| Meningioma is the most common intracranial tumor in humans.Most meningiomas occur in the skull,and more than 90% of meningiomas are benign.The incidence of meningiomas is female: approximately 2:1 in males,with a peak age of45 years.It is rare in children,and many asymptomatic meningiomas are incidental.At present,surgery is the most important clinical treatment of meningioma,most of the prognosis is good.Periumoral edema is a common concomitant symptom of meningioma,as high as 60-67.4%,and postoperative cerebral edema is easy to be complicated.Cerebral edema can be generally divided into cytotoxic cerebral edema and vasogenic cerebral edema.For patients without preoperative periumoral edema,severe postoperative cerebral edema and aggravation of postoperative cerebral edema near functional areas can seriously affect the prognosis of patients,prolong the length of hospital stay of patients,and cause great harm to both individual families and society.Peridumoral edema is a major cause of morbidity and mortality in patients with brain tumors.Uncontrolled cerebral edema may lead to intracranial hypertension(RICH),as well as severe neurological dysfunction and potentially fatal cerebral herniation.In a retrospective study,the researchers assessed 1995 between January2001 and January 376 consecutive in patients undergoing microsurgical resection of intracranial meningioma of clinical and related records,some patients showed aggravation of brain edema after surgery,and most of them could be treated conservatively by medicine,but including 13 cases(3.5%)patients with postoperative CT scan or MRI scans show a wide range of brain swelling and postoperative neurologic deterioration,these patients need further intervention treatment,some patients with tracheal intubation artificial ventilation,some patients even second line to bone disc decompression surgery,through these measures to alleviate postoperative edema,meningioma through edema peak and save the patients life,therefore,it is very important to establish a model to predict cerebral edema exacerbation(CEE)after operation in patients with meningioma,and to formulate the appropriate treatment plan.The concept of "artificial intelligence" was first put forward in the 1950 s.It has the advantage of strong computing power and can handle a large amount of data that human beings can't handle well.However,medical data has the characteristics of large amount of data,so artificial intelligence technology has a broad application prospect in the field of medical research.With the rapid development of artificial intelligence technology,the application of artificial intelligence in the field of medicine is more and more extensive,and has attracted more and more attention.Radiomics is a new method of machine learning,radiomics refers to the medical image by conventional image,by using the method of artificial intelligence,high flux to extract a large amount of image data,this data can often describe the related image characteristics,extract the data after processing,to convert image data,makes it has high resolution,can send the characteristics of spatial data.Finally,quantitative high-throughput analysis was performed on the data to obtain high-fidelity target information for comprehensive evaluation of tumor phenotypes,including tissue morphology,genetic inheritance,cell and molecular levels.The core theoretical basis of radiomics is theradiomics model,which contains the medical data information of the disease and can provide valuable information for the diagnosis,treatment and prognosis of the disease.It can be extracted from medical image information data reflect the characteristics of the important biological tissue,in recent years,studies have shown that in the central nervous system diseases(image screening,early accurate diagnosis,classification stage,molecular markers,treatment and prognosis has broad application prospects,and help to formulate individualized treatment strategies,in this study,we aim to develop a minimum image feature set based on MR image radiomics model to predict the increase of meningioma postoperative brain edema.This study is mainly elaborated through the following two parts:Part ? Analysis of related factors of postoperative cerebral edema exacerbation in patients with meningioma Objective: To investigate the related risk factors of postoperative cerebral edema exacerbation in patients with meningioma Methods: Review on January 1,2017 to December 30,2019 in our hospital neurosurgery received surgical treatment of 136 patients with meningioma,patients with meningiomas MRI image and the image acquisition,postoperative cerebral CT review,analysis of preoperative MRI images and postoperative CT images,analyzed related clinical data and surgical head edema is aggravating,statistics and analysis of influence of meningioma the related risk factors for postoperative cerebral edema exacerbation.Results: Among 136 patients included in this study,60 patients(44.1%)had postoperative cerebral edema exacerbation,while 76 patients(55.9%)had no postoperative cerebral edema exacerbation.There were significant correlations between preoperative periumoral edema,tumor size and location and postoperative cerebral edema exacerbation(P = 0.000-0.001).Patients with large tumor size,peri-tumor edema before surgery,para-sinus tumor and skull base tumor are more likely to develop postoperative brain edema exacerbation.In contrast,there were no significant differences in gender,age,presence or absence of hypertension,presence or absence of diabetes,and presence or absence of epilepsy(P = 0.076-0.810)between postoperative cerebral edema plus reorganization and non-edema plus reorganization.Part ? Predicting cerebral edema exacerbation after meningioma surgery based on radiomics Objective:To evaluate the predictive value of multi-parameter MRI three-dimensional imaging features in cerebral edema exacerbation after meningioma surgery.Methods:According to screening criteria,136 patients with meningioma with complete clinical and radiographic data were collected,and they were randomly assigned to a training set or a validation set.Firstly,the ITK-SNAP software was used to segment the MR images of the patients with meningioma in this study to obtain the Region of Interest(ROI).Then,the method of Py Radiomics was used to extract the three-dimensional imagomics features from the multi-sequence MR images.Then,by using Wilcoxon rank sum test,elastic net algorithm and recursive feature elimination algorithm for screening,radiomics labels were constructed.Combined with clinical and radiomics characteristics,a clinical-radiomics combination model was established for predicting postoperative cerebral edema exacerbation in individual meningioma patients.Calibration curves and Hosmer-Lemeshow tests were used to assess the similarity between predicted and observed postoperative CEE.By quantifying the net benefits under different threshold probabilities,decision curve analysis was performed to evaluate the clinical utility of the clinical-radiomics combination model.Results: On the basis of the above algorithm,select three significant image characteristics of radiomics,and to build the radiomics labels,AUC 0.86 in the training set,validation set of AUC 0.80,two clinical characteristics(with or without tumor weeks edema and tumor size)and select the image group characteristics is determined to establish the clinical,radiomics joint model,the training set of AUC is0.91,the AUC of validation set to 0.83.The combined clinical-radiomics model has good identification and calibration,and has good clinical application value for predicting postoperative cerebral edema exacerbation in patients with meningioma.Conclusions:The postoperative cerebral edema exacerbation in patients with meningioma is related to whether the patients have peritumor edema before operation,tumor size and location.By integrating clinical characteristics with radiomics signature,the clin-radiomics combined model could assist in postoperative CEE prediction before surgery,and provide a basis for surgical treatment decisions in patients with meningioma. |
PDF Full Text Request