Prediction Of P53,Ki-67 Labeling Index And MGMT Promoter Methylation Status In Glioblastomas Using Multimodality MRI

Objective:Earlier understanding the p53,Ki-67 labeling index and Oxygen⁶-methylguanine-DNA methyltransferase（MGMT）promoter methylation status in adult isocitrate dehydrogenase（IDH）wild-type glioblastomas is clinically relevant to the selection of therapeutic regimen and prognosis.We assessed whether conventional MRI,diffusion-weighted imaging（DWI）and dynamic susceptibility contrast-enhanced perfusion-weighted imaging（DSC-PWI）could noninvasively predict p53,Ki-67labeling index and MGMT promoter methylation status in adult IDH wild-type glioblastomas.Materials and Methods:We retrospectively reviewed consecutive 120 adult patients who had pathologically confirmed IDH wild-type glioblastoma from March 2015 to April 2020.All patients underwent conventional MRI,DWI and DSC-PWI before any intervention.The clinical demographic characteristics（sex and age）and conventional imaging features（tumor position,necrosis,edema,enhancement pattern and definition enhancing margin）between different molecular status of IDH wild-type glioblastomas were analyzed.The relative minimum ADC value(r ADC_min)and relative maximum CBV value(r CBV_max)were obtained from the solid component of the tumor.The study population was divided into“low p53 group（p53≤50%）versus high p53 group（p53>50%）”,“low Ki-67 group（Ki-67≤20%）versus high Ki-67 group（Ki-67>20%）”and“MGMT promoter methylated group versus MGMT promoter unmethylated group”.The efficacy of routine image features was qualitatively analyzed using chi-square tests or Fisher’s exact test.All quantitative data were compared by Mann-Whitney U test.The diagnostic efficacy of univariate and multivariate MR parameters was evaluated by the receiver operating characteristic（ROC）curves and logistic regression model.Area under the curves（AUCs）of different MR parameters were compared using the De Long method.Inter-observer reproducibility for c MRI features and functional parameters were assessed using the Kappa statistics and intraclass correlation coefficient（ICC）.Results:（1）Glioblastomas with high p53 expression were more likely to occur well-defined enhancing margin than those with low p53 labeling index（22/37 versus32/83,P=0.047）,the AUC of c MRI for predicting p53 expression status in IDH wild-type GBM was 0.605.（2）Compared with low Ki-67 labeling index group,glioblastomas with high Ki-67 labeling index demonstrated significantly lower r ADC_minvalues（1.12±0.15 versus 1.00±0.16,P<0.001）and higher r CBV_maxvalues（9.26±1.67 versus 10.12±1.51,P=0.001）,the AUCs were 0.715 for r ADC_minand 0.682 for r CBV_maxcould differentiate the status of Ki-67 labeling index in glioblastomas.（3）Glioblastomas without MGMT promoter methylated showed lower r ADC_minvalues（0.99±0.15 versus 1.11±0.15,P<0.001）and higher r CBV_maxvalues（10.62±1.36versus 8.71±1.25,P<0.001）than those with MGMT promoter methylated,the AUCs were 0.732 for r ADC_minand 0.858 for r CBV_maxcould differentiate the MGMT promoter methylation status in glioblastomas.There was a significant difference in the AUC between r ADC_minand r CBV_max（P=0.001）in predicting MGMT promoter methylation status.Conclusions:Our results suggest that DWI and DSC-PWI are helpful in predicting the labeling index of Ki-67 and MGMT promoter methylation status in adult IDH wild-type glioblastoma,while multimodality MRI is limited in predicting the expression status of p53.