Geometric molecular graph representation learning model for drug-drug interactions prediction.

Drug-drug interaction (DDI) can trigger many adverse effects in patients and has emerged as a threat to medicine and public health. Therefore, it is important to predict potential drug interactions since it can provide combination strategies of drugs for systematic and effective treatment. Existing deep learning-based methods often rely on DDI functional networks, or use them as an important part of the model information source. However, it is difficult to discover the interactions of a new drug. To address the above limitations, we propose a geometric molecular graph representation learning model (Mol-DDI) for DDI prediction based on the basic assumption that structure determines function. Mol-DDI only considers the covalent and non-covalent bond information of molecules, then it uses the pre-training idea of large-scale models to learn drug molecular representations and predict drug interactions during the fine-tuning process. Experimental results show that the Mol-DDI model outperforms others on the three datasets and performs better in predicting new drug interaction experiments.

An Objective Injury Threshold for the Maximum Principal Strain Criterion for Brain Tissue in the Finite Element Head Model and Its Application.

Although the finite element head model (FEHM) has been widely utilized to analyze injury locations and patterns in traumatic brain injury, significant controversy persists regarding the selection of a mechanical injury variable and its corresponding threshold. This paper aims to determine an objective injury threshold for maximum principal strain (MPS) through a novel data-driven method, and to validate and apply it. We extract the peak responses from all elements across 100 head impact simulations to form a dataset, and then determine the objective injury threshold by analyzing the relationship between the combined injury degree and the threshold according to the stationary value principle. Using an occipital impact case from a clinical report as an example, we evaluate the accuracy of the injury prediction based on the new threshold. The results show that the injury area predicted by finite element analysis closely matches the main injury area observed in CT images, without the issue of over- or underestimating the injury due to an unreasonable threshold. Furthermore, by applying this threshold to the finite element analysis of designed occipital impacts, we observe, for the first time, supra-tentorium cerebelli injury, which is related to visual memory impairment. This discovery may indicate the biomechanical mechanism of visual memory impairment after occipital impacts reported in clinical cases.

Are cities under bright lights more innovative? Evidence from China.

The prosperity of the nighttime economy is an important criterion for measuring urban development. Based on regression analysis of data from 330 prefecture-level cities in China, this study explores the potential link between urban lighting and urban innovation capacity. The results show that the relationship between nighttime light intensity and urban innovation capacity follows an inverted U-shaped curve. Moreover, urban nighttime light stability moderates the above relationship. In addition, the heterogeneity analysis indicates that low population density, weak economic foundation, and high administrative grade weaken the relationship between lighting intensity and urban innovation capacity. Finally, the mechanism analysis suggested that the inverted U-shaped relationship between urban lighting and innovation capacity may be caused by the encroachment of the service industry and the loss of urban attractiveness. By exploring the proxy effect of nighttime lighting on urban innovation, this study reveals the imbalance between extrinsic image and the intrinsic capacity of cities, which not only extends the research on urban construction and development but also provides guidance for local governments concerning how to achieve all-round coordination of cities.

Oral administration of herbal oligonucleotide drug JGL-sRNA-h7 ameliorates hyperglycemia in db/db mice and beagle dogs.

Type 2 diabetes mellitus is a prevalent metabolic disease, posing a considerable threat to public health. Oligonucleotide drugs have proven to be a promising field of therapy for the diseases. In this study, we reported that a herbal small RNA (sRNA), JGL-sRNA-h7 (B34735529, F1439.L002444.A11), could exhibit potent hypoglycemic effects by targeting glucose-6-phosphatase. Oral administration of sphingosine (d18:1)-JGL-sRNA-h7 bencaosomes ameliorated hyperglycemia and diabetic kidney injury better than metformin in db/db mice. Furthermore, glucose tolerance was also improved in sphingosine (d18:1)-JGL-sRNA-h7 bencaosomes-treated beagle dogs. Our study indicates that JGL-sRNA-h7 could be a promising hypoglycemic oligonucleotide drug.

Deep graph contrastive learning model for drug-drug interaction prediction.

Drug-drug interaction (DDI) is the combined effects of multiple drugs taken together, which can either enhance or reduce each other's efficacy. Thus, drug interaction analysis plays an important role in improving treatment effectiveness and patient safety. It has become a new challenge to use computational methods to accelerate drug interaction time and reduce its cost-effectiveness. The existing methods often do not fully explore the relationship between the structural information and the functional information of drug molecules, resulting in low prediction accuracy for drug interactions, poor generalization, and other issues. In this paper, we propose a novel method, which is a deep graph contrastive learning model for drug-drug interaction prediction (DeepGCL for brevity). DeepGCL incorporates a contrastive learning component to enhance the consistency of information between different views (molecular structure and interaction network), which means that the DeepGCL model predicts drug interactions by integrating molecular structure features and interaction network topology features. Experimental results show that DeepGCL achieves better performance than other methods in all datasets. Moreover, we conducted many experiments to analyze the necessity of each component of the model and the robustness of the model, which also showed promising results. The source code of DeepGCL is freely available at https://github.com/jzysj/DeepGCL.

Quantitative assessment of brain injury and concussion induced by an unintentional soccer ball impact.

BACKGROUND: Accidental impact on a player's head by a powerful soccer ball may lead to brain injuries and concussions during games. It is crucial to assess these injuries promptly and accurately on the field. However, it is challenging for referees, coaches, and even players themselves to accurately recognize potential injuries and concussions following such impacts. Therefore, it is necessary to establish a list of minimum ball velocity thresholds that can result in concussions at different impact locations on the head. Additionally, it is important to identify the affected brain regions responsible for impairments in brain function and potential clinical symptoms. METHODS: By using a full human finite element model, dynamic responses and brain injuries caused by unintentional soccer ball impacts on six distinct head locations (forehead, tempus, crown, occiput, face, and jaw) at varying ball velocities (10, 15, 20, 25, 30, 35, 40, and 60 m/s) were simulated and investigated. Intracranial pressure, Von-Mises stress, and first principal strain were analyzed, the ball velocity thresholds resulting in concussions at different impact locations were evaluated, and the damage evolution patterns in the brain tissue were analyzed. RESULTS: The impact on the occiput is most susceptible to induce brain injuries compared to all other impact locations. For a conservative assessment, the risk of concussion is present once the soccer ball reaches 17.2 m/s in a frontal impact, 16.6 m/s in a parietal impact, 14.0 m/s in an occipital impact, 17.8 m/s in a temporal impact, 18.5 m/s in a facial impact or 19.2 m/s in a mandibular impact. The brain exhibits the most significant dynamic responses during the initial 10-20 ms, and the damaged regions are primarily concentrated in the medial temporal lobe and the corpus callosum, potentially causing impairments in brain functions. CONCLUSIONS: This work offers a framework for quantitatively assessing brain injuries and concussions induced by an unintentional soccer ball impact. Determining the ball velocity thresholds at various impact locations provides a benchmark for evaluating the risks of concussion. The examination of brain tissue damage evolution introduces a novel approach to linking biomechanical responses with possible clinical symptoms.

Efficacy and safety of seven Chinese patent medicines combined with conventional triple/quadruple therapy for Helicobacter pylori-positive peptic ulcers: a systematic review and network meta-analysis.

OBJECTIVES: To compare the efficacy and safety of seven Chinese patent medicines (CPMs) combined with conventional triple/quadruple therapy (T/Q) for Helicobacter pylori-positive peptic ulcers. DESIGN: A systematic review and network meta-analysis. DATA SOURCES: China National Knowledge Infrastructure, VIP database, Wanfang database, ScienceDirect, EBSCO, EMBASE, Web of Science, Cochrane Library and PubMed were searched through 1 June 2022. ELIGIBILITY CRITERIA: Randomised controlled trials (RCTs) testing CPMs combined with T/Q for H. pylori-positive peptic ulcers were included. The CPMs included Anweiyang capsule, Jianweiyuyang tablets/capsule/granule, Jinghuaweikang capsule, Kangfuxin liquid, Puyuanhewei capsule, Weifuchun tablets/capsule and Weisu granule. At least one of the following outcome indicators was recorded: complete ulcer healing rate (CUHR), effective rate (ER), H. pylori eradication rate (HPER), rate of peptic ulcer recurrence (RPUR) and incidence of adverse reactions (IAR). DATA EXTRACTION AND SYNTHESIS: Two researchers independently conducted the study selection and extracted data for included studies. The risk of bias was assessed using the Cochrane risk of bias tool. A pairwise meta-analysis was performed using RevMan V.5.3. Network meta-analysis was performed using STATA/MP V.15.0. Confidence in the evidence was assessed using Grading of Recommendations, Assessment, Development and Evaluation. RESULTS: A total of 36 RCTs involving 3620 patients were included. Compared with T/Q alone, Weisu+T/Q, Weifuchun+T/Q and Puyuanhewei+T/Q had the highest CUHR, ER and HPER, respectively. Weisu+T/Q and Jianweiyuyang+T/Q had the lowest RPUR and IAR, respectively. The cluster analysis results showed Jianweiyuyang+T/Q might be the best choice concerning efficacy and safety simultaneously, followed by Kangfuxin+T/Q. CONCLUSION: Among the combination therapies with the CPMs, Jianweiyuyang+T/Q might be the most favourable option for H. pylori-positive peptic ulcers, followed by Kangfuxin+T/Q. Considering the limited quantity and quality of the included RCTs, the results should be interpreted with caution. PROSPERO REGISTRATION NUMBER: CRD42022327687.

Diagnosis of poorly differentiated adenocarcinoma of the stomach by confocal laser endomicroscopy: A case report.

BACKGROUND: In recent years, confocal laser endomicroscopy (CLE) has become a new endoscopic imaging technology at the microscopic level, which is extensively performed for real-time in vivo histological examination. CLE can be performed to distinguish benign from malignant lesions. In this study, we diagnosed using CLE an asymptomatic patient with poorly differentiated gastric adenocarcinoma. CASE SUMMARY: A 63-year-old woman was diagnosed with gastric mucosal lesions, which may be gastric cancer, in the small curvature of the stomach by gastroscopy. She consented to undergo CLE for morphological observation of the gastric mucosa. Through the combination of CLE diagnosis and postoperative pathology, the intraoperative CLE diagnosis was considered to be reliable. According to our experience, CLE can be performed as the first choice for the diagnosis of gastric cancer. CONCLUSION: CLE has several advantages over pathological diagnosis. We believe that CLE has great potential in the diagnosis of benign and malignant gastric lesions.

Influence of centroid acceleration acquisition and filtering class on head injury criterion evaluation.

BACKGROUND: Although the Head Injury Criteria (HIC) has been widely applied to assess head impact injuries, it faces two outstanding problems: 1) HIC is affected strongly by the cut-off frequency when processing acceleration signals. And these cut-off frequencies are experiential and lack unified guidelines; 2) If the head was impacted on a different part, should the corresponding HIC threshold be the same? If these problems are not resolved, it could potentially lead to a critical misinterpretation of the safety assessment. METHODS: Finite element method was used to reconstruct head impacts. The head model includes tissues like skull, brainstem, cerebrospinal fluid, etc. The head model was impacted in the frontal, occipital, parietal or lateral direction with different impact velocities. Acceleration signals of the head model were extracted directly from the skull and the head centroid node. To obtain a robust HIC, the filtering class of acceleration signals were analyzed carefully. Then, the relation between rigid body HIC and the centroid node HIC were studied systematically. RESULTS: When the filtering class of rigid body acceleration and centroid node acceleration reached the cut-off frequency, the corresponding derivative of HIC tended to change smoothly. Using these cut-off frequencies, robust HICs were obtained. The rigid body HIC far exceeded that of centroid node HIC, such as 8, 9, 14 and 31 times exceeded in the frontal, occipital, parietal and lateral impact conditions, respectively. Moreover, approximate linear relations were found between the rigid body HIC and the centroid node HIC in different impact directions, respectively. From these relations, the injury thresholds of rigid body HIC of various directions were given quantitatively. CONCLUSIONS: The rational filtering class like CFC 800 and CFC 700 were given for rigid body HIC and centroid node HIC, respectively. The rigid body HIC had a significant discrepancy from the centroid node HIC. Linear relations between the rigid body HIC and centroid node HIC were found, and their slopes changed with impact directions. From these relations, we can adjust the injury thresholds reasonably if the head receives different impacts. These findings can effectively enhance the applicability of HIC.

Uncovering lupus nephritis-specific genes and the potential of TNFRSF17-targeted immunotherapy: a high-throughput sequencing study.

Introduction: Lupus nephritis (LN) is a severe manifestation of systemic lupus erythematosus (SLE). This study aimed to identify LN specific-genes and potential therapeutic targets. Methods: We performed high-throughput transcriptome sequencing on peripheral blood mononuclear cells (PBMCs) from LN patients. Healthy individuals and SLE patients without LN were used as controls. To validate the sequencing results, qRT-PCR was performed for 5 upregulated and 5 downregulated genes. Furthermore, the effect of the TNFRSF17-targeting drug IBI379 on patient plasma cells and B cells was evaluated by flow cytometry. Results: Our analysis identified 1493 and 205 differential genes in the LN group compared to the control and SLE without LN groups respectively, with 70 genes common to both sets, marking them as LN-specific. These LN-specific genes were significantly enriched in the 'regulation of biological quality' GO term and the cell cycle pathway. Notably, several genes including TNFRSF17 were significantly overexpressed in the kidneys of both LN patients and NZB/W mice. TNFRSF17 levels correlated positively with urinary protein levels, and negatively with complement C3 and C4 levels in LN patients. The TNFRSF17-targeting drug IBI379 effectively induced apoptosis in patient plasma cells without significantly affecting B cells. Discussion: Our findings suggest that TNFRSF17 could serve as a potential therapeutic target for LN. Moreover, IBI379 is presented as a promising treatment option for LN.

Mouse auditory cortex sub-fields receive neuronal projections from MGB subdivisions independently.

Mouse auditory cortex is composed of six sub-fields: primary auditory field (AI), secondary auditory field (AII), anterior auditory field (AAF), insular auditory field (IAF), ultrasonic field (UF) and dorsoposterior field (DP). Previous studies have examined thalamo-cortical connections in the mice auditory system and learned that AI, AAF, and IAF receive inputs from the ventral division of the medial geniculate body (MGB). However, the functional and thalamo-cortical connections between nonprimary auditory cortex (AII, UF, and DP) is unclear. In this study, we examined the locations of neurons projecting to these three cortical sub-fields in the MGB, and addressed the question whether these cortical sub-fields receive inputs from different subsets of MGB neurons or common. To examine the distributions of projecting neurons in the MGB, retrograde tracers were injected into the AII, UF, DP, after identifying these areas by the method of Optical Imaging. Our results indicated that neuron cells which in ventral part of dorsal MGB (MGd) and that of ventral MGB (MGv) projecting to UF and AII with less overlap. And DP only received neuron projecting from MGd. Interestingly, these three cortical areas received input from distinct part of MGd and MGv in an independent manner. Based on our foundings these three auditory cortical sub-fields in mice may independently process auditory information.

Sodium butyrate restricts neutrophils migration and NETs formation through reducing macrophage-derived CXCL16 in calculous cholecystitis.

Background: Neutrophil extracellular traps (NETs) havebeen demonstrated to initiate gallstone formation. Cholecystitis is a common complication of gallstones. As short-chain fatty acids (SCFAs), Butyrate acid has anti-inflammatory effects and alleviates cholesterol gallstones. However, the role of Butyrate acid in NETs of calculous cholecystitis and the molecular mechanism remains unclear. The effect of Sodium butyrate on neutrophil migration and NETs formation involved in macrophages polarization and exosomalCXCL16 in calculous cholecystitis was explored in our study. Methods: The number of neutrophils and NETs, macrophages polarization and exosomal CXCL16 level were analyzed in clinic samples from patients. Exosomes were obtained and verified by gradient centrifugation, transmission electron microscopy, NanoSight analysis and Western blotting. Transwell, immunofluorescence and ELISA were used to detect neutrophil migration and NETs formation. Results: Our results demonstrated that a large number of neutrophils and NETs, as well as M1 macrophages and exosomal CXCL16, were found in the blood of gallstones patients, especially patients with acute calculous cholecystitis. Exosomal CXCL16 was upregulated in plasma of calculous cholecystitis patients or Lipopolysaccharide induced macrophages, and promoted neutrophil cell migration and NETs formation. Sodium butyrate reduced exosomal CXCL16 secretion through the inhibition of M1 macrophage polarization to suppress neutrophils migration and NETs formation. Conclusion: Our study suggested that Sodium butyrate may inhibit neutrophils migration and NETs formation to alleviate calculous cholecystitis by reducing exosomal CXCL16 secretion from macrophage and macrophage polarization. General significance: Our finding may provide a link between exosomes and neutrophils to serve as a potential therapeutic intervention in calculous cholecystitis.

Running Gait and Control of Quadruped Robot Based on SLIP Model.

Legged robots have shown great adaptability to various environments. However, conventional walking gaits are insufficient to meet the motion requirements of robots. Therefore, achieving high-speed running for legged robots has become a significant research topic. In this paper, based on the Spring-Loaded Inverted Pendulum (SLIP) model and the optimized Double leg-Spring-Loaded Inverted Pendulum (D-SLIP) model, the running control strategies for the double flying phase Bound gait and the Rotatory gallop gait of quadruped robots are designed. First, the dynamics of the double flying phase Bound gait and Rotatory gallop gait are analyzed. Then, based on the "three-way" control idea of the SLIP model, the running control strategy for the double flying phase Bound gait is designed. Subsequently, the SLIP model is optimized to derive the D-SLIP model with two touchdown legs, and its dynamic characteristics are analyzed. And the D-SLIP model is applied to the running control strategy of the Rotatory gallop gait. Furthermore, joint simulation verification is conducted using Adams virtual prototyping and MATLAB/Simulink control systems for the designed control strategies. Finally, experimental verification is performed for the double flying phase Bound gait running control strategy. The experimental results demonstrate that the quadruped robot can achieve high-speed and stable running.

Multifaceted Characterization of Human Embryonic Stem Cell-Derived Mesenchymal Stem/Stromal Cells Revealed Amelioration of Acute Liver Injury in NOD-SCID Mice.

Human embryonic stem cells (hESCs) are advantaged sources for large-scale and homogeneous mesenchymal stem/stromal cells (MSCs) generation. However, due to the limitations in high-efficiency procedures for hESC-MSCs induction, the systematic and detailed information of mesengenesis and early MSC development are largely obscure. In this study, we took advantage of the well-established twist-related protein 1 (TWIST1)-overexpressing hESCs and two small molecular cocktails (CHIR99021, decitabine) for high-efficient MSC induction. To assess the multidimensional biological and transcriptomic characteristics, we turned to cellular and molecular methods, such as flow cytometry (FCM), quantitative reverse transcription-polymerase chain reaction (qRT-PCR), in vitro tri-lineage differentiation, cytokine secretion analysis, in vivo transplantation for acute liver injury (ALI) management, and bioinformatics analyses (eg, gene ontology-biological processes [GO-BP], Kyoto Encyclopedia of Genes and Genomes [KEGG], HeatMap, and principal component analysis [PCA]). By combining TWIST1 overexpression (denoted as T) and the indicated small molecular cocktails (denoted as S), hESCs high-efficiently differentiated into MSCs (denoted as TS-MSCs, induced by T and S combination) within 2 weeks. TS-MSCs satisfied the criteria for MSC definition and revealed comparable tri-lineage differentiation potential and ameliorative efficacy upon ALI mice. According to RNA-sequencing (SEQ) analysis, we originally illuminated the gradual variations in gene expression pattern and the concomitant biofunctions of the programmed hESC-MSCs. Overall, our data indicated the feasibility of high-efficient generation of hESC-MSCs by TWIST1 and cocktail-based programming. The generated hESC-MSCs revealed multifaceted in vivo and in vitro biofunctions as adult BM-MSCs, which collectively suggested promising prospects in ALI management in future.

A comparative study of the predictive value of four models for death in patients with severe burns.

OBJECTIVE: To assess the prognostic value of the Ryan score, Belgian Outcome of Burn Injury (BOBI) score,revised Baux (rBaux) score, and a new model (a Logit(P)-based scoring method created in 2020) for predicting mortality risk in patients with extremely severe burns and to conduct a comparative analysis. METHODS: A retrospective analysis was conducted on 599 burn patients who met the inclusion criteria and were admitted to the burn unit of the First Affiliated Hospital of Nanchang University from 2017 to 2022. Relevant information was collected, and receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were plotted for each of the four models in assessing mortality in these burn patients using both age-stratified and unstratified forms. The ROC curve section was further compared with the area under the curve (AUC), optimal cutoff value, as well as its sensitivity and specificity. Additionally, the quality of the AUC was assessed using the Delong test. RESULT: Among the patients who met the inclusion criteria, 532 were in the survival group and 67 in the death group. Irrespective of age stratification, the novel model exhibited superior performance with an AUC of 0.868 (95% CI: 0.838-0.894) among all four models predicting mortality risk in included patients, and also demonstrated better AUC quality than other models; the calibration curves showed that the accuracy of all four models was good; the DCA curves showed that the clinical utility of the novel model and rBuax score were better. In the comparison of four scoring models across different age groups, the new model demonstrated the largest AUC in both 0-19 years (0.954, 95% CI 0.914-0.979) and 20-59 years groups (0.838, 95% CI 0.793-0.877), while rBuax score exhibited the highest AUC in ≥ 60 years group (0.708, 95% CI of 0.602-0.800). The calibration curves showed that the four models exhibited greater accuracy within the age range of 20-59 years, while the DCA curves indicated that both the novel model and rBuax score scale displayed better prediction in both the 20-59 and ≥ 60 years groups. CONCLUSIONS: All four models demonstrate accurate and effective prognostication for patients with severe burns. Both the novel model and rBaux score exhibit enhanced prediction utility. In terms of the model itself alone, the new model is not simpler than, for example, the rBaux score, and whether it can be applied clinicallyinvolves further study.

Cause of death during nasopharyngeal carcinoma survivorship: a population-based analysis.

Background: Recently, the survival rate of nasopharyngeal carcinoma (NPC) patients has improved greatly due to developments in NPC treatments. But cause-specific mortality in NPC patients remains unclear. This study aims to investigate the common causes of death in NPC patients. Methods: Eligible patients with NPC were included from the Surveillance, Epidemiology, and End Results (SEER) database. Standardized mortality ratios(SMRs) were calculated to compare death rates in NPC patients with those in the general population. Results: A total of 3475 patients with NPC were included, of whom 1696 patients died during the follow-up period. 52.83% of deaths were caused by NPC, followed by other cancers (28.13%) and non-cancer causes (18.46%). The proportion of patients who died of NPC decreased over survival time. Moreover, non-cancer causes of death increase from 12.94% to 51.22% over time after 10 years of diagnosis. Heart diseases was the most common non-cancer cause of death in NPC patients. Conclusions: Although NPC remains the leading cause of death after NPC diagnosis, other non-NPC causes of death represent an increased number of death in NPC patients. These findings support the involvement of multidisciplinary care for follow-up strategy in NPC patients.

Interactome profiling of Crimean-Congo hemorrhagic fever virus glycoproteins.

Crimean-Congo hemorrhagic fever virus (CCHFV) is a biosafety level-4 pathogen requiring urgent research and development efforts. The glycoproteins of CCHFV, Gn and Gc, are considered to play multiple roles in the viral life cycle by interactions with host cells; however, these interactions remain largely unclear to date. Here, we analyzed the cellular interactomes of CCHFV glycoproteins and identified 45 host proteins as high-confidence Gn/Gc interactors. These host molecules are involved in multiple cellular biological processes potentially associated with the physiological actions of the viral glycoproteins. Then, we elucidated the role of a representative cellular protein, HAX1. HAX1 interacts with Gn by its C-terminus, while its N-terminal region leads to mitochondrial localization. By the strong interaction, HAX1 sequestrates Gn to mitochondria, thus depriving Gn of its normal Golgi localization that is required for functional glycoprotein-mediated progeny virion packaging. Consistently, the inhibitory activity of HAX1 against viral packaging and hence propagation was further elucidated in the contexts of pseudotyped and authentic CCHFV infections in cellular and animal models. Together, the findings provide a systematic CCHFV Gn/Gc-cell protein-protein interaction map, but also unravel a HAX1/mitochondrion-associated host antiviral mechanism, which may facilitate further studies on CCHFV biology and therapeutic approaches.

Proliferative glomerulonephritis with monoclonal IgG Lambda deposits caused by plasmablastic lymphoma: a case report.

INTRODUCTION: As a very rare form of B-cell lymphoma, plasmablastic lymphoma (PBL) typically occurs in patients with underlying immunosuppression, including human immunodeficiency virus (HIV), organ transplantation, and autoimmune diseases. For HIV-positive patients, PBL normally originates in the gastrointestinal tract, especially from the oral cavity in most cases. It is extremely rare to find abdominal cavity involvement in PBL, and there has been no previously reported instance of proliferative glomerulonephritis with monoclonal immunoglobulin deposits (PGNMID) attributed to monoclonal IgG (MIgG) lambda secreted by PBL. CASE PRESENTATION: We report the case of an HIV-negative female with nephrotic syndrome, renal insufficiency, and multiple swollen lymph nodes. Ascitic fluid cytology revealed a high level of plasmablast-like lymphocytes with the restriction of lambda light chains. Besides, the renal biopsy revealed PGNMID, which could presumably be secondary to MIgG-lambda-secreting by PBL. MIgG-lambda-restricted expression was discovered earlier in the kidney tissue than in the blood. CONCLUSION: The diagnostic landscape for PBL is notoriously intricate, necessitating a multifaceted and nuanced approach to mitigate the risks of erroneous identification.

Exploring the challenge of early gastric cancer diagnostic AI system face in multiple centers and its potential solutions.

BACKGROUND: Artificial intelligence (AI) performed variously among test sets with different diversity due to sample selection bias, which can be stumbling block for AI applications. We previously tested AI named ENDOANGEL, diagnosing early gastric cancer (EGC) on single-center videos in man-machine competition. We aimed to re-test ENDOANGEL on multi-center videos to explore challenges applying AI in multiple centers, then upgrade ENDOANGEL and explore solutions to the challenge. METHODS: ENDOANGEL was re-tested on multi-center videos retrospectively collected from 12 institutions and compared with performance in previously reported single-center videos. We then upgraded ENDOANGEL to ENDOANGEL-2022 with more training samples and novel algorithms and conducted competition between ENDOANGEL-2022 and endoscopists. ENDOANGEL-2022 was then tested on single-center videos and compared with performance in multi-center videos; the two AI systems were also compared with each other and endoscopists. RESULTS: Forty-six EGCs and 54 non-cancers were included in multi-center video cohort. On diagnosing EGCs, compared with single-center videos, ENDOANGEL showed stable sensitivity (97.83% vs. 100.00%) while sharply decreased specificity (61.11% vs. 82.54%); ENDOANGEL-2022 showed similar tendency while achieving significantly higher specificity (79.63%, p < 0.01) making fewer mistakes on typical lesions than ENDOANGEL. On detecting gastric neoplasms, both AI showed stable sensitivity while sharply decreased specificity. Nevertheless, both AI outperformed endoscopists in the two competitions. CONCLUSIONS: Great increase of false positives is a prominent challenge for applying EGC diagnostic AI in multiple centers due to high heterogeneity of negative cases. Optimizing AI by adding samples and using novel algorithms is promising to overcome this challenge.