Browsing by Subject "Magnetic Resonance Imaging"

Now showing 1 - 7 of 7
  • Thumbnail Image
    Item
    Bridging Mri Reconstruction Across Eras: From Novel Optimization Of Traditional Methods To Efficient Deep Learning Strategies
    (2024-03) Gu, Hongyi
    Magnetic Resonance Imaging (MRI) has been extensively used as a non-invasive modality for imaging the human body. Despite substantial advances over the past decades, scan duration remains as a principal issue for MRI scanning, requiring novel techniques to accelerate data acquisition. Such techniques are poised to improve clinical patient throughput, reduce motion artifacts, enhance subject comfort, and allow higher resolution imaging in many applications. Several methods have been proposed to accelerate MRI scans. In parallel imaging (PI), k-space data was acquired at a sub-Nyquist rate with with multiple receiver coils, and the redundancy among these coils were used for image reconstruction. Following the clinical impact and success of PI methods, compressed sensing (CS) techniques were developed to reconstruct images by using compressibility of images in a pre-specified linear transform domain. Transform learning (TL) was another line of work that learned the linear transforms from data, while enforcing sparsity as in CS. Recently, deep learning (DL) has shown great promise for MRI reconstruction, especially at high acceleration rates where other traditional methods would fail. Specially, physics-guided DL (PG-DL) unrolls a traditional optimization algorithm for solving regularized least squares for a fixed number of iterations, and uses neural networks to implicitly perform regularization. These unrolled networks are trained end-to-end with large databases, using well-designed loss functions and advanced optimizers, usually using a reference fully-sampled image for supervised learning. Several approaches have noted the difficulty or impossibility of acquiring fully-sampled data in various MRI applications. Among these, self-supervised learning with data undersmapling (SSDU) was developed to allow training without fully-sampled data, and multi-mask SSDU was subsequently proposed for better reconstruction quality at high acceleration rates. Although PG-DL generally shows strong ability for excellent reconstruction performance, there are concerns for generalizabilty, interpretability and stability issues. In this thesis, we aimed to bridge the gap between traditional and DL methods, while also extending the utility of DL methods for non-Cartesian imaging. We first revisited l1-wavelet CS reconstruction for accelerated MRI by using modern data science tools similar to those used in DL for optimized performance. We showed that our proposed optimization approach improved traditional CS, and further performance boost was observed by incorporating wavelet subband processing and reweighted l1 minimization. The final version reached a performance similar to state-of-the-art PG-DL, while preserving better interpretability by solving a convex optimization problem in inference time. Second, we combined ideas from CS, TL and DL to enable the learning of deep linear convolutional transforms in a format similar to PG-DL. Our proposed method performed better than CS and TL, and gave similar performance as state-of-the-art PG-DL. It used a linear representation of image as regularization at inference time, and enabled convex sparse image reconstruction that may have better robustness, stability and generalizability properties. Third, we adapted a self-supervised PG-DL technique to non-Cartesian trajectories and showed its potential for reconstructing 10-fold accelerated spiral fMRI multi-echo acquisitions. Our proposed approach gave substantial improvements in reconstructed image quality over conventional methods. Furthermore, the blood oxygenation level dependent (BOLD) signal analysis of our proposed method provided meaningful sensitivities, with similar activation patterns and extent to the expected baselines.
  • Thumbnail Image
    Item
    Magnetic Resonance Based Electrical Properties Tomography (Ept) Using Multi-Channel Transmission For Imaging Human Brain And Animal Cancer Models
    (2018-05) Wang, Yicun
    Electrical properties (EPs) of biological tissues are determined by tissue constituents, and therefore may provide novel biomarkers for characterization of diseased tissues such as cancer. In addition, accurate quantification of tissue EPs is essential for understanding the biological effects of electromagnetic radiation involved in MRI exams as well as wireless communication. In this dissertation, non-invasive EP imaging methods are proposed based on inverse problems using a plurality of radiofrequency electromagnetic field maps (B1 maps) acquired with ultra-high-field MRI. For human brain imaging, an automatic seed selection strategy is developed for gradient-based Electrical Properties Tomography (gEPT) to provide objective EP values. Reconstruction results of twelve healthy subjects demonstrate that considerable intra- and inter- subject EP heterogeneity resides in the normal brain, which may provide rationale for subject-specific mapping of EPs for improved accuracy in electromagnetic safety evaluation. Furthermore, a generalized technology called “CONtrast Conformed Electrical Properties Tomography (CONCEPT)” is developed based on transmit B1 maps and image sparsity. Numerical simulations and phantom experiments have been performed to quantify its accuracy and sensitivity. For rodent cancer model imaging, Boundary Informed Electrical Properties Tomography (BIEPT) technology is proposed based on a constrained inverse problem that exploits prior information and image sparsity. The imaging platform and BIEPT reconstruction method have been evaluated using simulations, phantom experiments and in vivo cancer xenograft imaging experiments. The reconstructed EPs are compared to multiple conventional MR contrasts as well as histopathology slides to demonstrate their potential value for cancer diagnosis and staging.
  • Thumbnail Image
    Item
    Provable Deterministic Sampling Strategies for Fourier Encoding in Magnetic Resonance Imaging
    (2019-08) Udawat, Gamini
    There is a constant demand for acceleration of magnetic resonance (MR) imaging to alleviate motion artifacts, and more generally, due to the time sensitive nature of certain imaging applications. One way to speed up MR imaging is to reduce the image acquisition time by subsampling the data domain (k-space). There are several methods available to reconstruct the MR image from undersampled k-space, e.g., those based on the theory of Compressive Sensing. Standard methods employ random undersampling of k-space; however, these methods provide only probabilistic guarantees on the quality of reconstruction. We present a method to reconstruct MR images from deterministically undersampled k-space, and provide analytical guarantees on the quality of MR image reconstruction. Our approach uses sampling constructions formed by deterministic selection of rows of Fourier matrices; coupled with sparsity assumptions on the finite differences of MR images, we formulate the reconstruction problem as a Total Variation (TV) minimization problem. We demonstrate the utility of our TV minimization based approach for MR image reconstruction by reconstructing MR brain scan data, and compare our reconstructions with those obtained via random sampling. Our results suggest that accurate MR reconstructions are possible by deterministic undersampling the k-space, and the quality of deterministic reconstructions are on par with those of reconstructions from randomly acquired data.
  • Thumbnail Image
    Item
    Relationship between Unilateral Temporomandibular Joint Arthralgia and Disc Positions and Degenerative Joint Changes- A Cross Sectional Study
    (2021-05) SHRIVASTAVA, MAYANK
    Background: This study simultaneously assessed for both temporomandibular joint (TMJ) disc displacement (DD) and degenerative joint disease (DJD) in participants with unilateral TMJ arthralgia using TMJ MRI and CBCT.Methods: In the multi-center TMJ Impact Project, 401 subjects were examined by calibrated examiners that included rendering a diagnosis of TMJ arthralgia. All subjects had bilateral TMJ MRIs and CTs. Two radiologists rendered a consensus diagnosis of normal, DD with reduction (DDwR), or DD without reduction (DDw/oR) using MRI. CBCT consensus diagnoses included normal or grade I DJD and grade II DJD. Radiologist reliability was assessed by kappa. Descriptive analysis was performed using generalized linear mixed models. Models include a random intercept to account for correlations within subject. The level of significance is p< 0.05. Results: Of the 401 subjects, 58 subjects had a clinical diagnosis of unilateral arthralgia. In 58 joints with arthralgia, 11(19%) had normal disc position, 19 (33%) had DDwR and 28 (48%) had DDw/oR compared to 58 joints without arthralgia: 13 (22%) were normal, 25 (43%) had DDwR and 20 (34%) had DDw/oR (p=0.32). In joints with arthralgia, 25 (43%) had normal osseous morphology and 33 (57%) had DJD compared to joints without arthralgia 32 (55%) had normal osseous morphology and 26 (44%) had DJD (p=0.20). Radiologist reliability was kappa: 0.73 (CI: 0.64–0.83) for DD and 0.76 (CI: 0.68-0.83) for DJD. Conclusion: The presence of arthralgia is not significantly related to the radiographic findings of DD and DJD.
  • Thumbnail Image
    Item
    Self-Supervised Physics-Guided Deep Learning for Solving Inverse Problems in Imaging
    (2022-03) Yaman, Burhaneddin
    Inverse problems in computational imaging seek to recover an unknown image of interest from observed measurements acquired using a known forward model. These inverse problems are often ill-conditioned, requiring some form of regularization. The corresponding objective function for inverse problems in computational imaging can often be solved using iterative optimization approaches that alternate between two sub-problems that enforce data consistency and promote the regularization approach, respectively. Such inverse problems arise in a multitude of imaging modalities, in particular in magnetic resonance imaging (MRI), which is the main application area for this thesis. Lengthy scan times remain a challenge in MRI, thus accelerating MRI scans has remained an open research problem over decades. Conventional accelerated MRI techniques acceleration rate is limited by noise amplification and residual artifacts. Recently, deep learning has emerged as an alternative approach for accelerated MRI. Among deep learning techniques, physics-guided deep learning (PG-DL) has drawn great interest, as it incorporates the known physical forward model into the network architecture. PG-DL unrolls a conventional iterative algorithm for solving a regularized least squares problem for a fixed number of iterations, and replaces the proximal operation corresponding to the regularizer implicitly with neural networks. These unrolled networks are trained end-to-end, with the goal of minimizing the difference between the network output and the corresponding reference data. Most of the existing deep learning approaches in MRI reconstruction are based on supervised learning, which requires ground-truth/ fully-sampled data for training. However, acquisition of fully-sampled data is infeasible in many applications due to physiological constraints, such as organ motion, or physical constraints, such as signal decay. In several other scenarios, such as high-resolution anatomical brain imaging, it is impractical to acquire fully-sampled datasets as the scan time becomes extremely lengthy. Therefore, enabling the training of PG-DL reconstruction without fully-sampled data is essential for the integration of deep learning reconstruction into clinical MRI practice. The present thesis introduces novel frameworks to enable the training of deep learning reconstruction methods for inverse imaging problems in the absence of ground-truth/fully-sampled data. First, we introduce self-supervised learning via data undersampling (SSDU) approach to enable database training without fully-sampled data. Succinctly, SSDU partitions available measurements into two disjoint sets. One of these sets is used in the data consistency units of the unrolled network, while the other is used to define the loss in the measurement domain. Subsequently, we extend SSDU for processing 3D datasets and provide solutions for GPU memory constraints and data scarcity issues faced in 3D processing. To cope with potential performance degradation at very high acceleration rates, we develop a multi-mask self-supervised learning approach, which retrospectively splits available measurements into multiple 2-tuples of disjoint sets to perform training and define a loss function. Furthermore, we introduce a zero-shot self-supervised learning approach to enable training from a single scan/sample without any external training databases. ZS-SSL partitions the available measurements from a single scan into three disjoint sets. Two of these sets are used to enforce data consistency and define loss during training for self-supervision, while the last set serves to self-validate, establishing an early stopping criterion. Finally, we introduce a self-supervised learning algorithm for referenceless image denoising. Self-supervised deep learning algorithms split the pixels for each image into two disjoint sets to perform training and defining loss. In existent self-supervised denoising approaches which are purely data-driven, the set of pixels used as input to the network is not re-utilized in the end-to-end training since the network is only comprised of a neural network. Reusing the pixels within the network would promote consistency with acquired measurements, thus leading to a more robust and improved denoising performance. To tackle this challenge, we build upon existent self-supervised learning algorithms and recast the denoising problem into a regularized image inpainting framework which allows use of algorithm unrolling for denoising.
  • Thumbnail Image
    Item
    Solving Inverse Problems in Medical Imaging Using Statistical Signal Processing and Deep Learning
    (2020-02) Hosseini, Seyed Amir Hossein
    Solving inverse problems remain an active research area in various fields to study the cause of a phenomenon by observing the effects. In particular, such efforts are well grounded in medical imaging applications where inverse problems naturally arise due to the imaging target being either inaccessible or invisible to human eyes. Non-regularity or ill-conditioning is a major challenge in such situations which is a direct consequence of limited observations/measurements being available. Medical imaging applications have classically incorporated domain-specific knowledge about the forward encoding operator to regularize the pertaining inverse problem. Deep learning has recently received high interest as an alternative approach for data-driven regularization. In principle, such regularization can be achieved by acquiring more measurements. However, sufficient measurements are usually unavailable due to practical considerations. For instance, in electromagnetic source imaging (ESI) using electroencephalography (EEG), number of recording electrodes can not be excessively increased due to subsequent distortions that electrodes contacts can cause in the medium. This limitation adversely affects the spatial resolution of EEG-based ESI. Another example is magnetic resonance imaging (MRI) where increasing imaging resolution trades off with noise properties, temporal resolution and scan duration. High-resolution MRI can become prohibitively long especially when a sequence of images need to be obtained. Thus, accelerated MRI techniques are required to reconstruct high-resolution images using less data. The present thesis contributes to solving regularization issues pertaining to inverse problems in medical imaging with a focus on high-resolution electromagnetic source imaging (ESI) using electroencephalography (EEG) and accelerated high resolution magnetic resonance imaging (MRI) with high fidelity. The ESI spatial resolution and the brain full coverage is shown to significantly improve when simultaneous scalp EEG and intracranial EEG recordings are jointly utilized. Additionally, a new ESI technique is proposed that incorporates forward modeling uncertainties into inverse problem formulations, considerably improving localization accuracy and recovering underlying source dynamics and extents when such uncertainties are properly determined and modeled. A novel deep learning-based reconstruction technique is proposed for accelerated MRI using a recurrent neural network architecture that offers desired noise properties compared to conventional techniques, particularly at higher acceleration rates. Furthermore, a new deep learning methodology is developed that unrolls inverse problem optimization algorithms into dense recurrent neural networks for improved reconstruction performance without requiring extra computational power or increasing reconstruction time.
  • Thumbnail Image
    Item
    The use of biomarkers to determine the severity of osteoarthritis in the tarsus of an older horse population
    (2017-12) Coppelman, Elizabeth
    Background: Osteoarthritis (OA) is a group of diseases of different causes that ultimately lead to synovitis, subchondral bone remodeling, and articular cartilage degeneration. OA commonly develops in the distal intertarsal (DIT) and tarsometatarsal (TMT) joints of performance horses. Currently, the most accurate method of identifying OA in these joints is a combination of thorough physical, lameness, and radiographic examinations. However, many horses may have pain attributed or localized to these joints with minimal radiographic changes present. A novel way to identify and classify the degree of OA is through the measurement of molecular biomarkers. Molecular biomarkers of OA reflect quantitative and dynamic variations associated with joint metabolism. Objectives: (1) define the direct and indirect biomarker concentrations in synovial fluid from the tibiotarsal, distal intertarsal, and tarsometatarsal joints in horses with varying degrees of tarsal OA, and (2) validate/refute that the biomarker concentrations in these joints increase with severity of OA in the distal joints as determined by a radiographs (all joints), MRI (PIT, DIT, TMT joints), arthroscopic evaluation/ gross pathology (TT joint), and (3) determine if biomarker concentrations in the TT synovial fluid (SF) can be used to evaluated OA severity in the DIT and TMT joints. Methods: A cohort study of 11 older horses (>8 years) with variable amounts of OA in the tarsal joints identified on radiographs were included. The TT joints were examined by arthroscopy/gross examination. The distal tarsal joints were examined by MRI. Biomarkers BAP, CPII, C2C, CTX II, CS846 were examined in the distal tarsal joints; additional biomarkers C1,2C, IL-1β, IL-6, IL-8, IL-10, and TNFα were examined in the TT joint. Various statistical analyses were used to determine association between imaging modalities and biomarkers to degrees of OA severity. Results: In the TT joint, C2C and IL-6 were the best biomarkers distinguishing OA severity. There was more pathology present in TT joints than could be seen on radiographs, suggesting that arthroscopic surgery is still the best method to evaluate TT joint OA. In the distal tarsal joints, radiographs were better at distinguishing OA and correlated to the corresponding MRIs, but underestimated the degree of SCB bone sclerosis, and number and size of osteophytes in many of the cases. MRI also provided information about cartilage damage and SCB hyperintensity. The severity of SCB sclerosis and presence of SCB hyperintensity on MRI was a good indicator for separating moderate/severe from mild OA. Of the biomarkers evaluated, the best at determining OA severity in the DIT joint were BAP, CPII, and C2C. These biomarkers also correlated to subchondral bone hyperintensity on MRI. In the TMT joint, CPII was the best biomarker to determine OA severity. No biomarker was identified in the SF in the TT joint capable of identifying OA in the distal tarsal joints. Conclusions: Biomarkers have the potential to be a valuable source of information about the OA disease process in the tarsal joints. SF from the joint of interest must be collected. Further research is needed with more horses. To the author’s knowledge, this is the first study examining biomarkers in an older horse population and hopes it provides a template for forthcoming research.