Functional Magnetic Resonance Imaging Of Chronic Kidney Disease At 3.0T

Objective:To investigate the values of functional magnetic resonance imaging (MRI) techniques in diagnosis and stage of chronic kidney disease (CKD), and the values in evaluation the kidney function and the degree of tubulointerstitial injury. There were three parts included, the first part of blood oxygen level dependent (BOLD), the second part of diffusion-weighted imaging (DWI) and the third part of diffusion tensor imaging (DTI).Materials and Methods:Fifty-nine patients with chronic kidney disease were enrolled in this study from March 2010 to December 2010, and the serum creatinine level (Scr) were measured, the estimated glomerular filtration rate (eGFR) were calculated with MDRD equation, the stage of CKD were determined. Among the CKD patients, twenty-one received kidney tissue puncturation biopsy examination, and the score of tubulointerstitial injury degree were decided. The control group was composed of 27 healthy volunteers. All of the subjects underwent studies on 3.0T MR scanner(3.0THD-X, GE healthcare) as follows:①The first part were conventional MRI and BOLD examinations, mGRE sequence with 8 echoes were used for BOLD scan with a single breath-hold time of 18s;②The second part was DWI examinations with multiple b values, included 0、50 s/mm2,0、100 s/mm2,0、500 s/mm2 and 0、1000s/mm2 four group of b values, with a single breath-hold time of 21s for each scan;③The third part was DTI examination, with b factor of 0 and 300 s/mm2, with a single breath-hold time of 15s with 6 acquisition directions. Parameters evaluated and calculated consisted of:①Three parameters were evaluated in the first part, include cortical R2* value, medullary R2* value and medullary/cortical R2* ratio of control group and CKD patients;②Six group parameters were evaluated in the second part, include cortical and medullary ADC values with each group of b value, and cortical and medullary ADCd (pure diffusion coefficient) and Fp (perfusion fraction) obtained by biexponential fit calculation with IVIM model;③In the third part, six parameters were assessed, include cortical FA and RA values, medullary FA and RA values, and medullary/cortical FA and RA ratios. The following data were analyzed:①Independent-Samples T Test was used to compare mean R2* value between cortex and medulla in control group. One-Way ANOVA (analysis of variance) were used to compare mean cortical R2* value, medullary R2* value and medullary/ cortical R2* ratio among control group and different stage of CKD patients. Receiver operating characteristics (ROC) analysis was performed in order to evaluate the diagnostic performance of medullary/cortical R2* ratio in differentiating CKD kidney from normal kidney and different stage of CKD kidneys and extracted the optimal cutoff value;②In CKD patient group, Pearson and Spearman rank correlation coefficients were calculated to evaluate the correlations of the kidney R2* parameters with Scr, eGFR and CKD grade. In CKD patients who received kidney tissue biopsy examination, Spearman rank correlation coefficients were calculated to evaluate the correlations of the kidney R2* parameters with pathology score of tubulointerstitial injury of kidney tissue biopsy specimen;③ndependent-Samples T Tests were used to compare mean ADC value, ADCd and Fp between cortex and medulla in control group. One-Way ANOVA were used to compare mean cortical and medullary ADC value, cortical and medullary ADCd, cortical and medullary Fp among control group and different stage of CKD patients. Receiver operating characteristics (ROC) analysis was performed in order to evaluate the diagnostic performance of DWI parameters in differentiating CKD kidney from normal kidney and different stage of CKD kidneys and extracted the optimal cutoff value;④In CKD patient groups, Pearson and Spearman rank correlation coefficients were calculated to evaluate the correlations of the kidney DWI parameters with Scr, eGFR and CKD grade. In CKD patients who received kidney tissue biopsy examination, Spearman rank correlation coefficients were calculated to evaluate the correlations of the kidney DWI parameters with pathology score of tubulointerstitial injury;⑤Mean FA value and RA value between cortex and medulla in control group were compared. Mean kidney FA parameters and RA parameters were compared among control group and different stages of CKD patients by One-Way ANOVA. The diagnostic value of medullary FA and RA value in differentiating CKD kidney from normal kidney and different stages of CKD kidneys were compared by ROC under curve area Az and the optimal cutoff value were determined;⑥Pearson and Spearman rank correlation analyses were used to evaluate the correlations of the kidney DTI parameters with Scr, eGFR and CKD grade of CKD patients. Spearman rank correlation analyses were used to evaluate the correlations of the kidney DTI parameters with the score of tubulointerstitial injury.Results:The images of 27 healthy volunteers,23 minor CKD patients(CKD 1 and 2 stage),29 moderate/severe CKD patients(CKD 3～5stage) were finally analyzed. Among the CKD patients, twenty one received kidney tissue puncturation biopsy examination. The data analyses revealed:①In control group, the medullary R2* value was higher than cortical (p<0.001). There were statistical differences among the control group, minor CKD and moderate/severe CKD patients in the cortical R2* value, medullary R2* value and medullary/cortical R2* ratio(p<0.01), and the differences of medullary/cortical R2* ratio between every two groups were most significant(p<0.001). The area under the ROC curve for medullary/cortical R2* ratio in differentiating between minor CKD kidneys and normal kidneys and between minor CKD kidneys and moderate/severe CKD kidneys were 0.871,0.925 respectively, and the cutoff value were 1.858、1.577. With the threshold from ROC curve, the sensitivity and specificity were all above 86%;②Negative correlations were found between medullary R2* value with Scr level and CKD grade and also between medullary/cortical R2* ratio with Scr level and CKD grade (p<0.01). Positive correlations were found between medullary R2* value and medullary/ cortical R2* ratio with eGFR(p<0.01). Negative correlations were found between medullary R2* value and medullary/cortical R2* ratio with score of tubulointerstitial injury (p<0.05) The correlation coefficient of medullary/cortical R2* ratio was higher than medullary R2* value;③In control group, no significant difference was found between the cortical and medullary ADC50 value(p>0.05). However, the ADC100, ADC500 and ADC1000 of cortex were all higher than medulla in control group. The ADCd of cortex was higher than medulla in control group (p<0.05), however, the mean value of cortical Fp was higher than medullary but with no statistical significance (p>0.05). There were statistical differences among the control group, minor CKD and moderate/severe CKD patients in the cortical and medullary ADC value, cortical and medullary ADCd and cortical and medullary Fp (p<0.001). The differences of cortical and medullary ADC1000 value between two groups were most significant (p<0.001). Using the cortical and medullary ADC1000 value criteria to differentiate minor CKD kidneys from normal kidneys and discriminate minor CKD kidneys from moderate/severe CKD kidneys, the area under ROC curve were all larger than 0.76, with sensitivity and specificity were all above 65%;④Negative correlations were found between cortical and medullary ADC1000 value with Scr level and CKD grade (p<0.01). Positive correlations were found between cortical and medullary ADC1000 value with eGFR (p<0.01). Negative correlations were found between cortical and medullary ADC1000 value with score of tubulointerstitial injury (p<0.05);⑤The FA and RA value of medullary were higher than that of cortical in control group(p<0.001). There were statistical differences among the control group, minor CKD and moderate/severe CKD patients in the cortical FA value, medullary FA value, medullary/cortical FA ratio and cortical RA value, medullary RA value, medullary/cortical RA ratio (p<0.05). Among the six parameters mentioned above, the differences of medullary FA and medullary R.A value between two groups were most significant (p<0.001). From the ROC curve, the Az of medullary FA value were larger than medullary RA value in differentiating between minor CKD kidneys and normal kidneys and between minor CKD kidneys and moderate/severe CKD kidneys. Medullary FA value of 0.367 and 0.349 were determined as the threshold, with the sensitivity and specificity were all above 56%;⑥Negative correlations were found between medullary FA value and medullary/cortical FA ratio with Scr level and CKD grade (p<0.01). Positive correlations were found between medullary FA value and medullary/cortical FA ratio with eGFR (p<0.001). Negative correlations were found between medullary FA value and medullary/cortical FA ratio with score of tubulointerstitial injury (p<0.05)Conclusions:Functional magnetic resonance imaging of kidney could reveal the physiological characters and reflect the pathophysiological changes of CKD from different aspects, such as kidney oxygenation, water molecular diffusion isotropy and diffusion anisotropy.①In the first part, renal 3.0 T BOLD MRI demonstrated that the oxygenation level of cortex is higher than medulla in normal kidney. Medullary/ cortical R2* ratio is valuable in diagnosis of CKD and different stages of CKD. Medullary/cortical R2* ratio is a sensitive parameter to reflect the degree of kidney function and tubulointerstitial injury;②In the second part, ADCd value and Fp of cortex and medulla could be calculated with IVIM model and multiple b values. The extent of water molecular diffusion in normal cortex is higher than medulla, and the microperfusion of normal cortex is relatively high but with no significant difference between cortex and medulla. The ADC value, ADCd and Fp can indicate alternation of water diffusion and microcirculation in CKD kidneys. To some extent,the ADC1000 value of cortex and medulla could be used to diagnose and differentiate the stage of CKD and reflect kidney function and tubulointerstitial injury degree;③In the third part, it is feasible to apply DTI to kidney. The anisotropy of water diffusion in medullary is higher than cortical in normal kidney. Lower medullary FA value can reflect the decrease of water diffusion anisotropy in CKD patients. Medullary FA value can represent the kidney function and tubulointerstitial injury degree, and may provide valuable information with diagnosis and stage of CKD. Thus, DTI provides a new method for evaluating the CKD disease.