Study On Gene Heterogeneity And Prognosis Prediction Of Glioma

Background:Diffuse glioma is the most common malignant tumor of central nervous system with high recurrence rate and mortality.According to the 2016 World Health Organization(WHO)classification of central nervous system tumors,diffuse glioma mainly includes astrocytoma,oligodendroglioma,astrocytoma,oligodendroglioma and glioblastoma.Isocitrate dehydrogenase(IDH),TP53,ATRX and chromosome 1p19 q codeletion were added to classify subtypes.Diffuse gliomas were classified as grade II~IV,and grade II diffuse gliomas includes: diffuse astroglioma,IDH wild-type(IDH-wt);diffuse astroglioma IDH-mutated(IDH-mut);oligodendroglioma IDH-mut with 1p19 q codeletion;Grade III diffuse gliomas are classified into 1p/19 q codeleted anaplastic oligodendrogliomas and anaplastic astrocytoma with IDH-mut.Gliomas in grade IV including glioblastoma(GBM)IDH-wt,GBM IDH-mut,and diffuse the midline gliomas with H3-K27 M mutant.These glioma types combined with the gene mutant status suggest different prognosis of glioma.At present,IDH1/2 gene mutation,1p19 q codeletion and MGMT promoter methylation are important molecular indicators of good prognosis of glioma.In addition,some genes expressed in glioma,such as EGFR,PD-1 and PD-L1,have been used in gene targeted therapy.Although no breakthrough has been made in gene targeting therapy for glioma,more and more targeted drugs have entered clinical trials.This suggests that we need to classify gliomas more accurately and completely.However,genetic testing and pathological analysis currently require more than 1 week to get completely diagnosis report after surgery.However,preoperative evaluation of the prognosis of patients in the perioperative period is an important reference for the selection of treatment options and patient education.Currently,it is believed that high intra-heterogeneity of glioma is the key factor leading to resistance to postoperative chemoradiotherapy and early recurrence.Although many studies have confirmed the high heterogeneity of glioma using primary tumor cells culture,single cells sequencing,pathology and imageology,there is still a lack of clinical evaluation of the intra-heterogeneity.The heterogeneity of glioma was divided into intra-tumoral heterogeneity and inter-tumoral heterogeneity according to the study objects.And according to the effect,heterogeneity also contains: gene heterogeneity,protein heterogeneity,tissue composition heterogeneity and microenvironment heterogeneity.Among them,the genetic heterogeneity may be the root cause of others.With the development of gene testing technology,glioma whole exon sequencing(WES)technology has been widely used in clinical diagnosis.Based on the WES results,the mutant-allele tumor heterogeneity(MATH)value can be used to measure the genetic heterogeneity of glioma.The MATH value is calculated by the number of allele mutations and its distribution width.This approach quantifies the genetic heterogeneity of glioma with the results of total exon sequencing.The MATH value provides a good reference for us to evaluate the heterogeneity of glioma.Studies have shown that MATH values can reflect the prognosis of patients with head and neck cancer,colon cancer,breast cancer,and lung cancer.Therefore,to analyze the relationship between MATH value and prognosis in diffuse glioma,a highly heterogeneous tumor,is particularly meaningful.Surgery is currently the preferred treatment for glioma,and the maximum safe range of tumor resection is recommended to improve the prognosis of patients.After surgery,chemotherapeutic agent temozolomide and radiotherapy are followed.The recurrent gliomas are suggested by surgery,targeted drug therapy,electric field therapy and immunotherapy.Postoperative treatment of glioma depends on the results of pathological examination and genetic sequencing of biopsy tissue from tumor samples during surgery.However,Higher accurate therapies require higher accuracy on whether certain important genes alterations(such as IDH,MGMT promoter methylation,1p19 q combined deletion).By single-cell RNA sequencing,Patel et al.confirmed the existence of cell subtypes with significantly different gene expression in GBM.This suggests that different subgroups within GBM may have different responses to treatment.Therefore,how to evaluate glioma more accurately and comprehensively requires us to have a high degree of generality in the biopsy samples of glioma.So how to select biopsy sites is a difficult problem in clinical practice.Magnetic resonance imaging(MRI)is a milestone in the imaging diagnosis of intracranial tumors.Due to gene heterogeneity in glioma,the tumor showed cell proliferation,apoptosis,necrosis,edema and other pathological manifestations unevenly.These pathologies cause different signals on magnetic resonance.Contrast-enhanced MRI images based on gadolinium ions are widely used in the diagnosis of central nervous system tumors.Malignant gliomas,especially high-grade diffuse gliomas(grade III or IV),are destroyed in the blood-brain barrier,and contrast enhancer gadolinium ion can enter the tumor and be detected by MRI.In MRI T1-weighted contrast enhancement images,due to the heterogeneity of pathological changes of glioma,shows heterogeneous signal strength.Our study analyzed the degree of tumor heterogeneity weather is related to the MRI contrast-enhanced signal.It can provide an important reference for us to select biopsy sites during surgery.Method:Firstly,our study obtained the clinical data of 757 glioma patients in TCGA database,the whole exon sequencing information and the RNA-seq sequencing information of 703 glioma patients.We used univariate and multivariate COX regression analysis to scream the clinical factors which were significantly related to recurrence free survival(RFS)time and overall survival(OS)time of glioma.Nomogram predictive tables,including screened factors,were built.ROC curve and calibration curve were used to verify the nomogram prediction scale.Secondly,we used the whole exon sequencing information downloaded from TCGA database to calculate the MATH value of each patient.T-test and One-way Anova test were used to compare MATH values in different subtypes of glioma,respectively.We analyzed the relationship between MATH value levels and recurrence rate in different subtypes of glioma.And we also compared the difference of gene mutation frequency in different MATH values groups.Meanwhile,gene mutation enrichment and gene expression enrichment were studied between different MATH value groups.Combined with clinical information and gene mutation information,we used univariate and multivariate Cox regression analysis to screen out the key factors that affect the recurrence time of glioma.The factors than were included in the nomogram prediction table to predict the recurrence of glioma.In addition,we used ROC curve and calibration curve to verify the nomogram table.Finally,we enrolled high-grade glioma patients(grade III and grade IV glioma patients).They were studied by MRI,and the biopsy points were planed according to the MRI T1 weighted enhancement signal.Then,we preformed biopsy under neuro-navigation before resection.The biopsy tissues were studied by WES.According to the result of the WES,we calculated MATH value of each sample.Using T test and linear regression,we attempted to find relationship between MATH value and the intensity of MRI T1 enhanced signal.Result:By screening clinical data,we found that sex,age,WHO grade,tumor status,Karnofsky Performance Score(KPS),IDH1/2 mutation status,and radiotherapy were correlated with OS of glioma patients.WHO grade,histological type,tumor status,IDH1/2 mutation status and radiotherapy were correlated with the RFS rate of glioma.By incorporating screened factors,we formed two nomograms prediction table for OS and RFS respectively.We used receiver operating curve(ROC)and calibration curve showed the nomograms have good accuracy and consistency of ideal curve.MATH values were increased in patients with IDH-wt glioma(P=0.001)and glioblastoma(GBM;P=0.001).MATH values were negatively associated with the 2-and 5-year recurrence-free survival(RFS)rates in patients with glioma,particularly in the IDH1/2-wt and GBM cohorts(P=0.001 and P=0.017,respectively).Furthermore,glioma cases with different MATH levels had distinct patterns of gene mutation frequencies and gene expression enrichment.Finally,a nomogram table that contained MATH values could be used to accurately predict the probabilities of the 1-,2-and 5-year RFS of patients with glioma.High intratumoral gene heterogeneity exists in high-grade gliomas.Heterogeneity was observed in the number of base mutations,the number of gene mutations,MATH value and MRI T1 enhanced signal.There was no significant difference in MATH values between MRI T1 enhanced and non-enhanced sites.There was a positive linear relationship between MATH value and MRI T1 C enhanced signal in AO,but no significant linear relationship between MATH value and MRI T1 C enhanced signal in GBM.Conclusion:1.Some clinical characters can roughly predict the prognosis of glioma.Using nomogram table can quantifies and visualizes the risk of each factor in predicting prognosis of glioma.2.The MATH value of a patient may be an independent predictor that influences glioma recurrence.By incorporating MATH and other gene information,the nomogram model presented was an appropriate method to predict 1-,2-and 5-year RFS probabilities in patients with glioma.3.Highly intertumoral gene heterogeneity exists in high-grade glioma.In AO,higher T1 C enhancement site may have higher gene heterogeneity.However in AA and GBM,it needs further study of the relationship between the signal of MRI enhancement and the levels of gene heterogeneity at biopsy sites.