Clinical Investigation Of ¹¹C-MET PET/MRI For Glioma Diagnosis And Prognosis Prediction

Objective:Glioma is the most common type of malignant intracranial tumor in clinical practice.Non-invasive imaging diagnosis and monitoring are crucial in the clinical management of glioma.As the routine screening method,conventional Magnetic resonance imaging（MRI）mainly provides information on anatomic abnormalities and blood-brain barrier disruption,with certain limitations for differentiating brain tumors and non-neoplastic lesions and also for identifying residual/recurrent lesions after surgery and radiotherapy.In recent years,functional MRI and amino acid positron emission computed tomography（PET）represented by 11C-methyl-L-methionine（11C-MET）have been widely discussed for the imaging of brain tumors.This study is aimed to evaluate the clinical value of 11C-MET PET/MRI in glioma patients from three aspects:differential diagnosis between brain tumor and non-neoplastic lesions,monitoring of glioma recurrence and prognosis prediction of glioma patients after surgery.Method:This study was a retrospective clinical study including patients who received¹¹C-MET PET imaging for brain tumors in our hospital from July 2017 to May 2022.The study was approved by the Institutional Ethics Committee,and all patients had given written informed consent.This study was divided into three parts:（1）the use of ¹¹C-MET PET/MRI imaging to distinguish brain tumors from non-neoplastic lesions which enrolled patients with suspicious lesions by recent MRI;（2）the use of ¹¹C-MET PET/MRI imaging to distinguish recurrent/residual glioma from post-treatment changes which included post-treatment glioma patients who was suspected with recurrent or undetermined disease during routine followed up by enhanced MRI;（3）the role of ¹¹C-MET PET/MRI imaging for prognosis prediction of glioma patients after surgery,which enrolled histologically confirmed glioma patients who had postoperational contrast-enhanced MRI and received standard chemoradiotherapy.Postoperative pathological or clinical long-term follow-up results were used as the gold standard for the group of brain tumor diagnosis and recurrence surveillance.For postoperative prognosis prediction group,overall survival（OS）and progression-free survival（PFS）data were determined by long-term follow-up.Statistical analysis:Nonparametric test,receiver-operating characteristics curve（ROC）and Logistic regression analysis were performed to compare imaging parameters between different groups and evaluate their individual and combined diagnostic value.Log-rank test and COX regression model were used to analyze the relationship between image parameters,clinical features and patient survival.Least absolute shrinkage and selection operator（LASSO）regression model with R code was used to screen the parameters to establish nomogram prediction model for internal validation.Results:A total of 151 patients were included according to inclusion and exclusion criteria.For differential diagnosis of brain tumors and non-tumor lesions,A total of 36lesions from 35 patients were included.,¹¹C-MET PET/MRI had higher diagnostic accuracy,specificity,diagnostic consistency,and comparable sensitivity than CE-MRI for the diagnostic of brain tumors.ROC curve analysis showed that SUV_max,SUV_mean,TBR_max,TBR_mean,r CBF,Cho/NAA and Cho/Cr had high discriminative ability.As an individual parameter,TBR_maxhad the highest diagnostic value（sensitivity 82.6%,specificity 100%,AUC=0.953）.The combination of TBR_maxand r CBF achieved the best diagnostic efficacy（AUC 0.963）.For the differential diagnosis of recurrent or residual glioma and post-treatment changes,43 patients were included.¹¹C-MET PET/MRI had higher diagnostic accuracy,specificity and comparable sensitivity than CE-MRI for the differential diagnosis of recurrent or residual glioma and post-treatment changes.ROC curve analysis showed that SUV_max,SUV_mean,TBR_max,TBR_mean,CBF,r CBF,Cho/NAA and Cho/Cr had high discriminative ability.As an individual parameter,Cho/NAA had the highest diagnostic value（sensitivity 71.4%,specificity 88.9%,AUC=0.794）.The combination of TBR_max,CBF and Cho/Cr had the best diagnostic efficacy（AUC 0.853）.For postoperative prognosis prediction,73 patients were included.Multivariate COX regression analysis showed that SUV_max,SUV_mean,TBR_mean,MTV,TLMM,Cho/Cr and Cho/NAA can be used to predict PFS while MTV can be used to predict OS.A prediction model for overall survival was established including patient age,WHO grade,IDH mutation status and MTV,whose C-index could reach 0.900.Conclusions:¹¹C-MET PET/MRI had higher diagnostic accuracy than contrast-enhanced MRI in the differential diagnosis of brain tumors and glioma recurrence surveillance.Multiple parameters of ¹¹C-MET PET and MRI(SUV_max、SUV_mean、TBR_max、TBR_mean、r CBF、Cho/Cr、Cho/NAA)were proved to have high discriminative value.Besides,their combined use could gain incremental diagnostic value.The MTV of ¹¹C-MET PET with clinical indicators was useful for predicting the OS.SUV_max,SUV_mean,TBR_max,TBR_mean,MTV,TLMM,Cho/Cr,and Cho/NAA were helpful for predicting the PFS.