Interactions between myoblasts and macrophages under high glucose milieus result in inflammatory response and impaired insulin sensitivity.

BACKGROUND: Skeletal muscle handles about 80% of insulin-stimulated glucose uptake and become the major organ occurring insulin resistance (IR). Many studies have confirmed the interactions between macrophages and skeletal muscle regulated the inflammation and regeneration of skeletal muscle. However, despite of the decades of research, whether macrophages infiltration and polarization in skeletal muscle under high glucose (HG) milieus results in the development of IR is yet to be elucidated. C2C12 myoblasts are well-established and excellent model to study myogenic regulation and its responses to stimulation. Further exploration of macrophages' role in myoblasts IR and the dynamics of their infiltration and polarization is warranted. AIM: To evaluate interactions between myoblasts and macrophages under HG, and its effects on inflammation and IR in skeletal muscle. METHODS: We detected the polarization status of macrophages infiltrated to skeletal muscles of IR mice by hematoxylin and eosin and immunohistochemical staining. Then, we developed an in vitro co-culture system to study the interactions between myoblasts and macrophages under HG milieus. The effects of myoblasts on macrophages were explored through morphological observation, CCK-8 assay, Flow Cytometry, and enzyme-linked immunosorbent assay. The mediation of macrophages to myogenesis and insulin sensitivity were detected by morphological observation, CCK-8 assay, Immunofluorescence, and 2-NBDG assay. RESULTS: The F4/80 and co-localization of F4/80 and CD86 increased, and the myofiber size decreased in IR group (P < 0.01, g = 6.26). Compared to Mc group, F4/80+CD86+CD206- cells, tumor necrosis factor-α (TNFα), inerleukin-1ß (IL-1ß) and IL-6 decreased, and IL-10 increased in McM group (P < 0.01, g > 0.8). In McM + HG group, F4/80+CD86+CD206- cells, monocyte chemoattractant protein 1, TNFα, IL-1ß and IL-6 were increased, and F4/80+CD206+CD86- cells and IL-10 were decreased compared with Mc + HG group and McM group (P < 0.01, g > 0.8). Compered to M group, myotube area, myotube number and E-MHC were increased in MMc group (P < 0.01, g > 0.8). In MMc + HG group, myotube area, myotube number, E-MHC, GLUT4 and glucose uptake were decreased compared with M + HG group and MMc group (P < 0.01, g > 0.8). CONCLUSION: Interactions between myoblasts and macrophages under HG milieus results in inflammation and IR, which support that the macrophage may serve as a promising therapeutic target for skeletal muscle atrophy and IR.

Melatonin Improves Oxidative Stress Injury in Retinopathy of Prematurity by Targeting miR-23a-3p/Nrf2.

PURPOSE: Melatonin has promising protective effects for retinopathy. However, its roles in retinopathy of prematurity (ROP) and the underlying mechanisms remain unknown. We aimed to explore its roles and mechanisms in a ROP model. METHODS: Hematoxylin and eosin staining were used to observe the morphology of the retina. Immunofluorescence was used to detect positive (Nrf2+ and VEGF+) cells. Immunohistochemistry was used to detect the level of nuclear expression of PCNA in retinal tissue. Transmission electron microscope (TEM) was used to observe the morphology and structure of pigment cells. qRT-PCR was used to assay the expression of miR-23a-3p, Nrf2, and HO-1. Western blotting was used to detect the expression of Nrf2, HO-1, ß-actin, and Lamin B1. RESULTS: Melatonin or miR-23a-3p antagomir treatment could ameliorate the Oxygen-induced pathological changes, increased the expression of Nrf2 and HO-1, SOD, and GSH-Px, and decreased the expression of VEGF, miR-23a-3p, MDA and the apoptosis in the ROP model. Further target prediction and luciferase reporter assays confirmed the targeted binding relationship between miR-23a-3p and Nrf2. CONCLUSION: Our study showed that melatonin could ameliorate H2O2-induced apoptosis and oxidative stress injury in RGC cells by mediating miR-23a-3p/Nrf2 signaling pathway, thereby improving retinal degeneration.

Biomaterials-assisted cancer vaccine delivery: preclinical landscape, challenges, and opportunities.

INTRODUCTION: Cancer vaccines (protein and peptide, DNA, mRNA, and tumor cell) have achieved remarkable success in the treatment of cancer. In particular, advances in the design and manufacture of biomaterials have made it possible to control the presentation and delivery of vaccine components to immune cells. AREAS COVERED: This review summarizes findings from major databases, including PubMed, Scopus, and Web of Science, focusing on articles published between 2005 and 2024 that discuss biomaterials in cancer vaccine delivery. EXPERT OPINION: The development of cancer vaccines is hindered by several bottlenecks, including low immunogenicity, instability of vaccine components, and challenges in evaluating their clinical efficacy. To transform preclinical successes into viable treatments, it is essential to pursue continued innovation, collaborative research, and address issues related to scalability, regulatory pathways, and clinical validation, ultimately improving outcomes against cancer.

High-Performance Method and Architecture for Attention Computation in DNN Inference.

In recent years, The combination of Attention mechanism and deep learning has a wide range of applications in the field of medical imaging. However, due to its complex computational processes, existing hardware architectures have high resource consumption or low accuracy, and deploying them efficiently to DNN accelerators is a challenge. This paper proposes an online-programmable Attention hardware architecture based on compute-in-memory (CIM) marco, which reduces the complexity of Attention in hardware and improves integration density, energy efficiency, and calculation accuracy. First, the Attention computation process is decomposed into multiple cascaded combinatorial matrix operations to reduce the complexity of its implementation on the hardware side; second, in order to reduce the influence of the non-ideal characteristics of the hardware, an online-programmable CIM architecture is designed to improve calculation accuracy by dynamically adjusting the weights; and lastly, it is verified that the proposed Attention hardware architecture can be applied for the inference of deep neural networks through Spice simulation. Based on the 100nm CMOS process, compared with the traditional Attention hardware architectures, the integrated density and energy efficiency are increased by at least 91.38 times, and latency and computing efficiency are improved by about 12.5 times.

Synthesis of novel 4-substituted isatin Schiff base derivatives as potential autophagy inducers and evaluation of their antitumour activity.

Induction of autophagic death in cancer cells is one of the promising strategies for the development of anti-cancer therapeutics. In the present study, we designed and synthesized a series of isatin Schiff base derivatives containing thioether structures. After discovering the highly active target compound H13 (IC50 = 4.83 µM) based on in vitro antiproliferation, we also found it had a high safety against normal cells HEK293 with CC50 of 69.01 µM, indicating a sufficient therapeutic window. In addition, to provide reference for subsequent studies, a model was successfully constructed by Sybyl software. Preliminary mechanistic studies suggested that H13-induced apoptosis may be closely related to ROS accumulation and mitochondrial dysfunction. Subsequent studies revealed that H13 inhibited cell proliferation by inducing cellular autophagy mainly through blocking signal of the PI3K/AKT/mTOR pathway. Altogether, these results suggested that H13 was potentially valuable as a lead compound.

Memristive patch attention neural network for facial expression recognition and edge computing.

Facial expression recognition has made a significant progress as a result of the advent of more and more convolutional neural networks (CNN). However, with the improvement of CNN, the models continues to get deeper and larger so as to a greater focus on the high-level features of the image and the low-level features tend to be lost. Because of the reason above, the dependence of low-level features between different areas of the face often cannot be summarized. In response to this problem, we propose a novel network based on the CNN model. To extract long-range dependencies of low-level features, multiple attention mechanisms has been introduced into the network. In this paper, the patch attention mechanism is designed to obtain the dependence between low-level features of facial expressions firstly. After fusion, the feature maps are input to the backbone network incorporating convolutional block attention module (CBAM) to enhance the feature extraction ability and improve the accuracy of facial expression recognition, and achieve competitive results on three datasets CK+ (98.10%), JAFFE (95.12%) and FER2013 (73.50%). Further, according to the PA Net designed in this paper, a hardware friendly implementation scheme is designed based on memristor crossbars, which is expected to provide a software and hardware co-design scheme for edge computing of personal and wearable electronic products.

Effectiveness of inter-regional collaborative emission reduction for ozone mitigation under local-dominated and transport-affected synoptic patterns.

In recent years, the concentrations of ozone and the pollution days with ozone as the primary pollutant have been increasing year by year. The sources of regional ozone mainly depend on local photochemical formation and transboundary transport. The latter is influenced by different weather circulations. How to effectively reduce the inter-regional emission to control ozone pollution under different atmospheric circulation is rarely reported. In this study, we classify the atmospheric circulation of ozone pollution days from 2014 to 2019 over Central China based on the Lamb-Jenkinson method and the global analysis data of the fifth-generation European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis (ERA5) operation. The effectiveness of emission control to alleviate ozone pollution under different atmospheric circulation is simulated by the WRF-Chem model. Among the 26 types of circulation patterns, 9 types of pollution days account for 79.5% of the total pollution days and further classified into 5 types. The local types (A and C type) are characterized by low surface wind speed and stable weather conditions over Central China due to a high-pressure system or a southwest vortex low-pressure system, blocking the diffusion of pollutants. Sensitivity simulations of A-type show that this heavy pollution process is mainly contributed by local emission sources. Removing the anthropogenic emission of pollutants over Central China would reduce the ozone concentration by 39.1%. The other three circulation patterns show pollution of transport characteristics affected by easterly, northerly, or southerly winds (N-EC, EC, S-EC-type). Under the EC-type, removing anthropogenic pollutants of East China would reduce the ozone concentration by 22.7% in Central China.

Substrate binding and inhibition mechanism of norepinephrine transporter.

Norepinephrine transporter (NET; encoded by SLC6A2) reuptakes the majority of the released noradrenaline back to the presynaptic terminals, thereby affecting the synaptic noradrenaline level1. Genetic mutations and dysregulation of NET are associated with a spectrum of neurological conditions in humans, making NET an important therapeutic target1. However, the structure and mechanism of NET remain unclear. Here we provide cryogenic electron microscopy structures of the human NET (hNET) in three functional states-the apo state, and in states bound to the substrate meta-iodobenzylguanidine (MIBG) or the orthosteric inhibitor radafaxine. These structures were captured in an inward-facing conformation, with a tightly sealed extracellular gate and an open intracellular gate. The substrate MIBG binds at the centre of hNET. Radafaxine also occupies the substrate-binding site and might block the structural transition of hNET for inhibition. These structures provide insights into the mechanism of substrate recognition and orthosteric inhibition of hNET.

Disruption to de novo uridine biosynthesis alters ß-1,3-glucan masking in Candida albicans.

The uridine derivatives UDP-glucose and UDP-N-acetylglucosamine are important for cell wall construction as they are the precursors for the synthesis of ß-1,3-glucan and chitin, respectively. Previous studies have demonstrated attenuated virulence of uridine auxotrophs in mice, which has been attributed to insufficient uridine levels for growth in the host. We have discovered that uridine deprivation in the uridine auxotroph ura3ΔΔ disrupts cell wall architecture by increasing surface mannans, exposing ß-1,3-glucan and chitin, and decreasing UDP-sugar levels. Cell wall architecture and UDP-sugars can be rescued with uridine supplementation. The cell wall architectural disruptions in the ura3ΔΔ mutant also impact immune activation since the mutant elicited greater TNFα secretion from RAW264.7 macrophages than wild type. To determine if cell wall defects contributed to decreased virulence in the ura3ΔΔ mutant, we used a murine model of systemic infection. Mice infected with the ura3ΔΔ mutant exhibited increased survival and reduced kidney fungal burden compared with mice infected with wild type. However, suppression of the immune response with cyclophosphamide did not rescue virulence in mice infected with the ura3ΔΔ mutant, indicating the attenuation in virulence of uridine auxotrophs can be attributed to decreased growth in the host but not increased exposure of ß-1,3-glucan. Moreover, the ura3ΔΔ mutant is unable to grow on ex vivo kidney agar, which demonstrates its inability to colonize the kidneys due to poor growth. Thus, although uridine auxotrophy elicits changes to cell wall architecture that increase the exposure of immunogenic polymers, metabolic fitness costs more strongly drive the observed virulence attenuation.IMPORTANCECandida albicans is a common cause of bloodstream infections (candidemia). Treatment of these bloodstream infections is made difficult because of increasing antifungal resistance and drug toxicity. Thus, new tactics are needed for antifungal drug development, with immunotherapy being of particular interest. The cell wall of C. albicans is composed of highly immunogenic polymers, particularly ß-1,3-glucan. However, ß-1,3-glucan is naturally masked by an outer layer of mannoproteins, which hampers the detection of the fungus by the host immune system. Alteration in cell wall components has been shown to increase ß-1,3-glucan exposure; however, it is unknown how the inability to synthesize precursors to cell wall components affects unmasking. Here, we demonstrate how cell wall architecture is altered in response to a deficit in precursors for cell wall synthesis and how uridine is a crucial component of these precursors.

Association between social capital, mental health, and digital health literacy among the university students in China: a multigroup analysis based on major difference.

BACKGROUND: This study aimed to clarify medical-nonmedical difference on the relationship between social capital, mental health and digital health literacy of university students in China, and furtherly provide evidence-based suggestions on the improvement of the digital health literacy for the university students. METHODS: The snowball sampling method was used to collect data from the university students (including medical students and nonmedical students) through online questionnaires, and finally 1472 university students were included for the data analysis, of whom, 665 (45.18%) were medical students, 807 (54.82%) were nonmedical students; 462 (31.39%) were male, 1010 (68.61%) were female. Mean value of the age was 21.34 ± 2.33 for medical students vs. 20.96 ± 2.16 for nonmedical students. Descriptive analysis, chi-square test analysis, one-way Analysis of Variance (conducted by SPSS) and structural equation modeling (conducted by AMOS) were employed to explore the difference on the relationship between social capital, mental health and digital health literacy between the medical students and nonmedical students. RESULTS: The mean value of the digital health literacy was 36.27 (37.33 for medical students vs. 35.39 for nonmedical students). The SEM analysis showed that there was a statistically positive correlation between social capital and digital health literacy (stronger among the nonmedical students (0.317) than medical students (0.184)). Mental health had a statistically positive impact on the digital health literacy among medical students (0.242), but statistically significant correlation was not observed in nonmedical students (0.017). Social capital was negatively correlated with the mental health for both medical students and NMS (stronger among the nonmedical students (0.366) than medical students (0.255)). And the fitness indices of SEM were same between medical students and nonmedical students (GFI = 0.911, AGFI = 0.859, CFI = 0.922, RMSEA = 0.074). CONCLUSION: The digital health literacy of the university student was relatively high. Both social capital and mental health could exert a positive effect on digital health literacy, while social capital was found to be positively associated with mental health. Statistical difference was found between medical students and nonmedical students on the above correlations. Implications were given on the improvement of the digital health literacy among university students in China.

Dynamic histone modification signatures coordinate developmental programs in strawberry fruit ripening.

Chromatin structure plays a critical role in the regulation of dynamic gene expression in response to different developmental and environmental cues, but as yet their involvement in fruit ripening is not well understood. Here, we profile seven histone modifications in the woodland strawberry (Fragaria vesca) genome and analyze the histone modification signatures during ripening. Collectively, segments painted by the seven marks cover ~85% of the woodland strawberry genome. We report an eight-state chromatin structure model of the woodland strawberry based on the above histone marks, which reveals a diverse chromatin environment closely associated with transcriptional apparatus. Upon this model we build a chromatin-centric annotation to the strawberry genome. Expression of many genes essential for fruit ripening, such as abscisic acid catabolism, anthocyanin accumulation and fruit softening, are associated with shifts of active genic states and polycomb-associated chromatin states. Particularly, the expression levels of ripening-related genes are well correlated with histone acetylation, indicating a regulatory role of histone acetylation in strawberry ripening. Our identification of the chromatin states underpinning genome expression during fruit ripening not only elucidates the coordination of different pathways of morphological and metabolic development but also provides a framework to understand the signals that regulate fruit ripening.

Focal adhesion-mediated directional cell migration guided by gradient-stretched substrate.

Soft tissues experience strain under mechanical stresses, storing energy as residual stresses and strain energy. However, the specific impact of such strain on cell migration and its molecular mechanisms remains unclear. In this study, we investigated this by using polydimethylsiloxane (PDMS) membranes with varying prestrain levels but constant stiffness to mimic tissue-like conditions. Results showed that higher prestrain levels enhanced 3T3 fibroblast adhesion and reduced filopodia formation. Elevated prestrain also increased integrin and vinculin expression, which was associated with lower cell migration rates. Notably, both 3T3 fibroblasts and primary rat airway smooth muscle cells migrated faster toward higher prestrain areas on substrates with strain gradients. Knockdown of integrin or vinculin inhibited 3T3 cell migration directionality, highlighting their critical role. This research reveals a mechanobiological pathway where strain gradients direct cell migration, providing insight into a common mechanotransduction pathway influencing cellular responses to both stiffness and strain-related mechanical cues.

Unveiling potential sex-determining genes and sex-specific markers in autotetraploid Carassius auratus.

Autotetraploid Carassius auratus is a stable hereditary autotetraploid fish resulting from the hybridization of Carassius auratus red var. (RCC, ♀) × Megalobrama amblycephala (BSB, ♂), containing four sets of RCC chromosomes. However, the molecular mechanism underlying the determination of sex in this species remains largely unknown. Currently, there lacks a full understanding of the molecular mechanisms governing sex determination and specific molecular markers to differentiate sex in this species. In this study, 25,801,677 SNPs (Single-nucleotide polymorphism) and 6,210,306 Indels (insertion-deletion) were obtained from whole-genome resequencing of 100 individuals (including 50 female and 50 male). Further identification confirmed the candidate chromosomes as Chr46B, with the sex-determining region located at Chr46B: 22,500,000-22,800,000 bp. Based on the male-specific insertion (26 bp) within the candidate sex-determining region, a pair of sex-specific molecular markers has been identified. In addition, based on the screening of candidate sex-determining region genes and RT-qPCR validation analysis, ADAM10, AQP9 and tc1a were identified as candidate sex-determining genes. These findings provide a robust foundation for investigating sex determination mechanisms in fish, the evolution of sex chromosomes, and the development of monosex populations.

Bridging beauty and biodiversity: Coupling diversity and aesthetics through optimized plant communities in urban riverfront landscapes.

Urban waterfront areas are dynamic interfaces where human and natural systems converge, forming complex ecosystems that encompass social, economic, and environmental elements. These areas offer ecological benefits and aesthetic experiences. However, a disparity between social aesthetic preferences and vegetation diversity along riverbanks impedes the integration of ecological and aesthetic values. To address this, a plant community optimization strategy based on a coupling coordination degree model (CCDM) is proposed. Using the Xietang River in Suzhou, China as a case study, surveys were conducted on 33 woody plant plots and 60 herbaceous plant plots, assessing plant diversity with Shannon-Wiener, richness, and Pielou indices. Landscape beauty was evaluated by 87 respondents using the Scenic Beauty Estimation method. Using six representative plant communities as mediators, CCDM was applied to quantitatively analyze the coordination between plant diversity and aesthetics. Based on this analysis and considering factors influencing plant diversity and scenic beauty, plant community optimization strategies were devised to enhance the coordinated development of ecological diversity and aesthetics, fostering a synergistic improvement in ecological and aesthetic quality. Results revealed a range of coupling coordination across plant communities (0.203 to 0.947), encompassing various types. Linear regression analysis demonstrated a non-linear relationship between plant diversity and landscape beauty, influenced by independent yet partially overlapping factors. Hence, both aspects should be simultaneously considered in the planning and enhancement of riverbank areas. The coupling coordination degree offers a comprehensive understanding of harmonizing plant diversity and aesthetic value, providing a quantitative and objective approach to integrated research. This perspective extends beyond urban waterfront landscapes, holding significance for achieving dual goals of ecology and social services in urban design and landscape management.

Remimazolam has similar anesthetic effect and superior safety compared to propofol in elderly patients: A meta-analysis of randomized controlled trials.

BACKGROUND: The superiority between remimazolam and propofol for anesthesia is controversial in elderly patients (≥60 years). This meta-analysis aimed to systematically compare anesthetic effect and safety profile between remimazolam and propofol in elderly patients under any surgery. METHODS: Cochrane Library, Web of Science, and PubMed were searched until December 25, 2023 for relevant randomized controlled trials. RESULTS: Ten studies with 806 patients receiving remimazolam (experimental group) and 813 patients receiving propofol (control group) were included. Time to loss of consciousness [standard mean difference (SMD) (95% confidence interval (CI): 1.347 (-0.362, 3.055), p = 0.122] and recovery time [SMD (95% CI): -0.022 (-0.300, 0.257), p = 0.879] were similar between experimental and control groups. Mean arterial pressure at baseline minus 1 min after induction [SMD (95% CI): -1.800 (-3.250, -0.349), p = 0.015], heart rate at baseline minus 1 min after induction [SMD (95% CI): -1.041 (-1.537, -0.545), p < 0.001], incidences of hypoxemia [relative risk (RR) (95% CI): 0.247 (0.138, 0.444), p < 0.001], respiratory depression [RR (95% CI): 0.458 (0.300, 0.700), p < 0.001], bradycardia [RR (95% CI): 0.409 (0.176, 0.954), p = 0.043], hypotension [RR (95% CI): 0.415 (0.241, 0.714), p = 0.007], and injection pain [RR (95% CI): 0.172 (0.113, 0.263), p < 0.001] were lower in the experimental group compared to the control group. Postoperative nausea and vomiting was not different between groups [RR (95% CI): 1.194 (0.829, 1.718), p = 0.341]. Moreover, this meta-analysis displayed a low risk of bias, minimal publication bias, and good robustness. CONCLUSION: Remimazolam shows comparative anesthetic effect and better safety profile than propofol in elderly patients under any surgery.

The role of lactate-induced protein lactylation in gliomas: implications for preclinical research and the development of new treatments.

The most prevalent primary brain tumors in adults are gliomas. In addition to insufficient therapeutic alternatives, gliomas are fatal mostly due to the rapid proliferation and continuous infiltration of tumor cells into the surrounding healthy brain tissue. According to a growing body of research, aerobic glycolysis, or the Warburg effect, promotes glioma development because gliomas are heterogeneous cancers that undergo metabolic reprogramming. Therefore, addressing the Warburg effect might be a useful therapeutic strategy for treating cancer. Lactate plays a critical role in reprogramming energy metabolism, allowing cells to rapidly access large amounts of energy. Lactate, a byproduct of glycolysis, is therefore present in rapidly proliferating cells and tumors. In addition to the protumorigenesis pathways of lactate synthesis, circulation, and consumption, lactate-induced lactylation has been identified in recent investigations. Lactate plays crucial roles in modulating immune processes, maintaining homeostasis, and promoting metabolic reprogramming in tumors, which are processes regulated by the lactate-induced lactylation of the lysine residues of histones. In this paper, we discuss the discovery and effects of lactylation, review the published studies on how protein lactylation influences cancer growth and further explore novel treatment approaches to achieve improved antitumor effects by targeting lactylation. These findings could lead to a new approach and guidance for improving the prognosis of patients with gliomas.

Associations of amyloid-ß oligomers and plaques with neuropathology in the App NL-G-F mouse.

Amyloid-ß pathology and neurofibrillary tangles lead to glial activation and neurodegeneration in Alzheimer's disease. In this study, we investigated the relationships between the levels of amyloid-ß oligomers, amyloid-ß plaques, glial activation and markers related to neurodegeneration in the App NL-G-F triple mutation mouse line and in a knock-in line homozygous for the common human amyloid precursor protein (App hu mouse). The relationships between neuropathological features were characterized with immunohistochemistry and imaging mass cytometry. Markers assessing human amyloid-ß proteins, microglial and astrocytic activation and neuronal and synaptic densities were used in mice between 2.5 and 12 months of age. We found that amyloid-ß oligomers were abundant in the brains of App hu mice in the absence of classical amyloid-ß plaques. These brains showed morphological changes consistent with astrocyte activation but no evidence of microglial activation or synaptic or neuronal pathology. In contrast, both high levels of amyloid-ß oligomers and numerous plaques accumulated in App NL-G-F mice in association with substantial astrocytic and microglial activation. The increase in amyloid-ß oligomers over time was more strongly correlated with astrocytic than with microglia activation. Spatial analyses suggested that activated microglia were more closely associated with amyloid-ß oligomers than with amyloid-ß plaques in App NL-G-F mice, which also showed age-dependent decreases in neuronal and synaptic density markers. A comparative study of the two models highlighted the dependence of glial and neuronal pathology on the nature and aggregation state of the amyloid-ß peptide. Astrocyte activation and neuronal pathology appeared to be more strongly associated with amyloid-ß oligomers than with amyloid-ß plaques, although amyloid-ß plaques were associated with microglia activation.

Predictors of post-healing recurrence in patients with diabetic foot ulcers: A systematic review and meta-analysis.

BACKGROUND: Diabetic foot ulcer is one of the most prevalent, serious, and costly consequences of diabetes, often associated with peripheral neuropathy and peripheral arterial disease. These ulcers contribute to high disability and mortality rates in patients and pose a major challenge to clinical management. OBJECTIVE: To systematically review the risk prediction models for post-healing recurrence in diabetic foot ulcer (DFU) patients, so as to provide a reference for clinical staff to choose appropriate prediction models. METHODS: The authors searched five databases (Cochrane Library, PubMed, Web of Science, EMBASE, and Chinese Biomedical Database) from their inception to September 23, 2023, for relevant literature. After data extraction, the quality of the literature was evaluated using the Predictive Model Research Bias Risk and Suitability Assessment tool (PROBAST). Meta-analysis was performed using STATA 17.0 software. RESULTS: A total of 9 studies involving 5956 patients were included. The recurrence rate after DFU healing ranged from 6.2 % to 41.4 %. Nine studies established 15 risk prediction models, and the area under the curve (AUC) ranged from 0.660 to 0.940, of which 12 models had an AUC≥0.7, indicating good prediction performance. The combined AUC value of the 9 validation models was 0.83 (95 % confidence interval: 0.79-0.88). Hosmer-Lemeshow test was performed for 10 models, external validation for 5 models, and internal validation for 6 models. Meta-analysis showed that 14 predictors, such as age and living alone, could predict post-healing recurrence in DFU patients (p < 0.05). CONCLUSION: To enhance the quality of these risk prediction models, there is potential for future improvements in terms of follow-up duration, model calibration, and validation processes.

Machine Learning Approach in Dosage Individualization of Isoniazid for Tuberculosis.

INTRODUCTION: Isoniazid is a first-line antituberculosis agent with high variability, which would profit from individualized dosing. Concentrations of isoniazid at 2 h (C2h), as an indicator of safety and efficacy, are important for optimizing therapy. OBJECTIVE: The objective of this study was to establish machine learning (ML) models to predict the C2h, that can be used for establishing an individualized dosing regimen in clinical practice. METHODS: Published population pharmacokinetic (PopPK) models for adults were searched based on PubMed and ultimately four reliable models were selected for simulating individual C2h datasets under different conditions (demographics, genotype, ethnicity, etc.). Machine learning models were trained on simulated C2h obtained from the four PopPK models. Five different algorithms were used for ML model building to predict C2h. Real-world data were used for predictive performance evaluations. Virtual trials were used to compare ML-optimized doses with PopPK model-optimized doses. RESULTS: Categorical boosting (CatBoost) exhibited the highest prediction ability. Target C2h can be predicted using the ML model combined with the dosing regimen and three covariates (N-acetyltransferase 2 [NAT2] genotypes, weight and race [Asians and Africans]). Real-world data validation results showed that the ML model can achieve an overall prediction accuracy of 93.4%. Using the final ML model, the mean absolute prediction error value decreased by 45.7% relative to the average of PopPK models. Using the ML-optimized dosing regimen, the probability of target attainment increased by 43.7% relative to the PopPK model-optimized dosing regimens. CONCLUSION: Machine learning models were developed with great predictive performance, which can be used to determine the individualized initial dose of isoniazid in adult patients.