The Study Of Radiogenomics About Idh Genotype In Brain Gliomas

Part one The association between MRI characteristics and isocitrate dehydrogenase(IDH)genotype in brain gliomasPurpose To investigate the association between MR characteristics and isocitrate dehydrogenase(IDH)genotype in patients with brain gliomas.Materials and Methods 66 patients with pathological proved brain gliomas and receiving head MRI imaging before surgery were enrolled in this retrospective study.The mutation of IDH1/IDH2 gene statuses was by direct sequencing.MR imaging characteristics between IDH wild type(IDH-wt)and mutation(IDH-mt)gliomas were compared.Results Of 66 gliomas,there were 30 gliomas with IDH gene mutation.19(63.3%)of the IDH-mt gliomas were low grade and 11(37.3%)were high grade,while 6 of 36 IDH-wt gliomas(16.7%)were low grade and 30 were high grade(83.3%).The prevalence of IDH-mt was more common in LGG than that in HCC(73.2%vs 26.8%)?The difference was statistical significant(p=0.000).Tumors with IDH-mt preferably occurred in frontal lobe,followed by temporal lobe.There was no distribution dominant for IDH–wt gliomas.The location distribution showed statistical significant between IDH-wt and IDH-mt gliomas(p=0.032).Gliomas with IDH-mt were more like to be patch enhancement,while IDH-wt gliomas often showed with circle enhancement.The enhancement rates of IDH-mt gliomas was(41.2±42.0)%,which was less than that of IDH-wt gliomas'(82.4±57.0)%(p=0.001).The enhancement patterns between IDH-wt and IDH-mt gliomas were statistically significant(p=0.022).There were no statistical significant differences in diameter,edema,and ADC values between IDH-mt and IDH-wt gliomas.Conclusion Some MR imaging characteristics were associated with gliomas' IDH genotype,which showed imaging features might be potential biomarkers for IDH genotype.Part two Radiomics features based on multiparametric MRI to predict grade and IDH genotype of brain gliomasPURPOSE: To investigate the value of radiomics features based on multiparametric MRI to predict grade and IDH genotype of brain gliomas.MATERIALS AND METHODS: Sixty-one patients with pathologically proved gliomas and pre-treatment multi-parametric MRI were enrolled in this retrospective study.428 radiomic features were extracted from T2 WI,DWI(b=1000),ADCmaps and CE-T1 WI maps in every patient.The IDH1/IDH2 gene status was analyzed by direct sequencing.The association of radiomics features and clinical parameters with grade and IDH genotype of gliomas was firstly evaluated using univariate analysis.The variables with statistical significance were entered the least absolute shrinkage and selection operator(LASSO)regression model to identify the independent risk factors.Then radiomics signature score was calculated basee on LASSO coefficients.Nomogram was developed to predict gliomas' grade and IDH genotype.RESULTS:?Univariate analysis revealed that age,gender and 186 radiomics features were associated with grade of gliomas.Compared with females,males were more likely to have high grade gliomas.11 radiomics features were identified as independent predictors for HGG based on the LASSO regression.The nomogram based on age ?gender and radiomics signal score,arrived an AUC of 0.963,a sensitivity of 0.861,and a specificity of 0.960 for HGG.? Regarding to IDH genotype,age and 36 radiomics features were associated with IDH-mt.9 radiomics features were found to be independent predictors by LASSO regression.The nomogram based on age and radiomics signal score,could predict IDH-mt with an AUC of 0.913,a sensitivity of 0.897,and a specificity of 0.812.CONCLUSIONS: Radiomics features derived from multiparametric MRI were associated with gliomas' grade and IDH genotype.The nomogram based on radiomics signature and clinincal characters can predict the grade of glomas and IDH status with high accuracy.Therefore,radiomics signature derived from multiparametric MRI might be surrogate biomarkers to predict gliomas' grades and IDH genotype,provide decision support for personalized treatment of brain gliomas.