University of North Carolina Hospitals

Carolina Digital Repository
Not a member yet
    95038 research outputs found

    Breast tumor diagnosis via multimodal deep learning using ultrasound B-mode and Nakagami images

    Get PDF
    Purpose: We propose and evaluate multimodal deep learning (DL) approaches that combine ultrasound (US) B-mode and Nakagami parametric images for breast tumor classification. It is hypothesized that integrating tissue brightness information from B-mode images with scattering properties from Nakagami images will enhance diagnostic performance compared with single-input approaches. Approach: An EfficientNetV2B0 network was used to develop multimodal DL frameworks that took as input (i) numerical two-dimensional (2D) maps or (ii) rendered red-green-blue (RGB) representations of both B-mode and Nakagami data. The diagnostic performance of these frameworks was compared with single-input counterparts using 831 US acquisitions from 264 patients. In addition, gradient-weighted class activation mapping was applied to evaluate diagnostically relevant information utilized by the different networks. Results: The multimodal architectures demonstrated significantly higher area under the receiver operating characteristic curve (AUC) values (p < 0.05) than their monomodal counterparts, achieving an average improvement of 10.75%. In addition, the multimodal networks incorporated, on average, 15.70% more diagnostically relevant tissue information. Among the multimodal models, those using RGB representations as input outperformed those that utilized 2D numerical data maps (p < 0.05). The top-performing multimodal architecture achieved a mean AUC of 0.896 [95% confidence interval (CI): 0.813 to 0.959] when performance was assessed at the image level and 0.848 (95% CI: 0.755 to 0.903) when assessed at the lesion level. Conclusions: Incorporating B-mode and Nakagami information together in a multimodal DL framework improved classification outcomes and increased the amount of diagnostically relevant information accessed by networks, highlighting the potential for automating and standardizing US breast cancer diagnostics to enhance clinical outcomes

    Predictive machine learning algorithm for COPD exacerbations using a digital inhaler with integrated sensors

    Get PDF
    PURPOSE: By using data obtained with digital inhalers, machine learning models have the potential to detect early signs of deterioration and predict impending exacerbations of chronic obstructive pulmonary disease (COPD) for individual patients. This analysis aimed to determine if a machine learning algorithm capable of predicting impending exacerbations could be developed using data from an integrated digital inhaler. PATIENTS AND METHODS: A 12-week, open-label clinical study enrolled patients (≥40 years old) with COPD to use ProAir Digihaler, a digital dry powder inhaler with integrated sensors, to deliver their reliever medication (albuterol, 90 µg/dose; 1-2 inhalations every 4 hours, as needed). The Digihaler recorded inhaler use through timestamps, peak inspiratory flow (PIF), inhalation volume, inhalation duration, and time to PIF throughout the study. By applying machine learning methodology to data downloaded from the inhalers after study completion, along with clinical and demographic information, a model predictive of impending exacerbations was generated. RESULTS: The predictive analysis included 336 patients, 98 of whom experienced a total of 111 exacerbations. PIF and inhalation volume were observed to decline in the days preceding an exacerbation. Using gradient-boosting trees with data from the Digihaler and baseline patient characteristics, the machine learning model was able to predict an exacerbation over the following 5 days with a receiver operating characteristic area under curve of 0.77 (95% CI: 0.71-0.83). Features of the model with the highest weight were baseline inhalation parameters and changes in inhalation parameters before an exacerbation compared with baseline. CONCLUSION: We demonstrated the development of a proof-of-concept machine learning model predictive of impending COPD exacerbations using data from the integrated digital reliever inhaler. This approach may potentially support patient monitoring, help improve disease management, and enable pre-emptive interventions to minimise exacerbations. CLINICAL TRIAL REGISTRATION NUMBER: NCT03256695

    Unlocking the Conservation Potential of Managed Timberlands Through Understanding Avian Community Dynamics

    Get PDF
    This study investigates avian community composition across different forest stand stages within managed pine landscapes of the southeastern United States. It addresses two key research questions: (1) How do avian species richness and community composition vary across different successional stages within managed pine landscapes in the southeastern U.S.? and (2) How do stand-scale characteristics and landscape-scale features influence avian diversity in managed timberlands? A total of 69 bird species were documented, with natural stands supporting the highest species richness (55 species), followed by emergent and mature stands. The most abundant habitat-use groups were edge specialists and generalists, while insectivores dominated dietary guilds, accounting for 71.6% of detections. Among migration groups, non-migratory residents comprised 55% of detections, with long-distance migrants at 28% and short-distance migrants at 17%. Statistical analyses, including Jaccard Similarity Indices and PERMANOVA, revealed significant differences in bird community composition among stand stages, particularly between pre-thinned, natural, and emergent stands. Non-metric Multidimensional Scaling (NMDS) indicated that habitat structure and land cover characteristics—such as basal area, tree density, and herbaceous cover—strongly influenced avian community composition. These findings highlight the importance of both stand-scale and landscape-scale environmental gradients in shaping bird communities. An Indicator Species Analysis (ISA) identified 16 indicator species across stand stages, with mature and emergent stands supporting the highest numbers. Notably, the Blue Grosbeak, Common Yellowthroat, and Yellow-breasted Chat emerged as strong indicators of emergent habitats. Species Distribution Models (SDMs) for these species highlighted grassland-herbaceous cover, basal area, and proximity to water as key predictors of habitat presence. Jackknife analyses further confirmed the importance of vegetation structure and landscape features, with logistic regression models demonstrating classification accuracies ranging from 68.6% to 95.2%. Overall, this study underscores the significant influence of stand stage on avian community structure and habitat preferences. The results emphasize the ecological importance of natural and emergent stands in maintaining bird diversity, particularly for species associated with early successional and structurally complex habitats. These findings provide critical insights for forest management and conservation planning, supporting strategies aimed at promoting avian biodiversity across working timberlands.Master of City and Regional Plannin

    Understanding and Addressing Cancer Disparities Among American Indians in North Carolina: The Southeastern American Indian Cancer Health Equity Partnership (SAICEP)

    Get PDF
    Introduction American Indians and Alaska Natives (AIANs) experience significant cancer incidence and mortality disparities, with elevated cancer risk factor exposure, lower cancer screening rates, and poorer quality of cancer care relative to non-Hispanic Whites. To address these issues, the Southeastern American Indian Cancer health Equity Partnership (SAICEP) was formed to understand and address cancer disparities among southeastern American Indians (AIs). Methods SAICEP formed in 2021 through the Community Outreach and Engagement offices of the NCI-designated Comprehensive Cancer Centers in North Carolina (NC). The catchment areas for these cancer centers include the tribal homelands for eight state and federally recognized Tribes, representing the largest AI populations in the eastern US. SAICEP seeks to: (1) increase awareness of cancer health needs of AI populations; (2) expand access to cancer health education and build community capacity to address cancer health needs; (3) develop collaborative research relationships to better understand and address the AI cancer burden. Results For Aim 1, SAICEP created a virtual speakers’ series, featuring prominent AI cancer researchers and clinicians, hosted by the UNC Lineberger Cancer Network three times a year. To date, 10 webinars have been convened, with a total of 538 participants. For Aim 2, SAICEP participates in tribal events throughout the year, reaching over 3500 AIs and disseminating printed cancer educational materials and giveaways. For Aim 3, SAICEP secured funding to conduct analyses to assess cancer incidence, mortality, and care quality for NC AIs, to collect information to understand community cancer needs and culturally adapt and disseminate information on cancer screening and risk reduction. Conclusion Through its targeted research and engagement, SAICEP has successfully moved towards achieving its goal of understanding and addressing cancer disparities among AIs in NC. Future directions will involve the development of a community advisory board and collaborations with Tribes in other states.Plain Language Summary American Indians and Alaska Natives (AIANs) face significant challenges with cancer, including greater exposure to risk factors, lower screening rates, limited access to quality care compared to non-Hispanic Whites, and higher rates of cancer-related deaths. To address these issues, the Southeastern American Indian Cancer health Equity Partnership (SAICEP) was created. SAICEP’s mission is to understand and address the burden of cancer among American Indian (AI) people in our combined service area. SAICEP was formed in 2021 as a partnership between the three NCI designated cancer centers in North Carolina, which serve multiple tribal nations and Urban Indian organizations. SAICEP engages with Native people through initiatives like a virtual speakers’ series featuring AI cancer experts. So far, SAICEP has hosted 10 webinars, reaching 538 attendees. Additionally, SAICEP participates in tribal events, sharing educational materials and information to help communities understand and reduce cancer risks. Our team has also secured funding to study cancer trends, care quality, and community needs for AIs in North Carolina, which helps identify ways to improve care and support cancer prevention. SAICEP’s efforts have made meaningful progress toward understanding and addressing cancer disparities among AIs. Looking ahead, the group plans to create a community advisory board and expand its work to include partnerships with Tribes in other states. By combining research, education, and community outreach, SAICEP aims to ensure AI communities have the tools and resources they need to fight cancer

    A scoping review examining measurement of anti-transgender stigma in low- and middle-income countries

    Get PDF
    Globally, transgender and other gender diverse (trans) people experience widespread prejudice, discrimination, violence, and other forms of stigma, which contribute to negative health outcomes. Most anti-trans stigma research has been conducted in high-income countries. Measurement of anti-trans stigma in low- and middle-income countries (LMICs) is important for understanding and improving the health of trans populations globally. Accordingly, this scoping review explores the use of quantitative anti-trans stigma measures in LMICs. This scoping review follows the guidance of the PRISMA extension for Scoping Reviews (PRISMA-ScR) Checklist and examines empirical research with trans populations in LMICs published in English, Spanish, Arabic, and Russian between 2001-2024. Study eligibility criteria included: 1) trans study population, 2) LMIC study location, 3) quantitative or mixed-method study design, and 4) quantitative measurement of anti-trans stigma. The search yielded 82 articles (representing 65 unique studies) from 34 LMICs. Most articles were published since 2018. No articles focused exclusively on trans men. About 62% of articles included a primary focus on stigma; health outcomes primarily examined HIV and mental health. Nearly all articles (95%) measured enacted stigma; other forms of stigma (e.g., internalized and anticipated) were less commonly measured, and structural stigma was only measured in 4 articles. More than half of the articles (55%) measured stigma both broadly and within specific contexts (e.g., from family, in health care). More research exploring anti-trans stigma is needed, especially with trans-masculine and other gender diverse people, measuring outcomes beyond HIV and mental health, and measuring forms of stigma beyond enacted stigma. Expanding and improving measurement of anti-trans stigma in LMICs can improve our understanding of the mechanisms shaping health equity to inform context specific and tailored health interventions to support trans communities worldwide

    Relationship Between Holistic Admissions Criteria and Pharmacy Student Performance

    Get PDF
    This study examined how admissions criteria relate to student performance during the first year of pharmacy school. Data from the PharmD classes of 2025–2027 were analyzed, including undergraduate GPA, PCAT scores, MMI scores, and survey responses. Higher first-year GPAs were associated with higher undergraduate GPA, science/math GPA, PCAT scores, and MMI scores. Survey results indicated that students viewed prior academic performance as the best predictor of success, while heavy course load and time management challenges contributed most to academic struggles. These findings may help guide admissions decisions and inform student support strategies.Doctor of Pharmac

    Evaluation of Two Alloplastic Biomaterials in a Critical-Size Rat Calvarial Defect Model

    Get PDF
    Aim: to evaluate the bone regeneration capacity of two alloplastic biomaterials in a critical-size rat calvarial defect model. Methods: A total of 80 rats were randomized into 8 groups of 10 animals each. An Ø8 mm, critical-size calvarial defect was created, and the following treatments were randomly allocated: sham surgery, deproteinized bovine bone mineral (DBBM) + collagen membrane (CM), poly-(lactic-co-glycolic-acid) (PLGA)-coated pure phase β-tricalcium phosphate (β-TCP), or PLGA-coated 60% hydroxyapatite (HA):40%β-TCP. Animals were allowed to heal for 2 and 6 weeks. Microcomputed tomography (μCT) was used to evaluate mineralized tissue and biomaterial displacement. Histological samples were used to evaluate new bone formation. Results: μCT analysis showed no significant differences among groups for total volume of mineralized tissue or residual biomaterials. DBBM + CM showed significantly increased horizontal biomaterial displacement at 2 weeks but not at 6 weeks. Histological analysis showed that sham surgery had a significantly higher percentage of bone area fraction than the DBBM + CM and PLGA + β-TCP at 2 weeks, but not at 6 weeks. Residual biomaterial area fraction showed no significant differences among experimental groups at any healing time. Conclusions: The alloplastic biomaterials showed suitable construct integrity and retention in the defect. All biomaterials were associated with limited new bone formation comparable to the sham surgery control

    Tiny homes—big movement: building a permanent and affordable housing option for people with severe mental illness

    Get PDF
    Ensuring an adequate supply of affordable housing is one of the most pressing public health challenges facing the United States. This challenge is particularly pressing for people with severe mental illness living on incomes 25% below the federal poverty level, placing them at increased risk of housing insecurity. This paper presents a community case study of the Tiny Homes Village (THV) demonstration project. In this project a community partnership used tiny homes to create a new affordable housing option for people with severe mental illness. The THV built 15 tiny homes through a public/private cross-sector partnership consisting of a private non-profit organization, a university, a community mental health center, and construction companies. All 15 homes have the same floor plan and were constructed at the same time using a Blitz Build model in 90 business days at a cost of approximately $50,000 per home. Each home is built on a permanent foundation, and includes 416 square feet of interior, heated space and five living spaces: a full bathroom, a bedroom, an open-concept kitchen and living room, and a covered front porch that provides an additional 96 square feet of unheated space. The tiny homes are located within a village that offers several amenities and a range of community-based services. This community case study demonstrates the power of public-private partnerships to tackle some of our most complex and entrenched social problems while also providing a blueprint for how to expand the affordable housing options for people with severe mental illness

    A Transect Through the Living Environments of Slovakia’s Roma Population: Urban, Sub-Urban, and Rural Settlements, and Exposure to Environmental and Water-Related Health Risks

    Get PDF
    The Roma population is one of Europe’s largest ethnic minorities, often living in inadequate living conditions, worse than those of the majority population. They frequently lack access to essential services, even in high-income countries. This lack of basic services—particularly in combination with proximity to (stray) animals and human and solid waste—significantly increases environmental health risks, and leads to a higher rate of endoparasitic infections. Our study sheds light on the living conditions and health situation in Roma communities in Slovakia, focusing on the prevalence of intestinal endoparasitic infections across various settlement localisations. It highlights disparities and challenges in access to safe drinking water, sanitation, and hygiene (WASH) and other potentially disease-exposing factors among these marginalised populations. This study combines a comprehensive review of living conditions as per national data provided through the Atlas of Roma communities with an analysis of empirical data on parasitological infection rates in humans, animals, and the environment in settlements, applying descriptive statistical methods. It is the first study in Europe to provide detailed insights into how living conditions vary and cause health risks across Roma settlements, ranging from those integrated within villages (inside, urban), to those isolated on the outskirts (edge, sub-urban) or outside villages (natural/rural). Our study shows clear disparities in access to services, and in health outcomes, based on where people live. Our findings underscore the fact that (i) place—geographical centrality in particular—in an already challenged population group plays a major role in health inequalities and disease exposure, as well as (ii) the urgent need for more current and comprehensive data. Our study highlights persistent disparities in living conditions within high-income countries and stresses the need for greater attention and more sensitive targeted health-promoting approaches with marginalised communities in Europe that take into consideration any and all of the humans, ecology, and animals affected (=One Health)

    Manganese and iron in drinking water in three West-African countries: Implications for health, acceptability, and disinfection

    Get PDF
    Manganese and iron decrease aesthetic acceptability and chlorine disinfection performance in drinking water, and can be toxic at high doses. We present the first characterization of manganese and iron occurrence in drinking water relative to concentration benchmarks for these three outcomes. Manganese and iron concentrations were evaluated in 261 drinking water samples obtained from boreholes and small groundwater-fed piped systems across large rural regions of Ghana, Mali and Niger. One or both metals exceeded aesthetic benchmark concentrations in 30% of samples and reached concentrations likely to interfere with chlorine disinfection in 5% of samples. Manganese exceeded 2011 WHO health-based benchmarks for drinking water (400 µg/L) in 2% of samples, and exceeded updated provisional health-based benchmark concentrations (80 µg/L) in 13% of samples. Iron occurred at levels exceeding health-based benchmarks in 5% of samples. These results suggest that manganese and iron contribute directly and indirectly to public health problems in a substantive proportion of drinking water sources in the study setting. Implementation of rural drinking water systems reliant on groundwater sources should account for the occurrence of these metals during siting, design, construction, operation, and monitoring/surveillance. Strengthening these capacities, particularly with respect to sampling and testing water sources for these metals, may support management and regulatory efforts to manage the occurrence of these metals in drinking water and their potential adverse effects. Finally, generating and synthesizing additional evidence on the occurrence and effects of iron and manganese in drinking water will support national efforts to manage both contaminants, inform discussions regarding the suitability of current health-based Mn guidelines for protecting sensitive life stages, and underscore the value of monitoring Mn as a priority chemical contaminant under Sustainable Development Goal (SDG) target 6.1: “By 2030, achieve universal and equitable access to safe and affordable drinking water for all.

    90,565

    full texts

    95,038

    metadata records
    Updated in last 30 days.
    Carolina Digital Repository is based in United States
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇