The mental health management of laboratory animals is a critical factor in ensuring the reliability of scientific research data. However, due to the subtle nature of mental state alterations and the limitations of current assessment methods, the mental health of laboratory animals is often overlooked by researchers. Therefore, there is an urgent need to explore a new, objective, simple, and practical method for evaluating mental health. Stress, as a primary factor inducing alterations in the mental state of animals, can influence experimental results across multiple research fields through the neuroendocrine-immune network. This paper first elucidates the necessity of mental health management in laboratory animals from three perspectives: factors contributing to stress, neural mechanisms of stress, and research areas affected by stress. It highlights that excluding animals with abnormal mental states before experiments can enhance the efficiency and reproducibility of biomedical studies. Second, this paper briefly summarizes existing methods for assessing the health of laboratory animals, pointing out that approaches such as behavioral tests and metabolomics have many limitations in evaluating stress responses. Existing methods struggle to meet the demand for simple, objective, and non-invasive assessment methods in animal management. Therefore, this paper focuses on hair, a biological sample with the advantages of cumulative information and non-invasive collection, and systematically describes the theoretical basis and technological advances in using hair characteristics as indicators of stress responses from three perspectives: hair traits, the levels of substances in hair, and artificial intelligence (AI). It proposes and supports an innovative technical pathway that combines hair traits with AI-based image analysis, with the aim of offering an improved solution for non-invasive and objective stress assessment, while providing theoretical support for improving the health management system of laboratory animals.

ObjectiveTo study the effect of Bufei Tongbi decoction on pulmonary fibrosis in diabetic rats via the transforming growth factor-β₁ （TGF-β₁）/phosphorylated Smad family member 3 （p-Smad3） signaling pathway. MethodsStreptozotocin （60 mg·kg^-1） and bleomycin （24.80 U·kg^-1） were used to prepare the rat model of diabetes with pulmonary fibrosis by intratracheal injection. Sixty rats were randomly assigned into blank， model， low-， medium-， and high-dose （3.98， 7.95， and 15.90 g·kg^-1， respectively） Bufei Tongbi decoction， and pirfenidone （0.36 mg·kg^-1） groups （n=10）. The successfully modeled rats in each group were administrated with corresponding agents once per day for four consecutive weeks. After drug administration， fasting blood glucose and lung function indicators were measured. Chemical immunoassay was employed to determine the serum levels of hydroxyproline （Hyp）， hyaluronic acid （HA）， and laminin （LN）. The lung index was determined by the wet and dry methods. The pathological changes in the lung tissue were observed by hematoxylin-eosin （HE） staining， and the degree of fibrosis was detected by Masson staining. The mRNA and protein levels of TGF-β₁， p-Smad3， Smad3， α-smooth muscle actin （α-SMA）， collagen type Ⅰ alpha 1 （Col1A1）， and fibronectin were determined by PCR and Western blotting， respectively. ResultsCompared with the blank group， the model group showed alveolar septa thickening， obvious thickening of the basement membrane of pulmonary blood vessels， severe destruction of the alveolar structure， structural disarrangement of the lung parenchyma， and an increase in the proportion of inflammatory cell infiltration in the lung tissue， together with a large amount of blue collagen deposition and a large amount of collagen fibroplasia in the bronchial wall， vessel wall， interstitium， and alveolar wall， which indicated severe fibrosis. Bufei Tongbi decoction groups and the pirfenidone group showed lower fasting blood glucose level （P<0.05） and higher forced vital capacity （FVC）， cytoplasmic dynein （Cydn）， FEV_0.3/FEV ratio， and lung index （P<0.05） than the model group. Moreover， these groups demonstrated alleviated lung fibrosis， elevated Hyp， HA， and LN levels， down-regulated mRNA levels of α-SMA， Col1A1， and fibronectin， and down-regulated protein levels of TGF-β₁， Smad3， p-Smad3， α-SMA， Col1A1， and fibronectin （P<0.05）. ConclusionBufei Tongbi decoction can inhibit pulmonary fibrosis in diabetic rats by inhibiting the TGF-β₁/p-Smad3 signaling pathway.

Objective:To develop PixelSIFT-UNet, a novel semantic segmentation model that integrates deep learning with scale-invariant feature transform (SIFT) algorithm to improve the segmentation accuracy of oral mucosal lesions.Methods:This investigation utilized 838 standard clinical white light images of oral mucosal diseases acquired from January 2020 to December 2022 at the Stomatology Hospital Zhejiang University School of Medicine. Randomization was achieved through Python′s random.seed function implementation. The random sample function was subsequently applied for sampling distribution. The dataset was stratified into three subsets with a 6∶2∶2 ratio: training ( n=506), validation ( n=166), and testing ( n=166). Lesion boundaries were annotated using Labelme software, and a PixelSIFT-UNet-based deep learning model was developed with VGG-16 and ResNet-50 backbone networks. Model parameters were optimized using the validation set, and performance metrics [including Dice coefficient, mean intersection over union (mIoU), mean pixel accuracy (mPA), and Precision] were assessed on the test set. The model′s performance was benchmarked against conventional semantic segmentation frameworks (U-Net and PSPNet). Results:The developed PixelSIFT-UNet model could achieve precise segmentation of three common oral mucosal lesions: oral lichen planus, oral leukoplakia, and oral submucous fibrosis. Utilizing VGG-16 as the backbone network, the model achieved Dice coefficient, mIoU, mPA, and Precision values of 0.642, 0.699, 0.836, and 0.792, respectively. Implementation with ResNet-50 backbone network yielded metrics of 0.668, 0.733, 0.872 and 0.817, demonstrating significant improvements across all performance indicators compared to conventional U-Net model (relevant metrics: 0.662, 0.717, 0.861 and 0.809) and PSPNet model (relevant metrics: 0.671, 0.721, 0.858 and 0.813).Conclusions:The proposed PixelSIFT-UNet architecture demonstrates superior performance in oral mucosal lesion segmentation tasks, surpassing conventional semantic segmentation models and providing robust quantitative improvements in segmentation accuracy.

Objective:To develop PixelSIFT-UNet, a novel semantic segmentation model that integrates deep learning with scale-invariant feature transform (SIFT) algorithm to improve the segmentation accuracy of oral mucosal lesions.Methods:This investigation utilized 838 standard clinical white light images of oral mucosal diseases acquired from January 2020 to December 2022 at the Stomatology Hospital Zhejiang University School of Medicine. Randomization was achieved through Python′s random.seed function implementation. The random sample function was subsequently applied for sampling distribution. The dataset was stratified into three subsets with a 6∶2∶2 ratio: training ( n=506), validation ( n=166), and testing ( n=166). Lesion boundaries were annotated using Labelme software, and a PixelSIFT-UNet-based deep learning model was developed with VGG-16 and ResNet-50 backbone networks. Model parameters were optimized using the validation set, and performance metrics [including Dice coefficient, mean intersection over union (mIoU), mean pixel accuracy (mPA), and Precision] were assessed on the test set. The model′s performance was benchmarked against conventional semantic segmentation frameworks (U-Net and PSPNet). Results:The developed PixelSIFT-UNet model could achieve precise segmentation of three common oral mucosal lesions: oral lichen planus, oral leukoplakia, and oral submucous fibrosis. Utilizing VGG-16 as the backbone network, the model achieved Dice coefficient, mIoU, mPA, and Precision values of 0.642, 0.699, 0.836, and 0.792, respectively. Implementation with ResNet-50 backbone network yielded metrics of 0.668, 0.733, 0.872 and 0.817, demonstrating significant improvements across all performance indicators compared to conventional U-Net model (relevant metrics: 0.662, 0.717, 0.861 and 0.809) and PSPNet model (relevant metrics: 0.671, 0.721, 0.858 and 0.813).Conclusions:The proposed PixelSIFT-UNet architecture demonstrates superior performance in oral mucosal lesion segmentation tasks, surpassing conventional semantic segmentation models and providing robust quantitative improvements in segmentation accuracy.

Objective:To report the 5-year survival outcomes and recurrence patterns of robotic total gastrectomy (RTG) for locally advanced proximal gastric cancer in order to provide more valuable long-term follow-up results for clinical practice.Methods:This was a prospective, single-arm, open-label clinical trial (FUGES-014; Clinical-Trials.gov, NCT03524287). Patients with locally advanced proximal gastric cancer who underwent RTG at Fujian Medical University Union Hospital from March 5, 2018, to February 10, 2020, were included in the analysis. To evaluate the long-term efficacy of RTG in the most objective manner possible, we performed a propensity score-matched (1∶2) comparative analysis with historical control patients who had undergone laparoscopic total gastrectomy (LTG) from the FUGES-002 study (ClinicalTrials.gov, NCT02333721) in which the 5-year disease-free survival (DFS), 5-year overall survival (OS), and recurrence patterns were compared between the two groups.Results:Prior to matching, there were 48 cases in the RTG group and 263 cases in the LTG group; patients in the LTG group had more advanced cT and pT stages ( P=0.044 and 0.006, respectively) compared to the RTG group. After matching, there were 48 cases in the RTG group and 96 cases in the LTG group; however, no statistically significant differences were observed in the baseline clinical characteristics between the two groups (all P>0.05). Both groups had a median follow-up of 72 months. The 5-year DFS rates were 75.0% (95%CI: 63.7%- 88.3%) in the RTG group and 61.4% (95%CI: 52.5%-72.0%) in the LTG group ( P=0.116). Similarly, the 5-year OS rates were 79.2% (95%CI: 68.5%-91.5%) and 64.6% (95%CI: 55.7%-74.9%) in the RTG and LTG groups, respectively ( P=0.100). Within 5 years after surgery, tumor recurrence occurred in 10 patients (20.8%) in the RTG group and 33 patients (34.4%) in the LTG group ( P=0.124), and peritoneal recurrence was the predominant pattern in both groups (8.3%[4/48] vs. 10.4%[10/96]; risk difference: -0.02, P=0.554). Gastric cancer-related death was the predominant cause of death in both groups (16.7% [8/48] vs. 31.2% [30/96]; risk difference: -0.15, P=0.064). Among patients stratified by different pathological stages, no statistically significant differences were found in DFS, OS, or recurrence rates between the RTG and LTG groups (all P>0.05). Conclusions:We find the long-term oncological outcomes of RTG for locally advanced proximal gastric cancer to be noninferior to those of LTG. RTG should therefore be considered as a valid option for standardized minimally invasive surgery for locally advanced proximal gastric cancer.

Objective:This study aims to analyze the performance of five automatic cell type annotation methods in single cell RNA sequencing (scRNA-seq) data.Methods:Simulated data were generated using the Splatter package in R language, taking into account two data characteristics: the number of cells and the number of genes. The actual data came from the GSE10245 scRNA seq dataset of non-small cell lung cancer in Gene Expression Omnibus (GEO) database, the data had been pre-processed and batch effects had been eliminated. The automatic cell type recognition (ACTINN) of neural networks, the single-cell type annotation method based on deep learning (scDeepSort), the reference batch transcriptome annotation scRNA seq R-package (SingleR), the cross platform and cross species scRNA seq data classifier (SingleCellNet), and the cross scRNA seq dataset projection (scMap-cell) were implemented using the Tensorflow library in Python. The performance evaluation indicators for cell type annotation included accuracy (ACC), F1-score, and Matthews correlation coefficient (MCC). Each method was validated using ten fold cross validation, and the average value was taken after 50 repeated runs for performance comparison between methods. The Dunnett's t-test in the DescTools package of R language was used for multiple comparisons between ACTINN and other four methods. Results:Under 12 different scenarios (3 levels of cell numbers × 4 levels of gene numbers), simulated data analysis showed that compared with scDeepSort, SingleR, SingleCellNet, and scMap-cell, the percentage increase in ACC value of ACTINN ranged from 3.31% to 14.59%, 1.38% to 13.03%, 12.98% to 25.25%, and 20.72% to 29.62%, respectively; the range of F1 score improvement percentages were 2.75% - 22.74%, 2.46% - 23.68%, 5.07% - 27.47%, and 10.27% - 31.47%, respectively; the percentage increase ranges for MCC values were 3.42% - 9.75%, 2.26% - 7.61%, 5.41% - 11.11%, and 8.27% - 15.22%, respectively. Actual data analysis showed that the ACC value of ACTINN was 81.0%, which was increased by 2.1%, 5.2%, 7.9%, and 8.9% compared with the above four methods, respectively; the F1-score value was 80.5%, which was increased by 2.3%, 5.9%, 2.4%, and 6.0%, respectively; the MCC value was 83.3%, which was increased by 0.9%, 2.5%, 3.4%, and 11.2%, respectively. The results of Dunnett's t-test showed that the difference was not statistically significant in ACC values between scDeepSort and ACTINN ( P = 0.821), in F1-score values between scDeepSort and ACTINN ( P = 0.498), and in MCC values between scDeepSort, SingleCellNet and ACTINN ( P = 0.904, 0.134). However, the differences were statistically significant in other multiple comparisons ( P < 0.05). Conclusions:ACTINN and scDeepSort have good performance in cell type annotation, with ACTINN showing outstanding performance and SingleR showing robust performance, while SingleCellNet and scMap-cell have relatively limited performance. This suggests that self-attention mechanism algorithm based on Transformer framework is expected to promote further development of automatic cell annotation methods.

Objective:To report the 5-year survival outcomes and recurrence patterns of robotic total gastrectomy (RTG) for locally advanced proximal gastric cancer in order to provide more valuable long-term follow-up results for clinical practice.Methods:This was a prospective, single-arm, open-label clinical trial (FUGES-014; Clinical-Trials.gov, NCT03524287). Patients with locally advanced proximal gastric cancer who underwent RTG at Fujian Medical University Union Hospital from March 5, 2018, to February 10, 2020, were included in the analysis. To evaluate the long-term efficacy of RTG in the most objective manner possible, we performed a propensity score-matched (1∶2) comparative analysis with historical control patients who had undergone laparoscopic total gastrectomy (LTG) from the FUGES-002 study (ClinicalTrials.gov, NCT02333721) in which the 5-year disease-free survival (DFS), 5-year overall survival (OS), and recurrence patterns were compared between the two groups.Results:Prior to matching, there were 48 cases in the RTG group and 263 cases in the LTG group; patients in the LTG group had more advanced cT and pT stages ( P=0.044 and 0.006, respectively) compared to the RTG group. After matching, there were 48 cases in the RTG group and 96 cases in the LTG group; however, no statistically significant differences were observed in the baseline clinical characteristics between the two groups (all P>0.05). Both groups had a median follow-up of 72 months. The 5-year DFS rates were 75.0% (95%CI: 63.7%- 88.3%) in the RTG group and 61.4% (95%CI: 52.5%-72.0%) in the LTG group ( P=0.116). Similarly, the 5-year OS rates were 79.2% (95%CI: 68.5%-91.5%) and 64.6% (95%CI: 55.7%-74.9%) in the RTG and LTG groups, respectively ( P=0.100). Within 5 years after surgery, tumor recurrence occurred in 10 patients (20.8%) in the RTG group and 33 patients (34.4%) in the LTG group ( P=0.124), and peritoneal recurrence was the predominant pattern in both groups (8.3%[4/48] vs. 10.4%[10/96]; risk difference: -0.02, P=0.554). Gastric cancer-related death was the predominant cause of death in both groups (16.7% [8/48] vs. 31.2% [30/96]; risk difference: -0.15, P=0.064). Among patients stratified by different pathological stages, no statistically significant differences were found in DFS, OS, or recurrence rates between the RTG and LTG groups (all P>0.05). Conclusions:We find the long-term oncological outcomes of RTG for locally advanced proximal gastric cancer to be noninferior to those of LTG. RTG should therefore be considered as a valid option for standardized minimally invasive surgery for locally advanced proximal gastric cancer.

Objective:This study aims to analyze the performance of five automatic cell type annotation methods in single cell RNA sequencing (scRNA-seq) data.Methods:Simulated data were generated using the Splatter package in R language, taking into account two data characteristics: the number of cells and the number of genes. The actual data came from the GSE10245 scRNA seq dataset of non-small cell lung cancer in Gene Expression Omnibus (GEO) database, the data had been pre-processed and batch effects had been eliminated. The automatic cell type recognition (ACTINN) of neural networks, the single-cell type annotation method based on deep learning (scDeepSort), the reference batch transcriptome annotation scRNA seq R-package (SingleR), the cross platform and cross species scRNA seq data classifier (SingleCellNet), and the cross scRNA seq dataset projection (scMap-cell) were implemented using the Tensorflow library in Python. The performance evaluation indicators for cell type annotation included accuracy (ACC), F1-score, and Matthews correlation coefficient (MCC). Each method was validated using ten fold cross validation, and the average value was taken after 50 repeated runs for performance comparison between methods. The Dunnett's t-test in the DescTools package of R language was used for multiple comparisons between ACTINN and other four methods. Results:Under 12 different scenarios (3 levels of cell numbers × 4 levels of gene numbers), simulated data analysis showed that compared with scDeepSort, SingleR, SingleCellNet, and scMap-cell, the percentage increase in ACC value of ACTINN ranged from 3.31% to 14.59%, 1.38% to 13.03%, 12.98% to 25.25%, and 20.72% to 29.62%, respectively; the range of F1 score improvement percentages were 2.75% - 22.74%, 2.46% - 23.68%, 5.07% - 27.47%, and 10.27% - 31.47%, respectively; the percentage increase ranges for MCC values were 3.42% - 9.75%, 2.26% - 7.61%, 5.41% - 11.11%, and 8.27% - 15.22%, respectively. Actual data analysis showed that the ACC value of ACTINN was 81.0%, which was increased by 2.1%, 5.2%, 7.9%, and 8.9% compared with the above four methods, respectively; the F1-score value was 80.5%, which was increased by 2.3%, 5.9%, 2.4%, and 6.0%, respectively; the MCC value was 83.3%, which was increased by 0.9%, 2.5%, 3.4%, and 11.2%, respectively. The results of Dunnett's t-test showed that the difference was not statistically significant in ACC values between scDeepSort and ACTINN ( P = 0.821), in F1-score values between scDeepSort and ACTINN ( P = 0.498), and in MCC values between scDeepSort, SingleCellNet and ACTINN ( P = 0.904, 0.134). However, the differences were statistically significant in other multiple comparisons ( P < 0.05). Conclusions:ACTINN and scDeepSort have good performance in cell type annotation, with ACTINN showing outstanding performance and SingleR showing robust performance, while SingleCellNet and scMap-cell have relatively limited performance. This suggests that self-attention mechanism algorithm based on Transformer framework is expected to promote further development of automatic cell annotation methods.

Atherosclerosis （AS） is a chronic inflammatory pathological process in which lipid and/or fibrous substances are deposited in the intima of arteries， and it is one of the pathological bases of many cardiovascular and cerebrovascular diseases. Endoplasmic reticulum stress （ERS） is a protective mechanism of cell adaptation. Moderate ERS can reduce abnormal protein aggregation and increase the degradation of misfolded proteins to repair and stabilize the internal environment， while excessive ERS can cause unfolded protein reaction， activate inflammation， oxidative stress， apoptosis， autophagy， and other downstream pathways， and lead to cell damage， or even apoptosis. A large number of studies have shown that ERS mediates a variety of pathological processes related to AS， affects endothelial cells， smooth muscle cells， macrophages， endothelial progenitor cells， and other cell components closely related to its occurrence and development， influences the progress of AS by regulating cell function， and promotes the formation of AS plaque， the transformation of stable plaque to unstable plaque， and the rupture of unstable plaque. Regulation of ERS may be a key target for the prevention and treatment of AS， and it is a research hotspot at present. Traditional Chinese medicine （TCM） believes that the origin of AS is the imbalance of Yin and Yang， the disharmony of Zangfu organs， and the abnormal operation of Qi， blood， and body fluid， which leads to the accumulation of phlegm， blood stasis， and other pathological products in the pulse channels， making the blood flow blocked or misfunction and causing the disease， which belongs to the syndrome of deficiency in origin and excess in superficiality. As the pathogenesis of AS is complex， and the symptoms are diverse， TCM has significant advantages in treating AS because of its multiple targets， multiple pathways， stable efficacy， strong individualization， and high safety. This paper systematically elaborated on the role of ERS in the occurrence and development of AS and summarized the mechanism research on the regulation and control of ERS by Chinese herbal monomer， Chinese herbal extract， Chinese herbal compound， and proprietary medicine， so as to provide a theoretical basis for clinical research and drug development in the prevention and treatment of AS.

Objective:To investigate the risk factors and prognostic value of the textbook outcome (TO) in patients with advanced gastric cancer (AGC) who underwent neoadjuvant chemotherapy followed by surgical resection.Methods:This is a retrospective cohort study. A total of 253 patients with AGC who underwent neoadjuvant chemotherapy combined with gastrectomy and D2 lymphadenectomy in the Department of Gastric Surgery, Fujian Medical University Union Hospital from January 2010 to December 2019 were retrospectively included. There were 195 males and 58 females, aged (60.3±10.0) years (range: 27 to 75 years). The patients were then divided into the TO group ( n=168) and the non-TO group ( n=85). Multivariate Logistic regression was used to analyze the independent predictors of TO. Univariate and multivariate Cox analysis were used to analyze independent prognosis factors for overall survival (OS) and disease-free survival (DFS). Propensity score matching was performed to balance the TO and non-TO groups, and the Kaplan-Meier method was used to calculate survival rates and draw survival curves. Results:Among the 253 patients, 168 patients (66.4%) achieved TO. The Eastern Cooperative Oncology Group score ( OR=0.488, 95% CI: 0.278 to 0.856, P=0.012) and ypN stage ( OR=0.626, 95% CI:0.488 to 0.805, P<0.01) were independently predictive of TO. Multivariate analysis revealed that TO was an independent risk factor for both OS ( HR=0.662, 95% CI: 0.457 to 0.959, P=0.029) and DFS ( HR=0.687, 95% CI: 0.483 to 0.976, P=0.036). After matching, the 5-year OS rate (42.2% vs. 27.8%) and the 5-year DFS rate (37.5% vs. 27.8%) were significantly higher in the TO group than in the non-TO group (both P<0.05). Furthermore, patients in the non-TO group benefited significantly from postoperative chemotherapy (both P<0.05), but those in the TO group did not (both P>0.05). Conclusion:TO is an independent prognosis factor in patients undergoing neoadjuvant chemotherapy and surgery for AGC and is associated with postoperative chemotherapy benefits.