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 establish an analytical strategy covering multi-dataset processing, recursive feature elimination (RFE) screening and multi-model evaluation based on multiple machine learning algorithms, so as to screen radiation-sensitive genes and verify the feasibility of the evaluation strategy. Methods Qualified radiation transcriptome datasets were retrieved from public gene expression databases. Following standardized data preprocessing and feature preselection, 13 machine learning algorithms were adopted to construct models. The performance of each model was compared and validated in independent datasets. Results A total of 38 eligible datasets were included. Sixteen differentially expressed genes unreported in existing literature were screened out, among which ugcrhl, pdcl3, mct4, h2-g2 and fam120aos were correlated with radiation phenotypes. Ensemble learning algorithms including random forest and gradient boosting exhibited the optimal comprehensive performance. Independent dataset verification confirmed that the screened genes overlapped with known radiation-sensitive genes, and the model performance was consistent with the findings. Conclusion The machine learning strategy constructed in this study can effectively explore potential radiation-sensitive genes, and provides methodological support for subsequent relevant studies.

Objective To establish an analytical strategy covering multi-dataset processing, recursive feature elimination (RFE) screening and multi-model evaluation based on multiple machine learning algorithms, so as to screen radiation-sensitive genes and verify the feasibility of the evaluation strategy. Methods Qualified radiation transcriptome datasets were retrieved from public gene expression databases. Following standardized data preprocessing and feature preselection, 13 machine learning algorithms were adopted to construct models. The performance of each model was compared and validated in independent datasets. Results A total of 38 eligible datasets were included. Sixteen differentially expressed genes unreported in existing literature were screened out, among which ugcrhl, pdcl3, mct4, h2-g2 and fam120aos were correlated with radiation phenotypes. Ensemble learning algorithms including random forest and gradient boosting exhibited the optimal comprehensive performance. Independent dataset verification confirmed that the screened genes overlapped with known radiation-sensitive genes, and the model performance was consistent with the findings. Conclusion The machine learning strategy constructed in this study can effectively explore potential radiation-sensitive genes, and provides methodological support for subsequent relevant studies.

Objective To establish an analytical strategy covering multi-dataset processing, recursive feature elimination (RFE) screening and multi-model evaluation based on multiple machine learning algorithms, so as to screen radiation-sensitive genes and verify the feasibility of the evaluation strategy. Methods Qualified radiation transcriptome datasets were retrieved from public gene expression databases. Following standardized data preprocessing and feature preselection, 13 machine learning algorithms were adopted to construct models. The performance of each model was compared and validated in independent datasets. Results A total of 38 eligible datasets were included. Sixteen differentially expressed genes unreported in existing literature were screened out, among which ugcrhl, pdcl3, mct4, h2-g2 and fam120aos were correlated with radiation phenotypes. Ensemble learning algorithms including random forest and gradient boosting exhibited the optimal comprehensive performance. Independent dataset verification confirmed that the screened genes overlapped with known radiation-sensitive genes, and the model performance was consistent with the findings. Conclusion The machine learning strategy constructed in this study can effectively explore potential radiation-sensitive genes, and provides methodological support for subsequent relevant studies.

Lung cancer constitutes the most prevalent and lethal malignant tumor in China. Approximately 85% of lung cancer diagnoses correspond to the non-small cell histological subtype [non-small cell lung cancer (NSCLC)]. Despite surgery being the mainstay for early-stage disease, postoperative recurrence remains high and adjuvant chemotherapy offers limited benefit. In recent years, targeted therapy has demonstrated substantial advantages in driver mutation-positive NSCLC. To this end, the Lung Cancer Medical Education Committee of the Chinese Medical Education Association developed guidelines based on a systematic review of evidence through November 2025, using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach and a modified Delphi method. Focusing on epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK), and addressing ROS proto-oncogene 1 (ROS1), B-Raf proto-oncogene serine/threonine kinase (BRAF) V600E mutation, and mesenchymal-epithelial transition factor (MET) exon 14 (METex14) skipping, the guideline covers molecular testing, neoadjuvant/adjuvant therapy, perioperative strategies, minimal residual disease monitoring, and postoperative surveillance. It defines testing requirements, specifies stage-directed and subtype-specific treatments, and standardizes minimal residual disease monitoring. These recommendations emphasize precision and feasibility to improve survival and quality of life.

Objective To explore the effects and therapeutic mechanism of electroacupuncture on the levels of polarization markers and inflammatory factors interleukin(IL)-6,IL-4,tumor necrosis factor-alpha(TNF-α),and IL-10 in rats with para-chlorophenylalanine-induced insomnia(PCPA).Methods Fifty healthy specific-pathogen free grade Sprague-Dawley rats,half male and half female,were randomly divided into a blank group(n=10)and a model reserve group(n=40),in which insomnia was induced by intraperitoneal injection of a 500 mg/kg PCPA suspension.Using the random number table method,the 30 successfully modeled rats were divided into three treatment groups of 10 rats/group:model,electroacupuncture,and estazolam.The estazolam group was given estazolam 0.2 mg/(kg·d)by gavage;the electroacupuncture group was given once-daily electroacupuncture at the"Shenmen"and"Sanyinjiao"acupoints,and stimulation at the"Baihui"and"Benshen"acupoints,20 minutes each time,for 7 consecutive days.Following treatment,serum and hypothalamic levels of TNF-α,IL-6,IL-4,and IL-10 were detected using ELISA and Western blot,while immunofluorescence staining was used to detect the presence of Iba-1 in hypothalamic microglia and the co-expression of CD86 and CD163,which are markers for the M1 and M2 subtypes of microglial cells,respectively.Results Compared with the blank group,the model group exhibited prolonged sleep latency(SL)(P＜0.01),shortened sleep duration(ST)(P＜0.05),significantly higher serum and hypothalamic protein levels of IL-6 and TNF-α(P＜0.01),and significantly lower levels of IL-4 and IL-10(P＜0.01).Compared with the model group,the electroacupuncture and estazolam groups exhibited significantly shorter SL(P＜0.01),prolonged ST(P＜0.01),significantly lower serum and hypothalamic protein levels of IL-6 and TNF-α(P＜0.01),and significantly higher IL-4 and IL-10 levels(P＜0.01).IL-6 content was lower in the electroacupuncture group than in the estazolam group(P＜0.05).Compared with the blank group,the model group exhibited significantly enhanced Iba-1/CD86(M1 type)co-expression(P＜0.01)alongside significantly weakened Iba-1/CD163(M2 type)co-expression(P＜0.01).Under electroacupuncture or estazolam intervention,Iba-1/CD86 co-expression was significantly weakened(P＜0.01),and Iba-1/CD163 co-expression was significantly enhanced in the model group(P＜0.05).Conclusions Electroacupuncture effectively improved sleep disturbances in rats,with an underlying mechanism that may involve regulation of microglial polarization,downregulation of pro-inflammatory cytokine levels,upregulation of anti-inflammatory cytokine levels,and alleviation of neuroinflammation,thereby ameliorating sleep.

To evaluate the clinical application value of magnetic bead sorting technology (MBS) using the T-cell select kit combined with an automatic cell sorter for isolating peripheral blood mononuclear cells (PBMCs) to realize automated specimen pre-processing and process optimization in T-SPOT.TB assay. A cross-sectional study was conducted involving 300 patients recruited from the first affiliated hospital of Air Force Medical University from March 2023 to July 2024. Among them, there were 167 males and 133 females. The age range of the patients was 18 to 65 years, with a median age of 49.0 (33.3, 59.0) years and a mean age of (46.4±13.9) years. 4 ml of anticoagulated whole blood samples in duplicate were collected from each patient. One fresh sample (0-4 h) was processed immediately using density gradient centrifugation (DGC) for PBMC isolation, while the other was processed using MBS method at either 0-4 h or 34-54 h post-collection. The cycles for cell enrichment step of the automatic cell sorter were adjusted from the conventional 4 cycles to 2 cycles. Statistical analysis was performed using Cohen′s Kappa test to evaluate the concordance of T-SPOT.TB results across all processing groups. The results showed that for fresh samples (0-4 h), the T-SPOT.TB results for samples processed with 4-cycle and 2-cycle MBS enrichment steps demonstrated positive concordance rates of 91.7% and 93.1%, negative concordance rates of 100% for both, overall concordance rates of 98.0% and 97.9%, and Kappa values of 0.94 and 0.95, respectively, compared with the reference results for paired samples processed using DGC. Correspondingly, the results for samples processed with the 4-cycle and 2-cycle MBS methods at 34-54 h post-collection yielded positive concordance rates of 90.0% and 81.3%, negative concordance rates of 95.1% and 100%, overall concordance rates of 93.0% and 82.9%, and Kappa values of 0.86 and 0.43, respectively, compared with the reference results for fresh samples processed using DGC. In conclusion,for 0-4 h samples, the T-SPOT.TB results from both 4-cycle and 2-cycle MBS enrichment protocols were highly consistent with the reference results from conventional DGC methods. However, for 34-54 h stored samples, results from the 4-cycle MBS method rather than the 2-cycle protocol exhibited significantly superior concordance with the reference results. The MBS method using T-Cell Select kit achieves automated sample pre-processing for T-SPOT.TB assay, reduces hands-on time, and provides a viable alternative with reliable results for long-stored specimens.

Objectives:To explore the efficacy of robotic-assisted retroperitoneal benign tumor resection and to analyze its learning curve.Methods:This is a retrospective case series study. The data of patients who underwent robotic-assisted retroperitoneal benign tumor resection from August 2015 to February 2023 at the Department of Retroperitoneal Tumor Surgery was analyzed retrospectively. There were 24 males and 45 females, with an age of (46.3±10.6) years (range: 19 to 76 years). The perioperative data, postoperative pathological results, and follow-up data were recorded. The cumulative sum (CUSUM) method was used to analyze the robotic system setup time and operative time to plot the learning curve. A linear regression model was applied to determine the best-fit curve, selecting the model with the highest R2 value. Based on the vertex of the learning curve for surgical time, the patients were divided into a learning group and a mastery group. The general data and perioperative conditions of the two groups were compared. Independent sample t-tests, Mann-Whitney U tests, and χ2 tests were used for comparisons. Results:All 69 patients successfully completed the surgery without intraoperative complications. The diameter of tumors was (49.7±18.6) mm (range: 16 to 131 mm). The setup time for the robotic surgical system was (35.3±9.8) minutes (range: 20 to 61 minutes); the surgical time was (169.2±36.5) minutes (range: 70 to 305 minutes); intraoperative blood loss ( M(IQR)) was 10.0 (15.0) ml (range: 2.0 to 200.0 ml). The tumors in 32 patients (46.4%) were adherent to major blood vessels. All patients were discharged without complications. The follow-up period lasted until February 2024, and no patients required reoperation, readmission, or died due to retroperitoneal benign tumors. There were no severe long-term complications, and no radiological evidence of tumor recurrence was found. The best-fit equation for the learning curve based on surgical time was CUSUM=0.010X3-1.648X2-68.573X-61.091, and the best-fit equation for the learning curve based on robotic system setup time was CUSUM=0.0018X3-0.285X2+10.460X+57.541 (where X represents the number of surgeries). The R2 values of 2 learning curve models were 0.953 and 0.957, respectively, and the fit model tests had P<0.05. The inflection point of the learning curve based on surgical time was the 28th case, which is considered the minimum number of surgeries required to achieve proficiency in robotic-assisted retroperitoneal benign tumor resection. Based on this, the patients were divided into a learning group (cases 1 to 28) and a mastery group (cases 29 to 69). The surgical time for the learning group was significantly longer than that of the mastery group ((194.7±30.0) minutes vs. (151.9±31.4) minutes, t=4.126, P<0.01). No statistically significant differences were found for other parameters (all P>0.05). Conclusions:Robotic-assisted retroperitoneal benign tumor resection is feasible. The minimum number of surgeries required to achieve proficiency in overcoming the learning curve is about 28 cases.

Objective To explore and identify the nutritional/inflammatory indicator with the highest predictive potential for overall survival(OS)in postoperative tumor patients so as to provide guidance for postoperative rehabilitation of tumor patients.Methods Data from 3 191 surgical patients were collected,including 15 nutritional/inflammatory indicators.The maximum selection rank statistic method was used to calculate the optimal cut-off values for continuous indicators.The Kaplan-Meier method was used to assess OS,and Cox proportional hazards models were used to analyze the association between the aforementioned 15 indicators and survival.The predictive value of these 15 indicators was evaluated with receiver operating characteristic(ROC)curves and C-index.Results Multivariate analysis showed that all 15 indicators were significantly associated with poorer OS in surgical patients(P＜0.05 for all).Time-dependent area under the curve(AUC)and C-index analysis indicated that 3 indicators with the highest predictive potential in OS in postoperative tumor patients were the nutritional risk index(NRI)(C-index:0.597),C-reactive protein-to-albumin ratio(CAR)(C-index:0.587),and C-reactive protein-to-lymphocyte ratio(CLR)(C-index:0.587).The optimal cut-off value for NRI was determined to be 104.31(i.e.,NRI＜104.31 suggests malnutrition)with the maximum selection rank statistic method,the optimal cut-off value for CAR to be 0.05(i.e.,CAR≥0.05 suggests a strong inflammatory response,often accompanied by malnutrition),and the optimal cut-off value for CLR to be 1.18(i.e.,CLR≥1.18 suggests a strong inflammatory response).Subgroup analysis indicated that NRI,CAR,and CLR had good correlation with tumor staging,and there were significant differences between tumor node metastasis(TNM)Ⅲ/Ⅳ stage patients and TNM Ⅰ/Ⅱ stage patients when there was a strong inflammatory response or malnutrition.Conclusion In postoperative tumor patients,NRI,CLR,and CAR have high prognostic value.Combining these with the patient's clinical stage,it enables more precise guidance for clinical diagnosis and treatment strategies.

Objective:To establish a stable rat model of sequelae of pelvic inflammatory disease(SPID)with clinical characteristics,and to provide a reliable experimental model for the study of the pharmcological effect and mechanism of SPID.Methods:Twenty-four 7-week-old SD rats were divided into sham operation group,model-A(108 cfu/mL mixed bacterial solution,0.2 mL),model-B(109 cfu/mL mixed bacterial solution 0.2 mL),and model-C(108 cfu/mL E.coli 0.2 mL).The weight of the rat's uterine was weighed and the uterine index was calculated.The automatic hematology analyzer was used to detect the blood routine;hematoxylin-eosin staining(HE)and masson staining were used to detect uterine pathlogical changes in rats.Enzyme-linked immunosorbent assay(ELISA)was used to detect interleukin-1β(IL-1β),interleukin-6(IL-6)and tumor necrosis factor-α(TNF-α)in rat uterine tissue homogenates.Western blot was used to detect the expression of proteins related to NF-κB signaling pathway.Results:Compared with the sham operation group,the uterine index of model-A,model-B,and model-C were significantly increased(P＜0.05,P＜0.01).The levels of WBC and NE in the model-A increased significantly(P＜0.01).The level of LY in model-B decreased significantly(P＜0.01).The levels of IL-1β,TNF-α in model-A,model-B,and model-C were significantly increased(P＜0.01).The levels of IL-6 in model-A and model-B were significantly increased(P＜0.05,P＜0.01).The collagen volume fraction of model-A and model-B were significantly increased(P＜0.01).Mechanism study indicates that the expression levels of p-IKKβ/IKKβ,p-IκBα/IκBα and p-p65/p65 in model-A were significantly increased(P＜0.01),and the expression levels of IκBα/β-actin were significantly decreased(P＜0.01).The expression level of p-IKKβ/IKKβ in model-B was significantly increased(P＜0.01).Conclusions:A stable rat model of SPID that conforms to clinical characteristics can be successfully constructed by combining 0.2 mL of mixed bacterial solution with a concentration of 108 cfu/mL and mechanical injury.This modeling method intervened in the expression of the NF-κB inflammatory signaling pathway.