Doctor of Philosophy, Case Western Reserve University, 2016, Biomedical Engineering

The analysis of medical images--from magnetic resonance imaging (MRI) to digital pathology--for disease characterization typically involves extraction of hundreds of features, which may be used to predict disease presence, aggressiveness, or outcome. Unfortunately, the dimensionality of the feature space poses a formidable challenge to the construction of robust classifiers for predicting disease presence and aggressiveness. In this work we present novel strategies to facilitate the construction of robust, interpretable classifiers when the dimensionality of the feature space is high. In the context of prostate cancer, we demonstrate the benefit of our approach for identifying (a) radiomic features that are useful for detecting prostate cancer on multi-parametric MRI, (b) radiomic features that predict the risk of prostate cancer recurrence on T2-weighted MRI, and (c) histomorphometric features describing cellular and glandular architecture on digital pathology images that predict the risk of prostate cancer recurrence post-treatment. In the context of breast cancer, we identify histomorphometric features describing cancer patterns in estrogen receptor positive (ER+) breast cancer tissue slides that can predict (a) which cancer patients will have recurrence following treatment with tamoxifen and (b) risk category as determined by a 21 gene expression assay called Oncotype DX. Additionally, we also investigate whether radiomic features characterizing prostate tumors that manifest in the peripheral zone of the prostate are different from radiomic features characterizing transition zone tumors, and we develop a novel approach for pharmacokinetic modeling on dynamic contrast-enhanced MRI that relies exclusively on prostate voxels, with no reliance on an arterial input function or reference tissue.

Committee: Anant Madabhushi (Advisor) Subjects: Biomedical Engineering; Medical Imaging; Radiology

Doctor of Philosophy, The Ohio State University, 2015, Neuroscience Graduate Studies Program

The motor system consists of multiple regions of the central nervous system involved in the control of movement. Because each component of the motor system contributes to a specific motor function, clinical signs and symptoms of motor impairment can often be used to deduce the nature and location of a neurological lesion. In this way, a better understanding of neuroanatomical pathways and functional connections between motor areas leads directly to improvements in the diagnosis and treatment of motor system dysfunction. The purpose of this dissertation was to utilize electrophysiological and computational techniques to study the motor outputs of the pontomedullary reticular formation (PMRF) and the computer-aided diagnosis (CAD) of parkinsonism. Electrophysiological techniques are of particular usefulness in the study of motor function. In a research setting, electrical stimulation can be used to evoke neuronal action potentials. In the clinic, electroencephalography (EEG), nerve conduction studies, and electromyography (EMG) are used to assess motor system function for diagnosing disease and tracking its progression. Additionally, procedures such as transcranial direct-current stimulation and deep brain stimulation (DBS) can be used to modify brain activity during the treatment of certain disorders of the motor system. Computational methods are important in signal processing of electrophysiological recordings and modeling of motor pathways. Furthermore, machine learning algorithms are used in the CAD of neurological disorders. The first study in this dissertation used electrophysiological techniques to study the motor outputs of the PMRF. The PMRF is the origin of the reticulospinal tract, one of the major descending motor pathways. The reticulospinal system is of particular importance following damage to the corticospinal tract. Unilateral cortical injury and motor cortex stroke, which cause corticospinal neuron death, classically result in contra (open full item for complete abstract)

Committee: John Buford PhD (Advisor); Hojjat Adeli PhD (Committee Co-Chair); Dana McTigue PhD (Committee Member); Per Sederberg PhD (Committee Member) Subjects: Neurosciences

MS, University of Cincinnati, 2014, Engineering and Applied Science: Computer Science

Glioblastoma is a form of malignant brain tumor in humans involving glial or non-neuronal cells. The state-of-the-art diagnosis of Glioblastoma is predominantly based on subjective opinion of trained pathologists. However, with the availability of large-scale databases of Glioblastoma histopathology images, it is now possible, in principle, to objectively study and classify this class of tumors via image analysis and pattern recognition techniques. The objective of this work is to develop a quantitative framework for the analysis of Glioblastoma. The first, fundamental step in this process is the identification of histological structures in these images, that is, segmenting the constituent nuclei in the tissue. The work presents a two-step process of iterative thresholding and cleaving (ITC) to identify aforementioned structures. This improves significantly over standard color-based cell segmentation techniques in identifying cellular structures, giving 91.8% precision and 94.7% recall. Furthermore, using various architectural features obtained from each image, it ensures that the identification of regions important for the diagnosis process is distinctly clearer using the ITC approach than with standard approaches such as the Otsu method and adaptive thresholding.

Committee: Anca Ralescu Ph.D. (Committee Chair); Fred Annexstein Ph.D. (Committee Member); Bruce Aronow Ph.D. (Committee Member) Subjects: Computer Science

Doctor of Philosophy (PhD), Wright State University, 2009, Engineering PhD

We have established a modular virtual framework to design accurate, robust, efficient and cost-conscious autonomous target/object detection systems. Developed primarily for image-based detection problems, such as automatic target detection or computer-aided diagnosis, our approach is equally suitable for non-image-based pattern recognition problems. The framework features six modules: 1) the detection algorithm module accepts two-dimensional, spatially-coded sensor outputs; 2) the evaluation module uses our receiver operator characteristic (ROC)-like assessment tool to evaluate and fine-tune algorithm outputs; 3) the fusion module compares outputs combined under various fusion schemes; 4) the classifier selection module exploits the double-fault diversity measure (F2 DM) to identify the best classifier; 5) the weighting module judiciously weights the algorithm outputs to fine-tune classifiers, and 6) the cost-function analysis module determines the best detection parameters based on the trade-off between the costs of missed targets and false positive detections. Our solution can be generalized to facilitate detection system design in various applications, including target detection, medical diagnosis, biometrics, surveillance, machine vision, etc. For proof-of-principle, the framework was implemented for the autonomous detection of roadside improvised explosive devices (IEDs). From our set of nine multimodal detection algorithms that yield 1,536 possible classifiers, we identified the single best classifier to accomplish the detection task under a defined cost specification. System performance was tracked through each module and compared to standard approaches for system definition. Algorithm parameter optimization improved performance by an average of 18% (range of 3-32%). Our F2 DM-based classifier selection module predicted classifier performance with an average difference of 3% (standard deviation = ± 2%) from ROC area under the curve (AUC) predictions and an as (open full item for complete abstract)

Committee: Julie Skipper Ph.D. (Advisor); Kenneth Bauer Ph.D. (Committee Member); Fred Garber Ph.D. (Committee Member); Thomas Hangartner Ph.D. (Committee Member); Brian Rigling Ph.D. (Committee Member) Subjects: Biomedical Research; Electrical Engineering; Engineering; Health Care; Information Systems; Remote Sensing; Scientific Imaging; Systems Design

Master of Science in Engineering (MSEgr), Wright State University, 2008, Biomedical Engineering

Our challenge was to develop a semi-automatic target detection algorithm to aid human operators in locating potential targets within images. In contrast to currently available methods, our approach is relatively insensitive to image brightness, image contrast and object orientation. Working on overlapping image blocks, we used a sliding difference method of histogram matching. Incrementally sliding the histograms of the known object template and the image region of interest (ROI) together, the sum of absolute histogram differences was calculated. The minimum of the resultant array was stored in the corresponding spatial position of a response surface matrix. Local minima of the response surface suggest possible target locations. Because the template contrast will rarely perfectly match the contrast of the actual image contrast, which can be compromised by illumination conditions, background features, cloud cover, etc., we perform a random contrast manipulation, which we term ‘wobble', on the template histogram. Our results have shown improved object detection with the combination of the sliding histogram difference and wobble.

Committee: Julie Skipper PhD (Advisor); Daniel Repperger PhD (Committee Member); Thomas Hangartner PhD (Committee Member); S. Narayanan PhD (Other); Joseph F. Thomas, Jr. PhD (Other) Subjects: Biomedical Research; Engineering; Scientific Imaging

Doctor of Philosophy, The Ohio State University, 2008, Electrical and Computer Engineering

Dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) is considered to have great potential in cancer diagnosis and monitoring. During the DCE-MRI procedure, repeated MRI scans are used to monitor contrast agent movement through the vascular system and into tissue. By observing the vascular permeability characteristics, radiologists can detect and classify malignant tissues. When used for diagnostic purposes, the DCE-MRI procedure often requires manual detection, classification, and marking of tumor tissues. This process can be time consuming and fatiguing especially when multiple DCE-MRI procedures must be processed to monitor the progress of a cancer therapy. Manual analysis also suffers from inter- and intra-observer variations which can lead to lesion segmentation inconsistencies. The goal of this dissertation research is to design and develop a tool to aid radiologists, researchers, and clinicians in the detection, segmentation, and analysis of malignant lesions from DCE-MRI datasets. The diagnostic tool presented in this research is model independent, speeds analysis, and provides more consistent segmentations. The approach of the project is to apply statistical 4-D image texture analysis features along with a classifier (such as a neural network) to analyze DCE-MRI datasets. Performance of the computer aided diagnosis (CAD) tool for this project is demonstrated with breast and prostate DCE-MRI data. Training methodology is reported so that extension to other types of cancers and anatomical regions is made possible. Results from the computer assisted diagnostic tool are compared with manual analysis performed by radiologists. The specific research aims of this dissertation are: a) provide a tool for quantitative and quick DCE-MRI analysis by providing radiologists a segmentation (for semi-automatic or automatic application), b) quantify inter- and intra-observer variations that occur during manual lesion segmentation and compare performance with comp (open full item for complete abstract)

Committee: Bradley Clymer PhD (Advisor); Ashok Krishnamurthy PhD (Committee Member); Tahsin Kurc PhD (Committee Member) Subjects: Artificial Intelligence; Bioinformatics; Biomedical Research; Computer Science; Electrical Engineering; Radiology

Doctor of Philosophy, Case Western Reserve University, 2008, Biomedical Engineering

Optical coherence tomography (OCT) is an emerging optical technique based on low-coherence interferometry that provides noninvasive, subsurface, high-resolution imaging of biological microstructure. Endoscopic OCT (EOCT) differentiates the tissue layers of the gastrointestinal (GI) wall and can identify dysplasia in the mucosa. The objective of this research was the characterization of dysplasia in the GI tract, by EOCT in Barretts esophagus in the upper GI tract and by OCT in colonic crypts in the lower GI tract. Barretts esophagus (BE) and associated adenocarcinoma have emerged as a major health care problem over the last two decades. We developed computer-aided diagnosis (CAD) algorithms, utilizing dyplasia-related image features analysis with biopsy site classification, to aid in the identification of dysplasia in BE using EOCT imaging. A total of 96 EOCT image-biopsy pairs (63 non-dsyplastic, 26 low-grade and 7 high-grade dysplastic biopsy sites) were analyzed using our CAD algorithm. Non-dysplastic vs. dysplastic classification, using eight frames per biopsy site and requiring at least the first three frames to be positive to classify the site as positive, yielded a sensitivity of 0.76 (95% CI: 0.55-0.90) and a specificity of 0.85 (95% CI: 0.66-0.94). CAD has the potential to enable EOCT surveillance of large surface areas of Barretts mucosa to identify dysplasia. Colorectal cancer is the second leading cause of cancer-related death in the United States. Approximately 50% of these deaths may be prevented by earlier detection through screening. The relationship between colonic crypts' morphological patterns and histopathological diagnosis has shown close correlation. We conducted an in vitro colonic tissue study to quantify the morphological features of colonic crypts using our microscope-integrated bench-top OCT scanner. 2D microscopic colonic crypt images with correlated 3D OCT volume of ROI were segmented using marker-based watershed segmentation. Colonic cr (open full item for complete abstract)

Committee: Andrew M. Rollins PhD (Committee Chair); Michael V. Jr. Sivak MD (Committee Member); David L. Wilson PhD (Committee Member); Jiayang Sun PhD (Committee Member) Subjects: Biomedical Research