Diffusion Study Of Chronic Brain Damage In Type2Diabetes Mellitus Using Multi-component Diffusion Weighted Imaging

Purpose：To quantitative measure and analyze the white matter and gray nucleus inT2DM using multi-component diffusion weighted imaging. To explore the value ofexponential model, biexponential model and stretched-exponential model in ChronicBrain Damage and evaluate the correlation between cerebral abnormalities and bloodglucose. To find the quantitative diagnosis index of T2DM with cerebral abnormalitiesin early stage.Method and Material：This is a prospective Study. Approval of Ethics Committee was obtained. Subjects was divided into three groups: A, healthy volunteer(male=8female=7,60.43±2.61years old). B, T2DM patients with negativeMRI appearance(male=7female=8,60.67±1.67years old). C, T2DM patients with lacunar cerebral infarction, leukoaraiosis and/or encephalatrophy(male=6female=9,61.17±1.13years old). All subjects underwent conventional MRI examinations and multi-component diffusion weighted imaging（0-3000s/mm2altogether11b values）. Standard ADC value（ADCstan）,fast diffusion value（Df）,slow diffusion value（Ds）, Fraction of fast ADC，（f）, distributed diffusion coefficient（DDC） and stretched-exponential（a value）of bilateral anterior limbof internal capsule, posterior limb of internal capsule, lentiform nucleus, subcortex white matter of blood-supply of ACA，MCA，PCA were measured. Statistics package for social science (SPSS)17.0was used. ANOVA was used to compare the difference between ROIs and groups. Receiver operating characteristic curve (ROC) was used to determine the diagnosis ability of DWI parameters in T2DM with cerebral abnormalities. Pearson coefficient correlation analysis was used to analyze the relation between DWI parameters and blood glucose indexs. All Measurement data are shown as mean value and standard deviation. P <0.05was deemed as showing statistical significance.Results1. exponential modelCompared to group A：ADCstan of anterior limb of internal capsule, posterior limb ofinternal capsule, white matter of ACA，MCA，PCA were significantly increased ofgroup B and C（P＜0.05）. AUC of ADCstan of white matter of ACA was0.944whenthe cutoff value set as5.735×10-4mm2/s, and the sensitivity and specificity were93.3%and86.7%.2. biexponential model2.1Df valueCompared to group A：Df value of lentiform nucleus, subcortex white matter ofblood-supply of ACA，MCA，PCA were significantly decreased of group B and C（P＜0.05）. AUC of Df of lentiform nucleus was0.886when the cutoff value set as2.955×10-3mm2/s and the sensitivity and specificity were93.3%and86.7%.2.2Ds valueCompared to group A：Ds value of anterior limb of internal capsule, posterior limb ofinternal capsule, lentiform nucleus, subcortex white matter of blood-supply of ACA，MCA，PCA were significantly increased of group B and C（P＜0.05）. AUC of Ds ofwhite matter of ACA was0.987when the cutoff value set as3.905×10-4mm2/s andthe sensitivity and specificity were93.3%and100%.2.3f valueCompared to group A：f value of anterior limb of internal capsule, posterior limb ofinternal capsule, lentiform nucleus, subcortex white matter of blood-supply of ACA， MCA，PCA were significantly increased of group B and C（P＜0.05）. AUC of Ds ofwhite matter of PCA was0.938when the cutoff value set as0.373and the sensitivityand specificity were86.7%and90.0%.3stretched-exponential model3.1DDC valueCompared to group A：DDC value of anterior limb of internal capsule, posterior limbof internal capsule, lentiform nucleus, subcortex white matter of blood-supply ofACA，MCA，PCA were significantly increased of group B and C（P＜0.05）. AUCof DDC of white matter of ACA was0.938when the cutoff value set as6.886×10-4mm2/s and the sensitivity and specificity were93.3%and90.0%.3.2a valueCompared to group A：a value of subcortex white matter of blood-supply of ACA，MCA，PCA were significantly decreased of group B and C（P＜0.05）. AUC of a valueof white matter of PCA was0.860when the cutoff value set as0.928and thesensitivity and specificity were76.0%and83.3%.4correlation analysisCorrelation was found between DWI parameters of posterior limb of internal capsule,lentiform nucleus, subcortex white matter and HbA1C（P＜0.05）. Positive correlationwas found between ADCstan, Ds, f,DDC of white matter and HbA1C(r=0.212~0.521).Negative correlation was found between Df, a and HbA1C(r=-0.413~-0.141).Conclusion1.Parameters of multi-component diffusion weighted imaging based on exponentialmodel, biexponential model and stretched-exponential model can be used toquantitative evaluate the Chronic Brain Damage of T2DM.2.Compared to exponential model, biexponential model and stretched-exponentialmodel are more sensitively and accurately in Chronic Brain Damage detection ofT2DM. 3.Ds value of ACA areas is the best index in Chronic Brain Damage detection and canbe used to evaluate the state of T2DM.