The systematic review of clinical research papers is a labor-intensive and time-consuming process that often involves the screening of thousands of titles and abstracts. The accuracy and efficiency ...of this process are critical for the quality of the review and subsequent health care decisions. Traditional methods rely heavily on human reviewers, often requiring a significant investment of time and resources.
This study aims to assess the performance of the OpenAI generative pretrained transformer (GPT) and GPT-4 application programming interfaces (APIs) in accurately and efficiently identifying relevant titles and abstracts from real-world clinical review data sets and comparing their performance against ground truth labeling by 2 independent human reviewers.
We introduce a novel workflow using the Chat GPT and GPT-4 APIs for screening titles and abstracts in clinical reviews. A Python script was created to make calls to the API with the screening criteria in natural language and a corpus of title and abstract data sets filtered by a minimum of 2 human reviewers. We compared the performance of our model against human-reviewed papers across 6 review papers, screening over 24,000 titles and abstracts.
Our results show an accuracy of 0.91, a macro F
-score of 0.60, a sensitivity of excluded papers of 0.91, and a sensitivity of included papers of 0.76. The interrater variability between 2 independent human screeners was κ=0.46, and the prevalence and bias-adjusted κ between our proposed methods and the consensus-based human decisions was κ=0.96. On a randomly selected subset of papers, the GPT models demonstrated the ability to provide reasoning for their decisions and corrected their initial decisions upon being asked to explain their reasoning for incorrect classifications.
Large language models have the potential to streamline the clinical review process, save valuable time and effort for researchers, and contribute to the overall quality of clinical reviews. By prioritizing the workflow and acting as an aid rather than a replacement for researchers and reviewers, models such as GPT-4 can enhance efficiency and lead to more accurate and reliable conclusions in medical research.
Aging is associated with cognitive decline, which can critically affect quality of life. Examining the biology of cognitive aging across species will lead to a better understanding of the fundamental ...mechanisms involved in this process, and identify potential interventions that could help to improve cognitive function in aging individuals. This minireview aimed to explore the mechanisms and processes involved in cognitive aging across a range of species, from flies to rodents, and covers topics, such as the role of reactive oxygen species and autophagy/mitophagy in cognitive aging. Overall, this literature provides a comprehensive overview of the biology of cognitive aging across species, highlighting the latest research findings and identifying potential avenues for future research. Geriatr Gerontol Int 2024; 24: 15-24.
Plant cell death plays important roles during plant-pathogen interactions. To study pathogen-induced cell death, there is a need for cytological tools that allow determining not only host cell ...viability, but also cellular events leading to cell death with visualization of pathogen development. Here we describe a live cell imaging method to provide insights into the dynamics of cell death in rice (Oryza sativa). This method uses live-cell confocal microscopy of rice sheath cells mechanically damaged or invaded by fluorescently-tagged Magnaporthe oryzae together with fluorescent dyes fluorescein diacetate (FDA) and propidium iodide (PI). FDA stains the cytoplasm of live cells exclusively, thus also visualizing the vacuole, whereas PI stains nuclei of dead cells.
We first demonstrated that confocal microscopy of rice leaf sheaths stained with FDA and PI discriminated between live cells and mechanically-killed cells. FDA-derived fluorescein was confined to the cytoplasm of live cells, indicating the intact vacuolar and plasma membranes. We also observed previously unreported fluorescein patterns in mechanically damaged cells. These patterns include: (1) homogeneous distribution of fluorescein in the increased area of the cytoplasm due to the shrunken vacuole; (2) the increase of the fluorescein intensity; and (3) containment of the brighter fluorescein signal only in affected cells likely due to closure of plasmodesmata. We refer to these as novel fluorescein patterns in this study. Simultaneous imaging of fluorescently-tagged M. oryzae (red) and FDA staining (green) in rice cells revealed characteristic features of the hemibiotrophic interaction. That is, newly invaded cells are alive but subsequently become dead when the fungus spreads into neighbor cells, and biotrophic interfacial complexes are associated with the host cytoplasm. This also revealed novel fluorescein patterns in invaded cells. Time-lapse imaging suggested that the FDA staining pattern in the infected host cell progressed from typical cytoplasmic localization (live cell with the intact vacuole), to novel patterns (dying cell with closed plasmodesmata with the shrunken or ruptured vacuole), to lack of fluorescence (dead cell).
We have developed a method to visualize cellular events leading to host cell death during rice blast disease. This method can be used to compare and contrast host cell death associated with disease resistance and susceptibility in rice-M. oryzae and other host-pathogen interactions.
Background: Astaxanthin modulates immune response, inhibits cancer cell growth, reduces bacterial load and gastric inflammation, and protects against UVA-induced oxidative stress in in vitro and ...rodent models. Similar clinical studies in humans are unavailable. Our objective is to study the action of dietary astaxanthin in modulating immune response, oxidative status and inflammation in young healthy adult female human subjects. Methods: Participants (averaged 21.5 yr) received 0, 2, or 8 mg astaxanthin (n = 14/diet) daily for 8 wk in a randomized double-blind, placebo-controlled study. Immune response was assessed on wk 0, 4 and 8, and tuberculin test performed on wk 8. Results: Plasma astaxanthin increased (P < 0.01) dose-dependently after 4 or 8 wk of supplementation. Astaxanthin decreased a DNA damage biomarker after 4 wk but did not affect lipid peroxidation. Plasma C-reactive protein concentration was lower (P < 0.05) on wk 8 in subjects given 2 mg astaxanthin. Dietary astaxanthin stimulated mitogen-induced lymphoproliferation, increased natural killer cell cytotoxic activity, and increased total T and B cell subpopulations, but did not influence populations of T(helper), T(cytotoxic) or natural killer cells. A higher percentage of leukocytes expressed the LFA-1 marker in subjects given 2 mg astaxanthin on wk 8. Subjects fed 2 mg astaxanthin had a higher tuberculin response than unsupplemented subjects. There was no difference in TNF and IL-2 concentrations, but plasma IFN-γ and IL-6 increased on wk 8 in subjects given 8 mg astaxanthin. Conclusion: Therefore, dietary astaxanthin decreases a DNA damage biomarker and acute phase protein, and enhances immune response in young healthy females.
Age‐related memory impairment (AMI) occurs in many species, including humans. The underlying mechanisms are not fully understood. In wild‐type Drosophila (w1118), AMI appears in the form of a ...decrease in learning (3‐min memory) from middle age (30 days after eclosion DAE). We performed in vivo, DNA microarray, and behavioral screen studies to identify genes controlling both lifespan and AMI and selected mitochondrial Acon1 (mAcon1). mAcon1 expression in the head of w1118 decreased with age. Neuronal overexpression of mAcon1 extended its lifespan and improved AMI. Neuronal or mushroom body expression of mAcon1 regulated the learning of young (10 DAE) and middle‐aged flies. Interestingly, acetyl‐CoA and citrate levels increased in the heads of middle‐aged and neuronal mAcon1 knockdown flies. Acetyl‐CoA, as a cellular energy sensor, is related to autophagy. Autophagy activity and efficacy determined by the positive and negative changes in the expression levels of Atg8a‐II and p62 were proportional to the expression level of mAcon1. Levels of the presynaptic active zone scaffold protein Bruchpilot were inversely proportional to neuronal mAcon1 levels in the whole brain. Furthermore, mAcon1 overexpression in Kenyon cells induced mitophagy labeled with mt‐Keima and improved learning ability. Both processes were blocked by pink1 knockdown. Taken together, our results imply that the regulation of learning and AMI by mAcon1 occurs via autophagy/mitophagy‐mediated neural plasticity.
As aging progresses, the expression of the TCA enzyme mAcon1 decreases. Accordingly, the amounts of citrate and acetyl‐CoA increase, and the increased acetyl‐CoA inhibits autophagy/mitophagy. After that, synaptic plasticity is reduced by decreased autophagy/mitophagy, and as a result, AMI appears in middle‐aged flies.
Large language models (LLMs) like OpenAI's ChatGPT are powerful generative systems that rapidly synthesize natural language responses. Research on LLMs has revealed their potential and pitfalls, ...especially in clinical settings. However, the evolving landscape of LLM research in medicine has left several gaps regarding their evaluation, application, and evidence base.
This scoping review aims to (1) summarize current research evidence on the accuracy and efficacy of LLMs in medical applications, (2) discuss the ethical, legal, logistical, and socioeconomic implications of LLM use in clinical settings, (3) explore barriers and facilitators to LLM implementation in healthcare, (4) propose a standardized evaluation framework for assessing LLMs' clinical utility, and (5) identify evidence gaps and propose future research directions for LLMs in clinical applications.
We screened 4,036 records from MEDLINE, EMBASE, CINAHL, medRxiv, bioRxiv, and arXiv from January 2023 (inception of the search) to June 26, 2023 for English-language papers and analyzed findings from 55 worldwide studies. Quality of evidence was reported based on the Oxford Centre for Evidence-based Medicine recommendations.
Our results demonstrate that LLMs show promise in compiling patient notes, assisting patients in navigating the healthcare system, and to some extent, supporting clinical decision-making when combined with human oversight. However, their utilization is limited by biases in training data that may harm patients, the generation of inaccurate but convincing information, and ethical, legal, socioeconomic, and privacy concerns. We also identified a lack of standardized methods for evaluating LLMs' effectiveness and feasibility.
This review thus highlights potential future directions and questions to address these limitations and to further explore LLMs' potential in enhancing healthcare delivery.
This systematic review and meta‐analysis of randomised controlled trials (RCTs) aimed to evaluate the efficacy, safety, and tolerability of fluvoxamine for the outpatient management of COVID‐19. We ...conducted this review in accordance with the PRISMA 2020 guidelines. Literature searches were conducted in MEDLINE, EMBASE, International Pharmaceutical s, CINAHL, Web of Science, and CENTRAL up to 14 September 2023. Outcomes included incidence of hospitalisation, healthcare utilization (emergency room visits and/or hospitalisation), mortality, supplemental oxygen and mechanical ventilation requirements, serious adverse events (SAEs) and non‐adherence. Fluvoxamine 100 mg twice a day was associated with reductions in the risk of hospitalisation (risk ratio RR 0.75, 95% confidence interval CI 0.58–0.97; I
2 = 0%) and reductions in the risk of healthcare utilization (RR 0.68, 95% CI 0.53–0.86; I
2 = 0%). While no increased SAEs were observed, fluvoxamine 100 mg twice a day was associated with higher treatment non‐adherence compared to placebo (RR 1.61, 95% CI 1.22–2.14; I
2 = 53%). In subgroup analyses, fluvoxamine reduced healthcare utilization in outpatients with BMI ≥30 kg/m2, but not in those with lower BMIs. While fluvoxamine offers potential benefits in reducing healthcare utilization, its efficacy may be most pronounced in high‐risk patient populations. The observed non‐adherence rates highlight the need for better patient education and counselling. Future investigations should reassess trial endpoints to include outcomes relating to post‐COVID sequelaes. Registration: This review was prospectively registered on PROSPERO (CRD42023463829).
Background and Aims: People with diabetes on basal-bolus insulin regimens face challenges calculating bolus doses. Continuous glucose monitoring (CGM) systems can assist users with bolus dosing ...optimization through trend arrows. The purpose of this study was to demonstrate the safety and effectiveness of a CGM-informed insulin bolus calculator (IBC, Welldoc, Inc., Columbia, MD, U.S.A.) integrated into a mobile application that applies trend arrow and exercise adjustments to the bolus insulin dose recommendation. This investigational software also provided real-time coaching on CGM data to assist users in improving their time in range (TIR).
Methods: Fifty-four participants with diabetes using CGM (Dexcom G6, San Diego, CA, U.S.A.) were enrolled in a 30-day prospective study conducted at two research sites. Participants were asked to use the mobile application to monitor their CGM data and calculate their insulin doses. TIR during the prospective 30-day period was compared to that from 30 days of baseline data. Three populations were defined: 1) the intention-to-treat (ITT) group consisted of all enrolled subjects; 2) the complete cases (CC) group had ≥90% sensor wear time; and 3) the per-protocol (PP) group used the IBC at least 30 times during the study period.
Results: In the ITT group, the TIR improved from 68.4% to 71.8% (N=54, P=0.013). In the CC group, the TIR improved from 69.1% to 73.0% (N=49, P=0.005). In the PP group, the TIR improved from 69.2% to 73.0% (N=39, P=0.017). There was no increase in time with low glucose (<70 mg/dL) in all three populations. In a subgroup analysis, TIR increased from 74.7% to 81.4% in those with type 2 diabetes and from 64.4% to 65.2% in those with type 1 diabetes.
Conclusions: Use of a novel CGM-informed insulin bolus calculator with trend arrow and exercise dose adjustment by individuals with type 1 and type 2 diabetes was associated with significant improvement in TIR.
Disclosure
M.Shomali: Employee; WellDoc. C.Kelly: Consultant; WellDoc, Caladrius Biosciences, Inc., Epigenomics AG, Luna, Nevro Corp., Presidio Medical, Hitachi, Ltd., Resmed, Inc., Seaspine Hldgs, Inc., ViaCyte, Inc. A.K.Iyer: Employee; WellDoc. J.Y.Park: None. G.Aleppo: Advisory Panel; Medscape, Consultant; Bayer Inc., Insulet Corporation, Research Support; Dexcom, Inc., Eli Lilly and Company, Emmes, Insulet Corporation, Fractyl Health, Inc., WellDoc, Speaker's Bureau; Dexcom, Inc.
Funding
Welldoc Inc.
Sir2 is a NAD+-dependent protein deacetylase that extends lifespan in yeast and worms. This study examines seven human proteins homologous to Sir2 (SIRT1 through SIRT7) for cellular localization, ...expression profiles, protein deacetylation activity, and effects on human cell lifespan. We found that: 1) three nuclear SIRT proteins (SIRT1, SIRT6, and SIRT7) show different subnuclear localizations: SIRT6 and SIRT7 are associated with heterochromatic regions and nucleoli, respectively, where yeast Sir2 functions; 2) SIRT3, SIRT4, and SIRT5 are localized in mitochondria, an organelle that links aging and energy metabolism; 3) cellular p53 is a major in vivo substrate of SIRT1 deacetylase, but not the other six SIRT proteins; 4) SIRT1, but not the other two nuclear SIRT proteins, shows an in vitro deacetylase activity on histone H4 and p53 peptides; and 5) overexpression of any one of the seven SIRT proteins does not extend cellular replicative lifespan in normal human fibroblasts or prostate epithelial cells. This study supports the notion that multiple human SIRT proteins have evolutionarily conserved and nonconserved functions at different cellular locations and reveals that the lifespan of normal human cells, in contrast to that of lower eukaryotes, cannot be manipulated by increased expression of a single SIRT protein.
PDF
