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 explore CT manifestations of perivascular epithelioid cell tumor(PEComa)of liver and kidney.Methods Totally 18 hepatic PEComa and 5 renal PEComa confirmed by surgical pathology were retrospectively enrolled,and the preoperative CT manifestations were explored.Results Single lesion of liver or kidney was found in all 23 cases,and the main CT manifestations included low or slightly low density lesion(23/23,100％),with irregular morphology(16/23,69.57％),clear boundaries(17/23,73.91％),non-envelope(21/23,91.30％)and enhancement after administration of contrast agents(21/23,91.30％).Among 18 cases of liver PEComa,most lesions(13/18,72.22％)presented as uniform density,while some(5/18,27.88％)presented as non-uniform density lesions often contained with fat components(5/5,100％)and thickened blood vessels(4/5,80.00％)but rare hemorrhagic necrosis(1/5,20.00％)nor calcification(0/5,0).Fast in and fast out of contrast agents were observed in 16(16/18,88.89％)lesions.Uneven density and internal fat components were found in all 5(5/5,100％)renal PEComa,which rarely with hemorrhagic necrosis(1/5,20.00％),blood vessels orientation(1/5,20.00％)and calcification(0/5,0).After enhancement,fast in and fast out,progressive enhancement and non-enhancement was observed in 2(2/5,40.00％),1(1/5,20.00％)and 2(2/5,40.00％)cases,respectively.Conclusion CT manifestations of PEComa in liver and kidney had certain characteristics.

Objective:To compare the clinical effects of restrictive transfusion strategy and liberal transfusion strategy for cardiac valve surgery.Methods:This study employed a prospective, randomized controlled superiority design, enrolling 439 patients undergoing non-emergency cardiac valve surgery with cardiopulmonary bypass at Department of Cardiovascular Surgery, Guangdong Provincial People′s Hospital, Southern Medical University from June 2023 to October 2024 who met the inclusion and exclusion criteria. While all the patients appeared hematocrit (Hct)≤24% or hemoglobin (Hb)≤80 g/L during the cardiopulmonary bypass. A simple random design was adopted to generate a random sequence and participants were randomized into a restrictive transfusion group (restrictive criteria: Hct≤18% or Hb≤60 g/L during cardiopulmonary bypass, and Hct≤21% or Hb≤70 g/L postoperatively) or a liberal transfusion group (liberal criteria: Hct≤24% or Hb≤80 g/L during cardiopulmonary bypass and Hct≤30% or Hb≤100 g/L postoperatively). If Hb or Hct fell below the respective thresholds, 2 units of red blood cells were transfused, followed by re-evaluation. If levels remained below the threshold, an additional 2 units were transfused until the criteria were met. The primary outcome was a composite of postoperative 3-month mortality, infection, ischemic events, and new-onset renal failure requiring dialysis. Secondary outcomes included blood product utilization, length of stay in the ICU and so on. Intergroup comparisons were performed using independent sample t-test, Mann-Whitney U test, χ2 test, or Fisher′s exact test, and analyses were conducted using a binary multivariable Logistic regression model. Results:A total of 439 patients were included in this study. The restrictive roup included 221 patients, including 75 males and 146 females, aged ( M(IQR)) 57.0 (14.0) years (range: 21 to 76 years). The liberal group included 218 patients, including 67 males and 151 females, aged 56.0 (20.0) years (range: 19 to 74 years). No statistically significant difference was observed in the incidence of primary outcome (restrictive group: 10.9%(24/221) vs. liberal group: 9.6%(21/218), χ2=0.180, P>0.05), 2 patints in the restrictive group died and 3 patints in liberal group died ( P=0.684). The transfusion rate was significantly lower in the restrictive group(19.0%(42/221) vs. 100%(218/218), P<0.01), with no significant differences in other secondary outcomes (all P>0.05). Subgroup analysis revealed an interaction between sex and transfusion strategy ( P=0.023), suggesting that using liberal transfusion strategy in male patients might increase the risk of the primary outcome. Conclusion:The mid-term results do not show that the restrictive transfusion strategy is superior to the liberal transfusion strategy in reducing the incidence of postoperative outcome events in patients undergoing cardiac valve surgery.

Radiation-induced esophageal injury (RIEI) is a frequent complication following radiotherapy for thoracic and head-neck malignancies, which may lead to severe sequelae including esophageal stricture and perforation, adversely affecting patients' quality of life and therapeutic outcomes. With advancements in radiotherapy techniques — particularly the adoption of unconventional fractionation regimens, concurrent chemoradiotherapy, and combined molecular targeted / immunotherapy — the incidence of RIEI has been increasing. In this review, recent advances in understanding the pathogenesis, clinical manifestations, risk factors, and management strategies for RIEI were comprehensively summarized. Current therapeutic approaches have evolved beyond conventional anti-inflammatory and nutritional support to include novel interventions such as targeted therapy, free radical scavengers, and microbiota modulation, etc. Future research should prioritize the development of optimized, individualized prevention and treatment protocols to mitigate RIEI risk and improve patient prognosis.

Objective To construct an MRI radiomics model of cerebral damage following carbon monoxide(CO)poisoning.Methods Ninety patients with CO poisoning and 70 healthy controls(HC)were scanned by structural MRI.All patients were randomly divided into training set(n=128)and test set in a ratio of 8∶2.The radiomics features of 9 regions of interest(ROI)were automatically extracted,and each ROI was subjected to 5-fold cross-validation and 9 different types of classifiers to construct a diagnostic model.Receiver operating characteristic(ROC)curve and decision curve analysis(DCA)were used to evaluate the diagnostic efficacy and clinical value of the model.Results The model based on the radiomics features of the whole brain gray matter(GM)and the left globus pallidus(LGP)had better diagnostic efficacy and clinical application value.The area under the curve(AUC)in the training set and the test set were 0.813,0.815 and 0.806,0.806,respectively.Conclusion The GM-based radiomics model can effectively diagnose cerebral damage following CO poisoning.

Objective:To clarify the safety, repair rate, durability, and risk factors for recurrent mitral regurgitation(MR) in patients with Barlow disease(BD) who total thoracoscopic minimally invasive mitral valvuloplasty(TMVP).Methods:Clinical data, mid-term and long-term outcomes of BD patients who underwent TMVP at Guangdong Provincial People's Hospital from January 2009 to June 2022 were retrospectively analyzed. Patients were divided into a group with no MR recurrence(group A) and a group with MR recurrence(group B) according to whether recurrent MR appeared in the postoperative period, and the data of the two groups of patients were compared with each other for the risk factor analysis.Results:The repair rate of TMVP was 98.4%, and no patient died perioperatively. The median follow-up time was 3.1(1.7, 5.2) years, the follow-up rate was 95.8%, and there was no patient died. As of March 2023, 112 patients developed no recurrent MR(group A), 11 patients developed recurrent MR(group B), and 2 patients in group B underwent repeated mitral valve surgery. The left atrial diameter(LAD) and left ventricular end-systolic diameter(LVESD) were higher in group B than in group A patients[LAD: (50.9±7.7)mm vs.(43.7±8.7)mm, P=0.009; LVESD: (37.1±5.5)mm vs.(33.2±4.7)mm, P=0.011], and the percentage of tendon cord rupture was higher in group B than in group A( P=0.022), while the rest of the baseline data were not statistically significant. There was no statistically significant difference between two groups in terms of the use of different surgical techniques, aortic cross-clamp time, cardiopulmonary bypass time, and operative time. Postoperative LAD, postoperative LVESD, and postoperative left ventricular end-diastolic diameter of group B patients were higher than those of group A( P<0.05). There was no statistically significant difference in perioperative and long-term complication rates between the two groups. Multifactorial Cox regression analysis revealed that advanced age( HR=1.049, 95% CI: 0.997-1.103, P=0.066) and large preoperative LVESD( HR=1.168, 95% CI: 1.053-1.295, P=0.003) were the risk factors for postoperative recurrence MR. Conclusion:Total thoracoscopic minimally invasive BD repair is safe, which has a high success rate and good long-term results. Advanced age and large preoperative LVESD are risk factors for recurrent MR in the long term.

Objective To construct an MRI radiomics model of cerebral damage following carbon monoxide(CO)poisoning.Methods Ninety patients with CO poisoning and 70 healthy controls(HC)were scanned by structural MRI.All patients were randomly divided into training set(n=128)and test set in a ratio of 8∶2.The radiomics features of 9 regions of interest(ROI)were automatically extracted,and each ROI was subjected to 5-fold cross-validation and 9 different types of classifiers to construct a diagnostic model.Receiver operating characteristic(ROC)curve and decision curve analysis(DCA)were used to evaluate the diagnostic efficacy and clinical value of the model.Results The model based on the radiomics features of the whole brain gray matter(GM)and the left globus pallidus(LGP)had better diagnostic efficacy and clinical application value.The area under the curve(AUC)in the training set and the test set were 0.813,0.815 and 0.806,0.806,respectively.Conclusion The GM-based radiomics model can effectively diagnose cerebral damage following CO poisoning.

Objective To explore the value of spectral CT quantitative parameters combined with clinical and CT features for predicting lymphovascular invasion(LVI)of colorectal cancer.Methods Clinical,pathological and preoperative abdominal spectral CT data of 98 colorectal cancer patients were retrospectively analyzed.According to pathological results,the patients were divided into LVI group(n=36)and non-LVI group(n=62).Univariate and multivariate logistic regression were used to compared clinical,pathological,conventional CT manifestations and spectral CT quantitative parameters between groups to screen independent predictors for LVI of colorectal cancer,and then a regression model was constructed.Receiver operating characteristic(ROC)curves were drawn,and the area under the curves(AUC)were calculated to evaluate the predictive efficacy of each single independent predictor and regression model for predicting LVI of colorectal cancer.Results Serum carbohydrate antigen 724,CT showed periintestinal fat infiltration and effective atomic number(Zeff)at venous phase were all independent predictors of LVI of colorectal cancer(OR=4.723,7.301 and 18.912,all P＜0.05).AUC of the above independent predictors was 0.582,0.723 and 0.691,respectively,while of the regression model was 0.837.Conclusion Spectral CT quantitative parameters combined with clinical and CT features could effectively predict LVI of colorectal cancer.

Objective To explore the value of spectral CT quantitative parameters combined with clinical and CT features for predicting lymphovascular invasion(LVI)of colorectal cancer.Methods Clinical,pathological and preoperative abdominal spectral CT data of 98 colorectal cancer patients were retrospectively analyzed.According to pathological results,the patients were divided into LVI group(n=36)and non-LVI group(n=62).Univariate and multivariate logistic regression were used to compared clinical,pathological,conventional CT manifestations and spectral CT quantitative parameters between groups to screen independent predictors for LVI of colorectal cancer,and then a regression model was constructed.Receiver operating characteristic(ROC)curves were drawn,and the area under the curves(AUC)were calculated to evaluate the predictive efficacy of each single independent predictor and regression model for predicting LVI of colorectal cancer.Results Serum carbohydrate antigen 724,CT showed periintestinal fat infiltration and effective atomic number(Zeff)at venous phase were all independent predictors of LVI of colorectal cancer(OR=4.723,7.301 and 18.912,all P＜0.05).AUC of the above independent predictors was 0.582,0.723 and 0.691,respectively,while of the regression model was 0.837.Conclusion Spectral CT quantitative parameters combined with clinical and CT features could effectively predict LVI of colorectal cancer.

Objective To explore CT manifestations of perivascular epithelioid cell tumor(PEComa)of liver and kidney.Methods Totally 18 hepatic PEComa and 5 renal PEComa confirmed by surgical pathology were retrospectively enrolled,and the preoperative CT manifestations were explored.Results Single lesion of liver or kidney was found in all 23 cases,and the main CT manifestations included low or slightly low density lesion(23/23,100％),with irregular morphology(16/23,69.57％),clear boundaries(17/23,73.91％),non-envelope(21/23,91.30％)and enhancement after administration of contrast agents(21/23,91.30％).Among 18 cases of liver PEComa,most lesions(13/18,72.22％)presented as uniform density,while some(5/18,27.88％)presented as non-uniform density lesions often contained with fat components(5/5,100％)and thickened blood vessels(4/5,80.00％)but rare hemorrhagic necrosis(1/5,20.00％)nor calcification(0/5,0).Fast in and fast out of contrast agents were observed in 16(16/18,88.89％)lesions.Uneven density and internal fat components were found in all 5(5/5,100％)renal PEComa,which rarely with hemorrhagic necrosis(1/5,20.00％),blood vessels orientation(1/5,20.00％)and calcification(0/5,0).After enhancement,fast in and fast out,progressive enhancement and non-enhancement was observed in 2(2/5,40.00％),1(1/5,20.00％)and 2(2/5,40.00％)cases,respectively.Conclusion CT manifestations of PEComa in liver and kidney had certain characteristics.