University of South Florida
  • Tampa, United States
Recent publications
Microservice (MS) structures a service application as a collection of independently deployable service modules, making it particularly suitable for delivering complex applications in distributed computing systems. This paper investigates MS architecture over Mobile Edge Computing (MEC) networks (hereafter referred to as EdgeMS) and studies an EdgeMS placement problem that aims to deploy MS modules over the MEC network in a manner that maximizes the reward of MS application providers. A novel algorithm called Dual-GNN Deep Deterministic Policy Gradient (DG-DDPG) is proposed to establish an intelligent EdgeMS placement policy for optimizing the location of MS modules and performing fractional computing resource allocation. DG-DDPG leverages the graph neural network (GNN) to comprehend the graph-structured information encapsulated in the MS application structure and MEC network. A dual-GNN core is constructed in DG-DDPG, one GNN for MS applications to distill knowledge from intricate connections between MS modules, and the other GNN for MEC networks to capture complicated interactions between edge sites when providing EdgeMS. DG-DDPG embeds the dual-GNN core in a DDPG-based reinforcement learning framework, which not only handles temporal dependencies between EdgeMS placement decisions for maximizing long-term reward but also supports continuous action space for enabling fractional resource allocation. In particular, the learning process of DG-DDPG is tailored to address hard constraints (i.e., computing capacity and MS application completeness) in the EdgeMS placement problem. We design constraint-based regularization terms and add them to the objective of DG-DDPG, which facilitates the identification of feasible placement decisions during learning. We carry out systematic experiments to evaluate the performance of DG-DDPG, and the results show that DG-DDPG outperforms state-of-the-art benchmarks in terms of reward, service delay and deployment cost.
Federated Learning (FL) enables massive local data owners to collaboratively train a deep learning model without disclosing their private data. The importance of local data samples from various data owners to FL models varies widely. This is exacerbated by the presence of noisy data that exhibit large losses similar to important (hard) samples. Currently, there lacks an FL approach that can effectively distinguish hard samples (which are beneficial) from noisy samples (which are harmful). To bridge this gap, we propose the joint Federated Meta-Weighting based Client and Sample Selection (FedMW-CSS) approach to simultaneously mitigate label noise and hard sample selection. It is a bilevel optimization approach for FL client-and-sample selection and global model construction to achieve hard sample-aware noise-robust learning in a privacy preserving manner. It performs meta-learning based online approximation to iteratively update global FL models, select the most positively influential samples and deal with training data noise. To utilize both the instance-level information and class-level information for better performance improvements, FedMW-CSS efficiently learns a class-level weight by manipulating gradients at the class level, e.g. , it performs a gradient descent step on class-level weights, which only relies on intermediate gradients. Theoretically, we analyze the privacy guarantees and convergence of FedMW-CSS. Extensive experiments comparison against eight state-of-the-art baselines on six real-world datasets in the presence of data noise and heterogeneity shows that FedMW-CSS achieves up to 28.5% higher test accuracy, while saving communication and computation costs by at least 49.3% and 1.2%, respectively.
A shallow anterior chamber will occur when a patient has a short axial length, a small eye, hyperopic refraction, or a cataract that has been present too long before the patient is willing to undergo cataract surgery. A cataract is much like a tree. It will grow throughout life and can lay down a new ring of lens epithelial cells each year, growing and crowding the anterior chamber eventually causing a shallow anterior chamber and narrow-angle or angle closure glaucoma and sudden blindness if cataract surgery is not performed in a timely fashion.
Suturing has been a pillar of plastic surgery since its very foundation. From ancient civilizations to modern practices, hair sutures have resurfaced as vital material for wound closure and tissue approximation. This study aimed to explore the historical evolution and application of hair sutures, particularly in facial reconstruction, skin flaps and grafts, peripheral nerve surgery, oral/maxillofacial surgery, and oculoplastic surgery. Notable surgeons and scholars, from Susrata to Sir Harold Gillies, have cited unique properties of horsehair sutures, namely, their pliability, strength, and nonabsorbable nature in various applications. Despite the advent of synthetic materials, recent interest in hair sutures has emerged, driven by their renewable nature and potential for "green" surgical practices. Contemporary studies agree with the historical perspectives, suggesting that hair sutures offer favorable aesthetic outcomes, minimal tissue reaction, and cost-effectiveness. This study illustrates the legacy of hair sutures in plastic and reconstructive surgery.
Stimulator of interferon genes (STING) is a transmembrane endoplasmic reticulum (ER) resident protein involved in innate immunity. STING activation occurs by binding of cyclic guanosine-(2'→5')-monophosphate-adenosine-(3'→5')-monophosphate (2',3'-cGAMP) to STING, which leads to downstream production of type 1 interferons (IFN-1). We generated molecular dynamics (MD) equilibrated agonist and antagonist models of human STING (hSTING) for computer-based screening and now report the discovery of clonixeril (CXL) as the most potent non-nucleotide hSTING modulator discovered to date. We demonstrate in vitro and in cellulo that CXL has two modes of interaction with hSTING, one with an EC 50 above 1 nM and the other with an EC 50 in the 1 fM-100 aM range (10-15-10-16 M). In cell-based experiments, when CXL is titrated below 1 nM, it displays inverse dose dependent antagonistic behavior toward hSTING. We have substantiated that CXL displays this exceptionally strong inhibitory effect on hSTING mediated IFN-1 production using a THP-1 cell https://doi.org/10.26434/chemrxiv-2024-bxr73 ORCID: https://orcid.org/0000-0003-3353-4853 Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 2 luciferase reporter for interferon regulatory factor 3 (IRF3). Further characterization of CXL was performed in HEK293 cells and by using biophysical and biochemical techniques.
Introduction/Aims Prospective, randomized, controlled trials of intravenous immunoglobulin (IVIG) maintenance therapy in myasthenia gravis (MG) are lacking. In this trial, we evaluated the safety and efficacy of caprylate/chromatography‐purified IVIG; (IGIV‐C) in patients with generalized MG undergoing standard care. Methods Sixty‐two patients enrolled in this phase 2, multicenter, international, randomized trial (1:1 IGIV‐C [2 g/kg loading dose; 1 g/kg every 3 weeks through week 21] or placebo). Efficacy was assessed by changes in Quantitative MG (QMG) score at week 24 versus baseline (primary endpoint) and percentage of patients with clinical improvement in QMG, MG Composite (MGC), and MG‐Activities of Daily Living (MG‐ADL) scores (secondary endpoints). Safety assessments reported all adverse events (AEs). Results The change in QMG at 24 weeks was −5.1 for IGIV‐C and −3.1 for placebo ( p = .187). Seventy percent of patients in the IGIV‐C group had improvement in MG‐ADL (≥2‐point decrease) versus 40.6% in the placebo group ( p = .025). Patients showing clinical improvement in QMG and MGC (≥3‐point decrease) were 70.0% for IGIV‐C versus 59.4% for placebo ( p = .442) and 60.0% for IGIV‐C versus 53.1% for placebo ( p = .610). IGIV‐C was well tolerated; serious AEs were similar between arms. Three of four MG exacerbations requiring hospitalizations occurred in the IGIV‐C arm with one death. Discussion Several efficacy parameters showed numerical results greater than those seen in the placebo group. This was a small study and may have been underpowered to see significant differences. Additional studies may be warranted to fully determine the efficacy of IVIG maintenance therapy in MG.
We aimed to study transcriptional and phenotypic changes in circulating immune cells associated with increased risk of mortality in COVID-19, resolution of pulmonary fibrosis in post-COVID-19-Interstitial Lung Disease (ILD) and persistence of Idiopathic Pulmonary Fibrosis. Whole blood and Peripheral Blood Mononuclear cells (PBMC) were obtained from 227 subjects with COVID-19, post-COVID-19 Interstitial Lung Disease (ILD), IPF and controls. We measured a 50-gene signature (nCounter, Nanostring) previously found to be predictive of IPF and COVID-19 mortality along with plasma levels of several biomarkers by Luminex. Additionally, we performed single-cell RNA sequencing in PBMC (10X Genomics) to determine the cellular source of the 50-gene signature. We identified the presence of three genomic risk profiles in COVID-19 based on the 50-gene signature associated with low, intermediate, or high-risk of mortality and with significant differences in pro-inflammatory and pro-fibrotic cytokines. COVID-19 patients in the high-risk group had increased expression of seven genes in CD14 ⁺ HLA-DR low CD163 ⁺ Monocytic-Myeloid-Derived Suppressive cells (7Gene-M-MDSCs) and decreased expression of 43 genes in CD4 and CD8 T cell subsets. The loss of 7Gene-M-MDSC and increased expression of these 43 genes in T cells was seen in survivors with post-COVID-19-ILD. On the contrary, IPF patients had low expression of the 43 genes in CD4 and CD8 T cells. Collectively, we showed that a 50-gene, high-risk profile, predictive of IPF and COVID-19 mortality is characterized by a genomic imbalance in monocyte and T-cell subsets. This imbalance reverses in survivors with post-COVID-19-ILD highlighting genomic differences between post-COVID-19-ILD and IPF.
This research applied a critical quantitative approach to the 247sports.com recruiting website to consider whether the “Black quarterback” label systematically rooted in sport persisted as a mediated form of racial stacking for high school football quarterbacks. A 20-year content analysis (2001–2020) examined race, position code, star value, and position ranking for 3448 high school quarterbacks. The results indicated a pattern of racial stacking through the use of coded language, where 85.3% of “pro-style” quarterbacks were White, and Black quarterbacks occupied a majority (55.5%) in the “dual-threat” code. These findings are contextualized through a QuantCrit approach define here or see below where the greatest concern is the cyclical predictability of recruiting rankings that illustrate the centrality and permanence of racism through the constructed duality of two quarterback codes. This research identifies a racialization of ability that establishes “pro-style” as the property of Whiteness and showcases a fundamental relationship between recruiting websites and the discriminatory language drawn on by media, coaches, and others to demarcate quarterbacks by race. This study illuminates how power works through mediated practice that creates an ideological reservoir of racial marking of football players enmeshed in the historical stacking process.
The incidence of obstetric anal sphincter injuries (OASI) has increased in recent years, which may be due to improved recognition and documentation. There is limited evidence regarding the effects of thorough documentation of obstetric anal sphincter injury repairs on postpartum clinical outcomes. Our objectives were to (1) compare the incidence of perineal wound complications between documentation groups, (2) compare other adverse events, and (3) to describe factors associated with adequate documentation. We hypothesized that better documentation would be associated with improved clinical outcomes. This was a retrospective cohort study of 599 patients with OASI at a tertiary care referral center between January 2015 and December 2020. A priori definitions of documentation adequacy were utilized to stratify delivery notes. On the basis of these criteria, there were preferred, adequate, and inadequate documentation groups. Maternal characteristics, outcomes, and peripartum factors were compared between the groups. There were no significant differences in clinical outcomes between the groups. A higher degree of perineal laceration (p < 0.001), greater blood loss (p = 0.002), and the need for repairs in the operating room (p = 0.019) were significant factors associated with adequate documentation. Clinicians who were comprehensive in their documentation were more likely to refer patients to Urogynecology (p < 0.001) and to add OASI to the electronic medical record problem list (p = 0.005). While certain factors are associated with adequate documentation, this did not improve clinical outcomes for OASI and further research is warranted to explore the importance of medical documentation surrounding OASI.
Background Rapidly reversible weakness with neck flexion (McArdle sign) is common in patients with multiple sclerosis (MS). The pathophysiology is unknown. Objective To evaluate changes in central motor conduction time (CMCT) in patients with and without McArdle sign. Methods We measured McArdle sign with a torque cell and CMCT with neck flexed and extended in patients with MS, other causes of myelopathy, and healthy controls. Results CMCT was prolonged with neck flexion disproportionately in those with MS-associated myelopathy (MSAM) with prominent McArdle sign compared to MS patients with lesser degrees of McArdle sign, and to controls. Conclusion McArdle sign may result from stretch-induced slowing of conduction due to demyelination.
In photoacoustic imaging (PAI), a delay-and-sum (DAS) beamforming reconstruction algorithm is widely used due to its ease of implementation and fast execution. However, it is plagued by issues such as high sidelobe artifacts and low contrast, that significantly hinder the ability to differentiate various structures in the reconstructed images. In this study, we propose an adaptive weighting factor called spatial coherence mean-to-standard deviation factor (scMSF) in DAS, which is extended into the spatial frequency domain. By combining scMSF with a minimum variance (MV) algorithm, the clutter level is reduced, thereby enhancing the image contrast. Quantitative results obtained from the phantom experiment demonstrate that our proposed method improves contrast ratio (CR) by 30.15 dB and signal-to-noise ratio (SNR) by 8.62 dB compared to DAS while also improving full-width at half maxima (FWHM) by 56%. From the in-vivo experiments, the scMSF-based reconstruction image exhibits a higher generalized contrast-to-noise ratio (gCNR), indicating improved target detectability with a 25.6% enhancement over DAS and a 22.5% improvement over MV.
Introduction Optimal outcomes during childhood cancer treatment require effective management of toxicities, often called supportive care. A lack of agreement on what comprises supportive care limits the development and provision of comprehensive guidance (for this work, we have defined supportive care as any disease‐ or treatment‐related condition experienced by children with cancer, excluding psychosocial conditions, palliative care, survivorship, or procedural topics). To address this gap, we conducted a consensus‐building exercise among global experts to define and prioritize topics for supportive care. Methods Two rounds of brainstorming and prioritization exercises were conducted. A multidisciplinary panel nominated by professional societies and cooperative groups was formed to ensure geographic and resource representation using snowball sampling. An internal expert panel generated an initial list of supportive care topics. In round one, the multidisciplinary panel reviewed the initial list and recommended additional topics, followed by prioritization in round two using a seven‐point Likert scale. Results were summarized using descriptive statistics. Results The multidisciplinary panel consisted of 57 members representing 32 countries. The initial list included 46 topics; 161 additional topics were suggested. After removing duplicates and out‐of‐scope additions, the final list contained 62 topics. Febrile neutropenia, sepsis, bloodstream infections, and pain were ranked highest priority. Mortality, morbidity, and frequency of the event were identified as the most important factors influencing prioritization. Conclusion Through a multidisciplinary and globally representative process, we identified core supportive care topics and factors influencing their prioritization for childhood cancer. Outputs from this work will inform efforts to generate resource‐adapted recommendations for a global audience. This supports ongoing WHO CureAll work to develop a health systems‐level policy brief of supportive care requirements in the management of children with cancer.
Background Ten percent of patients with melanoma develop in-transit metastases (ITM). Isolated limb infusion (ILI) is a well-established therapy for unresectable ITM on the extremities, but the ideal sequencing/line of therapy of ILI has not been defined. This study evaluates ILI as first-line, second-line, or third or later-line therapy. Methods A retrospective review included all patients with unresectable ITM who underwent ILI from 2006-2023. Results A total of 130 patients were identified, 61% female, median age of 71 (31-89) years. Median follow-up was 37.5 months. ILI was first-line therapy in 80% (n = 104), second-line in 15% (n = 19), and third or later-line in 5.4% (n = 7). Overall response rate (ORR) and complete response (CR) rates for ILI as any line of therapy were 74% and 41%, respectively. ORR for ILI as first, second, or third or later-line therapy were 78%, 63%, and 57%, respectively. CR rates for ILI as first, second, or third or later-line therapy were 42%, 37%, and 43%, respectively. There were no significant differences in ORR, progression-free survival (PFS), overall survival, or disease-free survival between therapy lines. Median PFS for ILI as first, second, or third or later-line therapy were 6.9, 5.4, and 18 months, respectively. Conclusion Patients responded well to ILI, whether or not they received previous therapies for unresectable ITM. First-line ILI appears to have a better ORR than later lines of ILI. Although sample size limited statistical significance, there was a 21% improvement in ORR when ILI was used as first-line vs third-line therapy, which is clinically meaningful. ILI is effective for unresectable melanoma ITM and can be used as salvage therapy.
To standardize preparation for residency, the AAMC created 13 generalized Entrustable Professional Activities (EPA). To improve student preparedness for surgical residency, we developed surgery-specific EPAs by comparing attending surgeon and resident opinions regarding what EPAs should be addressed during the fourth year of medical school. A focus group of attending surgeons developed 29 surgery-specific EPAs based on the AAMC’s core EPAs. We then surveyed attending surgeons and current surgery residents. Data were collected using the Chen Scale for determining activity entrustability and then evaluated for agreement and discordance. Of those queried, 12 of 12 (100%) attending surgeons and 42 of 106 (40%) residents completed the survey. Those surveyed agreed that of the proposed EPAs 18 were entrustable, 6 were not, and there was discordance for 5 of the EPAs. Specialty-specific EPAs demonstrate the potential to bridge performance gaps between medical school and residency.
Background Immune checkpoint blocker (ICB) therapy has changed the treatment landscape of solid cancers, with remarkable upfront responses; however, most patients develop secondary resistance and relapse within a few months. Currently, biomarkers to predict secondary resistance to ICB are unidentified. We hypothesized that gut microbiome and microbial metabolome could predict and modulate secondary resistance to ICB therapy across solid cancers. Methods We conducted a prospective study to evaluate the differential composition of the gut microbiome and microbial metabolome in patients with durable responses-DR (defined as progression-free survival, PFS >1 year) versus those with progressive disease(PD) (PFS <6 months) and enrolled patients with metastatic renal cell cancer (RCC), melanoma, and head and neck cancer (HNSCC) receiving first-line ICB.We collected paired blood and stool samples at treatment initiation, at one year if ongoing response, or at disease progression. Additional samples were collected from a subset of durable responders who subsequently lost their response. Stool metagenomics and fecal and plasma metabolomics was evaluated, with focus on microbial metabolites. Complementarily, we conducted preclinical studies using mice models of HNSCC and RCC to elucidate the role of microbial metabolome in secondary ICB resistance. 30 mice were injected with MOC1 cells. When tumor volume reached 100mm³, they were treated with bi-weekly anti-PD1 for eight doses, with tumor measurement. A subset of ten mice with the slowest tumor growth and classified as durable responder, were randomized into two groups: one continued bi-weekly anti-PD1, while the other group also received daily Indole 3 Acetic Acid (I3A). Tumor growth and PFS were monitored for additional two weeks. Results 220 patients and their samples were eligible, with median follow-up of 26 months. Significant differences in microbial composition and metabolite abundance was noted in DR vs PD samples. Higher abundance of Ruminococcus Gnavus(Odds ration [OR] -22.4, p-0.01, Bacteroides uniformis (OR- 7.8,p-0.005) and significantly lower abundance of Immunosuppressive microbial metabolite- I3A and indole carboxylic acid (ICA) was associated with durable response(DR) samples.In preclinical models, administration of I3A accelerated tumor growth, with mean changes in tumor volume of 120% versus 25% (p=0.01). Conclusions This is one of the first prospective study to demonstrate the role of microbial metabolome in predicting and mediating secondary ICB resistance. Information gained can help predict which patients are more likely to develop secondary resistance and inform future trials aimed to modulate microbial metabolome to mitigate secondary ICB resistance.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
17,009 members
Yonggang Liu
  • College of Marine Science
Ulla Uusitalo
  • Department of Pediatrics
Ananda Shankar Bhattacharjee
  • Civil and Environmental Engineering
Ismail Kazem
  • Department of Oncologic Sciences
Ricardo Izurieta
  • College of Public Health
Information
Address
Tampa, United States
Head of institution
Steve Currall