This PDF file contains the front matter associated with SPIE Proceedings Volume 9417, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and Conference Committee listing.

Silicon-based nanoparticles are ideally suited as biomedical imaging agents, due to their biocompatibility, biodegradability, and simple surface chemistry that is amenable to drug loading and targeting. A method of hyperpolarizing silicon particles using dynamic nuclear polarization (DNP), which increases magnetic resonance imaging (MRI) signals by 4-5 orders of magnitude through enhanced nuclear spin alignment, has recently been developed and shown viable as a contrast agent for in vivo MRI. Naturally occurring electronic defects on the particle surface obviate the need for exogenous radicals, and the enhanced spin polarization lasts for significantly longer than other hyperpolarized agents (tens of minutes, instead of <1 minute for other species). We report our recent advances in determining the MR characteristics of hyperpolarized silicon particles, which could lead to non-invasive, non-radioactive molecular targeted imaging of various cancer systems. A variety of particle sizes (20 nm-2 μm) were found to have hyperpolarized relaxation times ranging from ~10-50 minutes. The addition of various functional groups to the particle surface, including biocompatible polymers, aptamers, and antibodies had no effect to the hyperpolarization dynamics or relaxation times, and appear to satisfactorily survive the harsh temperature conditions of DNP. Preliminary in vivo studies examined a variety of particle administration routes in mice, including intraperitoneal, tail vein, and rectal injections, as well as oral gavage. Ongoing experiments include targeted molecular imaging in orthotopic murine models of ovarian and colorectal cancers.

Evaluation of mechanical stiffness imaging biomarkers, through magnetic resonance elastography (MRE), has shown considerable promise for non-invasive assessment of liver stiffness to monitor hepatic fibrosis. MRE typically requires specialized externally-applied vibratory excitation and scanner-specific motion-sensitive pulse sequences. In this work, we have developed an elasticity imaging approach that utilizes natural diaphragmatic respiratory motion to induce deformation and eliminates the need for external deformation excitation hardware and specialized pulse sequences. Our approach uses clinically-available standard of care volumetric imaging acquisitions, combined with offline model-based post-processing to generate volumetric estimates of stiffness within the liver and surrounding tissue structures. We have previously developed a novel methodology for non-invasive elasticity imaging which utilizes a model-based elasticity reconstruction algorithm and MR image volumes acquired under different states of deformation. In prior work, deformation was external applied through inflation of an air bladder placed within the MR radiofrequency coil. In this work, we extend the methodology with the goal of determining the feasibility of assessing liver mechanical stiffness using diaphragmatic respiratory motion between end-inspiration and end-expiration breath-holds as a source of deformation. We present initial investigations towards applying this methodology to assess liver stiffness in healthy volunteers and cirrhotic patients. Our preliminary results suggest that this method is capable of non-invasive image-based assessment of liver stiffness using natural diaphragmatic respiratory motion and provides considerable enthusiasm for extension of our approach towards monitoring liver stiffness in cirrhotic patients with limited impact to standard-of-care clinical imaging acquisition workflow.

Changes in tissue mechanical properties, measured non-invasively by elastography methods, have been shown to be an important diagnostic tool, particularly for cancer. Tissue elasticity information, tracked over the course of therapy, may be an important prognostic indicator of tumor response to treatment. While many elastography techniques exist, this work reports on the use of a novel form of elastography that uses image texture to reconstruct elastic property distributions in tissue (i.e., a modality independent elastography (MIE) method) within the context of a pre-clinical breast cancer system.^1,2 The elasticity results have previously shown good correlation with independent mechanical testing.¹ Furthermore, MIE has been successfully utilized to localize and characterize lesions in both phantom experiments and simulation experiments with clinical data.^2,3 However, the reproducibility of this method has not been characterized in previous work. The goal of this study is to evaluate voxel-level reproducibility of MIE in a pre-clinical model of breast cancer. Bland-Altman analysis of co-registered repeat MIE scans in this preliminary study showed a reproducibility index of 24.7% (scaled to a percent of maximum stiffness) at the voxel level. As opposed to many reports in the magnetic resonance elastography (MRE) literature that speak to reproducibility measures of the bulk organ, these results establish MIE reproducibility at the voxel level; i.e., the reproducibility of locally-defined mechanical property measurements throughout the tumor volume.

Cutaneous (skin) oxygenation is an important prognostic factor for the treatment of chronic wounds, skin cancer, diabetes side effects, and limb amputation. Currently, there are no reliable methods for measuring this parameter. Oximetry, using electron paramagnetic resonance (EPR) spectroscopy, is emerging as a potential tool for clinical oximetry, including cutaneous applications. The problem with EPR oximetry, however, is that the conventional EPR design requires the use of a large magnet that can generate homogeneous field across the sample, making it unattractive for clinical practice. We present a novel approach that makes use of a miniature permanent magnet, combined with a small microwave resonator, to enable the acquisition of EPR signals from paramagnetic species placed on the skin. The instrumentation consists of a hand-held, modular, cylindrical probehead with overall dimensions of 36-mm diameter and 24-mm height, with 150-g weight. The probehead includes a Halbach array of 16 pieces (4×4×8 mm³) of Sm-Co permanent magnet and a loop-gap resonator (2.24 GHz). Preliminary measurements using a Hahn-echo pulse sequence (800 echos in 20 ms) showed a signalto- noise ratio of ~70 compared to ~435 in a homogenous magnet under identical settings. Further work is in progress to improve the performance of the probehead and to optimize the hand-held system for clinical use

Abdominal organ segmentation with clinically acquired computed tomography (CT) is drawing increasing interest in the medical imaging community. Gaussian mixture models (GMM) have been extensively used through medical segmentation, most notably in the brain for cerebrospinal fluid / gray matter / white matter differentiation. Because abdominal CT exhibit strong localized intensity characteristics, GMM have recently been incorporated in multi-stage abdominal segmentation algorithms. In the context of variable abdominal anatomy and rich algorithms, it is difficult to assess the marginal contribution of GMM. Herein, we characterize the efficacy of an a posteriori framework that integrates GMM of organ-wise intensity likelihood with spatial priors from multiple target-specific registered labels. In our study, we first manually labeled 100 CT images. Then, we assigned 40 images to use as training data for constructing target-specific spatial priors and intensity likelihoods. The remaining 60 images were evaluated as test targets for segmenting 12 abdominal organs. The overlap between the true and the automatic segmentations was measured by Dice similarity coefficient (DSC). A median improvement of 145% was achieved by integrating the GMM intensity likelihood against the specific spatial prior. The proposed framework opens the opportunities for abdominal organ segmentation by efficiently using both the spatial and appearance information from the atlases, and creates a benchmark for large-scale automatic abdominal segmentation.

Esophageal wall thickness is an important predictor of esophageal cancer response to therapy. In this study, we developed a computerized pipeline for quantification of esophageal wall thickness using computerized tomography (CT). We first segmented the esophagus using a multi-atlas-based segmentation scheme. The esophagus in each atlas CT was manually segmented to create a label map. Using image registration, all of the atlases were aligned to the imaging space of the target CT. The deformation field from the registration was applied to the label maps to warp them to the target space. A weighted majority-voting label fusion was employed to create the segmentation of esophagus. Finally, we excluded the lumen from the esophagus using a threshold of -600 HU and measured the esophageal wall thickness. The developed method was tested on a dataset of 30 CT scans, including 15 esophageal cancer patients and 15 normal controls. The mean Dice similarity coefficient (DSC) and mean absolute distance (MAD) between the segmented esophagus and the reference standard were employed to evaluate the segmentation results. Our method achieved a mean Dice coefficient of 65.55 ± 10.48% and mean MAD of 1.40 ± 1.31 mm for all the cases. The mean esophageal wall thickness of cancer patients and normal controls was 6.35 ± 1.19 mm and 6.03 ± 0.51 mm, respectively. We conclude that the proposed method can perform quantitative analysis of esophageal wall thickness and would be useful for tumor detection and tumor response evaluation of esophageal cancer.

The diaphragm is a sheet of muscle which separates the thorax from the abdomen and it acts as the most important muscle of the respiratory system. As such, an accurate segmentation of the diaphragm, not only provides key information for functional analysis of the respiratory system, but also can be used for locating other abdominal organs such as the liver. However, diaphragm segmentation is extremely challenging in non-contrast CT images due to the diaphragm's similar appearance to other abdominal organs. In this paper, we present a fully automatic algorithm for diaphragm segmentation in non-contrast CT images. The method is mainly based on a priori knowledge about the human diaphragm anatomy. The diaphragm domes are in contact with the lungs and the heart while its circumference runs along the lumbar vertebrae of the spine as well as the inferior border of the ribs and sternum. As such, the diaphragm can be delineated by segmentation of these organs followed by connecting relevant parts of their outline properly. More specifically, the bottom surface of the lungs and heart, the spine borders and the ribs are delineated, leading to a set of scattered points which represent the diaphragm's geometry. Next, a B-spline filter is used to find the smoothest surface which pass through these points. This algorithm was tested on a noncontrast CT image of a lung cancer patient. The results indicate that there is an average Hausdorff distance of 2.96 mm between the automatic and manually segmented diaphragms which implies a favourable accuracy.

Brown adipose tissue (BAT) is a small but significant tissue, which may play an important role in obesity and the pathogenesis of metabolic syndrome. Interest in studying BAT in adult humans is increasing, but in order to quantify BAT volume in a single measurement or to detect changes in BAT over the time course of a longitudinal experiment, BAT needs to first be reliably differentiated from surrounding tissue. Although the uptake of the radiotracer ¹⁸F-Fluorodeoxyglucose (¹⁸F-FDG) in adipose tissue on positron emission tomography (PET) scans following cold exposure is accepted as an indication of BAT, it is not a definitive indicator, and to date there exists no standardized method for segmenting BAT. Consequently, there is a strong need for robust automatic classification of BAT based on properties measured with biomedical imaging. In this study we begin the process of developing an automated segmentation method based on properties obtained from fat-water MRI and PET-CT scans acquired on ten healthy adult subjects.

Magnetic resonance imaging (MRI) of hyperpolarized (HP) agents has the potential to probe in-vivo metabolism with sensitivity and specificity that was not previously possible. Biological conversion of HP agents specifically for cancer has been shown to correlate to presence of disease, stage and response to therapy. For such metabolic biomarkers derived from MRI of hyperpolarized agents to be clinically impactful, they need to be validated and well characterized. However, imaging of HP substrates is distinct from conventional MRI, due to the non-renewable nature of transient HP magnetization. Moreover, due to current practical limitations in generation and evolution of hyperpolarized agents, it is not feasible to fully experimentally characterize measurement and processing strategies. In this work we use a custom Bloch-McConnell simulator with pharmacokinetic modeling to characterize the performance of specific magnetic resonance spectroscopy sequences over a range of biological conditions. We performed numerical simulations to evaluate the effect of sequence parameters over a range of chemical conversion rates. Each simulation was analyzed repeatedly with the addition of noise in order to determine the accuracy and reproducibility of measurements. Results indicate that under both closed and perfused conditions, acquisition parameters can affect measurements in a tissue dependent manner, suggesting that great care needs to be taken when designing studies involving hyperpolarized agents. More modeling studies will be needed to determine what effect sequence parameters have on more advanced acquisitions and processing methods.

In obesity, fat-water MRI (FWMRI) methods provide valuable information about adipose tissue (AT) distribution. AT is known to undergo complex metabolic and endocrine changes in association with chronic inflammation including iron overloading. Here, we investigate the potential for FWMRI parameters (fat signal fraction (FSF), local magnetic field offset, and T₂*) to be sensitive to AT inflammatory changes in an established diet-induced obesity mouse model. Male C57BL/6J mice were placed on a low fat (LFD) or a high fat diet (HFD). 3D multi- gradient-echo MRI at 15.2T was performed at baseline, 4, 8, 12, and 16 weeks after diet onset. A 3D fat-water separation algorithm and additional processing was used to generate FSF, local field offset, and T₂* maps. We examined these parameters in perirenal AT ROIs from HFD and LFD mice. Results: The data suggest that FSF, local field offset, and T₂* can differentiate time course behavior between inflamed and control AT (increasing FSF, decreasing local field offset, increasing followed by decreasing T₂*). The biophysical mechanisms of these observed changes are not well understood and require further study. To the best of our knowledge, we report the first evidence that FWMRI can provide biomarkers sensitive to AT inflammation, and that FWMRI has the potential for longitudinal non-invasive assessment of AT inflammation in obesity.

The neuroprotective effect of oxygen leads to recent interest in normobaric oxygen (NBO) therapy after acute ischemic stroke. However, the mechanism remains unclear and inconsistent outcomes were reported in human studies. Because NBO aims to improve brain tissue oxygenation by enhancing oxygen delivery to ischemic tissue, monitoring the oxygenation level changes in response to NBO becomes necessary to elucidate the mechanism and to assess the efficacy. Susceptibility weighted imaging (SWI) which provides a new MRI contrast by combining the magnitude and phase images is fit for purpose. SWI is sensitive to deoxyhemoglobin level changes and thus can be used to evaluate the cerebral metabolic rate of oxygen. In this study, SWI was used for in vivo monitoring of oxygenation changes in a rat model of permanent middle cerebral artery occlusion (MCAO) before, during and after 30 min of NBO treatment. Regions of interest in ischemic core, penumbra and contralateral normal area were generated based on diffusionweighted imaging and perfusion imaging. Significant differences in SWI indicating different oxygenation levels were generally found: contralateral normal > penumbra > ischemic core. Ischemic core showed insignificant increase in oxygenation during NBO and returned to pre-treatment level after termination of NBO. Meanwhile, the oxygenation levels slightly increased in contralateral normal and penumbra regions during NBO and significantly decreased to a level lower than pre-treatment after termination of NBO, indicating secondary metabolic disruption upon the termination of transient metabolic support from oxygen. Further investigation of NBO effect combined with reperfusion is necessary while SWI can be used to detect hemorrhagic transformation after reperfusion.

Amide proton transfer (APT) imaging is a specific form of chemical exchange saturation transfer (CEST) MRI that probes the pH-dependent amide proton exchange.The endogenous APT MRI is sensitive to tissue acidosis, which may complement the commonly used perfusion and diffusion scans for characterizing heterogeneous ischemic tissue damage. Whereas the saturation transfer asymmetry analysis (MTR_asym) may reasonably compensate for direct RF saturation, in vivo MTR_asym is however, susceptible to an intrinsically asymmetric shift (MTR’_asym). Specifically, the reference scan for the endogenous APT MRI is 7 ppm upfield from that of the label scan, and subjects to concomitant RF irradiation effects, including nuclear overhauser effect (NOE)-mediated saturation transfer and semisolid macromolecular magnetization transfer. As such, the commonly used asymmetry analysis could not fully compensate for such slightly asymmetric concomitant RF irradiation effects, and MTR_asym has to be delineated in order to properly characterize the pH-weighted APT MRI contrast. Given that there is very little change in relaxation time immediately after ischemia and the concomitant RF irradiation effects only minimally depends on pH, the APT contrast can be obtained as the difference of MTR_asym between the normal and ischemic regions. Thereby, the endogenous amide proton concentration and exchange rate can be solved using a dual 2-pool model, and the in vivo MTR’_asym can be calculated by subtracting the solved APT contrast from asymmetry analysis (i.e., MTR’_asym =MTR_asym-APTR). In addition, MTR’_asym can be quantified using the classical 2-pool exchange model. In sum, our study delineated the conventional in vivo pH-sensitive MTR_asym contrast so that pHspecific contrast can be obtained for imaging ischemic tissue acidosis.

Tissue oximetry studies using magnetic resonance imaging are increasingly contributing to advances in the imaging and treatment of cancer. The non-invasive measurement of tissue oxygenation (pO₂) may facilitate a better understanding of the pathophysiology and prognosis of diseases, particularly in the assessment of the extensive hypoxic regions associated with cancerous lesions. The availability of tumor hypoxia maps could help quantify and predict tumor response to intervention and therapy. The PISTOL (Proton Imaging of Siloxanes to map Tissue Oxygenation Levels) oximetry technique maps the T₁ of administered hexamethyldisiloxane (HMDSO), an ¹H NMR pO₂ reporter molecule in about 3 ½ min. This allows us to subsequently monitor static and dynamic changes in the tissue pO₂ (in response to intervention) at various locations due to the linear relationship between 1/T₁ and pO₂. In this work, an HMDSO-selective Look-Locker imaging sequence with EPI readout has been developed to enable faster PISTOL acquisitions. The new sequence incorporates the fast Look-Locker measurement method to enable T₁, and hence, pO₂ mapping of HMDSO in under one minute. To demonstrate the application of this pulse sequence in vivo, 50 μL of neat HMDSO was administered to the thigh muscle of a healthy rat (Fischer F344, n=4). Dynamic changes in the mean pO₂ of the thigh muscle were measured using both PISTOL and the developed LL oximetry sequence in response to oxygen challenge and compared. Results demonstrate the efficacy of the new sequence in rapidly mapping the pO₂ changes, leading to advances in fast quantitative ¹H MR oximetry.

The task of imaging is to gather spatiotemporal information which can be organized into a coherent map. Tomographic imaging in particular involves the use of multiple projections, or other interactions of a probe (light, sound, etc.) with a body, in order to determine cross-sectional information. Though the probes and the corresponding imaging modalities may vary, and though the methodology of particular imaging approaches is in constant ferment, the conceptual underpinnings of tomographic imaging have in many ways remained fixed for many decades. Recent advances in applied mathematics, however, have begun to roil this intellectual landscape. The advent of compressed sensing, anticipated in various algorithms dating back many years but unleashed in full theoretical force in the last decade, has changed the way imagers have begun to think about data acquisition and image reconstruction. The power of incoherent sampling and sparsity-enforcing reconstruction has been demonstrated in various contexts and, when combined with other modern fast imaging techniques, has enabled unprecedented increases in imaging efficiency. Perhaps more importantly, however, such approaches have spurred a shift in perspective, prompting us to focus less on nominal data sufficiency than on information content. Beginning with examples from MRI, then proceeding through selected other modalities such as CT and PET, as well as multimodality combinations, this paper explores the potential of newly evolving acquisition and reconstruction paradigms to change the way we do imaging in the lab and in the clinic.

CT plays an important role in the diagnosis of acute stroke patients. Dynamic contrast enhanced CT (DCE-CT) can estimate local tissue perfusion and extent of ischemia. However, hemodynamic information of the large intracranial vessels may also be obtained from DCE-CT data and may contain valuable diagnostic information. We describe a novel method to estimate intravascular blood velocity (IBV) in large cerebral vessels using DCE-CT data, which may be useful to help predict stroke outcome. DCE-CT scans from 34 patients with isolated M1 occlusions were included from a large prospective multi-center cohort study of patients with acute ischemic stroke. Gaussians fitted to the intravascular data yielded the time-to-peak (TTP) and cerebral-blood-volume (CBV). IBV was computed by taking the inverse of the TTP gradient magnitude. Voxels with a CBV of at least 10% of the CBV found in the arterial input function were considered part of a vessel. Mid-sagittal planes were drawn manually and averages of the IBV over all vessel-voxels (arterial and venous) were computed for each hemisphere. Mean-hemisphere IBV differences, mean-hemisphere TTP differences, and hemisphere vessel volume differences were used to differentiate between patients with good and bad outcome (modified Rankin Scale score <3 versus ≥3 at 90 days) using ROC analysis. AUCs from the ROC for IBV, TTP, and vessel volume were 0.80, 0.67 and 0.62 respectively. In conclusion, IBV was found to be a better predictor of patient outcome than the parameters used to compute it and may be a promising new parameter for stroke outcome prediction.

In PET brain imaging, patient motion can contribute significantly to the degradation of image quality potentially leading to diagnostic and therapeutic problems. To mitigate the image artifacts resulting from patient motion, motion must be detected and tracked then provided to a motion correction algorithm. Existing techniques to track patient motion fall into one of two categories: 1) image-derived approaches and 2) external motion tracking (EMT). Typical EMT requires patients to have markers in a known pattern on a rigid too attached to their head, which are then tracked by expensive and bulky motion tracking camera systems or stereo cameras. This has made marker-based EMT unattractive for routine clinical application. Our main contributions are the development of a marker-less motion tracking system that uses lowcost, small depth-sensing cameras which can be installed in the bore of the imaging system. Our motion tracking system does not require anything to be attached to the patient and can track the rigid transformation (6-degrees of freedom) of the patient’s head at a rate 60 Hz. We show that our method can not only be used in with Multi-frame Acquisition (MAF) PET motion correction, but precise timing can be employed to determine only the necessary frames needed for correction. This can speeds up reconstruction by eliminating the unnecessary subdivision of frames.

Functional magnetic resonance imaging allows to assess large scale functional integration of the brain. One of the leading techniques to extract functionally relevant networks is spatial independent component analysis (ICA). Spatial ICA separates independent spatial sources, many of whom are noise or imaging artifacts, whereas some do correspond to functionally relevant Spontaneous co-Activity Maps (SAMs). For research purposes, ICA is generally performed on group data. This strategy is well adapted to uncover commonly shared networks, e.g. resting-state networks, but fails to capture idiosyncratic functional networks which may be related to pathological activity, e.g. epilepsy, hallucinations. To capture these subject specific networks, ICA has to be applied to single subjects using a large number of components, from which a tenth are SAMs. Up to now, SAMs have to be selected manually by an expert based on predefined criteria. We aim to semi-automate the selection process in order to save time. To this end, some approaches have been proposed but none with the near 100 % sensitivity required for clinical purposes. In this paper, we propose a computerized version of the SAM's criteria used by experts, based on frequential and spatial characteristics of functional networks. Here we present a pre-selection method and its results at different resolutions, with different scanners or imaging sequences. While preserving a near 100 % sensitivity, it allows an average of 70 % reduction of components to be classified which save 55% of experts' time. In comparison, group ICA fails to detect about 25% of the SAMs.

The functional magnetic resonance imaging (fMRI) researches on working memory have found that activation of cortical areas appeared dependent on memory load, and event-related potentials (ERP) studies have demonstrated that amplitudes of P300 decreased significantly when working memory load increased. However, the cortical activities related with P300 amplitudes under different memory loads remains unclear. Joint fMRI and EEG analysis which fusions the time and spatial information in simultaneous EEG-fMRI recording can reveal the regional activation at each ERP time point. In this paper, we first used wavelet transform to obtain the single-trial amplitudes of P300 caused by a digital N-back task in the simultaneous EEG-fMRI recording as the ERP feature sequences. Then the feature sequences in 1-back condition and 3-back condition were introduced into general linear model (GLM) separately as parametric modulations to compare the cortical activation under different memory loads. The results showed that the average amplitudes of P300 in 3-back significantly decreased than that in 1-back, and the activities induced by ERP feature sequences in 3-back also significantly decreased than that in the 1-back, including the insular, anterior cingulate cortex, right inferior frontal gyrus, and medial frontal gyrus, which were relevant to the storage, monitoring, and manipulation of information in working memory task. Moreover, the difference in the activation caused by ERP feature showed a positive correlation with the difference in behavioral performance. These findings demonstrated the locations of P300 amplitudes differences modulated by the memory load and its relationship with the behavioral performance.

We explore a computational framework for functional connectivity analysis in resting-state functional MRI (fMRI) data acquired from the human brain for recovering the underlying network structure and understanding causality between network components. Termed mutual connectivity analysis (MCA), this framework involves two steps, the first of which is to evaluate the pair-wise cross-prediction performance between fMRI pixel time series within the brain. In a second step, the underlying network structure is subsequently recovered from the affinity matrix using non-metric network clustering approaches, such as the so-called Louvain method. Finally, we use convergent cross-mapping (CCM) to study causality between different network components. We demonstrate our MCA framework in the problem of recovering the motor cortex network associated with hand movement from resting state fMRI data. Results are compared with a ground truth of active motor cortex regions as identified by a task-based fMRI sequence involving a finger-tapping stimulation experiment. Our results regarding causation between regions of the motor cortex revealed a significant directional variability and were not readily interpretable in a consistent manner across subjects. However, our results on whole-slice fMRI analysis demonstrate that MCA-based model-free recovery of regions associated with the primary motor cortex and supplementary motor area are in close agreement with localization of similar regions achieved with a task-based fMRI acquisition. Thus, we conclude that our MCA methodology can extract and visualize valuable information concerning the underlying network structure between different regions of the brain in resting state fMRI.

Understanding when and how resting state brain functional activity begins in the human brain is an increasing area of interest in both basic neuroscience and in the clinical evaluation of the brain during pregnancy and after premature birth. Although fMRI studies have been carried out on pregnant women since the 1990's, reliable mapping of brain function in utero is an extremely challenging problem due to the unconstrained fetal head motion. Recent studies have employed scrubbing to exclude parts of the time series and whole subjects from studies in order to control the confounds of motion. Fundamentally, even after correction of the location of signals due to motion, signal intensity variations are a fundamental limitation, due to coil sensitivity and spin history effects. An alternative technique is to use a more parametric MRI signal derived from multiple echoes that provides a level of independence from basic MRI signal variation. Here we examine the use of R2* mapping combined with slice based multi echo geometric distortion correction for in-utero studies. The challenges for R2* mapping arise from the relatively low signal strength of in-utero data. In this paper we focus on comparing activation detection in-utero using T2W and R2* approaches. We make use a subset of studies with relatively limited motion to compare the activation patterns without the additional confound of significant motion. Results at different gestational ages indicate comparable agreement in many activation patterns when limited motion is present, and the detection of some additional networks in the R2* data, not seen in the T2W results.

Subject motion is a major challenge in functional magnetic resonance imaging studies (fMRI) of the fetal brain and placenta during maternal hyperoxia. We propose a motion correction and volume outlier rejection method for the correction of severe motion artifacts in both fetal brain and placenta. The method is optimized to the experimental design by processing different phases of acquisition separately. It also automatically excludes high-motion volumes and all the missing data are regressed from ROI-averaged signals. The results demonstrate that the proposed method is effective in enhancing motion correction in fetal fMRI without large data loss, compared to traditional motion correction methods.

Microcystic macular edema (MME) is a term used to describe pseudocystic spaces in the inner nuclear layer (INL) of the human retina. It has been noted in multiple sclerosis (MS) as well as a variety of other diseases. The processes that lead to MME formation and their change over time have yet to be explained sufficiently. The low rate at which MME occurs within such diverse patient groups makes the identification and consistent quantification of this pathology important for developing patient-specific prognoses. MME is observed in optical coherence tomography (OCT) scans of the retina as changes in light reflectivity in a pattern suggestive of fluid accumulations called pseudocysts. Pseudocysts can be readily identified in higher signal-to-noise ratio (SNR) images, however pseudocysts can be indistinguishable from noise in lower SNR scans. In this work, we expand upon our earlier MME identification methods on Spectralis OCT scans to handle lower quality Cirrus OCT scans. Our approach uses a random forest classifier, trained on manual segmentation of ten subjects, to automatically detect MME. The algorithm has a true positive rate for MME identification of 0.95 and a Dice score of 0.79. We include a preliminary longitudinal study of three patients over four to five years to explore the longitudinal changes of MME. The patients with relapsing-remitting MS and neuromyelitis optica appear to have dynamic pseudocyst volumes, while the MME volume appears stable in the one patient with primary progressive MS.

Early detection of oral cancer and its curable precursors can improve patient survival and quality of life. Hyperspectral imaging (HSI) holds the potential for noninvasive early detection of oral cancer. The quantification of tissue chromophores by spectral unmixing of hyperspectral images could provide insights for evaluating cancer progression. In this study, non-negative matrix factorization has been applied for decomposing hyperspectral images into physiologically meaningful chromophore concentration maps. The approach has been validated by computer-simulated hyperspectral images and in vivo tumor hyperspectral images from a head and neck cancer animal model.

Vessel branchings are critical vascular locations from the clinical point of view. In these sites, the arterial hemodynamic plays a relevant role in the progression of atherosclerosis, an important vascular pathology. In this paper, a fully automatic approach for the bifurcation classification in human Intravascular Optical Coherence Tomography (IV-OCT) sequences is introduced. Given the lumen contours, the method is capable of labeling the bifurcation slices. A geometric feature extraction was performed and the Forward Regression Orthogonal Least Squares method (FROLS) was applied to analyze the best features and to determine the appropriated weights in a binary classifier. A cross-validation scheme is applied in order to evaluate the performance of the classification approach and the results have shown a sensitivity of 86% and specificity of 92% to FROLS.

Chronic Kidney Disease (CKD) is characterized by a progressive loss in renal function over time. Pathology can provide valuable insights into the progression of CKD by analyzing the status of glomeruli and the uriniferous tubules over time. Optical coherence tomography (OCT) is a new procedure that can analyze the microscopic structure of the kidney in a non-invasive manner. This is especially important because there are significant artifacts associated with excision biopsies and immersion fixation procedures. Recently, we have shown that OCT can provide real time images of kidney microstructure and Doppler OCT (DOCT) can image glomerular renal blood flow in vivo without administrating exogenous contrast agents. In this study, we used OCT to evaluate CKD in a model induced by intravenous Adriamycin injection into Munich-Wistar rats. We evaluated tubular density and tubular diameter from OCT images at several post- Adriamycin induction time points and compared them with conventional light microscopic histological imaging. Proteinurea and serum creatinine were used as physiological markers of the extent of CKD. Preliminary OCT results revealed changes in tubular density due to tubular necrosis and interstitial fibrosis within the first 4 weeks following Adriamycin injection. From week 4 to 8 after Adriamycin induction, changes in tubular density and diameter occurred due to both tubular loss and tubular dilation. The results suggest OCT can provide additional information about kidney histopathology in CKD. DOCT revealed reduced blood flow in some glomeruli probably as a consequence of focal glomerularsclerosis.

Several obturation methods are available today to study the 3D filling of the root canal. There are also several methods capable to evaluate the ability to seal apically the root canals. However, the common methods of investigation are invasive; they also lead to the destruction of the samples. If the sectioning differs slightly from the desired area, the investigation is non-conclusive regarding the micro-leakages. Also, although the use of Cone-Beam Micro Computer Tomography (CBCT) appears to be most promising for endodontic purposes, its effective radiation doses are higher than with conventional intra-oral and panoramic imaging. In contrast, enface (ef) Optical Coherence Tomography (OCT) proves to be efficient for the investigation of material defects of dental restorations, dental materials, and micro-leakage at the interfaces, where the penetration depth depends on the material. Therefore, ef OCT has been proposed in our studies as a potential tool for in vivo endodontic imaging. Twenty five recently extracted human maxillary molars were selected for the study for caries or periodontal reasons. The pulp chambers were completely opened, the dental pulp was removed, and the root canals were shaped. Silver nanoparticles were used in half of the samples in order to increase the scattering of the adhesive material in comparison with the dental roots walls. The sample teeth were then probed using Time Domain (TD) OCT working at 1300 nm. A synchrotron radiation X-Ray microCT experiment was also performed. The imagistic results pointed out the efficiency of the silver nanoparticle layer used in order to increase the scattering of the root canal adhesive scattering for the OCT non-invasive investigation. MicroCT allowed for obtaining qualitative data related to the depth penetration of the root canal adhesive into the dentin walls.

Perfusion imaging is the most applied modality for the assessment of acute stroke. Parameters such as Cerebral Blood Flow (CBF), Cerebral Blood volume (CBV) and Mean Transit Time (MTT) are used to distinguish the tissue infarct core and ischemic penumbra. Due to lack of standardization these parameters vary significantly between vendors and software even when provided with the same data set. There is a critical need to standardize the systems and make them more reliable. We have designed a uniform phantom to test and verify the perfusion systems. We implemented a flow loop with different flow rates (250, 300, 350 ml/min) and injected the same amount of contrast. The images of the phantom were acquired using a Digital Angiographic system. Since this phantom is uniform, projection images obtained using DSA is sufficient for initial validation. To validate the phantom we measured the contrast concentration at three regions of interest (arterial input, venous output, perfused area) and derived time density curves (TDC). We then calculated the maximum slope, area under the TDCs and flow. The maximum slope calculations were linearly increasing with increase in flow rate, the area under the curve decreases with increase in flow rate. There was 25% error between the calculated flow and measured flow. The derived TDCs were clinically relevant and the calculated flow, maximum slope and areas under the curve were sensitive to the measured flow. We have created a systematic way to calibrate existing perfusion systems and assess their reliability.

Digital Subtraction Angiography (DSA) is the main diagnostic tool for intracranial aneurysms (IA) flow-diverter (FD) assisted treatment. Based on qualitative contrast flow evaluation, interventionists decide on subsequent steps. We developed a novel fully Retrievable Asymmetric Flow-Diverter (RAFD) which allows controlled deployment, repositioning and detachment achieve optimal flow diversion. The device has a small low porosity or solid region which is placed such that it would achieve maximum aneurysmal in-jet flow deflection with minimum impairment to adjacent vessels. We tested the new RAFD using a flow-loop with an idealized and a patient specific IA phantom in carotid-relevant physiological conditions. We positioned the deflection region at three locations: distally, center and proximally to the aneurysm orifice and analyzed aneurysm dome flow using DSA derived maps for mean transit time (MTT) and bolus arrival times (BAT). Comparison between treated and untreated (control) maps quantified the RAFD positioning effect. Average MTT, related to contrast presence in the aneurysm dome increased, indicating flow decoupling between the aneurysm and parent artery. Maximum effect was observed in the center and proximal position (~75%) of aneurysm models depending on their geometry. BAT maps, correlated well with inflow jet direction and magnitude. Reduction and jet dispersion as high as about 50% was observed for various treatments. We demonstrated the use of DSA data to guide the placement of the RAFD and showed that optimum flow diversion within the aneurysm dome is feasible. This could lead to more effective and a safer IA treatment using FDs.

Factor analysis of dynamic structures (FADS) is a methodology of extracting time-activity curves (TACs) for corresponding different tissue types from noisy dynamic images. The challenges of FADS include long computation time and sensitivity to the initial guess, resulting in convergence to local minima far from the true solution. We propose a method of accelerating and stabilizing FADS application to sequences of dynamic PET images by adding preliminary cluster analysis of the time activity curves for individual voxels. We treat the temporal variation of individual voxel concentrations as a set of time-series and use a partial clustering analysis to identify the types of voxel TACs that are most functionally distinct from each other. These TACs provide a good initial guess for the temporal factors for subsequent FADS processing. Applying this approach to a set of single slices of dynamic ¹¹C-PIB images of the brain allows identification of the arterial input function and two different tissue TACs that are likely to correspond to the specific and non-specific tracer binding-tissue types. These results enable us to perform direct classification of tissues based on their pharmacokinetic properties in dynamic PET without relying on a compartment-based kinetic model, without identification of the reference region, or without using any external methods of estimating the arterial input function, as needed in some techniques.

Myocardial CT perfusion (CTP) imaging is an application that should greatly benefit from spectral CT through the significant reduction of beam hardening (BH) artifacts using mono-energetic (monoE) image reconstructions. We used a prototype spectral detector CT (SDCT) scanner (Philips Healthcare) and developed advanced processing tools (registration, segmentation, and deconvolution-based flow estimation) for quantitative myocardial CTP in a porcine ischemia model with different degrees of coronary occlusion using a balloon catheter. The occlusion severity was adjusted with fractional flow reserve (FFR) measurements. The SDCT scanner is a single source, dual-layer detector system, which allows simultaneous acquisitions of low and high energy projections, hence enabling accurate projection-based material decomposition and effective reduction of BH-artifacts. In addition, the SDCT scanner eliminates partial scan artifacts with fast (0.27s), full gantry rotation acquisitions. We acquired CTP data under different hemodynamic conditions and reconstructed conventional 120kVp images and projection-based monoenergetic (monoE) images for energies ranging from 55keV-to-120keV. We computed and compared myocardial blood flow (MBF) between different reconstructions. With balloon completely deflated (FFR=1), we compared the mean attenuation in a myocardial region of interest before iodine arrival and at peak iodine enhancement in the left ventricle (LV), and we found that monoE images at 70keV effectively minimized the difference in attenuation, due to BH, to less than 1 HU compared to 14 HU with conventional 120kVp images. Flow maps under baseline condition (FFR=1) were more uniform throughout the myocardial wall at 70keV, whereas with 120kVp data about 12% reduction in blood flow was noticed on BH-hypoattenuated areas compared to other myocardial regions. We compared MBF maps at different keVs under an ischemic condition (FFR < 0.7), and we found that flow-contrast-to-noise-ratio (CNRf ) between LAD ischemic and remote healthy territories attains its maximum (2.87 ± 0.7) at 70keV. As energies diverge from 70keV, we noticed a steady decrease in CNRf and an overestimation of mean-MBF. Flow overestimation was also noticed for conventional 120kVp images in different myocardial regions.

Dynamic myocardial CT perfusion (CTP) can provide quantitative functional information for the assessment of coronary artery disease. However, x-ray dose in dynamic CTP is high, typically from 10mSv to >20mSv. We compared the dose reduction potential of advanced iterative reconstruction, Iterative Model Reconstruction (IMR, Philips Healthcare, Cleveland, Ohio) to hybrid iterative reconstruction (iDose⁴) and filtered back projection (FBP). Dynamic CTP scans were obtained using a porcine model with balloon-induced ischemia in the left anterior descending coronary artery to prescribed fractional flow reserve values. High dose dynamic CTP scans were acquired at 100kVp/100mAs with effective dose of 23mSv. Low dose scans at 75mAs, 50mAs, and 25mAs were simulated by adding x-ray quantum noise and detector electronic noise to the projection space data. Images were reconstructed with FBP, iDose⁴, and IMR at each dose level. Image quality in static CTP images was assessed by SNR and CNR. Blood flow was obtained using a dynamic CTP analysis pipeline and blood flow image quality was assessed using flow-SNR and flow-CNR. IMR showed highest static image quality according to SNR and CNR. Blood flow in FBP was increasingly over-estimated at reduced dose. Flow was more consistent for iDose⁴ from 100mAs to 50mAs, but was over-estimated at 25mAs. IMR was most consistent from 100mAs to 25mAs. Static images and flow maps for 100mAs FBP, 50mAs iDose⁴, and 25mAs IMR showed comparable, clear ischemia, CNR, and flow-CNR values. These results suggest that IMR can enable dynamic CTP at significantly reduced dose, at 5.8mSv or 25% of the comparable 23mSv FBP protocol.

This study explores changes in metabolic tumor volume, metabolic tumor diameter, and maximum standardized uptake value (SUV_max), for earlier and more accurate identification of lymphomas’ response to treatment using ¹⁸F- FDG PET/CT. Pre- and post-treatment PET/CT studies of 20 patients with Hodgkin disease (HL) and 7 patients with non- Hodgkin lymphoma (NHL) were retrospectively selected for this study. The diameter and volume of the metabolic tumor was determined by an in-house developed adaptive local thresholding technique based on a 50% threshold of the maximum pixel value within a region. Statistical analysis aimed at exploring associations between metabolic size measurements and SUV_max and the ability of the three biomarkers to predict the patients’ response to treatment as defined by the four classes in the European Organization for Research and Treatment of Cancer (EORTC) guidelines. Results indicated moderate correlations between % change in metabolic tumor volume and % change in metabolic tumor maximum diameter (R=0.51) and between % change in maximum diameter and % change in SUV_max (R=0.52). The correlation between % change in tumor volume and % change in SUV_max was weak (R=0.24). The % change in metabolic tumor size, either volume or diameter, was a “very strong” predictor of response to treatment (R=0.89), stronger than SUV_max (R=0.63). In conclusion, metabolic tumor volume could have important prognostic value, possibly higher than maximum metabolic diameter or SUV_max that are currently the standard of practice. Volume measurements, however, should be based on robust and standardized segmentation methodologies to avoid variability. In addition, SUV-peak or lean body mass corrected SUV-peak may be a better PET biomarker than SUV_max when SUV-volume combinations are considered.

The aim of this study was to evaluate if effective radiation dose due to PET component in adult whole-body tumor imaging with time-of-flight F-18 FDG PET/CT could be significantly reduced. We retrospectively analyzed data for 10 patients with the body mass index ranging from 25 to 50. We simulated F-18 FDG dose reduction to 25% of the ACR recommended dose via reconstruction of simulated shorter acquisition time per bed position scans from the acquired list data. F-18 FDG whole-body scans were reconstructed using time-of-flight OSEM algorithm and advanced system modeling. Two groups of images were obtained: group A with a standard dose of F-18 FDG and standard reconstruction parameters and group B with simulated 25% dose and modified reconstruction parameters, respectively. Three nuclear medicine physicians blinded to the simulated activity independently reviewed the images and compared diagnostic quality of images. Based on the input from the physicians, we selected optimal modified reconstruction parameters for group B. In so obtained images, all the lesions observed in the group A were visible in the group B. The tumor SUV values were different in the group A, as compared to group B, respectively. However, no significant differences were reported in the final interpretation of the images from A and B groups.

In conclusion, for a small number of patients, we have demonstrated that F-18 FDG dose reduction to 25% of the ACR recommended dose, accompanied by appropriate modification of the reconstruction parameters provided adequate diagnostic quality of PET images acquired on time-of-flight PET/CT.

Positron emission tomography (PET) using uorodeoxyglucose (18F-FDG) is commonly used in the assessment of breast lesions by computing voxel-wise standardized uptake value (SUV) maps. Simple metrics derived from ensemble properties of SUVs within each identified breast lesion are routinely used for disease diagnosis. The maximum SUV within the lesion (SUVmax) is the most popular of these metrics. However these simple metrics are known to be error-prone and are susceptible to image noise. Finding reliable SUV map-based features that correlate to established molecular phenotypes of breast cancer (viz. estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) expression) will enable non-invasive disease management. This study investigated 36 SUV features based on first and second order statistics, local histograms and texture of segmented lesions to predict ER and PR expression in 51 breast cancer patients. True ER and PR expression was obtained via immunohistochemistry (IHC) of tissue samples from each lesion. A supervised learning, adaptive boosting-support vector machine (AdaBoost-SVM), framework was used to select a subset of features to classify breast lesions into distinct phenotypes. Performance of the trained multi-feature classifier was compared against the baseline single-feature SUVmax classifier using receiver operating characteristic (ROC) curves. Results show that texture features encoding local lesion homogeneity extracted from gray-level co-occurrence matrices are the strongest discriminator of lesion ER expression. In particular, classifiers including these features increased prediction accuracy from 0.75 (baseline) to 0.82 and the area under the ROC curve from 0.64 (baseline) to 0.75.

Morbidity and complexity involved in lymph node staging via surgical resection and biopsy calls for staging techniques that are less invasive. While visible blue dyes are commonly used in locating sentinel lymph nodes, since they follow tumor-draining lymphatic vessels, they do not provide a metric to evaluate presence of cancer. An area of active research is to use fluorescent dyes to assess tumor burden of sentinel and secondary lymph nodes. The goal of this work was to successfully deploy and test an intra-nodal cancer-cell injection model to enable planar fluorescence imaging of a clinically relevant blue dye, specifically methylene blue along with a cancer targeting tracer, Affibody labeled with IRDYE800CW and subsequently segregate tumor-bearing from normal lymph nodes. This direct-injection based tumor model was employed in athymic rats (6 normal, 4 controls, 6 cancer-bearing), where luciferase-expressing breast cancer cells were injected into axillary lymph nodes. Tumor presence in nodes was confirmed by bioluminescence imaging before and after fluorescence imaging. Lymphatic uptake from the injection site (intradermal on forepaw) to lymph node was imaged at approximately 2 frames/minute. Large variability was observed within each cohort.

In the tumor microenvironment, the combination of compromised oxygen supply and high demand results in formation of regions of acute and chronic hypoxia, which promotes metastasis, proliferation, resistance to chemo and radiotherapy and poor prognosis. Targeted, non-invasive in vivo imaging of hypoxia has the potential to determine regions with poor oxygenation in the target and differentiate between normoxic vs hypoxic tissues. MRI provides a powerful platform for generating quantitative maps of hypoxia with the use of a novel pO₂ measuring technique PISTOL (Proton imaging of siloxanes to map tissue oxygenation levels) which could impact the therapeutic choices. In the present study, PISTOL was used to determine the changes in oxygenation of tumor in pre-clinical models of NSCLC (H1975) and epidermoid carcinoma (A431) in response to tirapzamine (TPZ), a hypoxia activated chemotherapeutic. The tumor volume measurements indicate that tirapazamine was more effective in slowing the tumor growth in H1975 as compared to A431 tumors, even though lower baseline pO₂ was observed in A431 as compared to H1975 tumors. These results indicate that other factors such as tumor perfusion (essential for delivering TPZ) and relative expression of nitroreductases (essential for activating TPZ) may play an important role in conjunction with pO₂.

Targeted nanobubbles have been reported to improve the contrast effect of ultrasound imaging due to the enhanced permeation and retention effects at tumor vascular leaks. In this work, the contrast enhancement abilities and the tumor targeting potential of a self-made VEGFR2-targeted nanobubble ultrasound contrast agent was evaluated in-vitro and in-vivo. Size distribution and zeta potential were assessed. Then the contrast-enhanced ultrasound imaging of the VEGFR2 targeted nanobubbles were evaluated with a custom-made experimental apparatus and in normal Wistar rats. Finally, the in-vivo tumor-targeting ability was evaluated on nude mice with subcutaneous tumor. The results showed that the target nanobubbles had uniform distribution with the average diameter of 208.1 nm, polydispersity index (PDI) of 0.411, and zeta potential of -13.21 mV. Significant contrast enhancement was observed in both in-vitro and in-vivo ultrasound imaging, demonstrating that the self-made target nanobubbles can enhance the contrast effect of ultrasound imaging efficiently. Targeted tumor imaging showed less promising result, due to the fact that the targeted nanobubbles arriving and permeating through tumor vessels were not many enough to produce significant enhancement. Future work will focus on exploring new imaging algorithm which is sensitive to targeted nanobubbles, so as to correctly detect the contrast agent, particularly at a low bubble concentration.

Pulmonary x-ray computed tomography (CT) may be used to characterize emphysema and airways disease in patients with chronic obstructive pulmonary disease (COPD). One analysis approach – parametric response mapping (PMR) utilizes registered inspiratory and expiratory CT image volumes and CT-density-histogram thresholds, but there is no consensus regarding the threshold values used, or their clinical meaning. Principal-component-analysis (PCA) of the CT density histogram can be exploited to quantify emphysema using data-driven CT-density-histogram thresholds. Thus, the objective of this proof-of-concept demonstration was to develop a PRM approach using PCA-derived thresholds in COPD patients and ex-smokers without airflow limitation. Methods: Fifteen COPD ex-smokers and 5 normal ex-smokers were evaluated. Thoracic CT images were also acquired at full inspiration and full expiration and these images were non-rigidly co-registered. PCA was performed for the CT density histograms, from which the components with the highest eigenvalues greater than one were summed. Since the values of the principal component curve correlate directly with the variability in the sample, the maximum and minimum points on the curve were used as threshold values for the PCA-adjusted PRM technique. Results: A significant correlation was determined between conventional and PCA-adjusted PRM with ³He MRI apparent diffusion coefficient (p<0.001), with CT RA₉₅₀ (p<0.0001), as well as with ³He MRI ventilation defect percent, a measurement of both small airways disease (p=0.049 and p=0.06, respectively) and emphysema (p=0.02). Conclusions: PRM generated using PCA thresholds of the CT density histogram showed significant correlations with CT and ³He MRI measurements of emphysema, but not airways disease.

Objectives: To develop and evaluate an automated registration and segmentation pipeline for regional lobar pulmonary structure-function measurements, using volume-matched thoracic CT and MRI in order to guide therapy. Methods: Ten subjects underwent pulmonary function tests and volume-matched ¹H and ³He MRI and thoracic CT during a single 2-hr visit. CT was registered to ¹H MRI using an affine method that incorporated block-matching and this was followed by a deformable step using free-form deformation. The resultant deformation field was used to deform the associated CT lobe mask that was generated using commercial software. ³He-¹H image registration used the same two-step registration method and ³He ventilation was segmented using hierarchical k-means clustering. Whole lung and lobar ³He ventilation and ventilation defect percent (VDP) were generated by mapping ventilation defects to CT-defined whole lung and lobe volumes. Target CT-³He registration accuracy was evaluated using region- , surface distance- and volume-based metrics. Automated whole lung and lobar VDP was compared with semi-automated and manual results using paired t-tests. Results: The proposed pipeline yielded regional spatial agreement of 88.0±0.9% and surface distance error of 3.9±0.5 mm. Automated and manual whole lung and lobar ventilation and VDP were not significantly different and they were significantly correlated (r = 0.77, p < 0.0001). Conclusion: The proposed automated pipeline can be used to generate regional pulmonary structural-functional maps with high accuracy and robustness, providing an important tool for image-guided pulmonary interventions.

SWeep Imaging with Fourier Transformation (SWIFT) with gradient modulation and DC navigator retrospective gating is introduced as a 3D cine magnetic resonance imaging (MRI) method for the lung. In anesthetized normal rats, the quasi-simultaneous excitation and acquisition in SWIFT enabled extremely high sensitivity to the fast-decaying parenchymal signals (TE=~4 μs), which are invisible with conventional MRI techniques. Respiratory motion information was extracted from DC navigator signals and the SWIFT data were reconstructed to 3D cine images with 16 respiratory phases. To test this technique’s capabilities, rats exposed to > 95% O₂ for 60 hours for induction of acute respiratory distress syndrome (ARDS), were imaged and compared with normal rat lungs (N=7 and 5 for ARDS and normal groups, respectively). SWIFT images showed lung tissue density differences along the gravity direction. In the cine SWIFT images, a parenchymal signal drop at the inhalation phase was consistently observed for both normal and ARDS rats due to lung inflation (i.e. decrease of the proton density), but the drop was less for ARDS rats. Depending on the respiratory phase and lung region, the lungs from the ARDS rats showed 1-24% higher parenchymal signal intensities relative to the normal rat lungs, likely due to accumulated extravascular water (EVLW). Those results demonstrate that SWIFT has high enough sensitivity for detecting the lung proton density changes due to gravity, different phases of respiration and accumulation of EVLW in the rat ARDS lungs.

Stereotactic ablative radiotherapy (SABR) is a treatment for early-stage lung cancer with local control rates comparable to surgery. After SABR, benign radiation induced lung injury (RILI) results in tumour-mimicking changes on computed tomography (CT) imaging. Distinguishing recurrence from RILI is a critical clinical decision determining the need for potentially life-saving salvage therapies whose high risks in this population dictate their use only for true recurrences. Current approaches do not reliably detect recurrence within a year post-SABR. We measured the detection accuracy of texture features within automatically determined regions of interest, with the only operator input being the single line segment measuring tumour diameter, normally taken during the clinical workflow. Our leave-one-out cross validation on images taken 2–5 months post-SABR showed robustness of the entropy measure, with classification error of 26% and area under the receiver operating characteristic curve (AUC) of 0.77 using automatic segmentation; the results using manual segmentation were 24% and 0.75, respectively. AUCs for this feature increased to 0.82 and 0.93 at 8–14 months and 14–20 months post SABR, respectively, suggesting even better performance nearer to the date of clinical diagnosis of recurrence; thus this system could also be used to support and reinforce the physician’s decision at that time. Based on our ongoing validation of this automatic approach on a larger sample, we aim to develop a computer-aided diagnosis system which will support the physician’s decision to apply timely salvage therapies and prevent patients with RILI from undergoing invasive and risky procedures.

Fourier-decomposition of free-breathing pulmonary magnetic resonance imaging (FDMRI) was recently piloted as a way to provide rapid quantitative pulmonary maps of ventilation and perfusion without the use of exogenous contrast agents. This method exploits fast pulmonary MRI acquisition of free-breathing proton (¹H) pulmonary images and non-rigid registration to compensate for changes in position and shape of the thorax associated with breathing. In this way, ventilation imaging using conventional MRI systems can be undertaken but there has been no systematic evaluation of fundamental image quality measurements based on linear systems theory. We investigated the performance of free-breathing pulmonary ventilation imaging using a Fourier-based linear system description of each operation required to generate FDMRI ventilation maps. Twelve subjects with chronic obstructive pulmonary disease (COPD) or bronchiectasis underwent pulmonary function tests and MRI. Non-rigid registration was used to co-register the temporal series of pulmonary images. Pulmonary voxel intensities were aligned along a time axis and discrete Fourier transforms were performed on the periodic signal intensity pattern to generate frequency spectra. We determined the signal-to-noise ratio (SNR) of the FDMRI ventilation maps using a conventional approach (SNR_C) and using the Fourier-based description (SNR_F). Mean SNR was 4.7 ± 1.3 for subjects with bronchiectasis and 3.4 ± 1.8, for COPD subjects (p>.05). SNR_F was significantly different than SNR_C (p<.01). SNR_F was approximately 50% of SNR_C suggesting that the linear system model well-estimates the current approach.

Osteoarthritis (OA) is associated with significant pain and 42.6% of patients with TMJ disorders present with evidence of TMJ OA. However, OA diagnosis and treatment remain controversial, since there are no clear symptoms of the disease. The subchondral bone in the TMJ is believed to play a major role in the progression of OA. We hypothesize that the textural imaging biomarkers computed in high resolution Conebeam CT (hr- CBCT) and μCT scans are comparable. The purpose of this study is to test the feasibility of computing textural imaging biomarkers in-vivo using hr-CBCT, compared to those computed in μCT scans as our Gold Standard. Specimens of condylar bones obtained from condylectomies were scanned using μCT and hr- CBCT. Nine different textural imaging biomarkers (four co-occurrence features and five run-length features) from each pair of μCT and hr-CBCT were computed and compared. Pearson correlation coefficients were computed to compare textural biomarkers values of μCT and hr-CBCT. Four of the nine computed textural biomarkers showed a strong positive correlation between biomarkers computed in μCT and hr-CBCT. Higher correlations in Energy and Contrast, and in GLN (grey-level non-uniformity) and RLN (run length non-uniformity) indicate quantitative texture features can be computed reliably in hr-CBCT, when compared with μCT. The textural imaging biomarkers computed in-vivo hr-CBCT have captured the structure, patterns, contrast between neighboring regions and uniformity of healthy and/or pathologic subchondral bone. The ability to quantify bone texture non-invasively now makes it possible to evaluate the progression of subchondral bone alterations, in TMJ OA.

Temporomandibular joint (TMJ) disorders are a group of conditions that cause pain and dysfunction in the jaw joint and the muscles controlling jaw movement. However, diagnosis and treatment of these conditions remain controversial. To date, there is no single sign, symptom, or test that can clearly diagnose early stages of osteoarthritis (OA). Instead, the diagnosis is based on a consideration of several factors, including radiological evaluation. The current radiological diagnosis scores of TMJ pathology are subject to misdiagnosis. We believe these scores are limited by the acquisition procedures, such as oblique cuts of the CT and head positioning errors, and can lead to incorrect diagnoses of flattening of the head of the condyle, formation of osteophytes, or condylar pitting. This study consists of creating and validating a methodological framework to simulate defects in CBCT scans of known location and size, in order to create synthetic TMJ OA database. User-generated defects were created using a non-rigid deformation protocol in CBCT. All segmentation evaluation, surface distances and linear distances from the user-generated to the simulated defects showed our methodological framework to be very precise and within a voxel (0.5 mm) of magnitude. A TMJ OA synthetic database will be created next, and evaluated by expert radiologists, and this will serve to evaluate how sensitive the current radiological diagnosis tools are.

High-resolution 3D bone-tissue structure measurements may provide information critical to the understanding of the bone regeneration processes and to the bone strength assessment. Tissue engineering studies rely on such nondestructive measurements to monitor bone graft regeneration area. In this study, we measured bone yield, fractal dimension and trabecular thickness through micro-CT slices for different grafts and controls. Eight canines underwent surgery to remove a bone volume (defect) in the canine’s jaw at a total of 44 different locations. We kept 11 defects empty for control and filled the remaining ones with three regenerative materials; NanoGen (NG), a FDA-approved material (n=11), a novel NanoCalcium Sulfate (NCS) material (n=11) and NCS alginate (NCS+alg) material (n=11). After a minimum of four and eight weeks, the canines were sacrificed and the jaw samples were extracted. We used a custombuilt micro-CT system to acquire the data volume and developed software to measure the bone yield, fractal dimension and trabecular thickness. The software used a segmentation algorithm based on histograms derived from volumes of interest indicated by the operator. Using bone yield and fractal dimension as indices we are able to differentiate between the control and regenerative material (p<0.005). Regenerative material NCS showed an average 63.15% bone yield improvement over the control sample, NCS+alg showed 55.55% and NanoGen showed 37.5%. The bone regeneration process and quality of bone were dependent upon the position of defect and time period of healing. This study presents one of the first quantitative comparisons using non-destructive Micro-CT analysis for bone regenerative material in a large animal with a critical defect model. Our results indicate that Micro-CT measurement could be used to monitor invivo bone regeneration studies for greater regenerative process understanding.

While the proximal femur is preferred for measuring bone mineral density (BMD) in fracture risk estimation, the introduction of volumetric quantitative computed tomography has revealed stronger associations between BMD and spinal fracture status. In this study, we propose to capture properties of trabecular bone structure in spinal vertebrae with advanced second-order statistical features for purposes of fracture risk assessment. For this purpose, axial multi-detector CT (MDCT) images were acquired from 28 spinal vertebrae specimens using a whole-body 256-row CT scanner with a dedicated calibration phantom. A semi-automated method was used to annotate the trabecular compartment in the central vertebral slice with a circular region of interest (ROI) to exclude cortical bone; pixels within were converted to values indicative of BMD. Six second-order statistical features derived from gray-level co-occurrence matrices (GLCM) and the mean BMD within the ROI were then extracted and used in conjunction with a generalized radial basis functions (GRBF) neural network to predict the failure load of the specimens; true failure load was measured through biomechanical testing. Prediction performance was evaluated with a root-mean-square error (RMSE) metric. The best prediction performance was observed with GLCM feature ‘correlation’ (RMSE = 1.02 ± 0.18), which significantly outperformed all other GLCM features (p < 0.01). GLCM feature correlation also significantly outperformed MDCTmeasured mean BMD (RMSE = 1.11 ± 0.17) (p< 10^-4). These results suggest that biomechanical strength prediction in spinal vertebrae can be significantly improved through characterization of trabecular bone structure with GLCM-derived texture features.

Phase contrast X-ray computed tomography (PCI-CT) has recently emerged as a novel imaging technique that allows visualization of cartilage soft tissue, subsequent examination of chondrocyte patterns, and their correlation to osteoarthritis. Previous studies have shown that 2D texture features are effective at distinguishing between healthy and osteoarthritic regions of interest annotated in the radial zone of cartilage matrix on PCI-CT images. In this study, we further extend the texture analysis to 3D and investigate the ability of volumetric texture features at characterizing chondrocyte patterns in the cartilage matrix for purposes of classification. Here, we extracted volumetric texture features derived from Minkowski Functionals and gray-level co-occurrence matrices (GLCM) from 496 volumes of interest (VOI) annotated on PCI-CT images of human patellar cartilage specimens. The extracted features were then used in a machine-learning task involving support vector regression to classify ROIs as healthy or osteoarthritic. Classification performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). The best classification performance was observed with GLCM features correlation (AUC = 0.83 ± 0.06) and homogeneity (AUC = 0.82 ± 0.07), which significantly outperformed all Minkowski Functionals (p < 0.05). These results suggest that such quantitative analysis of chondrocyte patterns in human patellar cartilage matrix involving GLCM-derived statistical features can distinguish between healthy and osteoarthritic tissue with high accuracy.

Our team has established a method to enable imaging of heterotopic ossification and bone growth locations in tissue using Stokes Raman signals with fast acquisition times. This technique relies on the unique Raman signatures of bone to capture parallel, full-field, 1 cm² field of view, without utilizing a spectrometer. This system was built in mind as a compact complementary tool for in vivo patient monitoring that can offer a high resolution optical characterization for early detection of failed wounds. Preliminary results of bone detection in flesh are presented here and pave the way for further development of this tool in clinical setting.

Chemical exchange saturation transfer (CEST) MRI is sensitive to dilute exchangeable protons and local properties such as pH and temperate, yet its susceptibility to field inhomogeneity limits its in vivo applications. Particularly, CEST measurement varies with RF irradiation power, the dependence of which is complex due to concomitant direct RF saturation (RF spillover) effect. Because the volume transmitters provide relatively homogeneous RF field, they have been conventionally used for CEST imaging despite of their elevated specific absorption rate (SAR) and relatively low sensitivity than surface coils. To address this limitation, we developed an efficient B₁ inhomogeneity correction algorithm that enables CEST MRI using surface transceiver coils. This is built on recent work that showed the inverse CEST asymmetry analysis (CESTR_ind) is not susceptible to confounding RF spillover effect. We here postulated that the linear relationship between RF power level and CESTR_ind can be extended for correcting B₁ inhomogeneity induced CEST MRI artifacts. Briefly, we prepared a tissue-like Creatine gel pH phantom and collected multiparametric MRI including relaxation, field map and CEST MRI under multiple RF power levels, using a conventional surface transceiver coil. The raw CEST images showed substantial heterogeneity due to B1 inhomogeneity, with pH contrast to noise ratio (CNR) being 8.8. In comparison, pH MRI CNR of the fieldinhomogeneity corrected CEST MRI was found to be 17.2, substantially higher than that without correction. To summarize, our study validated an efficient field inhomogeneity correction that enables sensitive CEST MRI with surface transceiver, promising for in vivo translation.

In an event where many thousands of people may have been exposed to levels of radiation that are sufficient to cause the acute radiation syndrome, we need technology that can estimate the absorbed dose on an individual basis for triage and meaningful medical decision making. Such dose estimates may be achieved using in vivo electron paramagnetic resonance (EPR) tooth biodosimetry, which measures the number of persistent free radicals that are generated in tooth enamel following irradiation. However, the accuracy of dose estimates may be impacted by individual variations in teeth, especially the amount and distribution of enamel in the inhomogeneous sensitive volume of the resonator used to detect the radicals. In order to study the relationship between interpersonal variations in enamel and EPR-based dose estimates, it is desirable to estimate these parameters nondestructively and without adding radiation to the teeth.

Magnetic Resonance Imaging (MRI) is capable of acquiring structural and biochemical information without imparting additional radiation, which may be beneficial for many EPR dosimetry studies. However, the extremely short T2 relaxation time in tooth structures precludes tooth imaging using conventional MRI methods. Therefore, we used zero echo time (ZTE) MRI to image teeth ex vivo to assess enamel volumes and spatial distributions. Using these data in combination with the data on the distribution of the transverse radio frequency magnetic field from electromagnetic simulations, we then can identify possible sources of variations in radiation-induced signals detectable by EPR. Unlike conventional MRI, ZTE applies spatial encoding gradients during the RF excitation pulse, thereby facilitating signal acquisition almost immediately after excitation, minimizing signal loss from short T2 relaxation times. ZTE successfully provided volumetric measures of tooth enamel that may be related to variations that impact EPR dosimetry and facilitate the development of analytical procedures for individual dose estimates.

Imaging lactate metabolism in vivo may improve cancer targeting and therapeutics due to its key role in the development, maintenance, and metastasis of cancer. The long acquisition times associated with magnetic resonance spectroscopic imaging (MRSI), which is a useful technique for assessing metabolic concentrations, are a deterrent to its routine clinical use. The objective of this study was to combine spectral editing and prospective compressed sensing (CS) acquisitions to enable precise and high-speed imaging of the lactate resonance. A MRSI pulse sequence with two key modifications was developed: (1) spectral editing components for selective detection of lactate, and (2) a variable density sampling mask for pseudo-random under-sampling of the k-space ‘on the fly’. The developed sequence was tested on phantoms and in vivo in rodent models of cancer. Datasets corresponding to the 1X (fully-sampled), 2X, 3X, 4X, 5X, and 10X accelerations were acquired. The under-sampled datasets were reconstructed using a custom-built algorithm in Matlab^TM, and the fidelity of the CS reconstructions was assessed in terms of the peak amplitudes, SNR, and total acquisition time. The accelerated reconstructions demonstrate a reduction in the scan time by up to 90% in vitro and up to 80% in vivo, with negligible loss of information when compared with the fully-sampled dataset. The proposed unique combination of spectral editing and CS facilitated rapid mapping of the spatial distribution of lactate at high temporal resolution. This technique could potentially be translated to the clinic for the routine assessment of lactate changes in solid tumors.

Magnetic resonance elastography (MRE) is a phase-contrast MRI based technique that allows quantitative, noninvasive assessment of the mechanical properties of tissues by the introduction of shear waves into the body and measurement of the resulting displacements. In MRE, the calculated stiffness values are affected by noise, which is amplified by the inversion process. It would be useful to know that beyond some SNR threshold, the stiffness values are accurate within some confidence limit. The most common methods to calculate SNR values in MRE are variations of displacement SNR, which estimate the noise in the measured displacement. However, the accuracy of stiffness determination depends not only on the displacement SNR, but also on the wavelength of the shear wave, in turn dependent on the stiffness of the underlying material. More recently, the SNR of the octahedral shear strain (OSS) has been proposed as a more appropriate measure, since shear deformation is the signal in MRE. We also propose here another measure based on the SNR of the Laplacian of the data, since this is the most noise sensitive quantity calculated when performing direct inversion of the Helmholtz equation. The three SNR measures were compared on simulated data for materials of different stiffness with varying amounts of noise using three inversion algorithms commonly used in MRE (phase gradient, local frequency estimation, and direct inversion). We demonstrate that the proper SNR measure for MRE depends on the inversion algorithm used, and, more precisely, on the order of derivatives used in the inversion process.

Neuroimaging studies of working memory training have identified the alteration of brain activity as well as the regional interactions within the functional networks such as central executive network (CEN) and default mode network (DMN). However, how the interaction within and between these multiple networks is modulated by the training remains unclear. In this paper, we examined the interaction of three training-induced brain networks during working memory training based on real-time functional magnetic resonance imaging (rtfMRI). Thirty subjects assigned to the experimental and control group respectively participated in two times training separated by seven days. Three networks including silence network (SN), CEN and DMN were identified by the training data with the calculated function connections within each network. Structural equation modeling (SEM) approach was used to construct the directional connectivity patterns. The results showed that the causal influences from the percent signal changes of target ROI to the SN were positively changed in both two groups, as well as the causal influence from the SN to CEN was positively changed in experimental group but negatively changed in control group from the SN to DMN. Further correlation analysis of the changes in each network with the behavioral improvements showed that the changes in SN were stronger positively correlated with the behavioral improvement of letter memory task. These findings indicated that the SN was not only a switch between the target ROI and the other networks in the feedback training but also an essential factor to the behavioral improvement.

Echo state networks (ESN) are recurrent neural networks where the hidden layer is replaced with a fixed reservoir of neurons. Unlike feed-forward networks, neuron training in ESN is restricted to the output neurons alone thereby providing a computational advantage. We demonstrate the use of such ESNs in our mutual connectivity analysis (MCA) framework for recovering the primary motor cortex network associated with hand movement from resting state functional MRI (fMRI) data. Such a framework consists of two steps - (1) defining a pair-wise affinity matrix between different pixel time series within the brain to characterize network activity and (2) recovering network components from the affinity matrix with non-metric clustering. Here, ESNs are used to evaluate pair-wise cross-estimation performance between pixel time series to create the affinity matrix, which is subsequently subject to non-metric clustering with the Louvain method. For comparison, the ground truth of the motor cortex network structure is established with a task-based fMRI sequence. Overlap between the primary motor cortex network recovered with our model free MCA approach and the ground truth was measured with the Dice coefficient. Our results show that network recovery with our proposed MCA approach is in close agreement with the ground truth. Such network recovery is achieved without requiring low-pass filtering of the time series ensembles prior to analysis, an fMRI preprocessing step that has courted controversy in recent years. Thus, we conclude our MCA framework can allow recovery and visualization of the underlying functionally connected networks in the brain on resting state fMRI.

Functional MRI (fMRI) is currently used to investigate structural and functional connectivity in human brain networks. To this end, previous studies have proposed computational methods that involve assumptions that can induce information loss, such as assumed linear coupling of the fMRI signals or requiring dimension reduction. This study presents a new computational framework for investigating the functional connectivity in the brain and recovering network structure while reducing the information loss inherent in previous methods. For this purpose, pair-wise mutual information (MI) was extracted from all pixel time series within the brain on resting-state fMRI data. Non-metric topographic mapping of proximity (TMP) data was subsequently applied to recover network structure from the pair-wise MI analysis. Our computational framework is demonstrated in the task of identifying regions of the primary motor cortex network on resting state fMRI data. For ground truth comparison, we also localized regions of the primary motor cortex associated with hand movement in a task-based fMRI sequence with a finger-tapping stimulus function. The similarity between our pair-wise MI clustering results and the ground truth is evaluated using the dice coefficient. Our results show that non-metric clustering with the TMP algorithm, as performed on pair-wise MI analysis, was able to detect the primary motor cortex network and achieved a dice coefficient of 0.53 in terms of overlap with the ground truth. Thus, we conclude that our computational framework can extract and visualize valuable information concerning the underlying network structure between different regions of the brain in resting state fMRI.

A challenging goal in neuroscience is to decode the mental states from brain activity. Recently, researchers have successfully deciphered the objective and static visual stimuli (such as orientation of stripes and category of objects) from brain activity recorded by functional magnetic resonance imaging (fMRI) technology. However, few studies focused on the decoding of the rotation direction perception of the actual three-dimensional world with two-dimensional representations. In this study, the brain activities when subjects viewed the animation of the spinning dancer in the front were recorded using fMRI, and subjects reported the viewing-from-bottom motion direction (clockwise or counterclockwise) by press different buttons. One multivariate pattern analysis method, support vector machine was trained to predict the rotation direction. The 5-fold cross-validation result showed that the subjective rotation direction reported by the subjects can be predicted from fMRI with a possibility above the chance level, which imply that fMRI activity of the brain contains detailed rotation direction information that can reliably predict the subjective perception.

The structures of developing human brain white matter (WM) tracts can be effectively quantified by DTI-derived metrics, including fractional anisotropy (FA), mean, axial and radial diffusivity (MD, AD and RD). However, dynamics of WM microstructure during very early developmental period from mid-fetal to perinatal stage is unknown. It is difficult to accurately measure microstructural properties of these WM tracts due to severe contamination from cerebrospinal fluid (CSF). In this study, high resolution DTI of fetal brains at mid-fetal stage (20 weeks of gestation or 20wg), 19 brains in the middle of 3^rd trimester (35wg) and 17 brains around term (40wg) were acquired. We established first population-averaged DTI templates at these three time points and extracted WM skeleton. 16 major WM tracts in limbic, projection, commissural and association tract groups were traced with DTI tractography in native space. The WM skeleton in the template space was inversely transformed back to the native space for measuring core WM microstructures of each individual tract. Continuous microstructural enhancement and volumetric increase of WM tracts were found from 20wg to 40wg. The microstructural enhancement from FA measurement is decelerated in late 3^rd trimester compared to mid-fetal to middle 3rd trimester, while volumetric increase of prefrontal WM tracts is accelerated. The microstructural enhancement from 35wg to 40wg is heterogeneous among different tract groups with microstructures of association tracts undergoing most dramatic change. Besides decreases of RD indicating active myelination, the decrease of AD for most WM tracts during late 3^rd trimester suggests axonal packing process.

CT perfusion (CTP) imaging allows for rapid diagnosis of ischemic stroke. Generation of perfusion maps from CTP data usually involves deconvolution algorithms providing estimates for the impulse response function in the tissue. We propose the use of a fast non-linear regression (NLR) method that we postulate has similar performance to the current academic state-of-art method (bSVD), but that has some important advantages, including the estimation of vascular permeability, improved robustness to tracer-delay, and very few tuning parameters, that are all important in stroke assessment. The aim of this study is to evaluate the fast NLR method against bSVD and a commercial clinical state-of-art method. The three methods were tested against a published digital perfusion phantom earlier used to illustrate the superiority of bSVD. In addition, the NLR and clinical methods were also tested against bSVD on 20 clinical scans. Pearson correlation coefficients were calculated for each of the tested methods. All three methods showed high correlation coefficients (>0.9) with the ground truth in the phantom. With respect to the clinical scans, the NLR perfusion maps showed higher correlation with bSVD than the perfusion maps from the clinical method. Furthermore, the perfusion maps showed that the fast NLR estimates are robust to tracer-delay. In conclusion, the proposed fast NLR method provides a simple and flexible way of estimating perfusion parameters from CT perfusion scans, with high correlation coefficients. This suggests that it could be a better alternative to the current clinical and academic state-of-art methods.

To develop and evaluate a novel processing framework for the relative quantification of myelin content in cerebral white matter (WM) regions from brain MRI data via a computed ratio of T1 to T2 weighted intensity values. We employed high resolution (1mm³ isotropic) T1 and T2 weighted MRI from 46 (28 male, 18 female) neonate subjects (typically developing controls) scanned on a Siemens Tim Trio 3T at UC Irvine. We developed a novel, yet relatively straightforward image processing framework for WM myelin content estimation based on earlier work by Glasser, et al. We first co-register the structural MRI data to correct for motion. Then, background areas are masked out via a joint T1w and T2 foreground mask computed. Raw T1w/T2w-ratios images are computed next. For purpose of calibration across subjects, we first coarsely segment the fat-rich facial regions via an atlas co-registration. Linear intensity rescaling based on median T1w/T2w-ratio values in those facial regions yields calibrated T1w/T2wratio images. Mean values in lobar regions are evaluated using standard statistical analysis to investigate their interaction with age at scan. Several lobes have strongly positive significant interactions of age at scan with the computed T1w/T2w-ratio. Most regions do not show sex effects. A few regions show no measurable effects of change in myelin content change within the first few weeks of postnatal development, such as cingulate and CC areas, which we attribute to sample size and measurement variability. We developed and evaluated a novel way to estimate white matter myelin content for use in studies of brain white matter development.

Over 50% of HIV+ individuals show significant impairment in psychomotor functioning, processing speed, working memory and attention [1, 2]. Patients receiving combination antiretroviral therapy may still have subcortical atrophy, but the profile of HIV-associated brain changes is poorly understood. With parametric surface-based shape analyses, we mapped the 3D profile of subcortical morphometry in 63 elderly HIV+ subjects (4 female; age=65.35 ± 2.21) and 31 uninfected elderly controls (2 female; age=64.68 ± 4.57) scanned with MRI as part of a San Francisco Bay Area study of elderly people with HIV. We also investigated whether morphometry was associated with nadir CD4+ (T-cell) counts, viral load and illness duration among HIV+ participants. FreeSurfer was used to segment the thalamus, caudate, putamen, pallidum, hippocampus, amygdala, accumbens, brainstem, callosum and ventricles from brain MRI scans. To study subcortical shape, we analyzed: (1) the Jacobian determinant (JD) indexed over structures’ surface coordinates and (2) radial distances (RD) of structure surfaces from a medial curve. A JD less than 1 reflects regional tissue atrophy and greater than 1 reflects expansion. The volumes of several subcortical regions were found to be associated with HIV status. No regional volumes showed detectable associations with CD4 counts, viral load or illness duration. The shapes of numerous subcortical regions were significantly linked to HIV status, detectability of viral RNA and illness duration. Our results show subcortical brain differences in HIV+ subjects in both shape and volumetric domains.

This effort is a continuation of development of a digital brain atlas of the common squirrel monkey, Saimiri sciureus, a New World monkey with functional and microstructural organization of central nervous system similar to that of humans. Here, we present the integration of histology with multi-modal magnetic resonance imaging (MRI) atlas constructed from the brain of an adult female squirrel monkey. The central concept of this work is to use block face photography to establish an intermediate common space in coordinate system which preserves the high resolution in-plane resolution of histology while enabling 3-D correspondence with MRI. In vivo MRI acquisitions include high resolution T2 structural imaging (300 μm isotropic) and low resolution diffusion tensor imaging (600 um isotropic). Ex vivo MRI acquisitions include high resolution T2 structural imaging and high resolution diffusion tensor imaging (both 300 μm isotropic). Cortical regions were manually annotated on the co-registered volumes based on published histological sections in-plane. We describe mapping of histology and MRI based data of the common squirrel monkey and construction of a viewing tool that enable online viewing of these datasets. The previously descried atlas MRI is used for its deformation to provide accurate conformation to the MRI, thus adding information at the histological level to the MRI volume. This paper presents the mapping of single 2D image slice in block face as a proof of concept and this can be extended to map the atlas space in 3D coordinate system as part of the future work and can be loaded to an XNAT system for further use.

Motion correction in Arterial Spin Labeling (ASL) is essential to accurately assess brain perfusion. Motion correction techniques are usually based on intensity-related information, which might be unreliable in ASL due to local intensity differences between control and labeled acquisitions and to non-uniform volume magnetization caused by background-suppressed acquisition protocols. Accordingly, a novel motion correction technique based only on brain contour points is presented and tested against a widely used intensity-based technique (MCFLIRT). The proposed Contour-Based Motion Correction (CBCM) technique relies on image segmentation (to extract brain contour point clouds) and on Iterative Closest Point algorithm (to estimate the roto-translation required to align them). At variance with other approaches based on point clouds alignment, the local 3D curvature is also computed for each contour point and used as an additional coordinate to increase the accuracy of the alignment. The technique has been tested along with MCFLIRT on a database of randomly roto-translated brain volumes. Several error metrics have been computed and compared between the two techniques. The results show that the proposed technique is able to achieve a higher accuracy than MCFLIRT without any intensity-dependent information.

The cognitive deficits of schizophrenia are largely resistant to current treatment, and are thus a life-long burden to patients. The MATRICS consensus cognitive battery (MCCB) provides a reliable and valid assessment of cognition across a comprehensive set of cognitive domains for schizophrenia. In resting-state fMRI, functional connectivity associated with MCCB has not yet been examined. In this paper, the interrelationships between MCCB and the abnormalities seen in two types of functional measures from resting-state fMRI—fractional amplitude of low frequency fluctuations (fALFF) and functional network connectivity (FNC) maps were investigated in data from 47 schizophrenia patients and 50 age-matched healthy controls. First, the fALFF maps were generated and decomposed by independent component analysis (ICA), and then the component showing the highest correlation with MCCB composite scores was selected. Second, the whole brain was separated into functional networks by group ICA, and the FNC maps were calculated. The FNC strengths with most significant correlations with MCCB were displayed and spatially overlapped with the fALFF component of interest. It demonstrated increased cognitive performance associated with higher fALFF values (intensity of regional spontaneous brain activity) in prefrontal regions, inferior parietal lobe (IPL) but lower ALFF values in thalamus, striatum, and superior temporal gyrus (STG). Interestingly, the FNC showing significant correlations with MCCB were in well agreement with the activated regions with highest z-values in fALFF component. Our results support the view that functional deficits in distributed cortico-striato-thalamic circuits and inferior parietal lobe may account for several aspects of cognitive impairment in schizophrenia.

Fluorescence from endogenous fluorophores has been emerging as a promising biomarker for tissue discrimination resulting a noninvasive screening methodology to understand the biochemical and morphological variations in tissues associated with cancer development. We have developed a scan based fiber optic probe system to image increased flavin adenine dinucleotide (FAD) fluorescence from epithelial tissues under conditions mimicking dysplasia surrounded by normal tissues. Experiments were conducted on optical phantoms mimicking epithelial tissues excited by 450nm LED source. The spectral emission from the sample is collected via optical fibers and the imaging is performed by scanning the sample using a translation stage at desired resolution. Monte Carlo simulations were also performed by devising an optical model corresponding to epithelial tissue and the results were correlated with experimental fluorescence measurements. This whole field imaging approach could be useful for in vivo assessment of tissue pathologies based on auto fluorescence and can give a better quantitative approach for estimation of tissue properties by correlating the experimental and simulated data.

Early screening of subsurface anomalies in breast can improve the patient survival rate. Clinically approved breast screening modalities may either have body ionizing effect/cause pain to the body parts/ involves body contact/ increased cost. In this paper, a non-invasive, whole field Digital Speckle Pattern Interferometry (DSPI) is used to study normal and abnormal breast mimicking tissue phantoms. While uniform fringes were obtained for a normal phantom in the out of plane speckle pattern interferometry configuration, the non uniformity in the observed fringes clearly showed the anomaly location in the abnormal phantom. The results are compared with deformation profiles using finite element analysis of the sample under similar loading conditions.

Determination of neoplasia is largely dependent on the state of cell growth. Infrared (IR) spectroscopy has the potential to measure differences between normal and cancerous cells. When analyzing biopsy sections using IR spectroscopy, careful analyses become important since biochemical variations may be misinterpreted due to variations in cell cycle. Processes like DNA replication, transcription and translation to produce proteins are important in determining if the cells are actively dividing but no studies on this aspect using IR spectroscopy have been conducted on isolated cell nuclei. Nuclei hold critical information about the phase of cell and its capacity to divide, but IR spectra of nuclei are often confounded by cytoplasmic signals during data acquisition from intact cells and tissues. Therefore, we sought to separate nuclear signals from cytoplasmic signals and identify spectral differences that characterize different phases of the cell cycle. Both cells and isolated nuclei were analyzed to assess the effect of the cytoplasmic background and to identify spectral changes in nuclei in different phases of cell cycle. We observed that signals of DNA could be obtained when imaging nuclei isolated from cells in different phases of cell cycle, which is in contrast to the oft-cited case in cells wherein nuclear contributions are obscured. The differences across cell cycle phases were more pronounced in nucleic acid regions of the spectra, showing that the use of nuclear spectrum can provide additional information on cellular state. These results can aid in developing computational models that extract nuclear spectra from whole cells and tissues for more accurate assessment of biochemical variations.

Recent studies have shown that the Bruch's membrane (BM) and retinal pigment epithelium (RPE), visualized on spectral-domain optical coherence tomography (SD-OCT), is deformed anteriorly towards the vitreous in patients with intracranial hypertension and papilledema. The BM/RPE shape has been quantified using a statistical-shape-model approach; however, to date, the approach has involved the tedious and time-consuming manual placement of landmarks and correspondingly, only the shape (and shape changes) of a limited number of patients has been studied. In this work, we first present a semi-automated approach for the extraction of 20 landmarks along the BM from an optic-nerve-head (ONH) centered OCT slice from each patient. In the approach, after the manual placement of the two Bruch's membrane opening (BMO) points, the remaining 18 landmarks are automatically determined using a graph-based segmentation approach. We apply the approach to the OCT scans of 116 patients (at baseline) enrolled in the Idiopathic Intracranial Hypertension Treatment Trial and generate a statistical shape model using principal components analysis. Using the resulting shape model, the coefficient (shape measure) corresponding to the second principal component (eigenvector) for each set of landmarks indicates the degree of the BM/RPE is oriented away from the vitreous. Using a subset of 20 patients, we compare the shape measure computed using this semi-automated approach with the resulting shape measure when (1) all landmarks are specified manually (Experiment I); and (2) a different expert specifies the two BMO points (Experiment II). In each case, a correlation coefficient ≥ 0.99 is obtained.

Micro-optical computed tomography (MOCT) is a method for performing image reconstruction using microscopic images to obtain tomographic images of small samples. Compared with conventional observation methods, it offers the possibility to obtain tomograpic images without distortion, and create three-dimensional images. However, MOCT system which developed previously outputs monochrome images, while useful color information could not be obtained from the analysis of the sample. Therefore, we focused on the features that simplify the wavelength measurement of visible light, and developed a color MOCT system that can obtain color tomographic images. In this study, we acquired tomographic images of phantom and biological samples, and evaluated its usefulness. In this system, a digital single-lens reflex camera was used as a detector that was connected to a stereoscopic microscope, and projection images were obtained by rotating the sample. The sample was fixed in the test tube by carrageenan. The projection images were obtained from various projection angles followed by decomposing the R, G and B components. Subsequently, we performed image reconstruction for each component using filtered back projection. Finally, color tomographic image was obtained by combining the three-color component images. In the experiments, we scanned a color phantom and biological samples and evaluated the color and shape reproducibility. As a result, it was found that the color and shape of the tomographic images were similar to those of the samples. These results indicate that the proposed system may be useful to obtain the three-dimensional color structure of biological samples.

Optical Projection Tomography (OPT) is a 3-Dimentional (3D) imaging technique for small specimens between 1mm and 10mm in size. Due to its high resolution and whole-body imaging ability, OPT has been widely used for imaging of small specimens such as murine embryos, murine organs, zebra fish, and plant sections. During an OPT imaging experiment, the ring artifacts are very common which severely impact the image quality of OPT. A ring artifact is caused by a bad pixel on the camera, or impurities on surface of lens and index matching vessel. Here we term these noises as coherent noise because they stay in the same image region during an OPT experiment. Currently, there is still no effective method to remove coherent noises. To address this problem, we propose a novel method to suppress the coherent noises before 3D OPT reconstruction. Our method consists of two steps: 1) find bad pixel positions on a blank image without specimen by using threshold segmentation, then fix the bad pixels on the projection image by using average of their neighbor pixels, 2) remove remained coherent noises on the sinogram by using Variational Coherent noise Remover (VSNR) method. After the two steps, lots of method can be used to generate the tomographic slices from the modified sinograms. We apply our method to a mouse heart imaging with our home-made OPT system. The experimental results show that our method has a good suppression on coherent noise and greatly improves the image quality. The innovation of our method is that we remove coherent noise automatically from both projection image and sinogram and they complement each other.

Optical projection tomography (OPT) is a mesoscopic scale optical imaging technique for specimens between 1mm and 10mm. OPT has been proven to be immensely useful in a wide variety of biological applications, such as developmental biology and pathology, but its shortcomings in imaging specimens containing widely differing contrast elements are obvious. The longer exposure for high intensity tissues may lead to over saturation of other areas, whereas a relatively short exposure may cause similarity with surrounding background. In this paper, we propose an approach to make a trade-off between capturing weak signals and revealing more details for OPT imaging. This approach consists of three steps. Firstly, the specimens are merely scanned in 360 degrees above a normal exposure but non-overexposure to acquire the projection data. This reduces the photo bleaching and pre-registration computation compared with multiple different exposures in conventional high dynamic range (HDR) imaging method. Secondly, three virtual channels are produced for each projection image based on the histogram distribution to simulate the low, normal and high exposure images used in the traditional HDR technology in photography. Finally, each virtual channel is normalized to the full gray scale range and three channels are recombined into one image using weighting coefficients optimized by a standard eigen-decomposition method. After applying our approach on the projection data, filtered back projection (FBP) algorithm is carried out for 3-dimentional reconstruction. The neonatal wild-type mouse paw has been scanned to verify this approach. Results demonstrated the effectiveness of the proposed approach.

It is clinically vital to devise a technique to evaluate regional functionality of the myocardium in order to determine the extent and intensity of local damage to the cardiac tissue caused by ischemic injuries. Such a technique can potentially enable cardiologists to discriminate between reversible and irreversible ischemic injuries and to devise appropriate revascularization therapy in case of reversible lesions. The technique is founded on the premise that sufficient contraction force generated by the cardiac tissue can be regarded as a direct and reliable criterion for regional analysis of tissue healthy functionality. To this end, a number of imaging techniques have been developed and, to our knowledge, none of them assess regional cardiac functionality based on a straightforward mechanical measure such as local cardiac contraction forces. . As such, a novel imaging technique is being developed on the basis of quantification and visualisation of local myocardial contraction forces. In this technique, cardiac contraction force distribution is attained through solving an inverse problem within an optimization framework which uses iterative forward mechanical modelling of the myocardium. Hence, a forward mechanical model of the myocardium which is computationally efficient, robust, and adaptable to diverse pathophysiological conditions is necessary for this development. As such, this paper is geared towards developing a novel mechanical model of the healthy and pathological myocardium which considers all aspects of the myocardial mechanics including hyperelasticity, anisotropy, and active contraction force. In this investigation, two major parts, including background tissue and reinforcement bars (fibers) have been considered for modelling the myocardium. The model was implemented using finite element (FE) approach and demonstrated very good performance in simulating normal and infarcted left ventricle (LV) contractile function.

Minimally invasive endovascular image-guided interventions (EIGIs) are the preferred procedures for treatment of a wide range of vascular disorders. Despite benefits including reduced trauma and recovery time, EIGIs have their own challenges. Remote catheter actuation and challenging anatomical morphology may lead to erroneous endovascular device selections, delays or even complications such as vessel injury. EIGI planning using 3D phantoms would allow interventionists to become familiarized with the patient vessel anatomy by first performing the planned treatment on a phantom under standard operating protocols. In this study the optimal workflow to obtain such phantoms from 3D data for interventionist to practice on prior to an actual procedure was investigated. Patientspecific phantoms and phantoms presenting a wide range of challenging geometries were created. Computed Tomographic Angiography (CTA) data was uploaded into a Vitrea 3D station which allows segmentation and resulting stereo-lithographic files to be exported. The files were uploaded using processing software where preloaded vessel structures were included to create a closed-flow vasculature having structural support. The final file was printed, cleaned, connected to a flow loop and placed in an angiographic room for EIGI practice. Various Circle of Willis and cardiac arterial geometries were used. The phantoms were tested for ischemic stroke treatment, distal catheter navigation, aneurysm stenting and cardiac imaging under angiographic guidance. This method should allow for adjustments to treatment plans to be made before the patient is actually in the procedure room and enabling reduced risk of peri-operative complications or delays.

Cerebral aneurysms may rarely coexist with a proximal artery stenosis. In that small percent of patients, such coexistence poses a challenge for interventional neuroradiologists and neurosurgeons to make the best treatment decision. According to previous studies, the incidence of cerebral aneurysms in patients with internal carotid artery stenosis is no greater than five percent, where the aneurysm is usually incidentally detected, being about two percent for aneurysms and stenoses in the same cerebral circulation. Those cases pose a difficult management decision for the physician. Case reports showed patients who died due to aneurysm rupture months after endarterectomy but before aneurysm clipping, while others did not show any change in the aneurysm after plaque removal, having optimum outcome after aneurysm coiling. The aim of this study is to investigate the intra-aneurysmal hemodynamic changes before and after treatment of stenotic plaque. Virtually created moderate stenoses in vascular models of internal carotid artery aneurysm patients were considered in a number of cases reconstructed from three dimensional rotational angiography images. The strategy to create those plaques was based on parameters analyzed in a previous work where idealized models were considered, including relative distance and stenosis grade. Ipsilateral and contralateral plaques were modeled. Wall shear stress and velocity pattern were computed from finite element pulsatile blood flow simulations. The results may suggest that wall shear stress changes depend on relative angular position between the aneurysm and the plaque.

A reconstruction method is proposed here to quantify the distribution of blood flow velocity fields inside the left ventricle from color Doppler echocardiography measurement. From 3D incompressible Navier- Stokes equation, a 2D incompressible Navier-Stokes equation with a mass source term is derived to utilize the measurable color flow ultrasound data in a plane along with the moving boundary condition. The proposed model reflects out-of-plane blood flows on the imaging plane through the mass source term. For demonstrating a feasibility of the proposed method, we have performed numerical simulations of the forward problem and numerical analysis of the reconstruction method. First, we construct a 3D moving LV region having a specific stroke volume. To obtain synthetic intra-ventricular flows, we performed a numerical simulation of the forward problem of Navier-Stokes equation inside the 3D moving LV, computed 3D intra-ventricular velocity fields as a solution of the forward problem, projected the 3D velocity fields on the imaging plane and took the inner product of the 2D velocity fields on the imaging plane and scanline directional velocity fields for synthetic scanline directional projected velocity at each position. The proposed method utilized the 2D synthetic projected velocity data for reconstructing LV blood flow. By computing the difference between synthetic flow and reconstructed flow fields, we obtained the averaged point-wise errors of 0.06 m/s and 0.02 m/s for u- and v-components, respectively.

Recent growth of genetic disease models in swine has presented the opportunity to advance translation of developed imaging protocols, while characterizing the genotype to phenotype relationship. Repeated imaging with multiple clinical modalities provides non-invasive detection, diagnosis, and monitoring of disease to accomplish these goals; however, longitudinal scanning requires repeatable and reproducible positioning of the animals. A modular positioning unit was designed to provide a fixed, stable base for the anesthetized animal through transit and imaging. Post ventilation and sedation, animals were placed supine in the unit and monitored for consistent vitals. Comprehensive imaging was performed with a computed tomography (CT) chest-abdomen-pelvis scan at each screening time point. Longitudinal images were rigidly registered, accounting for rotation, translation, and anisotropic scaling, and the skeleton was isolated using a basic thresholding algorithm. Assessment of alignment was quantified via eleven pairs of corresponding points on the skeleton with the first time point as the reference. Results were obtained with five animals over five screening time points. The developed unit aided in skeletal alignment within an average of 13.13 ± 6.7 mm for all five subjects providing a strong foundation for developing qualitative and quantitative methods of disease tracking.

The corpus callosum is the major brain structure responsible for inter{hemispheric communication between neurons. Many studies seek to relate corpus callosum attributes to patient characteristics, cerebral diseases and psychological disorders. Most of those studies rely on 2D analysis of the corpus callosum in the mid-sagittal plane. However, it is common to find conflicting results among studies, once many ignore methodological issues and define the mid-sagittal plane based on precary or invalid criteria with respect to the corpus callosum. In this work we propose a novel method to determine the mid-callosal plane using the corpus callosum internal preferred diffusion directions obtained from diffusion tensor images. This plane is analogous to the mid-sagittal plane, but intended to serve exclusively as the corpus callosum reference. Our method elucidates the great potential the directional information of the corpus callosum fibers have to indicate its own referential. Results from experiments with five image pairs from distinct subjects, obtained under the same conditions, demonstrate the method effectiveness to find the corpus callosum symmetric axis relative to the axial plane.

We propose a novel application of the sparse and low-rank (SLR) decomposition method to decode cognitive states for concept activity measured using fMRI BOLD. Current decoding methods attempt to reduce the dimensionality of fMRI BOLD signals to increase classification rate, but do not address the separable issues of multiple noise sources and complexity in the underlying data. Our feature transformation method extends SLR to separate task activity from the resting state and extract concept specific cognitive state. We show a significant increase in single trial decoding of concepts from fMRI BOLD using SLR to extract task specific cognitive state.

Content-based image retrieval (CBIR) offers the potential to identify similar case histories, understand rare disorders, and eventually, improve patient care. Recent advances in database capacity, algorithm efficiency, and deep Convolutional Neural Networks (dCNN), a machine learning technique, have enabled great CBIR success for general photographic images. Here, we investigate applying the leading ImageNet CBIR technique to clinically acquired medical images captured by the Vanderbilt Medical Center. Briefly, we (1) constructed a dCNN with four hidden layers, reducing dimensionality of an input scaled to 128x128 to an output encoded layer of 4x384, (2) trained the network using back-propagation 1 million random magnetic resonance (MR) and computed tomography (CT) images, (3) labeled an independent set of 2100 images, and (4) evaluated classifiers on the projection of the labeled images into manifold space. Quantitative results were disappointing (averaging a true positive rate of only 20%); however, the data suggest that improvements would be possible with more evenly distributed sampling across labels and potential re-grouping of label structures. This preliminary effort at automated classification of medical images with ImageNet is promising, but shows that more work is needed beyond direct adaptation of existing techniques.

Non-contact, imaging photoplethysmography uses cameras to facilitate measurements including pulse rate, pulse rate variability, respiration rate, and blood perfusion by measuring characteristic changes in light absorption at the skin’s surface resulting from changes in blood volume in the superficial microvasculature. Several factors may affect the accuracy of the physiological measurement including imager frame rate, resolution, compression, lighting conditions, image background, participant skin tone, and participant motion. Before this method can gain wider use outside basic research settings, its constraints and capabilities must be well understood. Recently, we presented a novel approach utilizing a synchronized, nine-camera, semicircular array backed by measurement of an electrocardiogram and fingertip reflectance photoplethysmogram. Twenty-five individuals participated in six, five-minute, controlled head motion artifact trials in front of a black and dynamic color backdrop. Increasing the input channel space for blind source separation using the camera array was effective in mitigating error from head motion artifact. Herein we present the effects of lower frame rates at 60 and 30 (reduced from 120) frames per second and reduced image resolution at 329x246 pixels (one-quarter of the original 658x492 pixel resolution) using bilinear and zero-order downsampling. This is the first time these factors have been examined for a multiple imager array and align well with previous findings utilizing a single imager. Examining windowed pulse rates, there is little observable difference in mean absolute error or error distributions resulting from reduced frame rates or image resolution, thus lowering requirements for systems measuring pulse rate over sufficient length time windows.

Cephalometric analysis is the study of the dental and skeletal relationship in the head, and it is used as an assessment and planning tool for improved orthodontic treatment of a patient. Conventional cephalometric analysis identifies bony and soft-tissue landmarks in 2D cephalometric radiographs, in order to diagnose facial features and abnormalities prior to treatment, or to evaluate the progress of treatment. Recent studies in orthodontics indicate that there are persistent inaccuracies and inconsistencies in the results provided using conventional 2D cephalometric analysis. Obviously, plane geometry is inappropriate for analyzing anatomical volumes and their growth; only a 3D analysis is able to analyze the three-dimensional, anatomical maxillofacial complex, which requires computing inertia systems for individual or groups of digitally segmented teeth from an image volume of a patient’s head. For the study of 3D cephalometric analysis, the current paper proposes a system for semi-automatically segmenting teeth from a cone beam computed tomography (CBCT) volume with two distinct features, including an intelligent user-input interface for automatic background seed generation, and a graphics processing unit (GPU) acceleration mechanism for three-dimensional GrowCut volume segmentation. Results show a satisfying average DICE score of 0.92, with the use of the proposed tooth segmentation system, by 15 novice users who segmented a randomly sampled tooth set. The average GrowCut processing time is around one second per tooth, excluding user interaction time.

In diagnostics and therapy control of cardiovascular diseases, detailed knowledge about the patient-specific behavior of blood flow and pressure can be essential. The only method capable of measuring complete time-resolved three-dimensional vector fields of the blood flow velocities is velocity-encoded magnetic resonance imaging (MRI), often denoted as 4D flow MRI. Furthermore, relative pressure maps can be computed from this data source, as presented by different groups in recent years. Hence, analysis of blood flow and pressure using 4D flow MRI can be a valuable technique in management of cardiovascular diseases. In order to perform these tasks, all necessary steps in the corresponding process chain can be carried out in our in-house developed software framework MEDIFRAME. In this article, we apply MEDIFRAME for a study of hemodynamics in the pulmonary arteries of five healthy volunteers. The study included measuring vector fields of blood flow velocities by phase-contrast MRI and subsequently computing relative blood pressure maps. We visualized blood flow by streamline depictions and computed characteristic values for the left and the right pulmonary artery (LPA and RPA). In all volunteers, we observed a lower amount of blood flow in the LPA compared to the RPA. Furthermore, we visualized blood pressure maps using volume rendering and generated graphs of pressure differences between the LPA, the RPA and the main pulmonary artery. In most volunteers, blood pressure was increased near to the bifurcation and in the proximal LPA, leading to higher average pressure values in the LPA compared to the RPA.

The tissue outcome prediction in acute ischemic stroke patients is highly relevant for clinical and research purposes. It has been shown that the combined analysis of diffusion and perfusion MRI datasets using high-level machine learning techniques leads to an improved prediction of final infarction compared to single perfusion parameter thresholding. However, most high-level classifiers require a previous training and, until now, it is ambiguous how many subjects are required for this, which is the focus of this work. 23 MRI datasets of acute stroke patients with known tissue outcome were used in this work. Relative values of diffusion and perfusion parameters as well as the binary tissue outcome were extracted on a voxel-by- voxel level for all patients and used for training of a random forest classifier. The number of patients used for training set definition was iteratively and randomly reduced from using all 22 other patients to only one other patient. Thus, 22 tissue outcome predictions were generated for each patient using the trained random forest classifiers and compared to the known tissue outcome using the Dice coefficient. Overall, a logarithmic relation between the number of patients used for training set definition and tissue outcome prediction accuracy was found. Quantitatively, a mean Dice coefficient of 0.45 was found for the prediction using the training set consisting of the voxel information from only one other patient, which increases to 0.53 if using all other patients (n=22). Based on extrapolation, 50-100 patients appear to be a reasonable tradeoff between tissue outcome prediction accuracy and effort required for data acquisition and preparation.

Evidence of leptin resistance is indicated by elevated leptin levels together with other hallmarks of obesity such as a defect in energy homeostasis.¹ As obesity is an increasing epidemic in the US, the investigation of mechanisms by which leptin resistance has a pathophysiological impact on energy is an intensive field of research.² However, the manner in which leptin resistance contributes to the dysregulation of energy, specifically thermoregulation,³ is not known. The aim of this study was to investigate whether the leptin receptor expressed in paraventricular nucleus (PVN) neurons plays a role in thermoregulation at different temperatures. Non-contact infrared (NCIR) thermometry was employed to measure surface body temperature (SBT) of nonanesthetized mice with a specific deletion of the leptin receptor in the PVN after exposure to room (25 °C) and cold (4 °C) temperature. Dorsal side infrared images of wild type (LepRwtwt/sim1-Cre), heterozygous (LepRfloxwt/sim1-Cre) and knock-out (LepRfloxflox/sim1-Cre) mice were collected. Images were input to an automated post-processing pipeline developed in MATLAB to calculate average and maximum SBTs. Linear regression was used to evaluate the relationship between sex, cold exposure and leptin genotype with SBT measurements. Findings indicate that average SBT has a negative relationship to the LepRfloxflox/sim1-Cre genotype, the female sex and cold exposure. However, max SBT is affected by the LepRfloxflox/sim1-Cre genotype and the female sex. In conclusion this data suggests that leptin within the PVN may have a neuroendocrine role in thermoregulation and that NCIR thermometry combined with an automated imaging-processing pipeline is a promising approach to determine SBT in non-anesthetized mice.

Mechanisms leading to Obstructive Sleep Apnea Syndrome (OSAS) in obese children are not well understood. We previously analyzed polysomnographic and demographic data to study the anatomical characteristics of the upper airway and body composition in two groups of obese children with and without OSAS, where object volume was evaluated. In this paper, in order to better understand the disease we expand the analysis considering a variety of features that include object-specific features such as size, surface area, sphericity, and image intensity properties of fourteen objects in the vicinity of the upper airway, as well as inter-object relationships such as distance between objects. Our preliminary results indicate several interesting phenomena: volumes and surface areas of adenoid and tonsils increase statistically significantly in OSAS. Standardized T2-weighted MR image intensities differ statistically significantly between the two groups, implying that perhaps intrinsic tissue composition undergoes changes in OSAS. Inter-object distances are significantly different between the two groups for object pairs (skin, oropharynx), (skin, fat pad), (skin, soft palate), (mandible, tongue), (oropharynx, soft palate), (left tonsil, oropharynx), (left tonsil, fat pad) and (left tonsil, right tonsil). We conclude that treatment methods for OSAS such as adenotonsillectomy should respect proportional object size relationships and spatial arrangement of objects as they exist in control subjects.

Glucose metabolism relates to biochemical processes in living organisms and plays an important role in diabetes and cancer-metastasis. Although many methods are available for measuring glucose metabolism-activities, from simple blood tests to positron emission tomography, currently there is no robust and affordable device that enables monitoring of glucose levels in real-time. In this study we tested feasibility of applying a unique resonance-frequency based electronic impedance spectroscopy (REIS) device that has been, recently developed to measure and monitor glucose metabolism levels using a phantom study. In this new testing model, a multi-frequency electrical signal sequence is applied and scanned through the subject. When the positive reactance of an inductor inside the device cancels out the negative reactance of the capacitance of the subject, the electrical impedance reaches a minimum value and this frequency is defined as the resonance frequency. The REIS system has a 24-bit analog-to-digital signal convertor and a frequency-resolution of 100Hz. In the experiment, two probes are placed inside a 100cc container initially filled with distilled water. As we gradually added liquid-glucose in increments of 1cc (250mg), we measured resonance frequencies and minimum electrical signal values (where A/D was normalized to a full scale of 1V). The results showed that resonance frequencies monotonously decreased from 243kHz to 178kHz, while the minimum voltages increased from 405mV to 793mV as the added amount of glucose increased from 0 to 5cc. The study demonstrated the feasibility of applying this new REIS technology to measure and/or monitor glucose levels in real-time in future.

Numerical modeling of sound propagation in the airways requires accurate knowledge of the airway geometry. These models are often validated using human and animal experiments. While many studies documented the geometric details of the human airways, information about the geometry of pig airways is scarcer. In addition, the morphology of animal airways can be significantly different from that of humans. The objective of this study is to measure the airway diameter, length and bifurcation angles in domestic pigs using computed tomography. After imaging the lungs of 3 pigs, segmentation software tools were used to extract the geometry of the airway lumen. The airway dimensions were then measured from the resulting 3 D models for the first 10 airway generations. Results showed that the size and morphology of the airways of different animals were similar. The measured airway dimensions were compared with those of the human airways. While the trachea diameter was found to be comparable to the adult human, the diameter, length and branching angles of other airways were noticeably different from that of humans. For example, pigs consistently had an early airway branching from the trachea that feeds the superior (top) right lung lobe proximal to the carina. This branch is absent in the human airways. These results suggested that the human geometry may not be a good approximation of the pig airways and may contribute to increasing the errors when the human airway geometric values are used in computational models of the pig chest.

Chronic obstructive pulmonary disease (COPD), is a major contributor to hospitalization and healthcare costs in North America. While the hallmark of COPD is airflow limitation, it is also associated with abnormalities of the cardiovascular system. Enlargement of the pulmonary artery (PA) is a morphological marker of pulmonary hypertension, and was previously shown to predict acute exacerbations using a one-dimensional diameter measurement of the main PA. We hypothesized that a three-dimensional (3D) quantification of PA size would be more sensitive than 1D methods and encompass morphological changes along the entire central pulmonary artery. Hence, we developed a 3D measurement of the main (MPA), left (LPA) and right (RPA) pulmonary arteries as well as total PA volume (TPAV) from thoracic CT images. This approach incorporates segmentation of pulmonary vessels in cross-section for the MPA, LPA and RPA to provide an estimate of their volumes. Three observers performed five repeated measurements for 15 ex-smokers with ≥10 pack-years, and randomly identified from a larger dataset of 199 patients. There was a strong agreement (r²=0.76) for PA volume and PA diameter measurements, which was used as a gold standard. Observer measurements were strongly correlated and coefficients of variation for observer 1 (MPA:2%, LPA:3%, RPA:2%, TPA:2%) were not significantly different from observer 2 and 3 results. In conclusion, we generated manual 3D pulmonary artery volume measurements from thoracic CT images that can be performed with high reproducibility. Future work will involve automation for implementation in clinical workflows.

Conversion of images at micro level of normal and with very early stage disease of the lung and quantitative analysis of morphology on CT image can contribute to the chest image diagnosis to the next generation. Previous, anatomy and pathology analysis of pulmonary lobule have been conducted to better understand the CT image of peripheral lung tissue disease. However, it is difficult to figure out three-dimensional (3D) conformation because of analyzing at the slice image. The purpose of this study is a 3D microstructual and quantitative analyses of pulmonary acinus with spatial resolution in the range of several micrometers by using a synchrotron radiation micro CT (SRμCT). In this paper, we present a semi-automatic method for segmenting the secondary pulmonary lobule into acinus or subacinus and extracting small vessel in human acinus imaged by the SRμCT.

Trabecular bone and its micro-architecture are of prime importance for health. Changes of bone micro-architecture are linked to different pathological situations like osteoporosis and begin now to be understood. In a previous paper, we started to investigate the relationships between bone and vessels and we also proposed to build a Bone Atlas. This study describes how to proceed for the elaboration and use of such an atlas. Here, we restricted the Atlas to legs (tibia, femur) of rats in order to work with well known geometry of the bone micro-architecture. From only 6 acquired bone, 132 trabecular bone volumes were generated using simple mathematical morphology tools. The variety and veracity of the created micro-architecture volumes is presented in this paper. Medical application and final goal would be to determinate bone micro-architecture with some angulated radiographs (3 or 4) and to easily diagnose the bone status (healthy, pathological or healing bone...).

Trabecular bone and its micro-architecture are of prime importance for health. Changes of bone micro-architecture are linked to different pathological situations like osteoporosis and begin now to be understood. In a previous paper [12], we started to investigate the relationships between bone and vessels and proposed some indices of characterization for the vessels issued from those used for the bone. Our main objective in this paper is to qualify the classical values used for bone as well as those we proposed for vessels according to different acquisition parameters and for several thresholding methods used to separate bone vessels and background. This study is also based on vessels perfusion by a contrast agent (barium sulfate mixed with gelatin) before euthanasia on rats. Femurs and tibias as well as mandibles were removed after rat’s death and were imaged by microCT (Skyscan 1272, Bruker, Belgium) with a resolution ranging from 18 to 3μm. The so obtained images were analyzed with various softwares (NRecon Reconstruction, CtAn, and CtVox from Bruker) in order to calculate bone and vessels micro-architecture parameters (density of bone/blood within the volume), and to know if the results both for bone and vascular micro-architecture are constant along the chosen pixel resolution. The result is clearly negative. We found a very different characterization both for bone and vessels with the 3μm acquisition. Tibia and mandibles bones were also used to show results that can be visually assessed. The largest portions of the vascular tree are orthogonal to the obtained slices of the bone. Therefore, the contrast agent appears as cylinders of various sizes.

Background: Cerenkov luminescence imaging (CLI) provides a great potential for clinical translation of optical molecular imaging techniques through using clinical approved radiotracers. However, it is difficult to obtain the Cerenkov luminescence signal of deeper biological tissues due to the small magnitude of the signal. To efficiently acquire the weak Cerenkov luminescence, we developed an endoscopic Cerenkov luminescence imaging (ECLI) system to reduce the in vivo imaging depth with minimum invasion, and validated the system on small animal tumor models. Methods: For the ECLI system, the laparoscope was connected to a high sensitive charge-couple device (CCD) camera (DU888+, Andor, UK) by a custom made adapter. We conducted a series of in vitro and in vivo experiments by use of the system. In the in vitro experiment, the endoscopic luminescence images of the ¹⁸F-FDG with various activities in EP tubes were acquired using ECLI system, and the sensitivity was compared with conventional CLI system. In the in vivo tumor experiment, ¹⁸F-FDG with the activity of 200μCi were intravenously injected into 3 tumor mice. Then the ECLI system was used to acquire the optical images for both non-invasive and invasive conditions. Conclusion: Experimental data showed the ECLI system could detect the ¹⁸F-FDG with the activity as low as 1μCi. Furthermore, our preliminary results indicated the possibility of ECLI technique for detecting Cerenkov signals inside the tumor tissue with deeper depth and guiding the surgical operation of tumor excision. We believe that this technique can help to accelerate the clinical translation of CLI.

To quantify low attenuation area (LAA) of emphysematous regions according to cluster size in 3D volumetric CT data of chronic obstructive pulmonary disease (COPD) patients and to compare these indices with their pulmonary functional test (PFT). Sixty patients with COPD were scanned by a more than 16-multi detector row CT scanner (Siemens Sensation 16 and 64) within 0.75mm collimation. Based on these LAA masks, a length scale analysis to estimate each emphysema LAA’s size was performed as follows. At first, Gaussian low pass filter from 30mm to 1mm kernel size with 1mm interval on the mask was performed from large to small size, iteratively. Centroid voxels resistant to the each filter were selected and dilated by the size of the kernel, which was regarded as the specific size emphysema mask. The slopes of area and number of size based LAA (slope of semi-log plot) were analyzed and compared with PFT. PFT parameters including DLco, FEV1, and FEV1/FVC were significantly (all p-value< 0.002) correlated with the slopes (r-values; -0.73, 0.54, 0.69, respectively) and EI (r-values; -0.84, -0.60, -0.68, respectively). In addition, the D independently contributed regression for FEV1 and FEV1/FVC (adjust R sq. of regression study: EI only, 0.70, 0.45; EI and D, 0.71, 0.51, respectively). By the size based LAA segmentation and analysis, we evaluated the Ds of area, number, and distribution of size based LAA, which would be independent factors for predictor of PFT parameters.

A revised version of this paper, published originally on 17 March 2015, was published on 2 July 2015, replacing the original paper. The text of the first paragraph of Section 2.2.2 has been revised and two additional references have been added. The text of the first paragraph of Section 2.3 has also been revised. The revised paper is available at http://dx.doi.org/10.1117/12.2081443.