Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2007. / Includes bibliographical references (leaves 138-142). / Phonetic variation is pervasive in everyday speech. Studying these variations is essential for building acoustic models and lexical representations that effectively capture the variability of speech. This thesis examines one of the commonly-occurring phonetic variations in English: the stop-like modification of the dental fricative /ð/. This variant exhibits a drastic change from the canonical /ð/; the manner of production is changed from one that is fricative to one that is stop-like. Furthermore, the place of articulation of stop-like /0/ has been a point of uncertainty, leading to the confusion between stop-like /ð/1 and /d/. This thesis aims to uncover the segmental context of stop-like /ð/, possible causes of the modification, whether the dental place of articulation is preserved despite modification, and if there are salient acoustic cues that distinguish between stop-like /ð/ and /d/. Word-initial /ð/ in the read speech of the TIMIT Database, the task-oriented spontaneous speech of the AEMT Corpus, and the non-task-oriented spontaneous speech of the Buckeye Corpus are examined acoustically. It is found that stop-like /ð/ occurs most often when it is preceded by silence or when preceded by a stop consonant. The occurrence is less frequent when /ð/ is preceded by a fricative or an affricate consonant. This modification rarely occurs when /ð/ is preceded by a vowel or liquid consonant. The findings suggest that possible factors that may contribute to the stop-like modification of /ð/include physiological mechanisms of speech production, prosody, and/or other aspects of speaking style and manner. Acoustic analysis indicates that stop-like /ð/ is significantly different from /d/ in burst amplitude, burst spectrum shape, burst peak frequency, and second formant at following- vowel onset. / (cont.) Moreover, the acoustic differences indicate that the dental place of articulation is preserved for stop-like /ð/. Automatic classification experiments involving these acoustic measures suggest that they are robust in distinguishing stop-like /ð/ from /d/. Applications of these findings may lie in areas of automatic speech recognition, speech transcription, and development of acoustic measures for speech disorder diagnosis. / by Sherry Y. Zhao. / Ph.D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2006. / Includes bibliographical references. / In 1992, Accelerated Approval, or Subpart H approval, was added to the NDA regulations so promising products that provide a meaningful therapeutic benefit for serious or life-threatening diseases could be introduced to the market sooner, particularly for diseases or conditions where there was a great unmet medical need. Accelerated Approval is based on either a surrogate endpoint that is reasonably likely to predict clinical benefit or a clinical endpoint other than survival or irreversible morbidity. After approval, the sponsor is required to perform post-marketing studies to demonstrate clinical benefit. Since the FDA expanded the use of the Accelerated Approval regulatory path to include oncology drugs in 1995, thirty drugs (both small molecule as well as biologics) have been granted accelerated approval in oncology. However, from various reports in the literature and the FDA site, it appears that only a small fraction of these approvals (four to six) have been converted into regular approvals, based on the demonstration of clinical benefit in the post-marketing studies that support the benefit seen in the pivotal studies. / (cont.) In my research, I examined the basis of approval for six drugs that were converted to full approval and compared this group to the seven drugs that received accelerated approval before 2000 but as yet have not converted to full approval. The six drugs that were converted to full approval, with the exception of dexrazoxane, completed their post-marketing requirements in 2.3 years after initial approval. The sponsors, who were all well-capitalized pharmaceutical companies, also pursued additional indications for these drugs. In the group of drugs that were designated as "not converted" by several sources, two of the drugs have been granted full approval within the past year. And in March 2006, the FDA withdrew its accelerated approval for one drug based on the results of a negative clinical trial. Six years after having received accelerated approval, two drugs in this group are still undergoing clinical trials. Due to the lack of information about the ongoing trials, it is difficult to assess the underlying reasons for the delay in attaining full approval. But the sponsors of these two drugs are small biotechnology companies, while all of the sponsors of the drugs that have been converted to full approval are major pharmaceutical companies. / (cont.) A majority of the drugs that converted to full approval were granted a broader label based on the post-marketing studies, which demonstrated clinical benefit in a wider patient population than originally tested. While the accelerated approval process holds many advantages in that companies can introduce their drug to the market sooner, the requirements for accelerated approval often result in the drugs having to meet 'a higher standard' in that they have to demonstrate "meaningful clinical benefit" over existing agents, which may in fact be a requirement for superiority, as was seen in the case of one agent, Doxil. The post-marketing studies can be expensive and difficult to complete, but companies with ample resources and sufficient incentives, such as additional potential indications, seem able to clear this hurdle easily. / by Jean Jinsun Kim. / S.M.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2005. / Includes bibliographical references (leaves 19-20). / Merging of clinical systems and medical databases, or aggregation of information from disparate databases, frequently requires a process where vocabularies are compared and similar concepts are mapped. Using a normalization phase followed by a novel alignment stage inspired by DNA sequence alignment methods, automated lexical mapping can map terms from various databases to standard vocabularies such as UMLS (Unified Medical Language System) and SNOMED (the Systematized Nomenclature of Medicine). This automated lexical mapping was evaluated using a real-world database of consultation letters from Children's Hospital Boston. The first phase involved extracting the reason for referral from the consultation letters. The reasons for referral were then mapped to SNOMED. The alignment algorithm was able to map 72% of equivalent concepts through lexical mapping alone. Lexical mapping can facilitate the integration of data from diverse sources and decrease the time and cost required for manual mapping and integration of clinical systems and medical databases. / by Jennifer Y. Sun. / S.M.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2011. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 110-117). / Basic scientific research over the past few decades has shown some light on the complex pathophysiology of SCI and has enhanced our understanding of some of the important factors that contribute to the lack of regeneration following the initial traumatic injury and secondary injury response in the adult spinal cord. These factors include both intrinsic limitations in the regeneration capacity of mature neurons, but also a host of environmental factors which inhibit spontaneous attempts of axon regeneration. These environmental factors include physical barriers to axon regeneration such as fibrous and glial scars as well as molecules which actively inhibit regeneration. Even when these barriers are removed, axons in the central nervous system still require the appropriate stimulatory signals in order for significant regeneration to occur. These signals may include a substrate to provide directional guidance of the extending axon growth cones as well as neurotrophic factors to promote axon growth and survival. Thus, to achieve a clinically meaningful regenerative response following spinal cord injury a combinatorial therapeutic approach is likely necessary. This thesis investigated the use of a porous collagen scaffold with aligned pores to serve as a substrate for axon guidance and a delivery vehicle for select therapeutic agents in both fulland hemi- resection injury models in adult rats. In the hemi-resection injury model, the treatment groups included: a dehydrothermally (DHT) crosslinked scaffold, a carbodiimide (EDAC) crosslinked scaffold, and an EDAC-treated scaffold delivering either soluble Nogo receptor (sNgR), chondroitinase ABC (ChABC), or bone marrow derived mesenchymal stem cells (MSCs). Improvement in hindlimb function over four weeks following injury was seen in the DHT scaffold, EDAC scaffold + ChABC, and EDAC scaffold + MSCs groups, but not in the untreated control group. Immunohistochemical evaluation of the tissue revealed a few regenerating axons reaching the center of the scaffold in the DHT scaffold and EDAC scaffold + ChABC groups at 4 week post injury. Histological evaluation at 4 weeks showed the defect area of all groups to be largely comprised of loosely organized fibrous tissue. Many macrophages were seen surrounding the defect area of all groups; however, there were significantly more macrophages within the defect area of the control group compared to the treatment groups. In the full-resection injury model, at 6 weeks post injury, implanted collagen scaffolds tended to reduce the number of cystic cavities in the defect and to better align the reparative tissue with the long axis of the spinal cord. These results show the potential benefit of collagen scaffolds alone or in combination with select therapeutic agents as a means to modulate the healing response following spinal cord injury. / by Rahmatullah H. Cholas. / Ph.D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2005. / Includes bibliographical references (leaves 34-36). / Objectives: There are several ongoing efforts aimed at developing formal models of medical knowledge and reasoning to design decision-support systems. Until now, these efforts have focused primarily on representing content of clinical guidelines and their logical structure. The present study aims to develop a computable representation of health-care providers' intentions to be used as part of a framework for implementing clinical decision-support systems. Our goal is to create an ontology that supports retrieval of plans based on the intentions or goals of the clinician. Methods: We developed an ontological representation of medical goals, plans, clinical scenarios and other relevant entities in medical decision-making. We used the resulting ontology along with an external ontology inference engine to simulate selection of clinical recommendations based on goals. The ontology instances used in the simulation were modeled from two clinical guidelines. Testing the design: Thirty-two clinical recommendations were encoded in the experimental model. Nine test cases were created to verify the ability of the model to retrieve the plans. For all nine cases, plans were successfully retrieved. Conclusion: The ontological design we developed supported effective reasoning over a medical knowledge base. / (cont.) The immediate extension of this approach to be fully developed in medical applications may be partially limited by the lack of available editing tools. Many efforts in this area are currently aiming to the development of needed technologies. / by Davide Zacacagnini [i.e. Zaccagnini]. / S.M.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2007. / Includes bibliographical references (leaves 36-39). / The goal of this study is to explore the economic conditions that favor the joint development of therapeutics and companion diagnostics. I hypothesize that predictive biomarkers can generate economic value in drug development by increasing success rates. I construct an economic model of the development of a hypothetical new therapy, and devote particular attention to parameters regarding safety, efficacy, cost, and market size, within a decision-theoretic framework. The results include a characterization of the dynamic net present value trade-offs between stratum size and biomarker success, as well as the identification of two complementary concepts of stratified medicine, namely, disease reclassification and value-based reimbursement. I also identify a strong potential incentive mechanism in the hands of public policy makers that could facilitate a resolution of the tension between patient interests and the interests of pharmaceutical sponsors. The conclusion is that a biomarker can compensate for smaller stratum by increasing success probabilities. However, the effects of longer development time due to biomarker inclusion counter the effects of improved success probabilities. Longer exclusivity periods for stratified medicine may be required in order to resolve the tension between patient interests and the interests of pharmaceutical sponsors. / by Amir Goren. / S.M.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references. / There is a constant demand for enormous numbers of high quality hepatocytes in the fields of cell transplantation, pharmacotoxicology, tissue engineering, and bioartificial assist devices. The scarcity of viable hepatocytes necessitates the use of suboptimal sources including damaged donor organs that are not transplantable. Many of these organs have potentially reversible pathologies however, that could be treated via ex vivo perfusion thereby increasing their cell yield. With the intent to translate organ recovery by perfusion into the clinic, we engineered a very simple room temperature-operated ex vivo organ perfusion system to test a rat liver model of uncontrolled non-heart beating donors. Seventeen times as many hepatocytes were recovered from livers exposed to an hour of warm ischemia (WI, 34*C) compared to untreated WI livers in only 3 hours of perfusion. Further, fresh liver hepatocyte yields were also increased by 32% postperfusion, demonstrating that both damaged and healthy donor livers could benefit from this methodology. A linear correlation between cell yield and tissue ATP content was established. This enables an accurate prediction of cell recovery during preservation and can be used as a direct measure of organ viability and the trajectory of organ recovery during perfusion resuscitation. Further, a strong correlation between perfusion flow rate and cell yield was also established supporting the use of flow rates as low as possible without causing hypoperfusion or oxygen deprivation. Morphologically and functionally, perfusion-isolated hepatocytes generally performed comparably or better than fresh hepatocytes in cell suspension and plate culture. Cumulatively, these findings strongly support the ubiquitous use of organ perfusion systems in the clinic for optimal enhancement of donor grafts. / by Maria-Louisa Izamis. / Ph.D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 56-58). / In this thesis I have developed a map of Healthcare Information Technology applications used in the United States for care delivery, healthcare enterprise management, clinical support, research and patient engagement. No attempt has previously been made to develop such a taxonomy for use by healthcare policy makers and on-the-spot decision makers. Using my own fifteen years of experience in HIT, along with an extensive set of literature reviews, interviews and on-site research I assembled lists of applications and organized them into categories based on primary workflows. Seven categories of HIT systems emerged, which are Practice Tools, Advisory Tools, Financial Tools, Remote Healthcare Tools, Clinical Research Tools, Health 2.0 Tools and Enterprise Clinical Analytics, each of which have different operational characteristics and user communities. The results of this pilot study demonstrate that a map is possible. The draft map presented here will allow researchers and investors to focus on developing the next generation of HIT tools, including software platforms that orchestrate a variety of healthcare transactions, and will support policy makers as they consider the impact of Federal funding for HIT deployment and adoption. Further studies will refine the map, adding an additional level of detail below the seven categories established here, thus supporting tactical decision making at the hospital and medical practice level. / by William Charles Richards Crawford. / S.M.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2006. / Includes bibliographical references (p. 183-198). / In this thesis we address the whole-brain tractography segmentation problem. Diffusion magnetic resonance imaging can be used to create a representation of white matter tracts in the brain via a process called tractography. Whole brain tractography outputs thousands of trajectories that each approximate a white matter fiber pathway. Our method performs automatic organization, or segmention, of these trajectories into anatomical regions and gives automatic region correspondence across subjects. Our method enables both the automatic group comparison of white matter anatomy and of its regional diffusion properties, and the creation of consistent white matter visualizations across subjects. We learn a model of common white matter structures by analyzing many registered tractography datasets simultaneously. Each trajectory is represented as a point in a high-dimensional spectral embedding space, and common structures are found by clustering in this space. By annotating the clusters with anatomical labels, we create a model that we call a high-dimensional white matter atlas. / (cont.) Our atlas creation method discovers structures corresponding to expected white matter anatomy, such as the corpus callosum, uncinate fasciculus, cingulum bundles, arcuate fasciculus, etc. We show how to extend the spectral clustering solution, stored in the atlas, using the Nystrom method to perform automatic segmentation of tractography from novel subjects. This automatic tractography segmentation gives an automatic region correspondence across subjects when all subjects are labeled using the atlas. We show the resulting automatic region correspondences, demonstrate that our clustering method is reproducible, and show that the automatically segmented regions can be used for robust measurement of fractional anisotropy. / by Lauren Jean O'Donnell. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2008. / Includes bibliographical references. / Otitis media (OM) is the most common specifically-treated childhood disease in the United States. The widespread use of systemic antibiotics against a disease of such high incidence is believed to be a driving force behind the observed increase in adaptive resistance among pathogenic bacteria in the nasopharynx. Local, sustained delivery of antimicrobial agents to the site of infection allows for higher drug concentrations and optimized release profiles than are permitted by systemic administration. Higher antimicrobial concentrations sustained for longer periods of time also allow for a faster and more complete eradication of OM bacteria (e.g., H. influenzae, S. pneumoniae), and minimize antibiotic exposure to other bacteria and natural flora in the nasopharynx and upper respiratory tract. We have developed in situ-forming hydrogels to serve as sustained release reservoirs for noninvasive trans-tympanic treatment of OM. A hydrogel that includes potentially synergistic chemical penetration enhancer (CPE) combinations and an antimicrobial sufficiently increases antimicrobial flux such that therapeutic levels can traverse the tympanic membrane (TM) within 12 hours, in vitro. We compare excised chinchilla TMs treated with ciprofloxacin (fluoroquinolone antibiotic) alone and with different combinations of sodium lauryl sulfate, limonene, and bupivacaine, with respect to resultant changes in TM electrical resistance and trans-TM ciprofloxacin flux. We also investigate the interactions of CPEs and local anesthetics with respect to both permeability enhancement and changes in nerve block potency and efficacy. Finally, we evaluate our hydrogel formulations in an in vivo chinchilla model of OM, and demonstrate early success in their ability to safely and effectively eradicate middle ear bacteria. / by Emmanuel J. Simons. / Ph.D.