic resonance imaging (fMRI) in rodents holds great promise for advancing our knowledge about human brain function. However, the use of anesthetics to immobilize rodents during fMRI experiments has restricted the type of questions that can be addressed using this technique. Here we describe an innovative procedure to train rats to be constrained without the need of any anesthesia during the whole procedure. We show that with 8-10 days of acclimation rats can be conscious and remain still during fMRI experiments under minimal stress. In addition, we provide fMRI results of conscious rodents in a variety of commonly used fMRI experimental paradigms, and we demonstrate the improved quality of these scans by comparing results when the same rodents were scanned under anesthesia. We confirm that the awake scanning procedure permits an improved evaluation of brain networks and brain response to external stimuli with minimal movement artifact. The present study further advances the field of fMRI in awake rodents, which provide more direct, forward and reverse, translational opportunities regarding brain functional correspondences between human and rodent research.

Quantitative magnetic resonance imaging in the transgenic APC(Δ468) mouse model of hereditary colon cancer.

In this study, we investigated the use of high-resolution magnetic resonance imaging (MRI) methods for in vivo detection and quantitative characterization of colorectal tumors in the transgenic APC(Δ468) mouse model. High-resolution T(1)-weighted (T(1)W) images, T(2)-weighted (T(2)W) images, and dynamic contrast-enhanced (DCE) measurements were performed using a 7.0 T small-animal imaging system (N = 10). Individual tumors were identified on both T(1)W and T(2)W images. Twenty-eight tumors (2.8 ± 0.9 mm, mean ± SD) were detected with high-resolution MRI across a distance of roughly 3 cm from the rectum to the splenic flexure, whereas 29 tumors were found within corresponding colon tissue samples examined at gross necropsy in the same area. T(2) values were significantly different between tumor, skeletal muscle, and normal intestinal wall tissues (p < .05). For analysis of the vascular characteristics of colon tumor tissues using DCE measurements, the initial area under the curve (IAUC) for Gd contrast concentration curve (time) (C(Gd) [t]) was calculated with integration times of 60 and 120 seconds post-contrast infusion; two integration times were selected to capture both tracer wash-in and wash-out characteristics. IAUC measurements were significantly larger in tumor tissues compared to both normal intestinal wall and skeletal muscle tissues (p < .001). In vivo anatomic and quantitative MRI measurements were readily feasible in the transgenic APC(Δ468) mouse model. These noninvasive methods should improve experimental efficiencies during longitudinal survival studies that otherwise would require single-end-point necropsy measurements.

Quantitative pharmacologic MRI in mice.

Pharmacologic MRI (phMRI) uses functional MRI techniques to provide a noninvasive in vivo measurement of the hemodynamic effects of drugs. The cerebral blood volume change (ΔCBV) serves as a surrogate for neuronal activity via neurovascular coupling mechanisms. By assessing the location and time course of brain activity in mouse mutant studies, phMRI can provide valuable insights into how different behavioral phenotypes are expressed in deferring brain activity response to drug challenge. In this report, we evaluate the utility of three different intravascular ultrasmall superparamagnetic iron oxide (USPIO) contrast agents for phMRI using a gradient-echo technique, with temporal resolution of one min at high magnetic field. The tissue half-life of the USPIOs was studied using a nonlinear detrending model. The three USPIOs are candidates for CBV weighted phMRI experiments, with r(2)/r(1) ratios ≥ 20 and apparent half-lives ≥ 1.5 h at the described doses. An echo-time of about 10 ms or longer results in a functional contrast to noise ratio (fCNR) > 75 after USPIO injection, with negligible decrease between 1.5-2 h. phMRI experiments were conducted at 7 T using cocaine as a psychotropic substance and acetazolamide, a global vasodilator, as a positive control. Cocaine acts as a dopamine-serotonin-norepinephrine reuptake inhibitor, increasing extracellular concentrations of these neurotransmitters, and thus increasing dopaminergic, serotonergic and noradrenergic neurotransmission. phMRI results showed that CBV was reduced in the normal mouse brain after cocaine challenge, with the largest effects in the nucleus accumbens, whereas after acetazolamide, blood volume was increased in both cerebral and extracerebral tissue.

Reversal of neuropathic pain is associated with corticostriatal functional reorganization after nerve repair in the spared nerve injury model.

ABSTRACT: Following surgical repair after peripheral nerve injury, neuropathic pain diminishes in most patients but can persist in a small proportion of cases, the mechanism of which remains poorly understood. Based on the spared nerve injury (SNI), we developed a rat nerve repair (NR) model, where a delayed reconstruction of the SNI-injured nerves resulted in alleviating chronic pain-like behavior only in a subpopulation of rats. Multiple behavioral measures were assayed over 11-week presurgery and postsurgery periods (tactile allodynia, pain prick responses, sucrose preference, motor coordination, and cold allodynia) in SNI (n = 10), sham (n = 8), and NR (n = 12) rats. All rats also underwent resting-state functional magnetic resonance imaging under anesthesia at multiple time points postsurgery, and at 10 weeks, histology and retrograde labeling were used to calculate peripheral reinnervation. Behavioral measures indicated that at approximately 5 weeks postsurgery, the NR group separated to pain persisting (NR persisting, n = 5) and recovering (NR recovering, n = 7) groups. Counts of afferent nerves and dorsal root ganglion cells were not different between NR groups. Therefore, NR group differences could not be explained by peripheral reorganization. By contrast, large brain functional connectivity differences were observed between NR groups, where corticolimbic reorganization paralleled with pain recovery (repeated-measures analysis of variance, false discovery rate, P < 0.05), and functional connectivity between accumbens and medial frontal cortex was related both to tactile allodynia (nociception) and to sucrose preference (anhedonia) in the NR group. Our study highlights the importance of brain circuitry in the reversal of neuropathic pain as a natural pain-relieving mechanism. Further studies regarding the therapeutic potentials of these processes are warranted.

Quantitative pharmacologic MRI: mapping the cerebral blood volume response to cocaine in dopamine transporter knockout mice.

The use of pharmacologic MRI (phMRI) in mouse models of brain disorders allows noninvasive in vivo assessment of drug-modulated local cerebral blood volume changes (ΔCBV) as one correlate of neuronal and neurovascular activities. In this report, we employed CBV-weighted phMRI to compare cocaine-modulated neuronal activity in dopamine transporter (DAT) knockout (KO) and wild-type mice. Cocaine acts to block the dopamine, norepinephrine, and serotonin transporters (DAT, NET, and SERT) that clear their respective neurotransmitters from the synapses, helping to terminate cognate neurotransmission. Cocaine consistently reduced CBV, with a similar pattern of regional ΔCBV in brain structures involved in mediating reward in both DAT genotypes. The largest effects (-20% to -30% ΔCBV) were seen in the nucleus accumbens and several cortical regions. Decreasing response amplitudes to cocaine were noted in more posterior components of the cortico-mesolimbic circuit. DAT KO mice had significantly attenuated ΔCBV amplitudes, shortened times to peak response, and reduced response duration in most regions. This study demonstrates that DAT knockout does not abolish the phMRI responses to cocaine, suggesting that adaptations to loss of DAT and/or retained cocaine activity in other monoamine neurotransmitter systems underlie these responses in DAT KO mice.

Simultaneous in vivo positron emission tomography and magnetic resonance imaging.

Positron emission tomography (PET) and magnetic resonance imaging (MRI) are widely used in vivo imaging technologies with both clinical and biomedical research applications. The strengths of MRI include high-resolution, high-contrast morphologic imaging of soft tissues; the ability to image physiologic parameters such as diffusion and changes in oxygenation level resulting from neuronal stimulation; and the measurement of metabolites using chemical shift imaging. PET images the distribution of biologically targeted radiotracers with high sensitivity, but images generally lack anatomic context and are of lower spatial resolution. Integration of these technologies permits the acquisition of temporally correlated data showing the distribution of PET radiotracers and MRI contrast agents or MR-detectable metabolites, with registration to the underlying anatomy. An MRI-compatible PET scanner has been built for biomedical research applications that allows data from both modalities to be acquired simultaneously. Experiments demonstrate no effect of the MRI system on the spatial resolution of the PET system and <10% reduction in the fraction of radioactive decay events detected by the PET scanner inside the MRI. The signal-to-noise ratio and uniformity of the MR images, with the exception of one particular pulse sequence, were little affected by the presence of the PET scanner. In vivo simultaneous PET and MRI studies were performed in mice. Proof-of-principle in vivo MR spectroscopy and functional MRI experiments were also demonstrated with the combined scanner.

Yttrium-90 Radioembolization and Tumor Hypoxia: Gas-challenge BOLD Imaging in the VX2 Rabbit Model of Hepatocellular Carcinoma.

RATIONALE AND OBJECTIVES: To use a rapid gas-challenge blood oxygen-level dependent magnetic resonance imaging exam to evaluate changes in tumor hypoxia after 90Y radioembolization (Y90) in the VX2 rabbit model. MATERIALS AND METHODS: White New Zealand rabbits (n = 11) provided a Y90 group (n = 6 rabbits) and untreated control group (n = 5 rabbits). R2* maps were generated with gas-challenges (O2/room air) at baseline, 1 week, and 2 weeks post-Y90. Laboratory toxicity was evaluated at baseline, 24 hours, 72 hours, 1 hours, and 2 weeks. Histology was used to evaluate tumor necrosis on hematoxylin and eosin and immunofluorescence imaging was used to assess microvessel density (CD31) and proliferative index (Ki67). RESULTS: At baseline, median tumor volumes and time to imaging were similar between groups (p = 1.000 and p = 0.4512, respectively). The median administered dose was 50.4 Gy (95% confidence interval:44.8-55.9). At week 2, mean tumor volumes were 5769.8 versus 643.7 mm3 for control versus Y90 rabbits, respectively (p = 0.0246). At two weeks, ΔR2* increased for control tumors to 12.37 ± 12.36sec-1 and decreased to 4.48 ± 9.00sec-1 after Y90. The Pearson correlation coefficient for ΔR2* at baseline and percent increase in tumor size by two weeks was 0.798 for the Y90 group (p = 0.002). There was no difference in mean microvessel density for control versus Y90 treated tumors (p = 0.6682). The mean proliferative index was reduced in Y90 treated tumors at 30.5% versus 47.5% for controls (p = 0.0071). CONCLUSION: The baseline ΔR2* of tumors prior to Y90 may be a predictive imaging biomarker of tumor response and treatment of these tumors with Y90 may influence tumor oxygenation over time.

A robust coregistration method for in vivo studies using a first generation simultaneous PET/MR scanner.

PURPOSE: Hybrid positron emission tomography (PET)/magnetic resonance (MR) imaging systems have recently been built that allow functional and anatomical information obtained from PET and MR to be acquired simultaneously. The authors have developed a robust coregistration scheme for a first generation small animal PET/MR imaging system and illustrated the potential of this system to study intratumoral heterogeneity in a mouse model. METHODS: An alignment strategy to fuse simultaneously acquired PET and MR data, using the MR imaging gradient coordinate system as the reference basis, was developed. The fidelity of the alignment was evaluated over multiple study sessions. In order to explore its robustness in vivo, the alignment strategy was applied to explore the heterogeneity of glucose metabolism in a xenograft tumor model, using 18F-FDG-PET to guide the acquisition of localized 1H MR spectra within a single imaging session. RESULTS: The alignment method consistently fused the PET/MR data sets with subvoxel accuracy (registration error mean = 0.55 voxels, < 0.28 mm); this was independent of location within the field of view. When the system was used to study intratumoral heterogeneity within xenograft tumors, a correlation of high 18F-FDG-PET signal with high choline/creatine ratio was observed. CONCLUSIONS: The authors present an implementation of an efficient and robust coregistration scheme for multimodal noninvasive imaging using PET and MR. This setup allows time-sensitive, multimodal studies of physiology to be conducted in an efficient manner.

Feasibility of Combination Intra-arterial Yttrium-90 and Irinotecan Microspheres in the VX2 Rabbit Model.

PURPOSE: To evaluate the combination of 90Y radioembolization (Y90) and drug-eluting bead irinotecan (DEBIRI) microspheres in the VX2 rabbit model. MATERIALS AND METHODS: An initial dose finding study was performed in 6 White New Zealand rabbits to identify a therapeutic but subcurative dose of Y90. In total, 29 rabbits were used in four groups: Y90 treatment (n = 8), DEBIRI treatment (n = 6), Y90 + DEBIRI treatment (n = 7), and an untreated control group (n = 8). Hepatic toxicity was evaluated at baseline, 24 h, 72 h, 1 week, and 2 weeks. MRI tumor volume (TV) and enhancing tumor volume were assessed baseline and 2 weeks. Tumor area and necrosis were evaluated on H&E for pathology. RESULTS: Infused activities of 84.0-94.4 MBq (corresponding to 55.1-72.7 Gy) were selected based on the initial dose finding study. Infusion of DEBIRI after Y90 was technically feasible in all cases (7/7). Overall, 21/29 animals survived to 2 weeks, and the remaining animals had extrahepatic disease on necropsy. Liver transaminases were elevated with Y90, DEBIRI, and Y90 + DEBIRI compared to control at 24 h, 72 h, and 1 week post-treatment and returned to baseline by 2 weeks. By TV, Y90 + DEBIRI was the only treatment to show statistically significant reduction at 2 weeks compared to the control group (p = 0.012). The change in tumor volume (week 2-baseline) for both Y90 + DEBIRI versus control (p = 0.002) and Y90 versus control (p = 0.014) was significantly decreased. There were no statistically significant differences among groups on pathology. CONCLUSION: Intra-arterial Y90 + DEBIRI was safe and demonstrated enhanced antitumor activity in rabbit VX2 tumors. This combined approach warrants further investigation.

Yttrium-90 Portal Vein Radioembolization in Sprague-Dawley Rats: Dose-Dependent Imaging and Pathological Changes in Normal Liver.

PURPOSE: Portal vein embolization (PVE) is an established neoadjuvant method to induce future liver remnant hypertrophy prior to surgical resection of hepatic tumors. The purpose of our study was to examine the feasibility of PVE with glass 90Y microspheres (Y90 PVE) in Sprague-Dawley rats. We tested the hypothesis that increased doses of Y90 PVE would increase target lobe fibrosis and atrophy. METHODS: Twenty-two rats were assigned to four groups for Y90 PVE to the right median lobe: very high- (273.8 MBq; n = 2), high- (99.9 MBq; n = 10), medium- (48.1 MBq; n = 5), and low-dose (14.8 MBq; n = 5). An untreated control group included seven rats. 90Y PET/CT of 90Y distributions confirmed lobar targeting. MRI volumes were measured at baseline, 2-, 4-, 8- and 12-weeks. Explanted hepatic lobes were weighed, sectioned, and stained for H&E and immunohistochemistry. Digitized slides allowed quantitative measurements of fibrosis (20 foci/slide). RESULTS: Ex vivo measurements confirmed 91-97% activity was localized to the target lobe (n = 4). The percent growth of the target lobe relative to baseline was - 5.0% (95% CI - 17.0-6.9%) for high-, medium dose rats compared to + 18.6% (95% CI + 7.6-29.7%) in the low-dose group at 12-weeks (p = 0.0043). Radiation fibrosis increased in a dose-dependent fashion. Fibrotic area/microsphere was 22,893.5, 14,946.2 ± 2253.3, 15,304.5 ± 4716.6, and 5268.8 ± 2297.2 µm2 for very high- (n = 1), high- (n = 4), medium- (n = 3), and low-dose groups (n = 5), respectively. CONCLUSION: Y90 PVE was feasible in the rat model, resulted in target lobe atrophy, and dose-dependent increases in hepatic fibrosis at 12 weeks. The onset of imaging-based volumetric changes was 8-12 weeks.

In vivo 19F magnetic resonance spectroscopy and chemical shift imaging of tri-fluoro-nitroimidazole as a potential hypoxia reporter in solid tumors.

PURPOSE: 2-Nitro-alpha-[(2,2,2-trifluoroethoxy)methyl]-imidazole-1-ethanol (TF-MISO) was investigated as a potential noninvasive marker of tissue oxygen levels in tumors using (19)F magnetic resonance spectroscopy (MRS) and (19)F chemical shift imaging. EXPERIMENTAL DESIGNS: In vitro data were obtained using high-performance liquid chromatography on tumor cells incubated under varying oxygen conditions to determine the oxygen-binding characteristics. In vivo data were obtained using a well-characterized hypoxic murine breast tumor (MCa), in addition to studies on a rat prostate tumor model (R3327-AT) implanted in nude mice. Detection of intratumor (19)F signal from TF-MISO was done using MRS for up to 10 h following a 75 mg/kg i.v. injection. Localized distribution of the compound in the implanted MCa tumor has been imaged using slice-selective two-dimensional chemical shift imaging 6 h after injection. RESULTS: The in vitro results showed that TF-MISO preferentially accumulates in cells incubated under anoxic conditions. The in vivo (19)F MR spectral features (line width and chemical shift) were recorded as a function of time after injection, and the results indicate that the fluorine atoms are indeed sensitive to changes in the local environment while still providing a detectable MR signal. Ex vivo spectra were collected and established the visibility of the (19)F signal under conditions of maximum hypoxia. Late time point (>6 h) tumor tissue concentrations, as obtained from (19)F MRS, suggest that TF-MISO is reduced and retained in hypoxic tumor. The feasibility of obtaining TF-MISO tumor distribution maps in a reasonable time frame was established. CONCLUSIONS: Based on the results presented herein, it is suggested that TF-MISO has the potential to be a valid magnetic resonance hypoxia imaging reporter for both preclinical hypoxia studies and hypoxia-directed clinical therapy.

The rabbit as a behavioral model system for magnetic resonance imaging.

BACKGROUND: fMRI requires that subjects not move during image acquisition. This has been achieved by instructing people not to move, or by anesthetizing experimental animal subjects to induce immobility. We have demonstrated that a surgically implanted headbolt onto the skull of a rabbit allows their brain to be imaged comfortably while the animal is awake. This article provides a detailed method for the preparation. NEW METHOD: We took advantage of the rabbit's tolerance for restraint to image the brain while holding the head at the standard stereotaxic angle. Visual stimulation was produced by flashing green LEDs and whisker stimulation was done by powering a small coil of wire attached to a fiber band. Blinking was recorded with an infrared emitter/detector directed at the eye with fiber-optic cabling. RESULTS: Results indicate that a single daily session of habituation is sufficient to produce adequate immobility on subsequent days to avoid movement artifacts. Results include high resolution images in the stereotaxic plane of the rabbit. COMPARISON WITH EXISTING METHOD(S): We see no degradation or distortion of MR signal, and the headbolt provides a means for rapid realignment of the head in the magnet from day to day, and across subjects. The use of rabbits instead of rodents allows much shorter periods of habituation, and the rabbit allows behavior to be observed during the day while the animal is in its normal wake cycle. CONCLUSIONS: The natural tolerance of the rabbit for restraint makes it a valuable subject for MRI studies of the brain.

MRI-guided interventional natural killer cell delivery for liver tumor treatment.

While natural killer (NK) cell-based adoptive transfer immunotherapy (ATI) provides only modest clinical success in cancer patients. This study was hypothesized that MRI-guided transcatheter intra-hepatic arterial (IHA) infusion permits local delivery to liver tumors to improve outcomes during NK-based ATI in a rat model of hepatocellular carcinoma (HCC). Mouse NK cells were labeled with clinically applicable iron nanocomplexes. Twenty rat HCC models were assigned to three groups: transcatheter IHA saline infusion as the control group, transcatheter IHA NK infusion group, and intravenous (IV) NK infusion group. MRI studies were performed at baseline and at 24 h, 48 h, and 8 days postinfusion. There was a significant difference in tumor R2* values between baseline and 24 h following the selective transcatheter IHA NK delivery to the tumors (P = 0.039) when compared to IV NK infusion (P = 0.803). At 8 days postinfusion, there were significant differences in tumor volumes between the control, IV, and IHA NK infusion groups (control vs. IV, P = 0.196; control vs. IHA, P < 0.001; and IV vs. IHA, P = 0.001). Moreover, there was a strong correlation between tumor R2* value change (∆R2*) at 24 h postinfusion and tumor volume change (∆volume) at 8 days in IHA group (R2  = 0.704, P < 0.001). Clinically applicable labeled NK cells with 12-h labeling time can be tracked by MRI. Transcatheter IHA infusion improves NK cell homing efficacy and immunotherapeutic efficiency. The change in tumor R2* value 24 h postinfusion is an important early biomarker for prediction of longitudinal response.