GFR Was Measured In A Unilateral Ureteral Obstruction Model In Pigs Using A Modified Rutland-Patlak Model

PurposeTo perform low dose dynamic contrast enhanced?DCE? magnetic resonance renography?MRR? on healthy and operatively established unilateral ureteral obstruction?UUO? model swine, then compare the modified Rutland-Patlak model with the classical Rutland-Patlak model to determine their accuracy of evaluating glomerular filtration rate of split kidney then try to explain it with the pathological analysis. Materials and methods20 male Chinese experimental pigs nurtured by the experimental animal breeding center of China Agricultural University were enrolled in this study, those pigs were subdivided into unilateral ureteral obstruction?UUO? group?n=10? and normal control group?n=10? according to the table of random numbers. During the experimental procedure one pig of the UUO group lost its dynamic nuclear renography result because of the bolus injection failure, one died from the tracheal rupture when establishing the endotracheal intubation pathway and one did not successfully established the UUO model and the rest 7 pigs of the UUO group finished DCE-MRI and dynamic nuclear renography examination. Two pigs of the normal control group did not finish the MRI examination because of the metal artifact in the intestine?metal residue mixed in the feed stuff? and the rest 8 pigs finished all of the examination. The subjects of the UUO group underwent renal surgery to establish the unilateral ureteral obstruction model and feed for one more weeks to wait for the hydronephrosis of the operated kidney and then performed the data acquisition. Both the UUO group and the normal control group subjects were venously anesthetized by liquid to undergo 99mTc-DTPA dynamic nuclear renography and low dose Gd-DTPA dynamic magnetic resonance renography?MRR? examination. For the dynamic nuclear renography, a dose of 10 m Ci?370 MBq?99mTc-DTPA was injected intravenously through the marginal ear vessel by bolus injection to perform the dynamic nuclear renography. MRR was performed on a 3.0 Tesla superconductor magnetic resonance scanner with 8-channel cardiac phased array coil using the 3D liver acceleration volume acquisition?LAVA? protocol to acquire the coronal dynamic contrast enhanced kidney image series. The MRR was accomplished in a breath-holding way to avoid the motion artifact. Before the MRR acquisition a vein passage was established through the marginal ear vessel and a detaining needle was imbedded in this vessel. A dose of 0.04 mmol/kg Gd-DTPA was injected by a high pressure syringe with a velocity of 3 m L/s bolus injection. The MRR acquisition was performed immediately after the Gd-DTPA injection and 16 continuous phases of LAVA acquisition were performed, then after a 30 seconds interval acquire another phase until 5 minutes. MRR data was processed using the Rutland-Patlak model and modified Rutland-Patlak model and GFR values was calculated using the Timing Lost software which was developed by Sino-Dutch biomedical engineering college to post process the data. After all the data successfully was acquired, a 20 m L dose of 10% potassium chloride was injected through the previously established vein passage to put the experimental pigs to death and both side of the kidneys were taken out and put into 10% formaldehyde solution to have the kidneys fixed, then the specimens were stained by haematoxylin eosin?H-E? and cut into slices for optical microscope observation to study the histological and pathological features.To evaluate the accuracy of the Rutland-Patlak model and modified Rutland-Patlak model, a comparison of the Pearson correlation between the GFR_R-P?GFR calculated by the Rutland-Patlak model? and GFR_mR-P?GFR calculated by the modified Rutland-Patlak model? calculated by these two methods and the reference method dynamic nuclear renography GFRSPECT was performed and the comparison of the consistency between the GFR_R-P, GFR_mR-P and the GFRSPECT by Bland-Altman analysis. To study the pathological change of the kidney of the UUO model, student t test was performed to compare the difference of the renal interstitial portion and the diameter of renal tubules between the UUO group and the normal control group. Results1.The GFR_R-P calculated by the Rutland-Patlak model of the 30 split kidneys of 15 pigs of the UUO group and the normal control group was 46.57±12.04 m L/min, which was not significantly correlated with the reference nuclear renography GFRSPECT of 44.02±17.76 m L/min? r=0.15,P=0.426?, and with the Bland-Altman consistency analysis of 3 points out of the 95% confidence interval. The GFR_R-P of the 16 split kidneys of the 8 pigs of the normal control group was 48.55±8.09 m L/min and the reference GFRSPECT was 47.64±12.02 m L/min and these two methods had significant correlation? r=0.71,P=0.002? and all of the points were within the 95% confidence interval of the Bland-Altman analysis. But the GFR_R-P of the 14 split kidneys of the 7 pigs of the UUO group were 44.31±15.41 m L/min which was not correlated with the reference GFRSPECT of 39.89±22.42 m L/min?r=-0.08,P=0.800? and the Bland-Altman analysis of consistency with 2 points out of the 95% confidence interval.2. The GFR_mR-P of 30 split kidneys of all the 15 pigs of both the UUO group and the normal control group calculated by the modified Rutland-Patlak model was 41.64±17.27 m L/min which was significantly correlated with the reference nuclear renography GFRSPECT of 44.02±17.76 m L/min?r=0.71,P<0.001?. Bland-Altman consistency analysis of GFR_mR-P and GFRSPECT result of one point out of the 95% confidence interval. The GFR_mR-P of the 16 split kidneys of the 8 pigs of the normal control group was 49.39±14.39 m L/min and the reference GFRSPECT was 47.64±12.02 m L/min and these two methods had significant correlation?r=0.84,P<0.001? and Bland-Altman analysis with one point out of the 95% confidence interval. The GFR_mR-P of the 14 split kidneys of the 7 pigs of the UUO group were 32.77±16.38 m L/min which was well correlated with the reference GFRSPECT of 39.89±22.42 m L/min?r=0.66,P=0.010? and the Bland-Altman analysis of consistency with one point out of the 95% confidence interval.3. The renal interstitial portion of normal kidneys were about 11.36% and that of the UUO kidneys were about 25.00% with statistical significance?P=0.00?. The diameter of renal tubules of the normal group was 38.5±2.75 ?m and the UUO group of 47.6±3.23 ?m and this difference was significant?P=0.00?. The proportion of Bowman's capsule to the renal corpuscle of normal control group was 40.87% and that of the UUO group was 44.70%, which was not significantly different?P=0.43?.Conclusion1. The GFR_R-P calculated by Rutland-Patlak model for normal control model showed good correlation and consistency with the reference method of nuclear renography. For unilateral ureteral obstruction model the Rutland-Patlak model showed poor correlation and consistency with the reference dynamic nuclear renography which indicated that the Rutland-Patlak model was not suitable for unilateral ureteral obstruction model.2. The GFR_mR-P calculated by the modified Rutland-Patlak model showed good correlation and consistency with the reference dynamic nuclear renography for both the normal control group and the unilateral ureteral obstruction swine model.3. The apparently dialated renal tubules and the significantly enlarged renal interstitial portion might be the possible reason for the bias of calculated GFR by the Rutland-Patlak model and the modified Rutland-Patlak model was more suitable for the obstructive kidney disease.