Objective To investigate the predictive value of combined radiomic features derived from chest CT scans with clinical characteristics for epidermal growth factor receptor(EGFR)gene mutations in non-small cell lung cancer(NSCLC).Methods A multi-center case-control study was conducted on the clinical data and CT images of 1 070 NSCLC patients from the radiology departments of the 3 medical institutions between January 2013 and October 2023.The 719 NSCLC patients from the First Affiliated Hospital of Army Medical University were randomly divided into a training set and an internal validation set in a ratio of 7∶3;The 173 patients in the Eastern Theatre General Hospital and the 178 patients in Army Medical Centre of PLA were assigned into the external validation set 1 and 2,respectively.Least absolute shrinkage and selection operator(LASSO)regression was employed to identify the optimal radiomic features,which were subsequently used to construct a radiomics model.Univariate and multivariate logistic regression analyses were applied to identify clinical features associated with EGFR mutation,thereby developing a clinical model.The radiomic and clinical features were subsequently combined to develop a comprehensive model.All the 3 classification models were built using random forest(RF)machine learning.The area under curve(AUC),accuracy,sensitivity and specificity were utilized to evaluate the predictive performance of the models.Calibration curve was plotted to assess the goodness of fit of the comprehensive model,while decision curve analysis was performed to assess the clinical utility of the model.Results The AUC value of the radiomics model was 0.762 4(95%CI:0.692 4～0.825 1),0.745 4(95%CI:0.671 1～0.814 3),and 0.724 7(95%CI:0.639 7～0.801 6),respectively,in the internal validation set,external validation set 1,and external validation set 2;The AUC value of the clinical prediction model was 0.691 7(95%CI:0.627 9～0.757 6),0.652 5(95%CI:0.576 7～0.729 1),and 0.779 2(95%CI:0.712 5～0.847 3),respectively in the above sets in turn;The comprehensive model constructed based on clinical features and radiomic features showed the best predictive efficacy,with an AUC value of 0.818 0(95%CI:0.757 7～0.874 3),0.782 4(95%CI:0.703 1～0.848 2),and 0.796 6(95%CI:0.718 1～0.868 6),respectively in the above sets.Calibration curve analysis indicated that the comprehensive model had a good fit,while decision curve analysis revealed that the model provided a favorable net benefit.Conclusion Our comprehensive model constructed based on chest CT radiomic features and clinical characteristics shows superior predictive performance for EGFR gene mutations in NSCLC across multiple center datasets,which may be helpful for clinical decision-making for treatment strategies.

Objective To evaluate the predictive value of deep learning features from chest CT images combined with clinical and radiomics features for epidermal growth factor receptor(EGFR)mutations in non-small cell lung cancer(NSCLC).Methods This case-control study retrospectively analyzed clinical and imaging data of 1 070 NSCLC patients from radiology departments at three hospitals(January 2013 to October 2023).Patients were divided into:a training set(n=502)and internal validation set(n=217)via 7∶3 randomization of 719 cases from the First Affiliated Hospital of Army Medical University;external validation set 1(n=173)from General Hospital of Eastern Theater Command;external validation set 2(n=178)from Daping Hospital of Army Medical University.Deep learning features were extracted using a 2.5D convolutional neural network(CNN)with ResNet101 backbone,radiomics features were derived from CT images,and clinical risk factors were identified to construct models.An integrated model combined deep learning,clinical,and radiomics features.All four models were developed using random forest(RF)classifiers.Calibration curves assessed goodness-of-fit,and decision curve analysis(DCA)evaluated clinical utility.Results The deep learning model achieved AUCs of 0.833 7(95%CI:0.770 6～0.884 7),0.815 1(0.741 6～0.882 8),and 0.810 1(0.745 2～0.873 6)in the internal and two external validation sets,respectively.Clinical models yielded AUCs of 0.731 0(0.660 2～0.802 1),0.746 0(0.666 4～0.824 9),and 0.813 4(0.743 1～0.883 6);radiomics models showed AUCs of 0.762 4(0.692 4～0.825 1),0.745 4(0.671 1～0.814 3),and 0.724 7(0.639 7～0.801 6).The integrated model demonstrated optimal performance with AUCs of 0.905 5(0.857 0～0.945 4),0.832 7(0.763 3～0.896 4),and 0.889 0(0.834 4～0.934 3).DCA indicated significant net benefit for EGFR prediction at threshold probabilities of 0.15～0.85 using the integrated model.Conclusion Deep learning features from CT images effectively predict EGFR mutation status in NSCLC.The integrated model combining deep learning,clinical,and radiomics features further enhances predictive performance.

Humans ; Urinary Bladder Neoplasms/pathology* ; Carcinoma, Transitional Cell/pathology* ; Genetic Markers ; Biomarkers, Tumor/genetics* ; Urothelium/pathology*

Objective To explore the predictive efficacy of three machine learning algorithms for carotid plaque stability in patients with cerebral infarction.Methods The clinical data of 500 pa-tients with cerebral infarction were retrospectively analyzed.Univariate analysis and multivariate anal-ysis were used to determine the predictive factors entering the model.The prediction model of carotid plaque stability in patients with cerebral infarction was constructed based on nomogram,decision tree and random forest respectively.The enrolled patients were randomly divided into training set and test set according to the ratio of 7∶3.Sensitivity,specificity,accuracy,recall,accuracy and area under the curve(AUC)were used to compare the application efficiency of the model.Results The AUC of the nomogram model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.910(95％CI,0.950 to 0.983),the sensitivity was 0.910,the specificity was 0.917,the accuracy was 0.886,the recall rate was 0.910,and the accuracy rate was 0.914.The AUC of the decision tree model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.932(95％CI,0.903 to 0.961),the sensitivity was 0.903,the specificity was 0.922,the accuracy was 0.891,the recall rate was 0.903,and the accuracy rate was 0.914.The AUC of the random forest model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.984(95％CI,0.970 to 0.998),the sensitivity was 0.972,the specificity was 0.995,the accuracy was 0.993,the recall rate was 0.972,and the ac-curacy was 0.986.Conclusion The model based on the random forest algorithm has a better pre-diction effect and stability in evaluating the stability of carotid plaque in patients with cerebral infarc-tion,and its prediction efficiency is better than that of the Nomogram and decision tree.

Objective To explore the predictive efficacy of three machine learning algorithms for carotid plaque stability in patients with cerebral infarction.Methods The clinical data of 500 pa-tients with cerebral infarction were retrospectively analyzed.Univariate analysis and multivariate anal-ysis were used to determine the predictive factors entering the model.The prediction model of carotid plaque stability in patients with cerebral infarction was constructed based on nomogram,decision tree and random forest respectively.The enrolled patients were randomly divided into training set and test set according to the ratio of 7∶3.Sensitivity,specificity,accuracy,recall,accuracy and area under the curve(AUC)were used to compare the application efficiency of the model.Results The AUC of the nomogram model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.910(95％CI,0.950 to 0.983),the sensitivity was 0.910,the specificity was 0.917,the accuracy was 0.886,the recall rate was 0.910,and the accuracy rate was 0.914.The AUC of the decision tree model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.932(95％CI,0.903 to 0.961),the sensitivity was 0.903,the specificity was 0.922,the accuracy was 0.891,the recall rate was 0.903,and the accuracy rate was 0.914.The AUC of the random forest model for evaluating the stability of carotid plaque in patients with cerebral infarction in the training set was 0.984(95％CI,0.970 to 0.998),the sensitivity was 0.972,the specificity was 0.995,the accuracy was 0.993,the recall rate was 0.972,and the ac-curacy was 0.986.Conclusion The model based on the random forest algorithm has a better pre-diction effect and stability in evaluating the stability of carotid plaque in patients with cerebral infarc-tion,and its prediction efficiency is better than that of the Nomogram and decision tree.

BACKGROUND: The KangDuo-Surgical Robot-01 (KD-SR-01) system is a new surgical robot recently developed in China. The aim of this study was to present our single-center experience and mid-term outcomes of urological procedures using the KD-SR-01 system. METHODS: From August 2020 to April 2023, consecutive urologic procedures were performed at Peking University First Hospital using the KD-SR-01 system. The clinical features, perioperative data, and follow-up outcomes were prospectively collected and analyzed. RESULTS: A total of 110 consecutive patients were recruited. Among these patients, 28 underwent partial nephrectomy (PN), 41 underwent urinary tract reconstruction (26 underwent pyeloplasty, 3 underwent ureteral reconstruction and 12 underwent ureterovesical reimplantation [UR]), and 41 underwent radical prostatectomy (RP). The median operative time for PN was 112.5 min, 157.0 min for pyeloplasty, 151.0 min for ureteral reconstruction, 142.5 min for UR, and 138.0 min for RP. The median intraoperative blood loss was 10 mL for PN, 10 mL for pyeloplasty, 30 mL for ureteral reconstruction, 20 mL for UR, and 50 mL for RP. All procedures were successfully completed without conversion, and there were no major complications in any patient. The median warm ischemia time of PN was 17.3 min, and positive surgical margin was not noted in any patient. The overall positive surgical margin rate of RP was 39% (16/41), and no biochemical recurrence was observed in any RP patient during the median follow-up of 11.0 months. The surgical success rates of pyeloplasty and UR were 96% (25/26) and 92% (11/12) during the median follow-up of 29.5 months and 11.5 months, respectively. CONCLUSION The KD-SR-01 system appears feasible, safe, and effective for most urological procedures, based on our single-center experience.
Male ; Humans ; Robotic Surgical Procedures/methods* ; Robotics ; Treatment Outcome ; Retrospective Studies ; Ureter/surgery* ; Urologic Surgical Procedures/methods* ; Laparoscopy/methods*

Objective:To evaluate the safety and effectiveness of Kangduo endoscopic surgical robot system for partial nephrectomy.Methods:Consecutive patients with stage T 1 renal tumor meeting the inclusion criteria from the Department of Urology, Peking University First Hospital from December 2020 to February 2021 were prospectively enrolled. All patients underwent partial nephrectomy with the Kangduo endoscopic surgical robot system after signing the informed consent. Clinical data including preoperative, perioperative and postoperative pathology and follow-up were collected. Results:Among the 26 patients, there were 16 males and 10 females, with a median age of 53(33-74) years, and a median body mass index of 25.99(20.90-32.91) kg/m 2. There were 12 cases of left kidney tumor and 14 cases of right kidney tumor. The median tumor diameter was 2.2(1.0-3.5) cm. The median time of warm ischemia was 17.7(7.1-29.2) minutes, and all of them were less than 30 minutes. The median docking time was 4.7(2.3-9.9) minutes, and the median time of robotic arm operation was 65.0 (37.0-155.0) minutes. The median National Aeronautics and Space Administration Task Load Index (NASA-TLX) score was 5.3 (2.0-28.0), and no instrument-related adverse events occurred intraoperatively. The median postoperative hospital stay was 4 (4-5) days. All tumor margins were negative on pathologic reports. No Clavien Ⅱ stage operative complications occurred in all patients during perioperative period and 1 month after the surgery. Conclusions:The partial nephrectomy using the kangduo endoscopic surgical robot system were completed successfully, and no instrument-related adverse events and complications occurred, showing that this surgical system used for partial nephrectomy is safe and effective.

Residency training is the mainstream model of doctor training around the world. Radiol-ogy teaching is an important part of the resident practice training. The purpose is improving and upgrading the quality of the residency training by establishing rigorous management system, intensifying the teachers disposition, setting tutorial system and reforming the examination and supervisor system, the residents.

Objective To analyze the correlation between white matter injury and cognitive dysfunction using diffusion tensor imaging (DTI).Methods Seventeen subjects with TBI hospitalized from October 2012 to September 2013 had Glasgow coma scale (GCS) score of ≥ 13 (mild injury group, 10 cases) and ≤ 12 (moderate-severe injury group, 7 cases).Another 17 healthy subjects were used as controls.All were submitted to DTI examination.Fractional anisotropy (FA) and apparent diffusion coefficient(ADC) values in genu corpus callosum, splenium corpus callosum, posterior internal capsule, anterior internal capsule, and cerebral peduncle were calculated using the Neuro 3D software.Correlations between FA and ADC with the mini-mental state examination (MMSE) score were evaluated.Results Moderate-severe injury group demonstrated significantly reduced FA values in genu corpus callosum and splenium corpus callosum, and significantly increased ADC values of genu corpus callosum, splenium corpus callosum, posterior internal capsule and cerebral peduncle when compared to control group (P ＜0.05 or 0.01).FA and ADC values in the regions of interest did not differ significantly between mild injury group and control group (P ＞ 0.05).In the genu corpus callosum and splenium corpus callosum, FA values were positively correlated with MMSE score (r =0.636, 0.601), while ADC values were negatively correlated with MMSE score (r =0.552, 0.660).Conclusions DTI reveals the cerebral white matter lesion that is undetectable using CT and conventional MRI.DTI is a helpful tool to evaluate the degree of cognitive function in patients with TBI, which provides the basic reference for the clinical treatment and prognosis.