Transmission X-ray tubes are relatively new devices characterized by portability,suitability for miniaturization,and low requirements for shielding,making them ideal radiation sources for kilovoltage X-ray therapy.However,their application in radiotherapy remains underexplored.An electron target model of a transmission X-ray tube is developed using the Monte Carlo toolkit TOPAS 3.8.1.The study investigates the effect of tungsten target thickness on X-ray output efficiency,finding that a tube voltage of 50 kV and a tungsten thickness of 1.4 μm yields the highest emission efficiency.Based on the energy spectrum at this optimal efficiency,polynomial fitting approach is applied to determine the corresponding aluminum filter thickness for mean energies ranging from 20 keV to 35 keV,achieving a mean fitting error of 0.91%.Next,the study simulates dose deposition in a water phantom for spectra with different mean energies and various source-to-surface distances,and plots percent-depth-dose curves,relative normalized dose-depth curves,and relative normalized dose histograms under each treatment condition.Finally,the simulated results are compared with experimental data from the intraoperative radiotherapy system Intrabeam and the superficial X-ray therapy unit SRT-100,obtaining average relative errors of 3.71%and 4.38%,respectively.These findings provide a theoretical foundation for further optimization of transmission X-ray tubes in kilovoltage radiotherapy.

Objective To propose a image processing strategy based on salient object detection algorithm for optimizing the presentation of prosthetic visual information at a limited resolution level,aiming to detect and enhance the salient objects in the scene and remove the background information.Methods A salient object detection model combining CNN and Transformer was used to extract salient objects.On this basis,methods such as object magnification,contour enhancement and contrast enhancement were utilized to optimize the image information.Psychophysical experiments were carried out at 5 resolution levels(16×16,24×24,32×32,48×48 and 64×64).Results In the simulated prosthetic vision,this image processing strategy had a remarkable effect on improving the object recognition ability of the subjects.Regardless of the resolutions of 16×16,24×24,32×32,48×48 and 64×64,the proposed strategy achieved the highest recognition accuracies,specifically 34%±6%,56%±9%,72%±9%,89%±4%and 96%±2%.Conclusion Using the salient object detection method and image processing strategy to extract and enhance salient objects can help prosthesis implant recipients effectively improve their object recognition ability.

The diagnosis of Meige's syndrome predominantly relies on the clinical assessment by physicians.Given the complexity and similarity of its symptoms to other neurological disorders,the diagnosis is crucial for both doctors and patients.Herein a detection dataset for Meige's syndrome is compiled from video recordings of 31 patients,and an automated diagnostic system for Meige's syndrome(MS-Net)applicable to untrimmed videos is developed.The system utilizes RetinaNet and UNet3+to construct temporal detection and segmentation branches for multi-scale feature extraction and temporal segmentation,obtains probability vectors for detection windows and the probability of disease onset per frame via the decoding of temporal detection and segmentation branches,and finally generates a refined probability for each window by processing the probability predictions from both branches using a multi-layer perceptron.The model performance is optimized using additional loss functions and data augmentation techniques,operating on features interpretable by clinical physicians.MS-Net can assist in the diagnosis of Meige's syndrome,improving the accuracy,convenience,and efficiency of the early diagnosis.The comparison of MS-Net with other state-of-the-art networks indicates that MS-Net achieves comparable performance in terms of average precision while utilizing interpretable features required in clinical practice.

Objective By establishing a model to screen out patients with potentially large positional deviations based on their abdominal circumference,personalized solutions can be taken to address setup errors in these patients and ensure treatment efficacy.Methods A total of 81 patients treated at Tianjin People's Hospital from May 2021 to June 2023 were selected as the study subjects.The correlations between setup errors in the lateral,longitudinal,and vertical directions and abdominal circumference were analyzed.Subsequently,linear regression was performed for the direction with a significant correlation to abdominal circumference to establish a linear regression model.Finally,the 81 patients were divided into a normal setup group and an abnormal setup group with 0.7 cm as the critical value.A receiver operating characteristic(ROC)curve was plotted,and the maximum Youden index was calculated to determine the optimal cutoff value for identifying patients prone to abnormal setup.Results The correlation analysis of abdominal circumference and setup errors in various directions showed that the correlation coefficient between abdominal circumference and longitudinal direction positioning error was 0.406 2,and the correlation was statistically significant and positively correlated.The correlation coefficients for abdominal circumference and the lateral as well as longitudinal directions were-0.117 5 and-0.067 47,respectively,with P values greater than 0.05,indicating no statistical significance.A linear regression model was established for abdominal circumference and longitudinal direction,and the results showed an R2of 0.165,with a regression coefficient B of 0.008(t=3.951,P<0.01),indicating that the model was well constructed.ROC curve analysis showed an area under the curve of 0.715,with a 95%confidence interval of 0.57-0.86 and a maximum Youden index of 0.478.The corresponding optimal cutoff value was determined to be at 87 cm,with sensitivity at 0.875 and specificity at 0.603.Conclusion The abdominal circumference is significant to diagnose whether the longitudinal setup error is abnormal.Patients with an abdominal circumference greater than 87 cm are more likely to experience abnormal positioning during radiotherapy,which is reflected in the longitudinal setup errors being the largest.It is recommended to provide individualized target area margins for patients with an abdominal circumference greater than 87 cm,or perform daily cone-beam computed tomography to correct setup errors,thereby ensuring target coverage and treatment efficacy.

Regarding the challenge of early identification of critically ill patients with acute exacerbation of chronic obstructive pulmonary disease(AECOPD),a radiomics-clinical fusion model is proposed based on non-contrast CT images of accessory respiratory muscles to predict life-threatening conditions.A retrospective study is conducted involving 233 AECOPD patients(153 non-life-threatening and 80 life-threatening cases).Patients are divided into a training set(n=186)and a test set(n=47)at a 4:1 ratio.A total of 1 874 radiomic features are extracted from the erector spinae and pectoralis muscle regions delineated by radiologists on non-contrast CT images,and the features selection is performed using maximum relevance minimum redundancy and least absolute shrinkage and selection operator(LASSO)algorithms.Meanwhile,clinical data are analyzed with t-test and LASSO for variable screening.The selected features are input into C-support vector classification,Logistic regression,random forest,adaptive boosting(AdaBoost),and extreme gradient boosting(XGBoost)to construct radiomics model,clinical model,and fusion model.Predictive performance and clinical practicality are evaluated in the test set using receiver operating characteristic curve,area under the curve(AUC),and decision curve analysis.The radiomics-clinical fusion model built with XGBoost outperformed standalone radiomics and clinical models,achieving an AUC of 0.902(95％CI 0.846,0.994),with accuracy,sensitivity,specificity,and precision of 0.837,0.933,0.786,and 0.7,respectively.Results demonstrate that the fusion model based on the non-contrast CT radiomics of accessory respiratory muscles and clinical data exhibits promising diagnostic performance,highlighting its potential clinical significance for stratified management and preemptive critical care intervention in AECOPD patients.

A self-attention generative adversarial network(SAGAN)is proposed to improve the accuracy of histological subtype prediction for lung cancer cases.After collecting and preprocessing the lung cancer image dataset and data augmentation,SAGAN model is trained,where the generator uses self-attention mechanism to strengthen feature extraction,while the discriminator optimizes the generation process.Experimental results show that SAGAN model achieves accuracies of 0.852 and 0.845 on the training and test sets,respectively,with recall rates of 0.833 and 0.829,outperforming the other models.Additionally,the narrow confidence intervals indicate the high stability of SAGAN model in classification.SAGAN is helpful for lung cancer image analysis,providing stronger support for clinical decision-making.

Stereotactic radiotherapy is widely favored because of its high treatment precision and less fractionations.ZND-A is a new domestic gamma-ray cone-beam focused stereotactic radiotherapy system.Herein the technical characteristics of ZND-A system are described in detail from the aspects of the treatment frame,gamma-ray module,collimator module,six-dimensional treatment couch module and image-guided system module,and the main parameters are compared with the mainstream gamma knife equipments at home and abroad.With reference to Response Evaluation Criteria in Solid Tumors(RECIST 1.1),the initial efficacy of the patients treated by the ZND-A system is analyzed to evaluate the advantages and disadvantages of the ZND-A system for providing a reference for the hospital clinical use of this type of gamma knife.

Objective To investigate the feasibility and accuracy of convolutional neural networks for automatically delineating the pancreatic cancer gross target volume(GTV)in pancreatic enhanced CT.Methods The localizable enhanced CT images of 114 patients with pancreatic cancer were retrospectively selected,in which the GTV was manually delineated using AccuContour.The imaging data were then import to AccuLearning and randomly divided as the training set,validation set and test set at a ratio of 8:1:1.Flex and Segresnet were used to train the automatic segmentation model,with each network structure trained continuously 3 times using fixed training parameters.The model was evaluated in terms of Dice similarity coefficient(DSC),95%Hausdorff distance(HD95),average symmetric surface distance(ASSD)and relative volume difference(RVD).Results In the model training phase,Flex-3 test results in Flex group were the worst,with a minimum DSC of 0.14%and an average DSC of 56.30%,while Flex-1 performed well,achieving a minimum DSC of 47.90%and an average DSC of 67.35%.Meanwhile,Segresnet-2 in Segresnet group had the worst test results,with a minimum DSC of 0.00%and an average DSC of 42.46%,while Segresnet-3 test results were better,with a minimum DSC of 42.65%and an average DSC of 63.28%.In the fixed testing phase,the best results among all were as follows:average DSC and RVD values of 63.88%and 29.41%in Segresnet-3 group,average ASSD value of 4.43 mm in Segresnet-2 group,and average HD95 value of 12.87 mm in Segresnet-1 group.Conclusion Both Flex and Segresnet architectures of convolutional neural network can be used for the automatic pancreatic tumor GTV segmentation training,with Segresnet performing better in comprehensive evaluation.

Objective To analyze the therapeutic efficacy of low-temperature plasma surgery for children with obstructive sleep apnea-hypopnea syndrome(OSAHS)and its effects on inflammatory response,immune function,pain,and sleep quality for providing a basis for the rational treatment of OSAHS.Methods A prospective study was conducted on 92 children with OSAHS from January 2021 to December 2023,and they were randomly divided into control group(n=46)and pilot group(n=46).Control group were given conventional tonsil and adenoidectomy,while observation group were given low-temperature plasma tonsil and adenoid ablation.These patients in both groups were followed-up for 6 months after surgery,and there was no lost case during the follow-up period.The two groups were compared for the efficacy at 6 months after surgery,surgery-related indexes,scores of visual analogue scale for pain at 1,2,and 3 days after surgery,sleep quality before and at 6 months after surgery,inflammatory response,immune function,and the occurrence of complications during the 6 month follow-up period.Results Compared with control group,pilot group had higher overall effective rate at 6 months postoperatively(93.48%vs 78.26%,P<0.05),less intraoperative bleeding,and shorter operation and hospitalization time(P<0.05).The visual analog scale score gradually decreased in both groups at 1,2,3 days postoperatively,and the scores were lower in pilot group than in control group(P<0.05).Compared with those before surgery,apnea hypoventilation index and respiratory disturbance index were lower in both groups at 6 months postoperatively,with lower indicator values in the pilot group(P<0.05);while the lowest blood oxygen saturation increased in the two groups,with higher indicator values in the pilot group(P<0.05).Compared with those before surgery,serum procalcitonin,interleukin-1β,interleukin-6,high sensitive C-reactive protein,tumor necrosis factor-α,and whole-blood CD8+were lower in both groups at 3 days postoperatively,with lower indicator values in pilot group(P<0.05);while whole-blood CD3+,CD4+,CD4+/CD8+were higher in both groups at 3 days postoperatively,with higher indicator values in pilot group(P<0.05).The total complication rate within 6 months of follow-up was lower in pilot group than in control group(4.35%vs 17.39%,P<0.05).Conclusion Low-temperature plasma surgery is effective and safe in children with OSAHS,and it is considered that it might be related to its contributions to improvements in surgery-related indexes,inflammatory response,immune function,pain and sleep quality.

Objective To explore and address the issue of insufficient segmentation accuracy in liver tumor segmentation using the traditional U-Net algorithm,which is caused by the lack of contextual information for both the liver and tumor,as well as the large morphological variability of tumors.Methods A cascaded liver tumor segmentation algorithm,DDR-UNet++,which integrated dilated convolutions and residual modules was proposed.Firstly,CT images from the LiTS-2017 dataset were preprocessed through window width/level adjustment,histogram equalization and Gaussian filtering to reduce noise and smooth edges.Then,a cascaded liver segmentation model was employed to enhance the liver region proportion,mitigate interference from surrounding tissues and address data imbalance issue.For liver tumor segmentation,deformable dilated convolutions and residual networks were introduced to expand the receptive field and improve feature extraction capability.Results DDR-UNet++outperformed the traditional U-Net on the LiTS-2017 dataset,achieving improvements of 4.7%,1.7%,and 8.5%in Dice similarity coefficient,relative volume difference,and Jaccard index,respectively.These enhancements contribute to overcoming the inefficiency and low accuracy issues in conventional tumor segmentation,thereby improving early tumor detection rates,enhancing patient survival outcomes,and alleviating the diagnostic burden on clinicians.Conclusion The proposed method improves the feature extraction capability to some extent by enhancing the model structure and segmentation strategy,effectively increases the accuracy and robustness of liver tumor segmentation,and provides a reliable technical reference for clinical auxiliary diagnosis.