Ultrafast Ultrasound Imaging for Super-Resolution Preclinical Cardiac PET

Purpose

Physiological motion and partial volume effect (PVE) significantly degrade the quality of cardiac positron emission tomography (PET) images in the fast-beating hearts of rodents. Several Super-resolution (SR) techniques using a priori anatomical information have been proposed to correct motion and PVE in PET images. Ultrasound is ideally suited to capture real-time high-resolution cine images of rodent hearts. Here, we evaluated an ultrasound-based SR method using simultaneously acquired and co-registered PET-CT-Ultrafast Ultrasound Imaging (UUI) of the beating heart in closed-chest rodents.

Procedures

The method was tested with numerical and animal data (n?=?2) acquired with the non-invasive hybrid imaging system PETRUS that acquires simultaneously PET, CT, and UUI.

Results

We showed that ultrasound-based SR drastically enhances the quality of PET images of the beating rodent heart. For the simulations, the deviations between expected and mean reconstructed values were 2 % after applying SR. For the experimental data, when using Ultrasound-based SR correction, contrast was improved by a factor of two, signal-to-noise ratio by 11 %, and spatial resolution by 56 % (~?0.88 mm) with respect to static PET. As a consequence, the metabolic defect following an acute cardiac ischemia was delineated with much higher anatomical precision.

Conclusions

Our results provided a proof-of-concept that image quality of cardiac PET in fast-beating rodent hearts can be significantly improved by ultrasound-based SR, a portable low-cost technique. Improved PET imaging of the rodent heart may allow new explorations of physiological and pathological situations related with cardiac metabolism.

     

VEGFR2-Targeted Molecular Imaging in the Mouse Embryo: An Alternative to the Tumor Model

As a tumor surrogate, the mouse embryo presents as an excellent alternative for examining the binding of angiogenesis-targeting microbubbles and assessing the quantitative nature of molecular ultrasound. We establish the validity of this model by developing a robust method to study microbubble kinetic behavior and investigate the reproducibility of targeted binding in the murine embryo. Vascular endothelial growth factor receptor 2 (VEGFR2)-targeted (MB_V), rat immunoglobulin G₂ (IgG₂) control antibody-targeted (MB_C) and untargeted (MB_U) microbubbles were introduced into vasculature of living mouse embryos. Non-linear contrast-specific and B-mode ultrasound imaging, performed at 21 MHz with a Vevo-2100 scanner, was used to collect basic perfusion parameters and contrast mean power ratios for all bubble types. We observed a twofold increase (p < 0.001) in contrast mean power ratios for MB_V (4.14 ± 1.78) compared with those for MB_C (1.95 ± 0.78) and MB_U (1.79 ± 0.45). Targeted imaging of endogenous endothelial cell surface markers in mouse embryos is possible with labeled microbubbles. The mouse embryo thus presents as a versatile model for testing the performance of ultrasound molecular targeting, where further development of quantitative imaging techniques may enable rapid evaluations of biomarker expression in studies of vascular development, disease and angiogenesis.     

Longitudinal Molecular Imaging with Single Cell Resolution of Disseminated Ovarian Cancer in Mice with a LED-based Confocal Microendoscope

Purpose  

We engineered a flexible fiber-optic microendoscope for longitudinal optical imaging studies in a mouse model of disseminated ovarian cancer.     

Diffraction-enhanced Synchrotron Imaging of Bovine Ovaries Ex Vivo

ObjectivesThe objective of this study was to test the hypothesis that diffraction-enhanced imaging (DEI), a synchrotron x-ray imaging technique, would provide greater contrast for evaluating bovine ovaries compared with conventional diagnostic ultrasonography.Materials and MethodsBovine ovaries were evaluated ex vivo as follows: fresh without radiographic arterial contrast (n = 2), fresh with contrast (n = 1), preserved in 10% formalin without contrast (n = 2), and preserved with contrast (n = 1). Each ovary was imaged with DEI and subsequently with ultrasonography and histology. The ability to visualize and differentiate preantral and antral follicles, corpora lutea (CL), and cumulus oocyte complexes (COCs) were compared using DEI, ultrasonography, and histology. The diameter of follicles and CL were measured and compared using ultrasonography, DEI, and histology. The diameter of the smallest follicle detected was reported using each of the three imaging methods. The number of antral follicles (antral follicle count ≥2 mm) was compared between ultrasonography and DEI.ResultsDEI enabled the detection of 71% of follicles and 67% of CL that were detected ultrasonographically. However, DEI did not allow the detection of COCs and cell layers of the follicle wall that were visualized histologically. Luteal tissues were not easily distinguished using DEI, and DEI was inferior for differentiating follicles and CL compared with ultrasonography. The mean follicle diameter was similar between DEI (4.00 ± 0.35 mm, fresh with contrast; 9.62 ± 2.43 mm, fresh without contrast) and ultrasonography (3.85 ± 0.28 mm, fresh with contrast; 8.97 ± 2.60 mm, fresh without contrast) (P > .05). However, the mean follicle diameter was greater using both DEI (4.00 ± 0.35 mm) and ultrasonography (3.85 ± 0.28 mm) compared with histology (2.21 ± 0.38 mm; P = .01, fresh ovaries with contrast). The mean CL diameter was similar between DEI (11.64 ± 1.67 mm), ultrasonography (9.34 ± 0.35 mm), and histology (9.59 ± 0.36 mm) (P > .05). The mean diameter of the smallest follicle detected was similar between DEI (3.06 ± 0.45 mm) and ultrasonography (2.95 ± 0.74 mm); both DEI and ultrasonographic measurements were greater than histology (0.39 ± 0.04 mm, P < .0001). The mean antral follicle count was similar between ultrasonography (6.50 ± 0.71 mm, fresh with no contrast; 6.50 ± 2.50 mm, preserved with no contrast) and DEI (4.50 ± 0.50 mm, fresh with no contrast; 6.50 ± 0.50 mm, preserved with no contrast) (P > .05).ConclusionsThe contrast resolution of antral follicles, CL, and COCs in bovine ovaries was inferior using DEI compared with ultrasonography and histology. Alternative synchrotron techniques, such as phase-contrast computed tomography and DEI computed tomography, may prove more effective than DEI for imaging ovaries ex vivo.     

Tibial Nerve Excursion During Lumbar Spine and Hip Flexion Measured with Diagnostic Ultrasound

Ultrasound imaging provides a method for non-invasive in vivo measurement of nerve motion resulting from joint movement. This study measured the proximal excursion of the tibial branch of the sciatic nerve at the popliteal fossa during forward bending in healthy subjects. Long-axis image sequences of the nerve were analysed using frame-by-frame cross-correlation software that calculated the longitudinal and axial movement of the nerve. Proximal excursion was calculated from the hypotenuse of these values. The mean proximal excursion recorded was 12.2 mm (SD 2.2 mm, n = 24). The reliability of three repeat measurements was found to be excellent (ICC 0.97, 95% CI 0.93–0.99; SEM 0.7 mm, n = 21). The protocol described provides a reliable method for analysing tibial nerve movement that could prove useful in future clinical studies.     

An In Vivo Multimodal Imaging Study Using MRI and PET of Stem Cell Transplantation after Myocardial Infarction in Rats

Purpose  The purpose of the study is to track iron-oxide nanoparticle-labelled adult rat bone marrow-derived stem cells (IO-rBMSCs) by magnetic resonance imaging (MRI) and determine their effect in host cardiac tissue using 2-deoxy-2-[F-18]fluoro-d-glucose-positron emission tomography (FDG-PET). Procedures  Infarcted rats were randomised to receive (1) live IO-rBMSCs by direct local injection, or (2) dead IO-rBMSCs as controls; (3) sham-operated rats received live IO-rBMSCs. The rats were then imaged from 2 days to 6 weeks post-cell implantation using both MRI at 9.4T and FDG-PET. Results  Implanted IO-rBMSCs were visible in the heart by MRI for the duration of the study. Histological analysis confirmed that the implanted IO-rBMSCs were present for up to 6 weeks post-implantation. At 1 week post-IO-rBMSC transplantation, PET studies demonstrated an increase in FDG uptake in infarcted regions implanted with live IO-rBMSC compared to controls. Conclusions  Noninvasive multimodality imaging allowed us to visualise IO-rBMSCs and establish their affect on cardiac function in a rat model of myocardial infarction (MI).     

Subharmonic,Non-linear Fundamental and Ultraharmonic Imaging of Microbubble Contrast at High Frequencies

There is increasing use of ultrasound contrast agent in high-frequency ultrasound imaging. However, conventional contrast detection methods perform poorly at high frequencies. We performed systematic in vitro comparisons of subharmonic, non-linear fundamental and ultraharmonic imaging for different depths and ultrasound contrast agent concentrations (Vevo 2100 system with MS250 probe and MicroMarker ultrasound contrast agent, VisualSonics, Toronto, ON, Canada). We investigated 4-, 6- and 10-cycle bursts at three power levels with the following pulse sequences: B-mode, amplitude modulation, pulse inversion and combined pulse inversion/amplitude modulation. The contrast-to-tissue (CTR) and contrast-to-artifact (CAR) ratios were calculated. At a depth of 8 mm, subharmonic pulse-inversion imaging performed the best (CTR = 26 dB, CAR = 18 dB) and at 16 mm, non-linear amplitude modulation imaging was the best contrast imaging method (CTR = 10 dB). Ultraharmonic imaging did not result in acceptable CTRs and CARs. The best candidates from the in vitro study were tested in vivo in chicken embryo and mouse models, and the results were in a good agreement with the in vitro findings.     

A New 15–50 MHz Array-Based Micro-Ultrasound Scanner for Preclinical Imaging

Most institutions now have a suite of imaging tools to follow mouse models of human disease. Micro-ultrasound is one of these tools and is second after whole-mouse fluorescence or bioluminescent imaging, in terms of installed systems. We report in this paper the first commercially available array transducer–based ultrasound imaging system that enables micro-ultrasound imaging at center frequencies between 15 and 50 MHz. At the heart of the new scanner is a laser-machined high-frequency 256 element, linear transducer array capable of forming dynamic diffraction limited beams. The power of the linear array approach is embodied in the uniform high resolution maintained over the full field of view. This leads to greatly expanded scope for real-time functional imaging that is demonstrated in this paper. The unprecedented images made with the new imaging system will enable many new applications not previously possible. These include real-time visualization of flow in the mouse placenta, visualization of flow development in the embryo, studies of embryonic to adult cardiac development/disease, and studies of real-time blood flow in mouse models of tumour angiogenesis. (E-mail: Stuart.foster@sunnybrook.ca)     

Transcranial Functional Ultrasound Imaging in Freely Moving Awake Mice and Anesthetized Young Rats without Contrast Agent

Functional ultrasound (fUS) imaging by ultrasensitive Doppler detection of blood volume was previously reported to measure adult rat brain activation and functional connectivity with unmatched spatiotemporal sampling (100 μm, 1 ms), but skull-induced attenuation of ultrasonic waves imposed skull surgery or contrast agent use. Also, fUS feasibility remains to be validated in mice, a major pre-clinical model organism. In the study described here, we performed full-depth ultrasensitive Doppler imaging and 3-D Doppler tomography of the entire mouse brain under anesthesia, non-invasively through the intact skull and skin, without contrast agents. Similar results were obtained in anesthetized young rats up to postnatal day 35, thus enabling longitudinal studies on postnatal brain development. Using a newly developed ultralight ultrasonic probe and an optimized ultrasonic sequence, we also performed minimally invasive full-transcranial fUS imaging of brain vasculature and whisker stimulation-induced barrel cortex activation in awake and freely moving mice, validating transcranial fUS for brain imaging, without anesthesia-induced bias, for behavioral studies.     

Propoxyphene and norpropoxyphene: pharmacologic and toxic effects in animals.

alpha-d-Propoxyphene and its principle metabolite, alpha-d-norpropoxyphene, were compared pharmacologically to establish their relative opioid profiles as defined by naloxone reversal. Propoxyphene exhibited opioid activity in the following tests: mouse abdominal constriction and rat tail heat analgesic tests, inhibition of the twitch of the guinea-pig ileum and acute lethality in rodents. Norpropoxyphene also showed opioid activity in the rat tail heat and guinea-pig ileum tests, but exhibited nonopioid activity in the mouse abdominal constriction and acute toxicity studies. Jumping in mice, precipitated by naloxone, suggests the following order for liability to produce physical dependence after repeated administration: morphine greater than codeine greater than propoxyphene greater than norpropoxyphene approximately saline. Propoxyphene and norpropoxyphene depressed axonal conduction in isolated peripheral nerve and were comparable in potency to standard local anesthetic agents. The nonopioid actions of norpropoxyphene might be due in part to its local anesthetic properties.     

Defining the Optimal Age for Focal Lesioning in a Rat Model of Transcranial HIFU

This study aimed at determining the optimal age group for high-intensity focused ultrasound (HIFU) experiments for producing lesions in rats. Younger rats have thinner skulls, allowing for the acoustic waves to propagate easily through the skull without causing burns of the skin and brain surface. Younger rats however, have a smaller brain that can make HIFU focusing in the brain parenchyma challenging because of the focus size. In this study, we conducted transcranial HIFU sonications in rat pups of different ages (from 9 to 43 d) with a 1.5MHz MR compatible transducer. The electric power was selected to always reach a target temperature of at least 50°C in the parenchyma. The thickness of the skull and of the brain parenchyma was measured using T2-weighted MR imaging. Results showed that the thickness of the brain parenchyma increased quickly from P9 to P12, reaching 8.5 mm at P16, and then increasing gradually along with age. The skull thickness increased gradually from P9 to P26, and then more quickly after P30. The ratio between brain parenchyma thickness and skull thickness decreased gradually with age. For the pups at 30 d, the temperature in the brain tissue adjacent to the skull increased to 48.9°C, and those from the rodents older than 33 d reached 60°C or higher, which can produce undesired irreversible damage in this location. We conclude that young rats aged 16–26 d are optimal for experiments producing transcranial HIFU lesions in rats with an intact skull.     

Small Animal Absorbed Radiation Dose from Serial Micro-Computed Tomography Imaging

Purpose To determine the radiation dose to mouse cancer xenografts from serial micro-computed tomography (CT) examinations. Procedures A nude mouse with a 15-mm subcutaneous pancreatic cancer xenograft in the rightflank was used. Radiation exposure to the subcutaneous tumor and the mouse pancreas (to simulate an orthotopic pancreatic tumor model) was measured using lithium fluoride thermoluminescent dosimeters. Ultrafast micro-CT was performed using 80 kVp, 0.26 mA, 0.156 mm slice thickness, 256 slices, 0.7 mm Al filtration, and 60-second image acquisition time (15 mA second). Micro-CT imaging acquisitions were repeated four times. Results We measured consistently low tumor doses (0.014 to 0.02 Gy; average = 0.017 Gy) per scan. Orthotopic doses in the region of the pancreas were also consistently low (0.014 to 0.018 Gy; average = 0.016 Gy) per scan. Conclusions Radiation doses delivered during ultrafast micro-CT serial imaging in the mouse are low and are likely below the threshold to affect tumor growth.     

CAN3D: Fast 3D medical image segmentation via compact context aggregation

Direct automatic segmentation of objects in 3D medical imaging, such as magnetic resonance (MR) imaging, is challenging as it often involves accurately identifying multiple individual structures with complex geometries within a large volume under investigation. Most deep learning approaches address these challenges by enhancing their learning capability through a substantial increase in trainable parameters within their models. An increased model complexity will incur high computational costs and large memory requirements unsuitable for real-time implementation on standard clinical workstations, as clinical imaging systems typically have low-end computer hardware with limited memory and CPU resources only. This paper presents a compact convolutional neural network (CAN3D) designed specifically for clinical workstations and allows the segmentation of large 3D Magnetic Resonance (MR) images in real-time. The proposed CAN3D has a shallow memory footprint to reduce the number of model parameters and computer memory required for state-of-the-art performance and maintain data integrity by directly processing large full-size 3D image input volumes with no patches required. The proposed architecture significantly reduces computational costs, especially for inference using the CPU. We also develop a novel loss function with extra shape constraints to improve segmentation accuracy for imbalanced classes in 3D MR images. Compared to state-of-the-art approaches (U-Net3D, improved U-Net3D and V-Net), CAN3D reduced the number of parameters up to two orders of magnitude and achieved much faster inference, up to 5 times when predicting with a standard commercial CPU (instead of GPU). For the open-access OAI-ZIB knee MR dataset, in comparison with manual segmentation, CAN3D achieved Dice coefficient values of (mean = 0.87 ± 0.02 and 0.85 ± 0.04) with mean surface distance errors (mean = 0.36 ± 0.32 mm and 0.29 ± 0.10 mm) for imbalanced classes such as (femoral and tibial) cartilage volumes respectively when training volume-wise under only 12G video memory. Similarly, CAN3D demonstrated high accuracy and efficiency on a pelvis 3D MR imaging dataset for prostate cancer consisting of 211 examinations with expert manual semantic labels (bladder, body, bone, rectum, prostate) now released publicly for scientific use as part of this work.     

Evaluation of a Multi-pinhole Collimator for Imaging Small Animals with Different Sizes

Purpose  

Our goal was to evaluate a multi-pinhole (MPH) collimator which allows changing configurations for mouse imaging and rat imaging.     

Longitudinal two-dimensional strain rate imaging: a potential approach to predict the response to cardiac resynchronization therapy

The purpose of our study was to test the usefulness of speckle-tracking two-dimensional echocardiography (in particular longitudinal strain and strain rate) in predicting the response to cardiac resynchronization therapy. The standard approach has been tissue Doppler-based echocardiographic imaging (TDI) has initially showed promising results in small clinical trials. However, recent larger, prospective randomized clinical trials (PROSPECT, ReTHINK) showed that TDI is inadequate to predict response from CRT in patients with heart failure. Altogether, these data suggest the need to identify alternative echocardiographic parameters to predict the response to CRT. We included 53 patients suffering from heart failure, who received CRT. TDI and two-dimensional speckle tracking imaging in addition to standard echocardiography were performed prior to CRT. The standard deviation of time to peak longitudinal strain in 12 LV segments (Tstrain-SD) and the standard deviation of time to the end of longitudinal systolic strain rate in six basal LV segments (Tsr-SD) were calculated. Standard echocardiography was performed 6 months after CRT. Patients were classified as echocardiographic responders if the LV end-systolic volume was reduced >15% compared with baseline volumes. No significant difference was seen in baseline Ts-SD, and Tstrain-SD between non-responders and responders. However, the Tsr-SD was much higher in responders than non-responders (95.9 ± 33.0% vs. 64.8 ± 39.6%, P < 0.05), and it showed a sensitivity of 73% and specificity of 65% for the defined echocardiographic response using a cutoff value of 70.7 ms. Our study demonstrates that longitudinal two-dimensional strain rate imaging is a promising potential echocardiographic parameter to predict benefit from CRT in patients with heart failure. This hypothesis needs to be further tested in prospective randomized clinical trials.     

Combined Vector Velocity and Spectral Doppler Imaging for Improved Imaging of Complex Blood Flow in the Carotid Arteries

Color flow imaging and pulsed wave (PW) Doppler are important diagnostic tools in the examination of patients with carotid artery disease. However, measurement of the true peak systolic velocity is dependent on sample volume placement and the operator's ability to provide an educated guess of the flow direction. Using plane wave transmissions and a duplex imaging scheme, we present an all-in-one modality that provides both vector velocity and spectral Doppler imaging from one acquisition, in addition to separate B-mode images of sufficient quality. The vector Doppler information was used to provide automatically calibrated (angle-corrected) PW Doppler spectra at every image point. It was demonstrated that the combined information can be used to generate spatial maps of the peak systolic velocity, highlighting regions of high velocity and the extent of the stenotic region, which could be used to automate work flow as well as improve the accuracy of measurement of true peak systolic velocity. The modality was tested in a small group (N = 12) of patients with carotid artery disease. PW Doppler, vector velocity and B-mode images could successfully be obtained from a single recording for all patients with a body mass index ranging from 21 to 31 and a carotid depth ranging from 16 to 28 mm.     

CINE-MR imaging of the normal and infarcted rat heart using an 11.7 T vertical bore MR system.

MR imaging is uniquely placed to non-invasively study rodent cardiac structure and function. High-field MR scanners commonly have a vertical bore, and the purpose of this work was to demonstrate CINE-MR imaging in normal and infarcted rat hearts after determining hemodynamic stability when positioned vertically for imaging. Optimisation of imaging parameters was carried out prior to assessment of cardiac function in a group of normal and infarcted rat hearts. Rat hemodynamics were unaltered when vertical for 90 minutes, compared with horizontal measurements and rat cardiac parameters were measured accurately and reproducibly with our optimized CINE-MR protocol. A flip angle of 17.5 degrees was shown to provide optimal contrast for the assessment of structure and function, and, in contrast to our findings in mice, respiratory gating was not found to be essential. Hence, we conclude that vertical bore MR systems can be used to measure in vivo cardiac function in normal and infarcted rat hearts.     

Rapid and Accurate Assessment of Aortic Arch Atherosclerosis Using Simultaneous Multi-Plane Imaging by Transesophageal Echocardiography

Transesophageal echocardiography (TEE) is widely used for the evaluation of aortic arch atherosclerosis which carries an increased risk of ischemic stroke. We investigated the feasibility of simultaneous multi-plane imaging by real-time 3-D TEE for the assessment of aortic arch plaques. In 152 patients, we assessed aortic arch plaques and measured their maximum thickness by both conventional TEE imaging and multi-plane TEE imaging. There was excellent correlation and good agreement between the two methods in the measurement of the maximum thickness of arch plaques (r = 0.95, mean difference, −0.1 ± 0.5 mm). The mean image acquisition time required for aortic arch assessment by multi-plane imaging was significantly shorter than that required for conventional imaging in all patients (p < 0.001), especially those with complex plaques. These findings suggest that simultaneous multi-plane TEE imaging enables rapid and accurate evaluation of arch plaques and is therefore a useful tool for the assessment of aortic arch plaques in the clinical setting.     

Hyperemic stress myocardial perfusion cardiovascular magnetic resonance in mice at 3 Tesla: initial experience and validation against microspheres

Background

Dynamic first pass contrast-enhanced myocardial perfusion is the standard CMR method for the estimation of myocardial blood flow (MBF) and MBF reserve in man, but it is challenging in rodents because of the high temporal and spatial resolution requirements. Hyperemic first pass myocardial perfusion CMR during vasodilator stress in mice has not been reported.

Methods

Five C57BL/6 J mice were scanned on a clinical 3.0 Tesla Achieva system (Philips Healthcare, Netherlands). Vasodilator stress was induced via a tail vein catheter with an injection of dipyridamole. Dynamic contrast-enhanced perfusion imaging (Gadobutrol 0.1 mmol/kg) was based on a saturation recovery spoiled gradient echo method with 10-fold k-space and time domain undersampling (k-t PCA). One week later the mice underwent repeat anaesthesia and LV injections of fluorescent microspheres at rest and at stress. Microspheres were analysed using confocal microscopy and fluorescence-activated cell sorting.

Results

Mean MBF at rest measured by Fermi-function constrained deconvolution was 4.1 ± 0.5 ml/g/min and increased to 9.6 ± 2.5 ml/g/min during dipyridamole stress (P = 0.005). The myocardial perfusion reserve was 2.4 ± 0.54. The mean count ratio of stress to rest microspheres was 2.4 ± 0.51 using confocal microscopy and 2.6 ± 0.46 using fluorescence. There was good agreement between cardiovascular magnetic resonance CMR and microspheres with no significant difference (P = 0.84).

Conclusion

First-pass myocardial stress perfusion CMR in a mouse model is feasible at 3 Tesla. Rest and stress MBF values were consistent with existing literature and perfusion reserve correlated closely to microsphere analysis. Data were acquired on a 3 Tesla scanner using an approach similar to clinical acquisition protocols, potentially facilitating translation of imaging findings between rodent and human studies.     

Robot-assisted catheter manipulation for intracardiac navigation

Objective  Manual navigation of intracardiac steerable catheters is inaccurate, requires dexterity for efficient manipulation of the catheter, and exposes the interventionalist to ionizing radiation. The objective of this research is to develop a system that replaces the interventionalists’s hands in catheter manipulation for accurate and semi-automatic tele-navigation of catheters. Methods  Based on a proposed kinematic model for the distal shaft of the catheter, a system has been developed for assisted navigation of intracardiac catheters. When the distal shaft of the catheter lies inside a cardiac chamber, a robotic apparatus is utilized for automatic steering of the catheter tip to reach designated targets within the chamber. Results  The catheter modeling was validated through the experiments on three swine. The robotic system could navigate the catheter tip to designated targets with a mean distance of 6.53 mm from the target. Conclusion  Preliminary in vivo studies demonstrate the feasible application of the system in catheter navigation and the validity of catheter modeling and control strategies.