Objective To identify immune cell-related biomarkers in lung adenocarcinoma (LUAD) using weighted gene co-expression network analysis (WGCNA). Methods Based on data from The Cancer Genome Atlas (TCGA) database, a gene co-expression network was constructed for the TCGA-LUAD dataset using the "WGCNA" R package, and genes were clustered into different modules. Concurrently, the Estimation of STromal and Immune cells in MAlignant Tumours using Expression data (ESTIMATE) algorithm was applied to the tumor samples in the TCGA-LUAD dataset. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed to evaluate the biological functions of genes within the most significantly correlated module. Candidate hub genes from the key module were intersected with a protein-protein interaction (PPI) network to identify the final hub genes. The prognostic performance of these hub genes and their correlation with immune cell infiltration were validated using Kaplan-Meier curves and the Tumor IMmune Estimation Resource (TIMER) algorithm. Finally, a multivariate Cox regression analysis was conducted on the identified hub genes to construct a prognostic risk model. Results In the co-expression network, the brown module was found to be highly correlated with the ImmuneScore, StromalScore, and ESTIMATE Score. Five immune-related hub genes were identified: CD53, PLEK, SPI1, IL10RA, and C3AR1. Enrichment analysis of the brown module revealed that its genes were primarily enriched in GO terms such as "regulation of innate immune response" and KEGG pathways like the "NF-kappa B signaling pathway". Furthermore, the expression levels of these five hub genes were significantly and positively correlated with the infiltration abundance of various immune cells. The immune relevance of the model was validated by the Immunophenoscore (IPS) and the Tumor Immune Dysfunction and Exclusion (TIDE) score. Moreover, the established RiskScore demonstrated significant potential in predicting the response to immunotherapy. Conclusion These five immune-related key genes may serve as novel and effective potential therapeutic targets for LUAD immunotherapy, facilitating the development of personalized diagnosis and treatment strategies for patients with LUAD.

With the continuous deepening of the concept of precision diagnosis and treatment for lung cancer, how to achieve higher efficiency and accuracy in the screening, diagnosis, and treatment pathways in clinical practice has become an important issue that urgently needs to be overcome. The current clinical difficulty lies in the fact that despite continuous advancements in imaging and molecular diagnostic technologies, there are still limitations in manual efficiency and subjective experience when it comes to massive data analysis and multi-scale feature extraction. Artificial intelligence (AI), especially algorithm systems based on deep learning, is an innovative technology capable of deeply empowering medical big data. This method utilizes algorithms such as convolutional neural networks, combined with radiomics, pathomics, and multi-modal data fusion analysis, demonstrating immense potential in early precise detection and benign-malignant differentiation of pulmonary nodules, digital pathological subtype recognition and non-invasive prediction of driver genes, precise 3D surgical planning and automatic delineation of radiotherapy target volumes, as well as dynamic risk warning during follow-up. This innovative technology provides a brand-new solution for realizing intelligent and individualized lung cancer diagnosis and treatment models. This consensus, based on the latest evidence from evidence-based medicine and combined with the development trends in the AI field and real-world clinical needs, was ultimately formed by gathering the consensus opinions of multidisciplinary experts in radiology, pathology, thoracic surgery, and other fields. The main content covers the application specifications of AI in the three core scenarios of lung cancer screening, diagnosis, and treatment, the technical standards for data collection and algorithm validation, as well as the ethical and regulatory challenges faced at the current stage. It aims to clarify the applicable boundaries of AI as a clinical auxiliary decision support tool, providing scientific guidance and standardized exploration directions for peers currently engaged in or planning to carry out AI-assisted clinical diagnosis, treatment, and translation of lung cancer.

Objective To propose a framework for risk assessment and stepwise management of pulmonary nodules based on existing guidelines and key studies, and to clarify the boundaries for triage, workup escalation, and referral in different detection scenarios. Methods A problem-oriented evidence synthesis approach was used, prioritizing Chinese expert consensus, Fleischner Society Guidelines, British Thoracic Society (BTS) Guidelines, and Lung-RADS. This was supplemented by studies on risk models, imaging measurements, liquid biopsy, tissue sampling, and surgical treatments. This paper is not presented as a formal consensus or a validated clinical pathway. Results A three-tiered framework consisting of basic non-invasive assessment, advanced assessment, and invasive diagnosis was established. It specified the applicable populations, boundaries for the use of risk models, imaging review, management of conflicting results, evaluation of histopathological consistency, and post-intervention follow-up. Liquid biopsy, genetic testing, and artificial intelligence were defined as investigational or adjunctive tools. Conclusion This framework can serve as a preliminary reference for multidisciplinary teams in developing local protocols. It cannot replace applicable guidelines, pathological diagnosis, shared decision-making with patients, or real-world validation.

Objective To observe the value of fetal MRI in diagnosis of duodenal obstruction(DO).Methods A total of 35 fetuses with suspected DO according to MRI were retrospectively included.The length and the maximum diameter of the dilated duodenum were measured,so as site(categorized in descending order from proximal to distal as descending segment,horizontal segment and ascending duodenum/proximal jejunum)and the degree of obstruction(complete or incomplete)were assessed.Taken findings of labor induction specimen,postnatal neonatal surgery or follow-up data as standards,the positive predictive value(PPV)of MRI for diagnosing fetal DO was calculated,while the correlations of the measured parameters of dilated duodenum and the confirmed obstruction site/degree were analyzed.Results Among 35 fetuses,DO was confirmed in 34 fetuses,yielding an overall PPV of 97.14％(34/35)for MRI.In 34 fetuses with confirmed DO,there were 23 with descending DO(DDO),4 with horizontal DO(HDO)and 7 with ascending DO/proximal jejunum obstruction(ADO/PJO),including 12 with complete DO and 22 with incomplete DO.PPV of MRI for diagnosing DDO,HDO and ADO/PJO was 87.50％(21/24),50.00％(2/4)and 100％(7/7),respectively,for diagnosing complete and incomplete DO was 90.00％(9/10)and 84.00％(21/25),respectively.Both the length and the maximum diameter of fetal proximal dilated duodenum showed on MRI were positively correlated with the actual obstruction site(from proximal to distal)(rs=0.736,P＜0.001;rs=0.424,P=0.011,respectively),but had no significant rank correlation with the degree of obstruction(rs=-0.216,P=0.212;rs=-0.285,P=0.097,respectively).Conclusion Fetal MRI could effectively evaluate the length and the maximum diameter of dilated duodenum hence indicating the level and degree of DO.

Objectives:To evaluate the incidence of neurological complications following left ventricular assist device(LVAD)implantation and to investigate related risk factors.Methods:A retrospective analysis was conducted on 151 patients who underwent LVAD implantation at Fuwai Hospital between June 2017 and September 2024.Clinical characteristics and postoperative survival outcomes were compared between patients with and without neurological complications.Results:Neurological complications occurred in 21 patients(13.9%)postoperatively,15 cases were ischemic strokes,5 cases were symptomatic intracranial hemorrhages or subarachnoid hemorrhages,and 1 case was transient ischemic attack(TIA).The total incidence of neurological complications was 0.08 events per person-year(EPPY),ischemic stroke was 0.06 EPPY and hemorrhagic stroke was 0.02 EPPY.Compared with patients without neurological complications,patients with neurological complications had a higher proportion of preoperative aortic regurgitation and tricuspid regurgitation,lower triglyceride levels,a lower rate of concurrent left atrial appendage resection and a higher rate of concurrent aortic valve replacement surgery.Multivariate cox regression analysis revealed that higher preoperative triglyceride levels(HR=0.21,95%CI:0.08-0.56,P=0.002)were associated with neurological complications.The median follow-up time was 508.0(186.5,931.5)days,12 out of 15 cases of ischemic stroke experienced no long-term sequelae,while 3 patients had varying degrees of residual deficits.All 5 patients with hemorrhagic stroke died,with 2 deaths directly attributed to hemorrhage.Kaplan-Meier survival curve analysis indicated that patients with neurological complications had a significantly lower survival rate(log-rank P=0.005).Conclusions:Neurological complications after LVAD implantation are predominantly ischemic strokes.Although less frequent,hemorrhagic strokes are associated with worse outcomes.Higher preoperative triglyceride levels is associated with neurological complications.

Objective To observe the value of fetal MRI in diagnosis of duodenal obstruction(DO).Methods A total of 35 fetuses with suspected DO according to MRI were retrospectively included.The length and the maximum diameter of the dilated duodenum were measured,so as site(categorized in descending order from proximal to distal as descending segment,horizontal segment and ascending duodenum/proximal jejunum)and the degree of obstruction(complete or incomplete)were assessed.Taken findings of labor induction specimen,postnatal neonatal surgery or follow-up data as standards,the positive predictive value(PPV)of MRI for diagnosing fetal DO was calculated,while the correlations of the measured parameters of dilated duodenum and the confirmed obstruction site/degree were analyzed.Results Among 35 fetuses,DO was confirmed in 34 fetuses,yielding an overall PPV of 97.14％(34/35)for MRI.In 34 fetuses with confirmed DO,there were 23 with descending DO(DDO),4 with horizontal DO(HDO)and 7 with ascending DO/proximal jejunum obstruction(ADO/PJO),including 12 with complete DO and 22 with incomplete DO.PPV of MRI for diagnosing DDO,HDO and ADO/PJO was 87.50％(21/24),50.00％(2/4)and 100％(7/7),respectively,for diagnosing complete and incomplete DO was 90.00％(9/10)and 84.00％(21/25),respectively.Both the length and the maximum diameter of fetal proximal dilated duodenum showed on MRI were positively correlated with the actual obstruction site(from proximal to distal)(rs=0.736,P＜0.001;rs=0.424,P=0.011,respectively),but had no significant rank correlation with the degree of obstruction(rs=-0.216,P=0.212;rs=-0.285,P=0.097,respectively).Conclusion Fetal MRI could effectively evaluate the length and the maximum diameter of dilated duodenum hence indicating the level and degree of DO.

Objectives:To evaluate the incidence of neurological complications following left ventricular assist device(LVAD)implantation and to investigate related risk factors.Methods:A retrospective analysis was conducted on 151 patients who underwent LVAD implantation at Fuwai Hospital between June 2017 and September 2024.Clinical characteristics and postoperative survival outcomes were compared between patients with and without neurological complications.Results:Neurological complications occurred in 21 patients(13.9%)postoperatively,15 cases were ischemic strokes,5 cases were symptomatic intracranial hemorrhages or subarachnoid hemorrhages,and 1 case was transient ischemic attack(TIA).The total incidence of neurological complications was 0.08 events per person-year(EPPY),ischemic stroke was 0.06 EPPY and hemorrhagic stroke was 0.02 EPPY.Compared with patients without neurological complications,patients with neurological complications had a higher proportion of preoperative aortic regurgitation and tricuspid regurgitation,lower triglyceride levels,a lower rate of concurrent left atrial appendage resection and a higher rate of concurrent aortic valve replacement surgery.Multivariate cox regression analysis revealed that higher preoperative triglyceride levels(HR=0.21,95%CI:0.08-0.56,P=0.002)were associated with neurological complications.The median follow-up time was 508.0(186.5,931.5)days,12 out of 15 cases of ischemic stroke experienced no long-term sequelae,while 3 patients had varying degrees of residual deficits.All 5 patients with hemorrhagic stroke died,with 2 deaths directly attributed to hemorrhage.Kaplan-Meier survival curve analysis indicated that patients with neurological complications had a significantly lower survival rate(log-rank P=0.005).Conclusions:Neurological complications after LVAD implantation are predominantly ischemic strokes.Although less frequent,hemorrhagic strokes are associated with worse outcomes.Higher preoperative triglyceride levels is associated with neurological complications.

Objective:In response to the subjectivity and poor real-time performance in methods for evaluating the quality of medical theory teaching, this study aims to develop an objective and real-time evaluation method for medical theory teaching quality.Methods:Classroom behavior data from both teachers and students in the course "medical image processing" were collected. An approach combining chain-like agent genetic algorithm with support vector regression machines was employed to analyze the collected classroom behavior data, and a classroom behavior-based evaluation model for medical theory teaching quality was established.Results:The predicted value generated by the model showed minimal error compared to students' actual answer accuracy, with an average absolute error of 4.84%. Additionally, the model demonstrated low time computation, with an average modeling time of 8.63 seconds and an average prediction time of 21.00 milliseconds.Conclusions:The constructed teaching quality assessment model shows high fitting precision and low time consumption.

Objective To analyze the application effects of artificial intelligence (AI) software and Mimics software in preoperative three-dimensional (3D) reconstruction for thoracoscopic anatomical pulmonary segmentectomy. Methods A retrospective analysis was conducted on patients who underwent thoracoscopic pulmonary segmentectomy at the Second People's Hospital of Huai'an from October 2019 to March 2024. Patients who underwent AI 3D reconstruction were included in the AI group, those who underwent Mimics 3D reconstruction were included in the Mimics group, and those who did not undergo 3D reconstruction were included in the control group. Perioperative related indicators of each group were compared. Results A total of 168 patients were included, including 73 males and 95 females, aged 25-81 (61.61±10.55) years. There were 79 patients in the AI group, 53 patients in the Mimics group, and 36 patients in the control group. There were no statistical differences in gender, age, smoking history, nodule size, number of lymph node dissection groups, postoperative pathological results, or postoperative complications among the three groups (P>0.05). There were statistical differences in operation time (P<0.001), extubation time (P<0.001), drainage volume (P<0.001), bleeding volume (P<0.001), and postoperative hospital stay (P=0.001) among the three groups. There were no statistical differences in operation time, extubation time, bleeding volume, or postoperative hospital stay between the AI group and the Mimics group (P>0.05). There was no statistical difference in drainage volume between the AI group and the control group (P=0.494), while there were statistical differences in operation time, drainage tube retention time, bleeding volume, and postoperative hospital stay (P<0.05). Conclusion For patients requiring thoracoscopic anatomical pulmonary segmentectomy, preoperative 3D reconstruction and preoperative planning based on 3D images can shorten the operation time, postoperative extubation time and hospital stay, and reduce intraoperative bleeding and postoperative drainage volume compared with reading CT images only. The use of AI software for 3D reconstruction is not inferior to Mimics manual 3D reconstruction in terms of surgical guidance and postoperative recovery, which can reduce the workload of clinicians and is worth promoting.

Objective:In response to the subjectivity and poor real-time performance in methods for evaluating the quality of medical theory teaching, this study aims to develop an objective and real-time evaluation method for medical theory teaching quality.Methods:Classroom behavior data from both teachers and students in the course "medical image processing" were collected. An approach combining chain-like agent genetic algorithm with support vector regression machines was employed to analyze the collected classroom behavior data, and a classroom behavior-based evaluation model for medical theory teaching quality was established.Results:The predicted value generated by the model showed minimal error compared to students' actual answer accuracy, with an average absolute error of 4.84%. Additionally, the model demonstrated low time computation, with an average modeling time of 8.63 seconds and an average prediction time of 21.00 milliseconds.Conclusions:The constructed teaching quality assessment model shows high fitting precision and low time consumption.