Arterial Spin Labeling And Diffusion Tensor Imaging Analysis Of Perfusion And Diffusion In Glioma And Its Correlation With HIF-1α Expression

Objective:To explore the correlation between conventional MRI characteristics,3D-ASL and DTI quantitative parameters and glioma HIF-1αexpression。Materials and Methods:2.1 Clinical DataPatients who were examined by MRI and pathologically diagnosed as glioma from January 2016 to November 2021 were collected.A total of 51 patients were collected,including 28 males and 23 females,with an age range of 18-76 years,a mean age of52.24 years,and a median age of 54 years.2.2 MRI EquipmentMRI was performed on GE Discovery MR 750W 3.0T scanner with scan sequences including T₁WI,T₂WI,T₁WI+C,3D-ASL,and DTI.2.3 MRI post-processingThe 3D-ASL and DTI raw data were imported into GE ADW 4.7 workstation,and the image post-processing was performed using Functool software.The area of interest was manually selected according to the lesion size,and the average value was obtained as the parameter value accordingly.2.4 Conventional MR image observation indexRoutine MR image observation indexes in 51 patients included:maximum tumor diameter（mm）,hemorrhage,necrosis volume,enhancement pattern,tumor border,peritumor edema,and whether the tumor crossed the midline and grew contralaterally.The diagnostic results were analyzed independently by two neurological diagnostic imaging physicians（with 2 and 10 years of neurological imaging experience,respectively）using a double-blind method,and when the assessment results of the two physicians did not agree,the two physicians discussed together and came to a consensus result.2.5 Pathological examination and HIF-1αimmunohistochemical stainingPathological grading of gliomas was recorded,and pathological sections for HIF-1α,immunohistochemical staining and assessment of HIF-1αpositive cell rates and staining intensity were performed with the assistance of pathologists,and HIF-1αscores were recorded.2.6 Observation indicators（1）Conventional MRI features:maximum tumor diameter（mm）,hemorrhag,percentage of necrotic volume,enhancement pattern,tumor boundary,peritumor edema,whether the tumor grows across the midline to the contralateral side.（2）3D-ASL quantitative parameters:CBF_max、CBF_mean、rCBF_max、rCBF_mean;（3）DTI quantitative parameters:ADC_min、ADC_mean、rADC_min、rADC_mean;FAmax、FAmean、rFAmax、rFAmean;（4）Glioma pathological classification:WHO grade I-IV.（5）HIF-1αexpression score.2.7 Statistical AnalysesSPSS 26.0 and Graph pad prism 7 were used for statistical analysis.Normality tests were performed for continuous variables,and independent sample t-tests were used for continuous variables that met the normality test,and Mann-Whitney U-tests were used for continuous variables that did not meet the normality test.Categorical variables were tested using theχ2 test orFisher’s exact test.Interobserver agreement tests were performed for two observers,and multiple interpolation was used to supplement for missing data.Correlation analysis of 3D-ASL and DTI parameter values with HIF-1αpercentiles was performed using the Spearman test.The area under the curve（AUC）was calculated using ROC for assessing the diagnostic efficacy of imaging parameters alone or in combination on HIF-1αexpression and to obtain diagnostic thresholds,specificity and sensitivity.Single-factor and multifactor logistic regression analysis were used to analyze screening imaging HIF-1αcorrelates,and age and gender variables were included for correction,and P<0.05 differences were statistically significant.Results:3.1 Analysis of case data and MRI imaging features and pathological features of gliomaStatistical analysis of MRI routine signs and pathological features of 51 glioma patients showed that there were significant differences in necrotic volume,tumor enhancement pattern and pathological grade between HIF-1αhigh expression group and HIF-1αlow expression group.3.2 Correlation analysis of 3D-ASL,DTI parameters and HIF-1αCBF_max,rCBF_max values were positively correlated with HIF-1αexpression（r=0.401,P=0.004;r=0.3,P=0.033）.HIF-1αexpression was negatively correlated with ADC_min,rADC_minand ADC_mean values（r=-0.401,P=0.004;r=-0.280,P=0.047;r=-0.281,P=0.046）.HIF-1αexpression did not correlate with CBF_mean,rCBF_mean,rADCmean,FAmax,rFAmax,FAmean,and rFAmean.3.3 Parameter analysis of 3D-ASL,DTI and HIF-1αexpressionCBF_max,rCBF_max,ADC_min and rADC_min values were significantly different between the HIF-1αlow expression group and HIF-1αhigh expression group（P<0.05）.c CBF_mean,rCBF_mean,ADC_mean,rADC_mean,FA_max,rFA_mean values were statistically different between the HIF-1αlow expression group and HIF-1αhigh expression group.rFA_max,FA_mean,and rFA_mean values were not statistically different.3.4 Diagnostic efficacy of 3D-ASL and DTI alone and in combination to assess HIF-1αexpressionROC analysis showed that the AUC values of CBF_max,rCBF_max,ADC_min and rADC_min in evaluating the diagnostic efficacy of glioma HIF-1αwere 0.784,0.746,0.835 and 0.742,respectively.The AUC value of combined CBF_max-ADC_min in evaluating HIF-1αexpression was 0.869,and the AUC value of combined test was larger than that of single test,the difference was statistically significant.3.5 Analysis of influencing factors of HIF-1αexpression in gliomaUnivariate logistic regression analysis showed that pathological grade,necrosis,enhancement mode,CBF_max,rCBF_max,ADC_min and rADC_min were correlated with HIF-1αexpression.Statistically significant factors in univariate analysis were taken as covariables,and age and gender were added to establish multivariate stepwise regression model.The results showed that:CBF_max and ADC_min were correlated with HIF-1αexpression（P<0.05）,and were independent influencing factors of HIF-1αexpression.Conclusion:1.There were significant differences in tumor necrosis volume,enhancement pattern and pathological grade between high and low HIF-1αexpression groups.HIF-1αhigh expression was more common among tumor necrosis volume≥50%,ring enhancement and high grade pathology.2.The values of CBF_max,rCBF_max,ADC_min and rADC_min were significantly different between HIF-1αlow expression group and HIF-1αhigh expression group.HIF-1αwas positively correlated with the values of CBF_max and rCBF_max,but negatively correlated with the values of ADC_min,rADC_min and ADC_mean.3.CBF_max,ADC_min were independent influencing factors of HIF-1αexpression.The diagnostic efficacy of combined CBF_max-ADC_min evaluation of HIF-1αexpression was higher than that of CBF_max and ADC_min evaluation of HIF-1αexpression alone.