| ObjectiveTo determine the accuracy of unilateral glomerular filtration rate (GFR) measured by low dose Gd-DTPA dynamic enhanced magnetic resonance imaging (MRI) and the Patlak plot technique in left renal artery stenosis (RAS) swines models.Materials and MethodsThirty healthy swines (aged four months, mean weight21.93±2.31kg) were randomly divided into RAS group (n=20) and control group (n=10). Four swines in RAS group died from narcotic drug anaphylactic shock, pulmonary edema or asphyxia, three failed in RAS surgery, and two failed in MRI examination because of serious bowel mentallic artifacts. Two swines in control group died from pulmonary edema or asphyxia, another two failed in MRI examination because of serious bowel mentallic artifacts. Finally, eleven swines enrolled in RAS group and another six enrolled in control group. All seventeen swines underwent gadopentetate dimeglumine (Gd-DTPA) dynamic enhanced MR (DCE-MR) and99mTc-DTPA scintigraphy a week after left RAS surgery. DCE-MR accomplished by3D liver acceleration volume acquisition (LAVA) on a3.0T scanner equipped with body pahse array coil. Bolus injection of Gd-DTPA was perfomed with low dose of0.04mmol/kg at3ml/sec. Bilateral renal3D images series were acquired after initial of intravenous injection and continuously repeat for16times, the acquisition time was3.0seconds for each time, followed by one acquisition every30seconds till6min. The images were post-processed and calculated offline by a self-developed software from Northeasthern University. Firstly, to determine the optimal time window (ti-t2) for Patlak plot calculating GFR, three time windows were set up to compare with each other, namely, giving aorta enhancement peak time as ti, t2were choose as30seconds,39seconds and48seconds after injection. Secondly, to determine the appropoxiate input ROI of calculating GFR, ROIs were respectively drawn over renal cortex and parenchyma to contrast. Thirdly, to determine the optimal method of calculating GFR, two-point and fitting methods were introduced and compared as well. In summary, to estimate renal contrast agent clearance rate of bilateral kidneys by Patlak plot, ROIs were set to renal cortex and parenchyma respectively, three different time windows were chosen, and both two-point and fitting methods were tested. GFRs, which was derived from contrast agent clearance rate with haematocrit (Hct) correction. Pearson correlation analysis and Bland-Altman analysis were performed to test the correlation and the consistence between MRI results (GFRMR) and scintigraphic renograph resultes (GFRspect).Finally, to evaluate the effect of enlarged renal interstitium volume on GFRMR measurement errors, swines were sacrificed immediately after MR and radionuclide examinations, both kidneys were stained by hematoxylin-erosion, The percentage of renal interstitial volume was measured microscopicly, then Pearson correlation analysis was performed between the GFRMR-GFRspect deviation and interstitial volume percentages.Result1. All GFRMR from34kidneys in17pigs in both RAS group and control group had correlated well with GFRspect whenever t2was30s,39s or48s, however, when t2was set to39s, GFRMr (49.90±17.01ml/min) obtained from renal parenchyma with two-point method and GFRspect (50.10±20.04ml/min) had best correlation (R2=0.53, p±0) than other two time windows. B-A analysis found one value located outside the95%confidence interval. GFRMR of cortex with fitting method (48.4317.41ml/min) was also close to GFRspect and had good correlation (R2=0.45, p=0). GFRMR of cortex with two-point method was44.2515.82ml/min (R2=0.52, p=0), GFRMR of parenchyma with fitting method was54.75±18.98ml/min (R2=0.45,p=0).2. The accuracy of GFRMR in normal kidneys, RAS kidneys and their contralateral kidneys:(1) In control group (n=12), when t2was set to39s, GFRMR (46.3±18.72ml/min) obtained from renal parenchyma with two-point method was close to and correlated well with GFRspect (51.04±12.31ml/min)(R2=0.66,P<0.01). B-A analysis found all vaules were within the95%confidence interval.(2) In RAS kidneys (n=11), when t2was set to39s, GFRMR (38.33±15.46ml/min) obtained from cortex with two-point method was close to and correlated well with GFRspect (37.97±16.29 ml/min)(R2=0.76, p<0.01). B-A analysis found all vaules were within the95%confidence interval.(3) Of the contralateral kidneys in RAS group, when t2was set to30s, GFRMR (63.36±14.22ml/min) from cortex with two-point method was close to and correlated with GFRspect (61.91±23.91ml/min)(R2=0.44, p<0.05). B-A analysis found one vaule located outside the95%confidence interval.3. The effect of renal interstitial volume percentage on GFR measurement:the renal interstitiums of five mild to moderate RAS kidneys and their contralateral kidenys presented with inflammatory cells infiltration, and the percentages of interstitial volumes in RAS kidneys (18.42±2.34%) and their contralateral kidneys (15.72±3.56%) increased compared with one pair of normal kidneys (13.79±1.57%). The major change of the renal interstitium of RAS kidney in one severe RAS kidney was infiltrated by both inflammatory cells and fibrocytes, renal tubular necrosis, loss and apoptosis, which percentage of renal interstitial volume (24.98%) was increased. The contralateral kidney of severe RAS pig showed inflammatory cells infiltrating in renal interstitium too, of which percentage (15.58%) was higher than that of normal. There was no correlation between percentages of fourteen renal interstitial volumes and the GFRMR-GFRspect deviation (r=-0.035, p=0.905).Conclusion1. When the time window was set between aorta peak enhancement time and39seconds after injection, GFRMR and GFRspect had excellent agreement.2. The most optimal time window for contralateral kidneys in RAS swine shifted to30seconds.3. Renal cortex would be the optimal ROI for Patlak plot calculating GFR.4. Two-point method was better than fitting method to calculate GFR by Patlak plot.5. Histopathology of RAS kidneys demonstrated inflammatory cells infiltration and early fibosis in renal interstitium, however, the impact of the enlarged renal interstitial volume percentage on GFR measurement remained unclear. |
PDF Full Text Request