Histologic And Genetic Features Assessment Of Gliomas With Diffusion And Perfusion Magnetic Resonance Imaging

Part ? Grading and proliferation assessment of gliomas with diffusion and perfusion magnetic resonance imagingPurpose: To compare the main parameters derived from monoexponential diffusionweighted imaging(DWI),biexponential DWI,stretched-exponential DWI,diffusion kurtosis imaging(DKI),and arterial spin labeling(ASL)imaging with respect to diagnostic performance for tumor grading and proliferation assessment in gliomas and its subtype astrocytomas.Materials and Methods: Ninety pathologically confirmed glioma patients who underwent DWI,DKI and ASL on a 3.0T scanner were retrospectively reviewed.Measurements including the apparent diffusion coefficient(ADC),true diffusion coefficient(D),pseudodiffusion coefficient(D*),perfusion fraction(f),distributed diffusion coefficient(DDC),heterogeneity index(?),mean diffusivity(MD),fractional anisotropy(FA),mean kurtosis(MK),and cerebral blood flow(CBF)were compared between tumor grades by using the Jonckheere-Terpstra test,Kruskal-Wallis test,and Mann-Whitney U test.Receiver operating characteristic(ROC)curves were used to assess the diagnostic efficacy of these parameters,the area under the curves(AUCs)were compared by using the Z test.The parametric values were also normalized to the contralateral normal-appearing white matter for the same analyses.Spearman's rho with the Ki-67 labeling index(LI)was calculated for each parameter.The intraclass correlation coefficient(ICC)was calculated to assess the interobserver variation.Results: D,?,and MK values differed significantly among grade ?,?,and ? and between high-and low-grade astrocytomas,ADC,f,DDC,MD,CBF values differed significantly between grade ? and ?/?,and between high-and low-grade astrocytomas,D* values differed significantly between grade ?/? and ?,and between high-and low-grade astrocytomas,FA value differed significantly between high-and low-grade astrocytomas.There was a significant tendency to higher values of D*,FA,CBF,MK and lower values of ADC,D f,DDC,?,MD for higher grade(all P < 0.05).In the ROC analyses,only D* and FA showed significant smaller AUCs in differentiating between grade ? and ? and between high-and low-grade astrocytomas(P < 0.05),the rest parameters showed no significant difference in the differentiation of astrocytoma grades.With the inclusion of oligodendrogliomas,D,?,and MK were no longer significantly differed between grade ? and ? tumors,FA was no longer significantly differed between high-and low-grade tumors,CBF showed significantly smaller AUC in discriminating between high-and low-grade tumors compared with DDC,MD,and MK(P < 0.05),the rest conclusions were consistent.D and MK showed a strong correlation with the Ki-67 LI(? =-0.762 and 0.745,respectively,P < 0.001).ADC,f,DDC,?,MD,CBF showed a moderate correlation(|?| ranged from 0.509 to 0.669,P < 0.05),D* and FA showed no significant correlation(P > 0.05).Conclusion: D,MK,and ? could effectively characterize microstructural changes throughout the malignant transformation of astrocytomas,making them promising markers for tumor grading.D and MK have great potential for noninvasively evaluating cellular proliferation of gliomas.Part ? Genetic features assessment of gliomas with diffusion and perfusion magnetic resonance imagingPurpose: To explore the association between the main parameters derived from monoexponential diffusion-weighted imaging(DWI),biexponential DWI,stretchedexponential DWI,diffusion kurtosis imaging(DKI),and arterial spin labeling(ASL)imaging and common genetic variations of gliomas.Materials and Methods: Thirty-two glioma patients were consecutively collected.Preoperative diffusion and perfusion imaging were performed on a 3.0T unit.The solid component of the tumors was delineated for parametric measurements and guiding the tissue collection.The tissues were examined for 71 genetic variations,only those with the number of cases reaching six were further analyzed.Measurements including the apparent diffusion coefficient(ADC),true diffusion coefficient(D),pseudodiffusion coefficient(D*),perfusion fraction(f),distributed diffusion coefficient(DDC),heterogeneity index(?),mean diffusivity(MD),fractional anisotropy(FA),mean kurtosis(MK),and cerebral blood flow(CBF)were compared between tumors with and without a genetic variation by using a Mann-Whitney U test.Receiver operating characteristic(ROC)curves were used to assess the diagnostic efficacy of these parameters.The parametric values were also normalized to the contralateral normal-appearing white matter for the same analyses.Result: Except DDC and CBF,all parameters showed a significant difference between IDHmutant and-wildtype tumors.No parameter differed significantly between 1p/19q-codeleted positive and negative tumors.CBF differed significantly between TERT promoter-mutant and-wildtype tumors.ADC,f,?,and FA showed a significant difference between ATRXmutant and-wildtype tumors.Between tumors with and without MGMT promotermethylation,only FA showed a significant difference.D,f,?,MD,and CBF differed significantly between TP53-mutant and-wildtype tumors.f,MD,FA,and CBF differed significantly between EGFR-amplificated and EGFR-normal tumors.MD and CBF differed significantly between PTEN-mutant and-wildtype tumors(all P < 0.05).All parameters with a significant difference showed an acceptable or better discrimination(area under ROC > 0.7,P < 0.05)in the differentiation of genetic variations.Conclusion: Gliomas with different genetic features have different tendencies in diffusion and perfusion within the tissue.Diffusion and perfusion magnetic resonance imaging may have potential to assess genetic features of gliomas noninvasively.