OBJECTIVE To investigate the apoptosis-inducing effect of esculetin （Esc） on acute myeloid leukemia （AML） HL-60 cells by regulating the protein kinase B （AKT）/S-phase kinase-associated protein 2 （SKP2）/MutT homolog 1 （MTH1） pathway. METHODS AML HL-60 cells were randomly divided into control group （routine culture）， Esc low-concentration group （L-Esc group， 25 μmol/L Esc）， Esc medium-concentration group （M-Esc group， 50 μmol/L Esc）， Esc high-concentration group （H-Esc group， 100 μmol/L Esc）， and high-concentration of Esc+ SC79 （AKT agonist） group （100 μmol/L Esc+5 μmol/L SC79）. Cell proliferation in each group was detected by MTT assay and colony formation assay. The level of reactive oxygen species （ROS） in cells was measured by using the CM-H2DCFDA fluorescent probe. Cell apoptosis was analyzed by flow cytometry. Western blot assay was performed to detect the expression levels of apoptosis-related proteins [B-cell lymphoma 2 （Bcl-2）， Bcl-2-associated X protein （Bax）， cleaved caspase-3]， AKT/SKP2/MTH1 pathway-related proteins （p-AKT， AKT， SKP2， MTH1）， along with the upstream and downstream proteins of AKT phosphatidylinositol 3-kinase （PI3K）， cyclin-dependent kinase inhibitor 1 （P21） and cyclin-dependent kinase inhibitor 1B （P27）. RESULTS Compared with control group， the cell viability， colony number， and the phosphorylation levels of AKT and PI3K proteins as well as protein expressions of SKP2， MTH1 and Bcl-2 were significantly decreased （P＜0.05）， while ROS level， apoptosis rate， and the expression levels of Bax， cleaved caspase-3， P21 and P27 proteins were significantly increased （P＜0.05）. Moreover， the effects of Esc exhibited concentration-dependence （P＜0.05）. Compared with H-Esc group， above indexes of high-concentration of Esc+ SC79 group were reversed significantly （P＜0.05）. CONCLUSIONS Esc may promote massive ROS production and induce activation of apoptosis in HL-60 cells by inhibiting the AKT/SKP2/MTH1 pathway， thus inhibiting the proliferation of HL-60 cells.

With the widespread adoption of low-dose CT screening and the extensive application of high-resolution CT, the detection rate of sub-centimeter lung nodules has significantly increased. How to scientifically manage these nodules while avoiding overtreatment and diagnostic delays has become an important clinical issue. Among them, lung nodules with a consolidation tumor ratio less than 0.25, dominated by ground-glass shadows, are particularly worthy of attention. The therapeutic challenge for this group is how to achieve precise and complete resection of nodules during surgery while maximizing the preservation of the patient's lung function. The "watershed topography map" is a new technology based on big data and artificial intelligence algorithms. This method uses Dicom data from conventional dose CT scans, combined with microscopic (22-24 levels) capillary network anatomical watershed features, to generate high-precision simulated natural segmentation planes of lung sub-segments through specific textures and forms. This technology forms fluorescent watershed boundaries on the lung surface, which highly fit the actual lung anatomical structure. By analyzing the adjacent relationship between the nodule and the watershed boundary, real-time, visually accurate positioning of the nodule can be achieved. This innovative technology provides a new solution for the intraoperative positioning and resection of lung nodules. This consensus was led by four major domestic societies, jointly with expert teams in related fields, oriented to clinical practical needs, referring to domestic and foreign guidelines and consensus, and finally formed after multiple rounds of consultation, discussion, and voting. The main content covers the theoretical basis of the "watershed topography map" technology, indications, operation procedures, surgical planning details, and postoperative evaluation standards, aiming to provide scientific guidance and exploration directions for clinical peers who are currently or plan to carry out lung nodule resection using the fluorescent microscope watershed analysis method.

Inflammatory diseases (IDs) are a general term of diseases characterized by chronic inflammation as the primary pathogenetic mechanism, which seriously affect the quality of patient′s life and cause significant social and medical burden. Current drugs for IDs include nonsteroidal anti-inflammatory drugs, corticosteroids, immunomodulators, biologics, and antioxidants, but these drugs may cause gastrointestinal side effects, induce or worsen infections, and cause non-response or intolerance. Given the outstanding performance of metal polyphenol network (MPN) in the fields of drug delivery, biomedical imaging, and catalytic therapy, its application in the diagnosis and treatment of IDs has attracted much attention and significant progress has been made. In this paper, we first provide an overview of the types of IDs and their generating mechanisms, then sort out and summarize the different forms of MPN in recent years, and finally discuss in detail the characteristics of MPN and their latest research progress in the diagnosis and treatment of IDs. This research may provide useful references for scientific research and clinical practice in the related fields.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

ObjectiveTransfer RNA-derived fragments (tRFs) are a recently characterized and rapidly expanding class of small non-coding RNAs, typically ranging from 13 to 50 nucleotides in length. They are derived from mature or precursor tRNA molecules through specific cleavage events and have been implicated in a wide range of cellular processes. Increasing evidence indicates that tRFs play important regulatory roles in gene expression, primarily by interacting with target messenger RNAs (mRNAs) to induce transcript degradation, in a manner partially analogous to microRNAs (miRNAs). However, despite their emerging biological relevance and potential roles in disease mechanisms, there remains a significant lack of computational tools capable of systematically predicting the interaction landscape between tRFs and their target mRNAs. Existing databases often rely on limited interaction features and lack the flexibility to accommodate novel or user-defined tRF sequences. The primary goal of this study was to develop a machine learning based prediction algorithm that enables high-throughput, accurate identification of tRF:mRNA binding events, thereby facilitating the functional analysis of tRF regulatory networks. MethodsWe began by assembling a manually curated dataset of 38 687 experimentally verified tRF:mRNA interaction pairs and extracting seven biologically informed features for each pair: (1) AU content of the binding site, (2) site pairing status, (3) binding region location, (4) number of binding sites per mRNA, (5) length of the longest consecutive complementary stretch, (6) total binding region length, and (7) seed sequence complementarity. Using this dataset and feature set, we trained 4 distinct machine learning classifiers—logistic regression, random forest, decision tree, and a multilayer perceptron (MLP)—to compare their ability to discriminate true interactions from non-interactions. Each model’s performance was evaluated using overall accuracy, receiver operating characteristic (ROC) curves, and the corresponding area under the ROC curve (AUC). The MLP consistently achieved the highest AUC among the four, and was therefore selected as the backbone of our prediction framework, which we named tRF Prospect. For biological validation, we retrieved 3 high-throughput RNA-seq datasets from the gene expression omnibus (GEO) in which individual tRFs were overexpressed: AS-tDR-007333 (GSE184690), tRF-3004b (GSE197091), and tRF-20-S998LO9D (GSE208381). Differential expression analysis of each dataset identified genes downregulated upon tRF overexpression, which we designated as putative targets. We then compared the predictions generated by tRF Prospect against those from three established tools—tRFTar, tRForest, and tRFTarget—by quantifying the number of predicted targets for each tRF and assessing concordance with the experimentally derived gene sets. ResultsThe proposed algorithm achieved high predictive accuracy, with an AUC of 0.934. Functional validation was conducted using transcriptome-wide RNA-seq datasets from cells overexpressing specific tRFs, confirming the model’s ability to accurately predict biologically relevant downregulation of mRNA targets. When benchmarked against established tools such as tRFTar, tRForest, and tRFTarget, tRF Prospect consistently demonstrated superior performance, both in terms of predictive precision and sensitivity, as well as in identifying a higher number of true-positive interactions. Moreover, unlike static databases that are limited to precomputed results, tRF Prospect supports real-time prediction for any user-defined tRF sequence, enhancing its applicability in exploratory and hypothesis-driven research. ConclusionThis study introduces tRF Prospect as a powerful and flexible computational tool for investigating tRF:mRNA interactions. By leveraging the predictive strength of deep learning and incorporating a broad spectrum of interaction-relevant features, it addresses key limitations of existing platforms. Specifically, tRF Prospect: (1) expands the range of detectable tRF and target types; (2) improves prediction accuracy through multilayer perceptron model; and (3) allows for dynamic, user-driven analysis beyond database constraints. Although the current version emphasizes miRNA-like repression mechanisms and faces challenges in accurately capturing 5'UTR-associated binding events, it nonetheless provides a critical foundation for future studies aiming to unravel the complex roles of tRFs in gene regulation, cellular function, and disease pathogenesis.

BACKGROUND: This study aimed to investigate programmed death-ligand 1 tumor proportion score in predicting the safety and efficacy of PD-1/PD-L1 antibody-based therapy in treating patients with advanced non-small cell lung cancer (NSCLC) in a real-world setting. METHODS: This retrospective, multicenter, observational study enrolled adult patients who received PD-1/PD-L1 antibody-based therapy in China and met the following criteria: (1) had pathologically confirmed, unresectable stage III-IV NSCLC; (2) had a baseline PD-L1 tumor proportion score (TPS); and (3) had confirmed efficacy evaluation results after PD-1/PD-L1 treatment. Logistic regression, Kaplan-Meier analysis, and Cox regression were used to assess the progression-free survival (PFS), overall survival (OS), and immune-related adverse events (irAEs) as appropriate. RESULTS: A total of 409 patients, 65.0% ( n = 266) with a positive PD-L1 TPS (≥1%) and 32.8% ( n = 134) with PD-L1 TPS ≥50%, were included in this study. Cox regression confirmed that patients with a PD-L1 TPS ≥1% had significantly improved PFS (hazard ratio [HR] 0.747, 95% confidence interval [CI] 0.573-0.975, P = 0.032). A total of 160 (39.1%) patients experienced 206 irAEs, and 27 (6.6%) patients experienced 31 grade 3-5 irAEs. The organs most frequently associated with irAEs were the skin (52/409, 12.7%), thyroid (40/409, 9.8%), and lung (34/409, 8.3%). Multivariate logistic regression revealed that a PD-L1 TPS ≥1% (odds ratio [OR] 1.713, 95% CI 1.054-2.784, P = 0.030) was an independent risk factor for irAEs. Other risk factors for irAEs included pretreatment absolute lymphocyte count >2.5 × 10 9 /L (OR 3.772, 95% CI 1.377-10.329, P = 0.010) and pretreatment absolute eosinophil count >0.2 × 10 9 /L (OR 2.006, 95% CI 1.219-3.302, P = 0.006). Moreover, patients who developed irAEs demonstrated improved PFS (13.7 months vs. 8.4 months, P <0.001) and OS (28.0 months vs. 18.0 months, P = 0.007) compared with patients without irAEs. CONCLUSIONS A positive PD-L1 TPS (≥1%) was associated with improved PFS and an increased risk of irAEs in a real-world setting. The onset of irAEs was associated with improved PFS and OS in patients with advanced NSCLC receiving PD-1/PD-L1-based therapy.
Humans ; Carcinoma, Non-Small-Cell Lung/metabolism* ; Male ; Female ; Retrospective Studies ; Middle Aged ; Lung Neoplasms/metabolism* ; Aged ; B7-H1 Antigen/metabolism* ; Programmed Cell Death 1 Receptor/metabolism* ; Adult ; Aged, 80 and over ; Immune Checkpoint Inhibitors/therapeutic use*