This thesis compares and contrasts currents- and varifolds-based diffeomorphic image registration approaches for registering tree-like structures in the lung and surface of the lung. In these approaches, curve-like structures in the lung—for example, the skeletons of vessels and airways segmentation—and surface of the lung are represented by currents or varifolds in the dual space of a Reproducing Kernel Hilbert Space (RKHS). Currents and varifolds representations are discretized and are parameterized via of a collection of momenta. A momenta corresponds to a line segment via the coordinates of the center of the line segment and the tangent direction of the line segment at the center. A momentum corresponds to a mesh via the coordinates of the center of the mesh and the normal direction of the mesh at the center. The magnitude of the tangent vector for the line segment and the normal vector for the mesh are the length of the line segment and the area of the mesh respectively. A varifolds-based registration approach is similar to currents except that two varifolds representations are aligned independent of the tangent (normal) vector orientation. An advantage of varifolds over currents is that the orientation of the tangent vectors can be difficult to determine especially when the vessel and airway trees are not connected. In this thesis, we examine the image registration sensitivity and accuracy of currents- and varifolds-based registration as a function of the number and location of momenta used to represent tree like-structures in the lung and the surface of the lung. The registrations presented in this thesis were generated using the Deformetrica software package, which is publicly available at www.deformetrica.org.

Lung disease affects tens of millions of Americans, making it one of the most common medical conditions in the United States. Many of these lung diseases are classified as chronic airway disease. Because of this, it is important to be able to catch the development early so as to begin treatment as soon as possible to delay the progression and subsequently monitor that progression. One method of doing so is the use of quantitative computed tomography (CT). Study of the airway anatomy can be quantified using such measures as minor inner diameter (MinD), major inner diameter (MajD), wall thickness (WT), inner area (IA), and outer area (OA). Changes in these measures can then be tracked over time to determine how the airways are being affected by disease. The challenge with the desired longitudinal imaging is that prolonged radiation exposure can be dangerous to the patient. In order to make longitudinal imaging more feasible, it is important to determine what quantitative measures can reliably be made at different radiations doses so as to optimize radiation dose and quantitative assessment. Working to make this determination, three different radiation doses were tested to evaluate their quantitative outputs. A high dose (14.98 mGy), medium dose (6.00), and low dose (0.74 mGy) were used to image six different porcine subjects. Images were collected at these doses both while the lungs were in-vivo and once the lungs had been fixed and excised ex-vivo. All of the scans were then processed using APOLLO (VIDA Diagnostics). From the complete airway trees, quantitative measures were collected from thirty-five airways. For the whole lung analysis, the medium and low dose in-vivo scans and the high dose ex-vivo scans were compared to the high dose in-vivo scans to compare assess MinD, MajD, WT, IA, and OA. Then, in order to determine how well the CT measures represent the actual anatomy, a total of thirteen cube samples containing airways were segmented from one of the lungs (based on volume analysis of the lung pre- and post-fixation and visual inspection). The cubes were imaged in CT, for the purpose of aiding in the establishment of original location and studying the effect of a narrowed imaging window, and microscopic CT (μCT). Since μCT can have a resolution on the scale of microns, the values measured in these images were considered ground-truth. The CT and μCT cubes were then registered to the high dose ex-vivo scan so as to compare the cube values with the ex-vivo values from each of the three doses. The same five measures were collected and analyzed. The MinD, MajD, WT, IA, OA were statistically analyzed between the three in-vivo radiation dose scan sets, the high dose in- and ex-vivo scans, and the µCT cube, CT cube, and the three ex-vivo radiation dose sets. Preliminary results for the in-vivo scans show that the low dose and medium dose scans can reliably (< 5% error) be used to evaluate airways with minor diameters between 3.42 mm and 10.34 mm and major diameters between 3.98 mm and 12.06 mm. Comparison of the high-dose in-vivo and ex-vivo scans showed that the fixation and excision of the lungs did not significantly affect the ex-vivo lungs' ability to be used as a model for the in-vivo lungs. Finally, analysis of the various forms of the ex-vivo airways showed that there were few statistically significant differences between the datasets. These results support the use of using the low (0.74 mGy) radiation dose when studying airway disease affecting airways with minor diameters between 3.42 mm and 10.34 mm and major diameters between 3.98 mm and 12.06 mm. They also show that the quantitative measures from CT are representative of the true measures of the airways.

Graphics Processing Units (GPUs) have grown in popularity beyond the original video game enthusiast audience. They have been embraced by the high-performance computing community due to their high computational throughput, low cost, low energy demands, wide availability, and ability to dramatically improve application performance. In addition, as hybrid computing continues into mainstream applications, the use of GPUs will continue to grow. However, due to architectural difference between the CPU and GPU, adapting CPU-based scientific computing applications to fully exploit the potential speedup that GPUs offer is a non-trivial task. Algorithms must be designed with the architecture benefits and limitations in mind in order to unlock the full performance gains afforded by the use GPU. In this work, we develop fast GPU methods to improve the performance of two important components in computational lung modeling - image registration and particle tracking. We first propose a novel method for multi-level mass-preserving deformable image registration. The strength of this method is that it allows for flexibility of choice for the similarity criteria to be used by the registration method, making possible the implementation of simple and complex similarity measures on the GPU with excellent performance results. The method is tested using three similarity criteria for registering two CT lung datasets - the commonly used sum of squared intensity differences (SSD), the sum of squared tissue value differences (SSTVD), and a symmetric version of SSTVD currently being developed by our research group. The GPU method is validated against a previously validated single-threaded CPU counterpart using six healthy human subjects, and demonstrated strong agreement of results. Separately, three GPU methods were developed for tracking particle trajectories and deposition efficiencies in the human airway tree, including a multiple-GPU method. Though parallelization was straightforward, the complex geometry of the lungs and use of an unstructured mesh provided challenges that were addressed by the GPU methods. The results of the GPU methods were tested for various numbers of particles and compared to a previously validated single-threaded CPU version and demonstrated dramatic speedup over the single-threaded CPU version and 12-threaded CPU versions.

Abstract Purpose: The purpose of this paper is to quantify if using an average intensity projection (AIP) scan or a 30% phase (mid-ventilation surrogate, MidV) scan as the reference image for patient position verification affects reproducibility of lung cancer patient alignment under free-breathing cone beam computed tomography (CBCT) image guidance and to analyze the effects of common clinical issues on registration variability. Methods: AIPs were retrospectively created for 16 lung patients (14 SBRT, 2 conventional treatments) originally planned/treated using the 30% phase MidV surrogate scan as reference. The study included 3-5 CBCTs from each patient. Registrations were performed between the AIP-CBCT and between the MidV-CBCT by 5 individuals (student, medical physics resident, medical resident, medical physicist, and attending physician) using MIM 6.2 image registration platform (Beachwood, OH). The images were rigidly registered, internal tumor volume (ITV) contours were displayed, and no rotational adjustments were allowed to reflect real treatment conditions. Additionally, the registrations for AIP-CBCT and MidV-CBCT were repeated 3 times by one individual for intra-observer variability assessment. Patient setup rotations, tumor volume, tumor motion, and breathing variability were estimated for correlation with registration variability. Results: The magnitude of the average intra-observer standard deviations from the lateral (LAT), anterior-posterior (AP), and superior-inferior (SI) directions for the AIP/CBCT and MidV/CBCT registrations were 0.9 mm and 1.2 mm, respectively. The magnitude of the average inter-observer standard deviations for the AIP/CBCT and MidV/CBCT were 1.7 mm and 1.8 mm, respectively. Average discrepancies over the whole population were found to be small; however, some individual patients presented high variability. Patient-specific cases with high variability were analyzed and observations on its potential causes are discussed. Conclusion: The differences in alignment using AIP versus MidV as the reference images are, when averaged over the population studied, very small and clinically irrelevant for PTV margins > 5mm; however, individual patients may be impacted in a clinically relevant manner if smaller margins, 3 mm and below, are used instead.

The millennial generation has become the largest generation in the United States. Yet as more members of this generation reach voting age, their propensity to vote remains stagnant. For instance, in the 2016 U.S. presidential election, less than 50% of eligible millennials voted, in comparison to the 69% of baby boomers and 63% of Generation X. Voting is a civic duty essential to a successful democracy; therefore, it is imperative to find solutions to increase millennial political engagement. As millennials represent the largest proportion of users of social media, the purpose of this quantitative study was to examine the relationships between voter registration and voting rates and social media usage. To provide clarification on the issue of millennial voting and voter registration, a conceptual framework was used to explore whether a connection exists between millennial political participation and social media because existing theory was insufficient to address this issue. Using secondary data from the 2016 Millennial Impact Report, 1,050 millennial survey responses were gathered on millennial social media usage, intent to vote, and voter registration. A 2 proportions z-test was used to conclude that there was no difference in voter registration and voting rates between millennials who posted 1 to 3 times per week and those who posted 4 to 7 times per week on social media. This study may promote social change by informing those who seek solutions to increase millennial voting and voter registration rates for the continuation of the American democratic system.

Registration of medical imaging is essential for aligning in time and space different modalities and hence consolidating their strengths for enhanced diagnosis and for the effective planning of treatment or therapeutic interventions. The primary objective of this study is to develop an integrated registration method that is effective for registering both brain and whole-body images. We seek in the proposed method to combine in one setting the excellent registration results that FMRIB Software Library (FSL) produces with brain images and the excellent results of Statistical Parametric Mapping (SPM) when registering whole-body images. To assess attainment of these objectives, the following registration tasks were performed: (1) FDG_CT with FLT_CT images, (2) pre-operation MRI with intra-operation CT images, (3) brain only MRI with corresponding PET images, and (4) MRI T1 with T2, T1 with FLAIR, and T1 with GE images. Then, the results of the proposed method will be compared to those obtained using existing state-of-the-art registration methods such as SPM and FSL. Initially, three slices were chosen from the reference image, and the normalized mutual information (NMI) was calculated between each of them for every slice in the moving image. The three pairs with the highest NMI values were chosen. The wavelet decomposition method is applied to minimize the computational requirements. An initial search applying a genetic algorithm is conducted on the three pairs to obtain three sets of registration parameters. The Powell method is applied to reference and moving images to validate the three sets of registration parameters. A linear interpolation method is then used to obtain the registration parameters for all remaining slices. Finally, the aligned registered image with the reference image were displayed to show the different performances of the 3 methods, namely the proposed method, SPM and FSL by gauging the average NMI values obtained in the registration results. Visual observations are also provided in support of these NMI values. For comparative purposes, tests using different multi-modal imaging platforms are performed.

Extreme Learning Machine (ELM) is a training algorithm for Single-Layer Feed-forward Neural Network (SLFN). The difference in theory of ELM from other training algorithms is in the existence of explicitly-given solution due to the immutability of initialed weights. In practice, ELMs achieve performance similar to that of other state-of-the-art training techniques, while taking much less time to train a model. Experiments show that the speedup of training ELM is up to the 5 orders of magnitude comparing to standard Error Back-propagation algorithm. ELM is a recently discovered technique that has proved its efficiency in classic regression and classification tasks, including multi-class cases. In this thesis, extensions of ELMs for non-typical for Artificial Neural Networks (ANNs) problems are presented. The first extension, described in the third chapter, allows to use ELMs to get probabilistic outputs for multi-class classification problems. The standard way of solving this type of problems is based 'majority vote' of classifier's raw outputs. This approach can rise issues if the penalty for misclassification is different for different classes. In this case, having probability outputs would be more useful. In the scope of this extension, two methods are proposed. Additionally, an alternative way of interpreting probabilistic outputs is proposed. ELM method prove useful for non-linear dimensionality reduction and visualization, based on repetitive re-training and re-evaluation of model. The forth chapter introduces adaptations of ELM-based visualization for classification and regression tasks. A set of experiments has been conducted to prove that these adaptations provide better visualization results that can then be used for perform classification or regression on previously unseen samples. Shape registration of 3D models with non-isometric distortion is an open problem in 3D Computer Graphics and Computational Geometry. The fifth chapter discusses a novel approach for solving this problem by introducing a similarity metric for spectral descriptors. Practically, this approach has been implemented in two methods. The first one utilizes Siamese Neural Network to embed original spectral descriptors into a lower dimensional metric space, for which the Euclidean distance provides a good measure of similarity. The second method uses Extreme Learning Machines to learn similarity metric directly for original spectral descriptors. Over a set of experiments, the consistency of the proposed approach for solving deformable registration problem has been proven.

There is notable heterogeneity in clinical presentation of patients with chronic obstructive pulmonary disease (COPD). Classification of COPD is usually based on the severity of airflow limitation (pre- and post- bronchodilator FEV1), which may not sensitively differentiate subpopulations with distinct phenotypes. A recent advance of quantitative medical imaging and data analysis techniques allows for deriving quantitative computed tomography (QCT) imaging-based metrics. These imaging-based metrics can be used to link structural and functional alterations at multiscale levels of human lung. We acquired QCT images of 800 former and current smokers from Subpopulations and Intermediate Outcomes in COPD Study (SPIROMICS). A GPU-based symmetric non-rigid image registration method was applied at expiration and inspiration to derived QCT-based imaging metrics at multiscale levels. With these imaging-based variables, we employed a machine learning method (an unsupervised clustering technique (K-means)) to identify imaging-based clusters. Four clusters were identified for both current and former smokers. Four clusters were identified for both current and former smokers with meaningful associations with clinical and biomarker measures. Results demonstrated that QCT imaging-based variables in patients with COPD can derive statistically stable and clinically meaningful clusters. This sub-grouping can help better categorize the disease phenotypes, ultimately leading to a development of an efficient therapy.

The aim of this study is to investigate the functional and structural differences between normal subjects and asthmatics via image registration and computational fluid dynamics (CFD), together with pulmonary function test's (PFT) and one-image-based variables. We analyzed three populations of CT images: 50 normal, 42 non-severe asthmatic and 52 severe asthmatic subjects at total lung capacity (TLC) and functional residual capacity (FRC). A mass preserving image registration technique was employed to match CT images at TLC and FRC for assessments of regional volume change and anisotropic deformation. Instead of existing threshold-based air-trapping measure, a fraction-based air-trapping measure was proposed to account for inter-site and inter-subject variations of CT density. We also analyzed structural alterations of asthmatic airways, including bifurcation angle, hydraulic diameter, luminal area and wall area. CFD and particle tracking simulations are employed with physiologically-consistent boundary condition. As compared with normal subjects, severe asthmatics exhibit reduced air volume change (consistent with air-trapping) and more isotropic deformation in the basal lung regions, but increased air volume change associated with increased anisotropic deformation in the apical lung regions. In the multi-center study, the traditional air-trapping measure showed the significant site-variability due to the differences of scanners and coaching methods. The proposed fraction-based air-trapping measure is able to overcome the inter-site and inter-subject variations, allowing analysis of large data sets collected from multiple centers. We further demonstrate alterations of bifurcation angle, constriction, wall thickness and non-circularity at local branch level in severe asthmatics. The bifurcation angle, non-circularity and especially reduced hydraulic diameter significantly affect the increase of particle deposition in severe asthmatics. In summary, the two-image registration-based deformation provides a tool for distinguishing differences in lung mechanics among populations. The new fraction-based air-trapping measure significantly improves the association of air-trapping with the presence and severity of asthma and the correlation with forced expiratory volume in 1 second over forced vital capacity (FEV1/FVC) than existing approaches. The altered functions and structures such as air-volume change, branching angles, non-circular shapes, wall thickness and hydraulic diameters that found in asthmatics are strongly associated with the flow structures and particle depositions.

Predicting breast tissue deformation is of great significance in several medical applications such as surgery, biopsy and imaging. In breast surgery, surgeons are often concerned with a specific portion of the breast, e.g., tumor, which must be located accurately beforehand. Also clinically it is important for combining the information provided by images from several modalities or at different times, for the planning and guidance of interventions. Multi-modality imaging of the breast obtained by mammography, MRI and PET is thought to be best achieved through some form of data fusion technique. However, images taken by these various techniques are often obtained under entirely different tissue configurations, compression, orientation or body position. In these cases some form of spatial transformation of image data from one geometry to another is required such that the tissues are represented in an equivalent configuration. We constructed the 3D biomechanical models for this purpose using Finite Element Methods (FEM). The models were based on phantom and patient MRIs and could be used to model the interrelation between different types of tissue by applying displacements of forces and to register multimodality medical images.