Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2007. / Includes bibliographical references (p. 157-168). / Spectrally encoded confocal microscopy (SECM) is a technique that facilitates the incorporation of confocal microscopy into small, portable clinical instruments. This would allow in vivo evaluation of cellular and sub-cellular features in a non-destructive, minimally invasive manner. Prior studies have demonstrated the potential of the techniques as well as highlighted the need for faster acquisition rates and higher sensitivity. In this thesis, new laser sources, optical fiber arrangements and probe designs are explored to ultimately evaluate SECM's relevance as a clinical tool. Clinical imaging at cellular scales requires imaging rates on the order of tens of frames per second to reduce motion artifacts from unavoidable patient movements. Rapid SECM imaging was achieved through the development of a novel wavelength swept laser which simultaneously provided high output power (> 10mrW), narrow linewidth (10GHz), broad wavelength tuning (80 nm centered at 1310 nm) and fast repetition rates (up to 16,000 Hz), while being compact and environmentally stable. Imaging with a wavelength swept SECM system was characterized by coupling the laser to a tabletop imaging arm comprising a high density holographic grating, a galvanometer mounted mirror and a 0.9 NA water immersion microscope objective. / (cont.) Rapid SECM imaging is performed at a transverse resolution of 1.4 microns, axial resolution of 6 microns over a field of view of 440x440 microns and allows subcellular imaging ex vivo (excised specimens) and in vivo (human skin). A study on 40 excised head and neck specimens showed that SECM has the potential to perform tissue identification, but also revealed the presence of speckle noise due to the coherent nature of the illumination and collection schemes through a single mode optical fiber. A partially coherent system based on single mode fiber for illumination and multimode fiber for detection was simulated, implemented and tested to find adequate balance between attenuation of speckle noise and conservation of resolution. A coupling of 20 modes was found to reduce speckle by a factor 4.5 with a minimal sectioning penalty of 0.25, while allowing a signal increase of 8dB. This improvement in sensitivity allowed SECM table top system to be used for investigations in developmental biology where Dual clad fibers (DCF) were previously shown to allow partially coherent endoscopic imaging, using the single mode core for illumination and inner clad for multimodal collection. / (cont.) Commercially available DCF's which propagate thousands of modes are ill suited for confocal endoscopes as collecting such a number of modes would destroy the axial resolution. Based on results from the previous section and through modal analysis, a DCF was designed, drawn - via a collaboration with Boston University Photonics Center -, and tested for use with SECM. The prototype DCF yielded promising results (3 fold speckle attenuation, optical sectioning degradation of 0.85), and showed the need for implementation of better coupling mechanisms to take advantage of increased signal collection. Finally, a portable SECM system was built for in vivo evaluation of pediatric vocal fold. A preliminary study on porcine and cadaveric tissue showed that SECM can distinguish between epithelium, superior and intermediate layers of the lamina propria, which could help elucidate the development mechanism of the voice apparatus if performed in vivo. The handheld instrument comprises a custom grating scanner imaging the scanning pivot onto the back pupil of a high NA microscope objective. The imaging tube can easily be interchanged to accommodate geometrical constraints imposed by different age groups. / (cont.) The probe, currently under review by the biomedical engineering committee, revealed cellular and sub cellular details of human skin in vivo at depth and acquisition rates sufficient to capture blood cells flowing through capillaries. Through major improvements in acquisition speeds, sensitivity, and speckle appearance, this work established SECM as a potent clinical and biological imaging tool. Ultimate confirmation will be revealed through in vivo studies to come, but limitations are likely to be of engineering nature rather than from physical considerations. Future work should explore the possibility to combine SECM with other contrast mechanisms to provide imaging with increased specificity. / by Caroline Boudoux. / Ph.D.

Thesis (Sc. D.)--Harvard University--Massachusetts Institute of Technology Division of Health Sciences and Technology, Program in Medical Engineering and Medical Physics, 1990. / Includes bibliographical references (leaves 233-249). / by Robert Lie-Yuan Sah. / Sc.D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2009. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 53-56). / In recent years, significant advances have been made in the realm of in vitro diagnostics with the development of novel tests which are able to meaningfully impact the course of a patients' disease management. This transformation has strained the traditional in vitro diagnostic business model and raised questions as to whether the economics support the commercial development of these tests. The goal of this study is to evaluate the economics of in vitro diagnostics from development to commercialization, with a focus on a specific a class of novel and complex tests called In Vitro Diagnostic Multivariate Index Assays (IVDMIA). My hypothesis is that the current dynamics of the market can only sustain a small number of such novel tests. To evaluate this hypothesis, I construct an economic model of the development of a hypothetical new in vitro diagnostics which focuses on both the cost of development and commercialization together with market potential and adoption. The analysis reviews specific break-even scenarios to determine the parameters which would allow for an economically viable complex in vitro diagnostic. The conclusion I reach based on this analysis is that only a very small number of medical conditions could economically support the development of a novel in vitro diagnostic. The medical conditions which could support the development of a novel test are governed by complexity, severity and prevalence of the disease. Given the dramatic impact these new tests may have on disease management, incentives may be required to offset the risks associated with expanding novel diagnostics into smaller but medically significant disease areas. / by Adrian A. Bignami. / S.M.

Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, 2014. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 66-74). / Despite the fact that mammalian hair cells and neurons do not naturally regenerate in vivo, progenitor cells exist within the postnatal inner ear that can be manipulated to generate hair cells and neurons. This work reveals the differentiation capabilities of distinct inner ear progenitor populations and pinpoints cell types that can become cochlear hair cells, vestibular hair cells, neurons, and CNS glia. We expanded and differentiated cochlear and vestibular progenitors from mice (postnatal days 1-3) and analyzed the cells for expression of mature properties by RT-PCR, immunostaining, and patch clamping. Whereas previous reports suggested that inner ear stem cells may be pluripotent and/or revert to a more neural stem cell fate, we find that cells from each organ type differentiated into cells with characteristics of the respective organ. Only cochlear-derived cells expressed the outer-hair-cell protein, prestin, while only vestibular derived cells expressed the vestibular extracellular matrix marker, otopetrin. Since Atohi expression is consistently found in new hair cells, we used an Atohl-nGFP mouse line to identify hair cell candidates. We find that cells expressing Atohl also expressed key transduction, hair bundle, and synaptic genes needed for proper function. Whole-cell patch clamp recordings showed that Atoh1-nGFP+ cells derived from both cochlear and vestibular tissue had voltage gated ion channels that were typical of postnatal hair cells. Only vestibular-derived AtohinGFP+ cells, however, had Ih, a hyperpolarization-activated current typical of native vestibular hair cells but not native cochlear hair cells. Lineage tracing studies with known supporting cell and glial cell markers showed that progenitor capacity of cochlear supporting cells positive for Lgr5 (Lgr5+ cells) was limited to differentiation into hair cell-like cells but not neuron-like cells. In contrast, glial cells positive for PLP (PLP1+ cells) from the auditory nerve differentiated into multiple cell types, with properties of neurons, astrocytes, or mature oligodendrocytes but not hair cells. Thus, PLP+ progenitor cells within the auditory nerve are limited to neuronal or glial fates but have greater potency than Lgr5+ progenitors, which only formed hair cell-like cells. In summary, this work identifies distinct populations of post-natal inner ear progenitors and delineates their capacity for differentiation and maturation. / by Will McLean. / Ph. D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2007. / Includes bibliographical references (p. 201-207). / In the past decade, biology has been revolutionized by an explosion in the availability of data. Translating this new wealth of information into meaningful biological insights and clinical breakthroughs will require a complete overhaul both in the questions being asked, and the methodologies used to answer them. One of the largest challenges in organizing and understanding the data coming from genome sequencing, microarray experiments, and other high-throughput measurements, will be the ability to find large-scale structure in biological systems. Ideally, this would lead to a simplified representation, wherein the thousands of genes in an organism can be viewed as a much smaller number of dynamic modules working in concert to accomplish cellular functions. Toward demonstrating the importance of higher-level, modular structure in biological systems, we have performed the following analyses: 1. Using computational techniques and pre-existing protein-protein interaction (PPI) data, we have developed general tools to find and validate modular structure. We have applied these approaches to the PPI networks of yeast, fly, worm, and human. / (cont.) 2. Utilizing a modular scaffold, we have generated predictions that attempt to explain existing system-wide experiments as well as predict the function of otherwise uncharacterized proteins. 3. Following the example of comparative genomics, we have aligned biological networks at the modular level to elucidate principles of how modules evolve. We show that conserved modular structure can further aid in functional annotation across the proteome. In addition to the detection and use of modular structure for computational analyses, experimental techniques must be adapted to support top-down strategies, and the targeting of entire modules with combinations of small-molecules. With this in mind, we have designed experimental strategies to find sets of small-molecules capable of perturbing fimctional modules through a variety of distinct, but related, mechanisms. As a first test, we have looked for classes of small-molecules targeting growth signaling through the phosphatidyl-inositol-3-kinase (PI3K) pathway. This provides a platform for developing new screening techniques in the setting of biology relevant to diabetes and cancer. In combination, these investigations provide an extensible computational approach to finding and utilizing modular structure in biological networks, and experimental approaches to bring them toward clinical endpoints. / by Gopal Ramachandran. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2006. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Includes bibliographical references (p. 145-147). / Nasal codas in English and Standard Chinese (SC) are compared to distinguish between the acoustic correlates of language-universal distinctive features and language-specific enhancing attributes. The distinctive feature theory and the theory of enhancement provide a framework for quantifying the acoustic and articulatory patterns observed in the two languages. An acoustic model of nasalization is first presented, in which the area of the velopharyngeal port and the place of oral constriction are varied, in order to observe the behavior of the acoustic correlates for the feature [nasal] and to establish a quantal relation between the continuous displacement of the primary articulator and the acoustic consequence of this displacement. The first two experiments identify differences in the distribution of acoustic correlates of nasalization contained in the vowel transition and the murmur regions in vowel-nasal environments in English and SC. Results for the low vowel /a/ show a mapping based on vowel rather than coda similarity. Acoustic analysis shows that the SC vowel /a/ shifts in the frequency of the second formant (F2) depending on the nasal coda, while the English vowel does not. / (cont.) The SC mid vowel /e/ shifts in F2 while the SC high vowel /i/ does not. Furthermore, analysis of syllable-initial nasals in Chinese and English shows that the SC nasals behave like the English nasals. The third experiment is a perceptual study in which subjects are asked to make judgments of the place of articulation based on limited portions of stimuli that can be either nasal or non-nasal and contain one of the three vowels. The nasal place of articulation was identified best when the nasal was preceded by the mid vowel /e/, was identified less when followed by the low vowel /a/, and was identified the worst when the nasal was preceded by the high vowel /i/. Together, the results of these experiments suggest that language-specific constraints play an important role in determining the enhancing attributes that occur alongside languageuniversal features. The interactions of the distinctive features and the enhancing gestures may lead to differences in the acoustic manifestation of the same feature in different languages. / by Xiaomin Mou. / Ph.D.

Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 164-176). / The immune system has proven valuable in the fight against cancer. Therapies that unleash a T cell response against tumors have led to durable remissions in multiple cancers. Specifically, antibodies blocking the programmed death (PD)-1 pathway have been approved for the treatment of metastatic melanoma, non-small cell lung cancer, and renal cell carcinoma, amongst others. However, only a limited number of patients respond to these therapies. The field is now trying to determine combination strategies and biomarkers to extend the benefits of these therapies to additional patients in a rationale manner. A fundamental challenge towards this goal is that the cellular and molecular mechanisms underlying the efficacy of PD-1 pathway blockade are not well understood. In this thesis, we dissected the role of PD-1 and its ligands on multiple cell types in the tumor microenvironment. PD-1 is a receptor expressed on T cells upon activation, amongst other cells. Its ligands, PD-L1 and PD-L2, can be expressed on many cell types, including tumor cells. In the first section, we show that PD-1 pathway blockade can effectively combine with another therapy targeted at tumor cells themselves, BRAF inhibitors. This work provided support for ongoing clinical trials. In the second section, we show that tumor cells can protect themselves from immune eradication by expressing PD-L1, which directly suppresses the cytotoxicity of CD8* T cells. This establishes a key mechanism by which the PD-1 pathway prevents effective antitumor immunity. In the third section, we show that the inhibition of CD8* T cell cytotoxicity through PD-1 signaling is due in part to cell-intrinsic and cell-extrinsic suppression of T cell metabolism. Removing the inhibitory PD-1 signal on a fraction of cells enhances their metabolic state and allows them to become more cytotoxic. In turn, this creates a tumor microenvironment that allows additional CD8* T cells to become more functional. We show that pharmacologic agents that mimic these effects of metabolism can enhance CD8* T cell cytotoxicity. These mechanistic insights will assist in developing cancer therapies that combine PD-1 blockade with other approaches to broaden the benefit of PD-1 immunotherapy. / by Vikram R. Juneja. / Ph. D. in Medical Engineering and Medical Physics

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, February 2011. / "February 2011." Cataloged from PDF version of thesis. / Includes bibliographical references. / Biological membranes form transient, conductive pores in response to elevated transmembrane voltage, a phenomenon termed electroporation. These pores facilitate electrical and molecular transport across cell membranes that are normally impermeable. By applying pulsed electric fields to cells, electroporation can be used to deliver nucleic acids, drugs, and other molecules into cells, making it a powerful research tool. Because of its widely demonstrated utility for in vitro applications, researchers are increasingly investigating related in vivo clinical applications of electroporation, such as gene delivery, drug delivery, and tissue ablation. In this thesis, we describe a quantitative, mechanistic model of electroporation and concomitant molecular transport that can be used for guiding and interpreting electroporation experiments and applications. The model comprises coupled mathematical descriptions of electrical transport, electrodiffusive molecular transport, and pore dynamics. Where possible, each of these components is independently validated against experimental results in the literature. We determine the response of a discretized cell system to an applied electric pulse by assembling the discretized transport relations into a large system of nonlinear differential equations that is efficiently solved and analyzed with MATLAB. We validate the model by replicating in silico two sets of experiments in the literature that measure electroporation-mediated transport of fluorescent probes. The model predictions of molecular uptake are in excellent agreement with these experimental measurements, for which the applied electric pulses collectively span nearly three orders of magnitude in pulse duration (50 ts -20 ms) and an order of magnitude in pulse magnitude (0.3 -3 kV/cm). The advantages of our theoretical approach are the ability to (1) analyze in silico the same quantities that are measured by experimental studies in vitro, (2) simulate electroporation dynamics that are difficult to assess experimentally, and (3) quickly screen a wide array of electric pulse waveforms for particular applications. We believe that our approach will contribute to a greater understanding of the mechanisms of electroporation and provide an in silico platform for guiding new experiments and applications. / by Kyle Christopher Smith. / Ph.D.

Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 225-251). / A central goal in genomics is to understand the genetic variants that underlie molecular changes and lead to disease. Recent studies have identified thousands of genetic loci associated with human phenotypes. These have primarily analyzed point mutations, ignoring more complex types of variation. Here we focus on Short Tandem Repeats (STRs) as a model for complex variation. STRs are comprised of repeating motifs of 1-6bp that span over 1% of the human genome. The level of STR variation and its effect on phenotypes remains mostly uncharted, mainly due to the difficulty in accurately genotyping STRs on a large scale. To overcome bioinformatic challenges in STR genotyping, we developed lobSTR, an algorithm for profiling STRs from high throughput sequencing data. lobSTR employs a unique mapping strategy to rapidly align repetitive reads, and uses statistical learning techniques to account for STR-specific noise patterns. We applied lobSTR to generate the largest and highest quality STR catalog to date. This provided the first characterization of more than a million loci and gave novel insights into population-wide trends of STR variation. We used our catalog to conduct a genome-wide analysis of the contribution of STRs to gene expression in humans. This revealed that STRs explain 10-15% of the cis heritability of expression mediated by common variants and potentially play a role in various clinically relevant conditions. Overall these studies highlight the contribution of STRs to the genetic architecture of quantitative traits. We anticipate that integrating repetitive elements, specifically STRs, into genome-wide analyses will lead to the discovery of new genetic variants relevant to human conditions. / by Melissa A. Gymrek. / Ph. D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2009. / "June 2009." Cataloged from PDF version of thesis. / Includes bibliographical references (p. 49-52). / It is widely believed that healthcare information technology (health IT) can improve care and lower costs. However, the pattern and uptake of beneficial features of health IT is poorly understood, and is an important part of realizing the full benefits of health IT. This thesis examines the factors relating to adoption and use of reporting features within an outpatient practice management system. A retrospective observational study was performed utilizing web log data from a practice management and electronic health record system vendor. Two years of data were analyzed on the use of features within the system in two different scenarios: the use of a newly released custom reporting feature among existing clients, and the use of a physician-level monthly report among new clients. Among these two different populations and features, the first use and subsequent utilization exhibited similar patterns. Using the Bass model of technology diffusion to quantify the adoption of these features, it was found that adoption had a low social component (coefficient of imitation) and a high personal component (coefficient of innovation). One physician's use of a feature in his practice did not appear to influence whether a new physician joining the same practice would use the feature. In addition, the earliest users of a feature tended to utilize that feature more often. Practices and providers that used these features performed better across three of four operational and financial metrics. The purchase and installation of a health IT system by an organization does not ensure that individuals within it will fully utilize the system and realize all the benefits. / (cont.) Incentives for health IT should focus on the advantages gained from these systems, and not merely on their purchase. Health IT vendors should be cognizant of the way they introduce new functionality to their clients in order to ensure maximal use. / by Adam Weinstein. / S.M.