22689 research outputs found
Sort by
Advancing Total Body Photography for Early Detection and Spatio-temporal Monitoring of Skin Cancer
No full textTotal body photograph (TBP), a method that photographs the entire cutaneous surface of a person through a series of images of various body sectors, is becoming commercially available to aid in the early detection of skin cancer in high-risk individuals. Despite the potential of TBP, several challenges are hindering its widespread applications at the bedside, including limited system resolution, limited image sharpness, lack of robust longitudinal tracking methods, and absence of a standardized data format.
The first part of the dissertation introduces a 3D-vision-based method to improve TBP image fidelity in terms of system resolution and image sharpness. We formulate the TBP scanning problem as a multi-view capture when restricted to determining the camera’s depth of field. Leveraging multi-view geometry in 3D vision, we propose a shape-aware focusing method to optimize the in-focus surface coverage of the subject in a TBP scan. Starting from the Expectation-Minimization method, we develop a -view algorithm to mitigate the local minima issue and show the effectiveness of the s-view algorithm in extensive simulations. The proposed shape-aware focus is validated with a TBP system prototype, demonstrating superior image fidelity over existing image capture protocols.
The second part presents a 3D-geometry-based framework to tackle the longitudinal tracking of skin lesions in TBP and a data format to standardize the storage of TBP data. The proposed framework uses a flow-field refinement of template-based correspondence maps to achieve state-of-the-art matching and mapping accuracy in lesion tracking. Finally, we propose a unified and DICOM-compliant data format to accommodate various TBP data (2D images and/or 3D meshes) for the temporal comparison.
We hope that the proposed methods can benefit automated frameworks and systems for more accurate and efficient skin lesion monitoring, and can eventually be used to create more reliable tools for the early detection of skin cancer
Real-Time Byzantine-Resilient Power Grid Infrastructure
Get PDFThe power grid, critical to society and the economy, is increasingly targeted by sophisticated cyber attacks, especially from nation-state actors. These threats, at both system and network levels, aim to compromise key grid components, risking severe disruptions and blackouts. While much of the existing research focuses on isolated security concerns, it neglects complex threats to the broad grid infrastructure, especially in substations. This gap undermines grid resilience and endangers both lives and billions of dollars.
This thesis takes a step towards resilient grid infrastructure by introducing a novel comprehensive threat model and pioneering real-time, Byzantine-resilient solutions for grid infrastructure. We present the first real-time Byzantine-resilient architecture and protocols for the substation, ensuring correct protective operations even in the face of protective relay compromises and network attacks. We evaluate our implementation across a comprehensive range of fault-free and faulty operating conditions in relevant testbeds, demonstrating its ability to meet strict real-time latency requirements even in the worst operating conditions.
We introduce the first end-to-end Byzantine-resilient system framework for the broad grid infrastructure from the control center to the substation and field devices under the comprehensive threat model. We demonstrate the proposed system frame- work’s ability to support real-time grid operations in a Byzantine-resilient manner. To enhance situational awareness, we integrate unsupervised machine learning models for anomaly detection.
The solutions we propose satisfy other critical domain needs, including continu- ous availability over a long system lifetime and seamless integration with the grid. We implemented all modules and protocols and made them available to the commu- nity within the open-source Spire Toolkit. The system has successfully withstood a purple team exercise and has been transitioned to SCADA manufacturers GE and Siemens, as well as two national laboratories PNNL and SANDIA. Finally, we pro- pose a practical incremental deployment strategy for large-scale real-world power grid topologies
ADVANCES IN COMPUTATIONAL MODELING AND SURGICAL PLANNING OF CARDIOVASCULAR REPAIRS
No full textCongenital heart disease affects approximately 1% of live births globally, often necessitating complex vascular reconstruction. Traditional repair methods and standard biomaterials frequently fall short because each patient’s unique anatomy requires a highly customized approach. Consequently, many patients experience suboptimal postoperative outcomes, often resulting in the need for additional surgeries.
Recent advancements in surgical planning and patient-specific design for vascular repairs offer the potential for improved postoperative outcomes and a reduced need for reintervention. They enable the simulation of personalized biomedical devices used in reconstructive surgeries, allowing doctors to visualize and adjust repairs pre-surgery while aiming to provide insights into their postoperative performance. However, current progress is constrained by the lack of comprehensive simulation of the entire repair process, particularly in incorporating virtual tools and materials, accurately modeling complex vessel deformations during surgery, and thoroughly comparing predicted outcomes with actual surgical results.
This thesis presents a comprehensive framework aimed at improving patient-specific surgical planning for reconstructive vascular surgeries in patients with congenital heart disease. The framework focuses on 1) the development of high-fidelity computational fluid dynamics models to non-invasively predict parameters that indicate the need for intervention, 2) assessment of blood flow in repaired arteries to evaluate the influence of postoperative geometry on hemodynamics, providing insights to guide the final arterial shape, 3) integration of virtual surgical tools in patient-specific repair planning to isolate the design region, supporting the parameterization and optimization of vascular graft shapes, 4) manufacture of optimized tubular and branched vascular grafts using tissue-engineering methods, 5) evaluation of tissue-engineered graft performance through implantation in porcine models, 6) development of a finite element model to simulate vessel deformation during graft implantation and analyze placement accuracy informed by in vivo postoperative data from animal studies, 7) refined biomechanical modeling of native vessels using imaging data and ex-vivo measurements. This study represents a crucial first step toward advancing medical care for patients with congenital heart disease by leveraging computational modeling and tissue-engineered solutions to enhance patient outcomes and quality of life
NEXT-GENERATION SEQUENCING FOR PATHOGEN EVOLUTION AND EPIDEMIOLOGY
Get PDFSince its inception in the early 2000s, next-generation sequencing (NGS) has played a pivotal role in our understanding of molecular genomics. Researchers did not take long to embrace NGS platforms due to the combination of parallel-sequencing and data reliability. Today, the entire human genome can be sequenced in one day - a task that researchers took 13 years to complete using traditional sequencing techniques. In this thesis, we utilize the highly versatile NGS platforms GridION and PrometION by Oxford Nanopore Technologies to characterize the evolution of seasonal influenza and challenge the prevalence of macrolide-resistant Mycoplasma pneumoniae in the Johns Hopkins Health System.
The 2024-2025 influenza season exhibited high incidence and hospital admissions not observed since 2009-2010. Using left-over positive clinical influenza samples, we identified the co-circulation of H3N2 and H1N1pdm09 and a shift in dominant influenza subclades. Additionally, we observed an increase in S247N in the neuraminidase gene of H1N1pdm09, an amino acid substitution associated with oseltamivir resistance. Similarly, Mycoplasma pneumoniae exhibited higher positivity rates in 2024, which was not seen since 2020. We sequenced left-over positive samples to establish the prevalence of macrolide resistance-associated single nucleotide polymorphisms A2063G and A2064G. Previous studies that used Sanger sequencing or PCR estimated the prevalence to be ~8%. Through a binary alignment map analysis, we established the prevalence to be 26.4%. Together, our data highlight the utility of NGS in clinical medicine and public health
Evaluating Oral Selective Estrogen Receptor Degraders as Therapy for Endocrine Resistant ER+ Breast Cancer
Get PDFOral selective estrogen receptor degraders (SERDs) are the newest generation of endocrine therapy for estrogen receptor positive (ER+) breast cancer with more potent and efficacy against common mechanisms of endocrine resistance compared to current standard-of-care therapies. In this study, we compare a panel of oral SERDs to fulvestrant, both as monotherapies and with combinatorial agents, using viability assays to assess their efficacy on known mechanisms of resistance in invasive ductal and lobular carcinoma models. We determined that oral SERDs vary in ER degradation and growth inhibition in MCF7, T47D, MDA-MB-134VI, and SUM44PE cell lines. Camizestrant and giredestrant were consistently the most potent in treating mechanisms of resistance, achieving greatest efficacy when combined with kinase inhibitors, but most alterations conferring resistance to fulvestrant were also resistant to oral SERDs. Findings from transcriptional and epigenetic sequencing can uncover cellular pathways to potentially target and prevent resistance or resensitize tumors to endocrine therapy
Computational Systems Pharmacology of Antibody-Drug Conjugates
No full textDesigned as targeted cancer therapeutics, antibody-drug conjugates (ADCs) comprise a monoclonal antibody base attached to one or more cytotoxic agents via chemical linkers. This strategy enables specific targeting of antigens expressed on the cancer cell surface, resulting in receptor-mediated endocytosis and delivery of toxic payload to the cells of interest while sparing healthy tissues. Despite these promising mechanisms of action, ADCs may still exhibit low efficacy and considerable toxicity, as preclinical efficacy and safety do not always translate to clinical settings. Thus, developing a quantitative understanding of ADC mechanisms and pharmacokinetics/ pharmacodynamics (PK/PD) is important to design safe and effective therapies. Quantitative systems pharmacology (QSP) modeling combines mechanistic knowledge with preclinical and clinical data to generate computational simulations, enabling predictions of efficacy and toxicity.
In this thesis, I detail the development of multiscale, computational systems pharmacology models of antibody-drug conjugates, specifically those carrying pyrrolobenzodiazepine (PBD) payloads. First, I built a mechanistic computational model platform reflecting the cellular mechanisms of PBD ADCs, using ordinary differential equations to track changes in concentration over time in the system. I parameterized the model using in vitro experimental data for PBD ADCs targeting B cell maturation antigen (BCMA), intended to treat multiple myeloma, and PBD ADCs targeting human epidermal growth factor receptor 2 (HER2), intended to treat HER2-positive solid tumors. Using the model, I conducted simulations of cancer cell culture experiments, seeking the factors most important to cell killing and exploring how changes in ADC design and systemic parameters impacted the predicted efficacy and potential toxicity.
Next, I extended the mechanistic model into a compartmental model to represent tumor xenografts in mice by adding mouse PK/PD. By designing and incorporating a novel tracking module into the models, I was able to identify the recent location history of the cytotoxic payload, which facilitated estimates of both the on-target and off-target (bystander) potential for cell killing. Using this multiscale, mechanistic, computational model platform, I can generate insights for optimization of ADC design and determine which factors are most critical to efficacy and toxicity, leading to more informed and rational development of cancer therapies to ultimately improve patient lives
A microwave stimulated inelastic neutron scattering technique and its first use exploring spin relaxation in Cr8
Get PDFWhen we consider the 2nd law of thermodynamics, that all systems flow towards equilibrium with their surroundings, we do not concern ourselves with how they get there or what paths through phase space they traverse. Answering these questions is a central part of the rapidly growing field of nonequilibrium condensed matter. Additionally, interest in interacting many body systems out of equilibrium has sky-rocketed in recent years due to the ability for quasiparticle interactions to drive strong non-linearities that could result in new transient phases and symmetries. As for technological applications, hard matter materials are typically operated out of equilibrium, while our exploration and understanding of them is overwhelmingly of their equilibrium and near equilibrium properties. Thus, developing tools and techniques to drive systems out of equilibrium and measure their properties as a function of time is highly worthwhile. We describe a new time-resolved microwave-stimulated inelastic neutron scattering technique and its first use driving a molecular magnet out of equilibrium and monitoring its slow return to a singlet ground state that requires high order effects to reach
Beyond the Countryside: Unraveling the Tapestry of Academic Advising Support and Sense of Belonging for Rural First-Generation Scholars
Get PDFU.S. postsecondary degree attainment remains a critical issue, with research showing that rural, first-generation students are less likely to persist once enrolled. This needs assessment revealed the importance of faculty support to persisting students at a rural-serving institution. This qualitative study explored the academic advising experiences of rural, first-generation students and faculty advisors at rural-serving institutions. Data were from semi-structured interviews of rural, first-generation students (n = 2) and faculty advisors serving rural students (n = 13). Data were analyzed using open coding and thematic analysis. Findings suggest that rural, first-generation students encounter diverse experiences with academic advising, which can profoundly influence their college paths. The study showed contrasting views between students and advisors regarding advisors’ impact on student retention. Faculty members expressed doubt about their influence, but students viewed their advisors as vital to success. Faculty members acknowledged their gaps in advising expertise and the need for professional development. Findings highlighted the critical roles of a nurturing campus atmosphere and diligent guidance in fostering a sense of belonging. The findings uncovered the critical need for institutions to recognize and prioritize academic advising through institutional-wide professional development on advising approaches and strategies and the unique needs of rural, first-generation students. Advising and targeted programming will support the persistence and success of rural, first-generation students at institutions
CELL DENSITY AND PHAGOSOMAL EFFECTS ON THE METABOLISM OF CRYPTOCOCCUS NEOFORMANS
No full textHere, we explored the metabolic response of C. neoformans to infection. Long-term survival within the macrophage phagosome is important for C. neoformans pathogenesis. For the first time, we were able to isolate the host phagosome containing cryptococcal cells. This achievement allowed us to identify proteomics candidates directly relevant to the phagosomal environment. Specifically, we are interested in the host-pathogen interaction differences between macrophage polarization states that lead to cell transfer. This dataset will begin to elucidate the differences and conserved fungal proteomic milieu within the phagosome. Additionally, we argue that great emphasis has been placed on identifying novel proteins involved in fungal-host interactions, but conserved signaling pathways between eukaryotic organisms may influence host pathogen interactions. Understanding the extent to which cryptococcal proteins can influence homologous host responses may provide important context for these interactions.
Characterization of a proteomic candidate revealed issues with our 2015 gene deletion library collection. Using PCR and RNAseq validation, we discovered that our 2015 plates were directly swapped with 2008 library plates. Validation of our library uncovered a quick, straightforward method for testing the fidelity of library strains using NAT insertion cassette specific primers.
Our work on the peroxisomal β-oxidation pathway, a pathway upregulated during nickel exposure and in many infection datasets [1-5], uncovered interesting cell-density-dependent growth and virulence factor phenotypes. We implicate mitochondrial retrograde signaling (RTG), which has not been characterized in C. neoformans, in this response.
Finally, two appendix chapters containing previously published papers are included. The first characterizes unique methods for isolating cryptococcal capsular polysaccharides, an aspect of virulence that has evaded detailed characterization due to methodological challenges. The second characterizes the cuticle of insecticide resistant Aedes aegypti mosquitoes using techniques from C. neoformans melanization. Given the breadth of topics covered, these chapters are appendices to improve clarity
WHAT'S IN A MEASURE? THE ROLE OF MORTALITY ESTIMATION IN POPULATIONS IN DISTRESS: EVIDENCE FROM NEGLECTED CRISES IN CHAD AND THE CENTRAL AFRICAN REPUBLIC
No full textMortality rates are crucial indicators of a population’s health, especially in humanitarian crises, where accurate estimates can drive life-saving interventions. Despite the long-standing use of mortality thresholds to trigger responses, questions remain regarding the adequacy and relevance of these methods. This dissertation critically examines the application of mortality estimation methods in crisis settings, where the complexity of conflicts, displacements, and natural disasters complicates data collection and analysis. Through a series of studies, this dissertation assesses the reliability and validity of various methods—retrospective household surveys, community-based surveillance (CBS), key informant surveillance (KIS), and burial site surveillance (BSS)—in capturing mortality rates. The research reveals that retrospective surveys are not without limitations, including biases and delays, they remain an important tool in their ease of use and capacity to provide valid estimates. Conversely, CBS and KIS, though useful in capturing real-time data, face challenges related to bias, resource demands, and sustainability. The dissertation also explores an innovative approach by use of Multiple Systems Estimation (MSE) to evaluate the validity of outcomes estimated by each method. By comparing methods across two settings in neglected crises, this research underscores the need for actors to commit to regular mortality estimation in the areas they serve, rigorously evaluate the results, and invest in research to refine these methods. Accurate mortality estimation is crucial for ensuring that humanitarian aid is targeted where it is needed most, and that interventions are both timely and effective. The findings highlight the vital role of precise mortality data in guiding public health actions, advocating for affected populations, and ultimately reducing excess mortality in crisis settings