Research On Risk Assessment Technology Of Robot Assisted Laparoscopic Radical Prostatectomy Positive Margins In Prostate Cancer Based On Image Intelligent Semantic Model

As people’s living conditions improve day by day,the demand for maintaining health will also gradually increase.Prostate cancer is a common malignant tumor of the male urinary system,and the treatment mainly includes endocrine therapy and radical prostate cancer surgery.Positive margins for radical prostatectomy are a common problem encountered after radical prostatectomy,and positive margins directly determine whether the patient’s lesion is completely resected.Therefore,the correct understanding and treatment of positive margins and reducing the incidence of positive margins are the key issues to improve the survival of patients.In thesis,using clinical data and medical images,a positive prediction method based on image intelligence semantic model is proposed,and the specific research content is as follows:(1)In thesis,the clinical data in the patient dataset was statistically analyzed,and the clinical data that were helpful for positive prediction of margins were included in the feature vector.At the same time,the complexity and correlation of clinical data are analyzed to provide reference for clinical testing.In addition,the clinical dataset was trained using automatic machine learning to train an optimized margin positive prediction machine learning model.(2)This thesis proposes an image generation method to supplement the patient MRI medical image dataset and provide input data for subsequent image semantic analysis.This method is explained in several aspects,including the idea of generating prostate cross-sectional images based on variational autoencoder,relevant key technologies and improvement methods,design of generation network structure,and results of testing on the dataset.Experiments have shown that this method can effectively solve the problem of insufficient medical imaging data.(3)This thesis proposes a method for predicting positive margin status,mainly using three-dimensional convolutional neural networks for feature extraction and knowledge distillation optimization of patient MRI medical image datasets.The following experiments proved its effectiveness and analyzed some image features in medical images,providing reference for the inspection of key areas in medical imaging.(4)Based on the above method,a prototype of the margin positive prediction system was designed.The system mainly includes role management module,medical image generation module and margin positive prediction module.