HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Daghini, E. Articles by Lerman, L. O. Search for Related Content PubMed PubMed Citation Articles by Daghini, E. Articles by Lerman, L. O. Related Collections Related Article

Comparison of Mathematic Models for Assessment of Glomerular Filtration Rate with Electron-Beam CT in Pigs¹

¹ From the Division of Nephrology and Hypertension (E.D., J.D.K., L.O.L.) and Department of Physiology and Biomedical Engineering (L.J., J.A.H., J.C.R.), Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905. Received December 30, 2005; revision requested February 24, 2006; revision received March 24; accepted May 2; final version accepted July 6. Supported in part by National Institutes of Health grant nos. HL63282, DK73608, and HL77131, the Mayo Foundation, the Société de Néphrologie, Philippe Foundation and Hospices Civils de Lyon, and Università degli Studi di Pisa. Address correspondence to L.O.L. (e-mail: lerman.lilach{at}mayo.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 
Purpose: To prospectively compare in pigs three mathematic models for assessment of glomerular filtration rate (GFR) on electron-beam (EB) computed tomographic (CT) images, with concurrent inulin clearance serving as the reference standard.

Materials and Methods: This study was approved by the institutional animal care and use committee. Inulin clearance was measured in nine pigs (18 kidneys) and compared with single-kidney GFR assessed from renal time-attenuation curves (TACs) obtained with EB CT before and after infusion of the vasodilator acetylcholine. CT-derived GFR was calculated with the original and modified Patlak methods and with previously validated extended gamma variate modeling of first-pass cortical TACs. Statistical analysis was performed to assess correlation between CT methods and inulin clearance for estimation of GFR with least-squares regression analysis and Bland-Altman graphical representation. Comparisons within groups were performed with a paired t test.

Results: GFR assessed with the original Patlak method indicated poor correlation with inulin clearance, whereas GFR assessed with the modified Patlak method (P < .001, r = 0.75) and with gamma variate modeling (P < .001, r = 0.79) correlated significantly with inulin clearance and indicated an increase in response to acetylcholine.

Conclusion: CT-derived estimates of GFR can be significantly improved by modifications in image analysis methods (eg, use of a cortical region of interest).

© RSNA, 2007

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 
In patients with asymmetric renal diseases—such as renal artery stenosis, ureteral obstruction, or renal atrophy—measurement of single-kidney glomerular filtration rate (GFR) is critical because changes in the plasma creatinine level or in inulin clearance do not reflect the functional impairment of individually affected kidneys (1). The reference standard for measurement of renal function, inulin clearance, cannot be used to assess single-kidney GFR without invasive ureteral catheterization. Furthermore, alternative methods to derive an index of single-kidney function, such as renal scintigraphy, are only semiquantitative (2).

Development of minimally invasive computed tomographic (CT) techniques provided the impetus to couple thin-section CT with evaluation of organ function and blood perfusion. Previous study results have demonstrated the applicability and reproducibility of estimating GFR on electron-beam (EB) CT images with a gamma variate model; applicability and reproducibility were validated by means of comparison with a reference standard (3,4). The feasibility of the graphic model described by Patlak and Blasberg (5) for estimation of GFR from renal time-attenuation curves (TACs) has also been demonstrated (6–11); however, its application mandates further consideration of the anatomic and functional heterogeneity of the kidney. In previous studies, regions of interest (ROIs) encompassed the entire kidney, including the medulla, which is not physiologically involved in glomerular filtration (6,8,9,11). To our knowledge, GFR, as derived with the model of Patlak and Blasberg, has not been directly compared with inulin clearance (7,8,12).

The purpose of our study was to prospectively compare in pigs three mathematic models for assessment of GFR on EB CT images, with concurrent inulin clearance serving as the reference standard.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 
Animals and Inulin Clearance
The institutional animal care and use committee approved this study. In nine domestic pigs (weight range, 30–70 kg), GFR was measured in terms of inulin clearance and assessed with EB CT and three mathematic models.

The pigs were not fed for 12 hours before the study but had free access to water. They were then anesthetized with 500 mg of ketamine administered intramuscularly, intubated, and mechanically ventilated. Anesthesia was maintained with ketamine and xylazine (0.2 and 0.03 mg · kg^–l · min^–1, respectively). An intravenous bolus of heparin (5000 U) was followed by continuous infusion of heparin (1000 U/h) in saline. An 8-F guide catheter was positioned with fluoroscopic guidance in the upper abdominal aorta and used to measure blood pressure; a tracker catheter was positioned above the level of the renal arteries and used to infuse (a) saline at a rate of 1.5–2.0 mL/min or (b) the vasodilator acetylcholine at a rate of 4.5 µg · kg^–l · min^–1. A central venous pigtail catheter was placed in the superior vena cava and used for contrast medium injection.

A 40-mL primer bolus of 2% inulin was administered through the venous sheath and followed with constant infusion at a rate of 1 mL/min and with collection of urine from a bladder catheter (J.D.K.). Spontaneous contractions occur in the pig ureter (13), and it is very sensitive to manipulation (14). Therefore, ureteral catheterization was avoided to minimize functional perturbations that could interfere with measurement of GFR.

EB CT Studies
EB CT (C-150; Imatron, South San Francisco, Calif) studies were performed during respiratory suspension at end expiration. All CT levels that contained both kidneys were identified with localization scans, whereas two adjacent midhilar CT levels that contained both kidneys were selected for measurement of intrarenal vascular and tubular flow (E.D., L.J.). Urine was collected during a 10-minute control period before each EB CT flow study, and a peripheral blood sample was collected in the middle of this period.

To assess GFR, the kidneys were scanned in the multisection flow mode (50 msec per scan), resulting in two contiguous 8-mm-thick CT sections through the hilar regions of both kidneys. The field of view was 26 cm; the matrix, 360 x 360 pixels; and the resultant pixel size, 0.72 mm². Scanning was initiated 3 seconds after bolus injection of the nonionic low-osmolar contrast medium iopamidol (Isovue-370; Squibb Diagnostics, Princeton, NJ) (concentration, 0.5 mL/kg; rate of injection, 15 mL/sec; osmolarity, 796 mOsm/kg water) into the central venous catheter with a power injector.

Forty consecutive scans were performed over the preselected levels: The first 20 scans were performed at a rate of one scan every 0.6–2.5 seconds. The second 20 scans were performed at 6–8-second intervals. Total scanning time was 3 minutes. Animals received assisted ventilation during the second scanning period.

After a 15-minute recovery period, a 20-minute infusion of acetylcholine (4.5 µg · kg^–l · min^–l) (3,15) was initiated in six pigs to induce vasodilatation and increase GFR. During the last 10 minutes of this infusion, urine and blood samples were again collected, and EB CT was repeated (3).

Renal Volume
The flow study was followed by a volume study, in which the kidneys were scanned from pole to pole in the continuous volume scanning mode (6-mm section thickness). Timing for initiation of scanning relative to contrast material injection was predetermined from findings of the preceding flow study (that characterized bolus dynamics) to correspond to bolus arrival in the kidney. Scanning was performed during a short central venous infusion of iopamidol (0.5 mL/kg over 5 seconds) to sustain corticomedullary differentiation (vascular phase) throughout volume scanning (16,17) while mostly preceding glomerular filtration of contrast material.

Image Analysis
Reconstructed images were displayed on a workstation (Ultra-80; Sun Microsystems, Mountain View, Calif) with the Analyze (Biomedical Imaging Resource, Mayo Clinic, Rochester, Minn) software package. ROIs were traced in the opacified peripheral zone representing the cortex and in the darker enclosed area of the medulla (3,18–20). The outer boundaries of both the cortex and the medulla were used to define the region representing the entire kidney (Fig 1).

View larger version (85K): [in this window] [in a new window] [Download PPT slide]   Figure 1a: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

View larger version (85K): [in this window] [in a new window] [Download PPT slide]   Figure 1b: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 1c: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 1d: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 1e: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 1f: Analysis of CT images was performed with the original and the modified Patlak methods. EB CT images of a transverse section of the same kidney, with an ROI drawn on (a) the entire kidney, including the cortex and medulla, and (b) only the opacified cortex. TACs derived from the ROI, including (c) both the cortex and the medulla and (d) only the cortex. = Data included in the corresponding Patlak plot, = data not included in the corresponding Patlak plot. Patlak graphical plots obtained by using (e) all points from the entire kidney as originally described or (f) only the cortical points that precede the beginning of contrast medium displacement toward the medulla.

For functional measurements, ROIs were selected by two observers working independently (L.J., J.A.H.). These observers manually traced the aorta, cortex, and whole kidney (Fig 1) on vascular phase images, which show the best corticomedullary differentiation (21). The CT numbers (Hounsfield units) of each ROI were averaged for each of the 40 consecutive images and tabulated. After subtraction of background attenuation values, the increase in enhancement value from the newly-defined baseline was replotted as the TAC, describing the attenuation change subsequent to transit of contrast material through the regional vascular compartments, tubular compartments, or both (Figs 1, 2).

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 2a: Graphs were obtained by analyzing CT images with the gamma variate method. TACs obtained in a cortical ROI (a) similar to that in Figure 1 were subsequently fitted with (b) a gamma variate curve algorithm. The series of peaks represents displacement of the contrast medium in the cortical vascular, proximal tubular, and distal tubular compartments.

View larger version (15K): [in this window] [in a new window] [Download PPT slide]   Figure 2b: Graphs were obtained by analyzing CT images with the gamma variate method. TACs obtained in a cortical ROI (a) similar to that in Figure 1 were subsequently fitted with (b) a gamma variate curve algorithm. The series of peaks represents displacement of the contrast medium in the cortical vascular, proximal tubular, and distal tubular compartments.

In previous studies in which the original Patlak plot was used, the ROI encompassed the entire kidney observed on a CT image (Fig 1). The advantage of such an approach is that it complies with the assumption of the Patlak method that the indicator must stay in the ROI during the relevant time (5). Therefore, this type of plot depicts all the renal tissue attenuation measurements starting at the time of contrast agent appearance. Contrarily, for the modified Patlak method, the ROI included only the cortex (Fig 1), with sampling discontinued after the peak of the proximal tubular curve, when tubular flow of the contrast material toward the medulla results in indicator outflow from the ROI. The time element was corrected for the delay between the arrival of contrast material in the aorta and the kidney, as previously described (8).

Assessment of GFR with Extended Gamma Variate Modeling
GFR was also evaluated (E.D., L.J., J.A.H.) with parameters derived from extended gamma variate modeling of the cortical TAC, as previously validated (3,4). Briefly, this TAC typically exhibits three sequential peaks, consistent with transit of the contrast agent bolus in the cortical vascular, proximal tubular, and distal tubular compartments (Fig 2). The TAC was fitted by an extension of the standard gamma variate curve-fitting algorithm, as previously shown (Fig 2) (3). Although all 40 scans were needed for application of the extended gamma variate model, GFR (measured in milliliters per minute per cubic centimeter of cortex) was derived from the initial scans that defined the slope of the proximal tubular curve, which represents the rate of contrast medium accumulation subsequent to glomerular filtration. The aortic TAC was fitted by using a standard gamma variate (21).

For all three methods (original and modified Patlak graphical methods and extended gamma variate analysis), absolute GFR (measured in milliliters per minute) was assessed by multiplying the normalized GFR by the corresponding volume (whole kidney or cortex) to enable comparison with inulin clearance.

Data Analysis
Blood and urine samples were analyzed (J.D.K.) with a spectrophotometer for inulin concentration and with a graduated cylinder for urinary flow rate. Inulin clearance (C) for both kidneys was calculated as follows: C = UV/P, where U is urinary inulin concentration, V is urinary flow rate, and P is inulin plasma concentration. We assumed that in healthy subjects the relative contribution of each kidney would be approximately proportional to its size; thus, inulin clearance was calculated in each kidney as the fractional volume (right or left renal volume divided by their sum) multiplied by total inulin clearance, as previously shown (3,4,12).

Statistical Analysis
Results are presented as the mean ± the standard error of the mean. The GFR values estimated with the different methods were compared within the group with a paired t test and considered significant if the P value was less than .05. The correlation and agreement between CT methods and inulin clearance in estimating GFR were tested by two authors working independently (E.D., L.J.) by using least-squares regression analysis and Bland-Altman graphical representation. A standard statistical software package (StatView 5.1; SAS Institute, Cary, NC) was used for analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 
The images showed excellent corticomedullary differentiation owing to the high cortical enhancement during the vascular phase (Fig 1).

The vascular, proximal, and distal tubule peaks were smaller in TACs obtained in the entire kidney ROI than in TACs obtained in the cortical ROI, and this resulted in a substantial difference in slope in the Patlak analysis (Fig 1). The TACs obtained in the same cortical ROI were successfully fitted with the gamma variate method (Fig 2).

Basal GFR obtained by means of inulin clearance was underestimated with the original Patlak method (40.1 mL/min ± 2.2 vs 26.6 mL/min ± 2.2, P = .001) and overestimated with both the modified Patlak method and the gamma variate analysis (69.4 mL/min ± 4.4 and 51.3 mL/min ± 2.2, respectively; P < .05 for both). No significant correlation was observed between inulin clearance and GFR obtained with the original Patlak method (r = 0.12, P = .62; Fig 3). In contrast, GFR obtained with the modified Patlak method and with extended gamma variate modeling correlated significantly with inulin clearance (r = 0.75, P < .001 and r = 0.79, P < .001, respectively; Fig 3). Bland-Altman analysis revealed agreement between GFR derived by means of calculation of inulin clearance and GFR derived with either (a) the modified Patlak method (mean difference, –29.26 mL/min ± 10.76; Fig 4) or (b) the gamma variate model (mean difference, –12.47 mL/min ± 8.04; Fig 4), with some overestimation of GFR at high filtration rates. On the other hand, the original Patlak method showed consistent underestimation of GFR (Fig 4).

View larger version (8K): [in this window] [in a new window] [Download PPT slide]   Figure 3a: Statistical analysis was performed with least-squares regression analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. Graphs show correlation between inulin clearance and GFR derived with the (a) original Patlak (r = 0.12, P = .62), (b) modified Patlak (r = 0.75, P < .001), and (c) extended gamma variate (r = 0.79, P < .001) methods. The thin line is the regression line, and the thicker line that runs through zero is the line of identity.

View larger version (8K): [in this window] [in a new window] [Download PPT slide]   Figure 3b: Statistical analysis was performed with least-squares regression analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. Graphs show correlation between inulin clearance and GFR derived with the (a) original Patlak (r = 0.12, P = .62), (b) modified Patlak (r = 0.75, P < .001), and (c) extended gamma variate (r = 0.79, P < .001) methods. The thin line is the regression line, and the thicker line that runs through zero is the line of identity.

View larger version (9K): [in this window] [in a new window] [Download PPT slide]   Figure 3c: Statistical analysis was performed with least-squares regression analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. Graphs show correlation between inulin clearance and GFR derived with the (a) original Patlak (r = 0.12, P = .62), (b) modified Patlak (r = 0.75, P < .001), and (c) extended gamma variate (r = 0.79, P < .001) methods. The thin line is the regression line, and the thicker line that runs through zero is the line of identity.

View larger version (8K): [in this window] [in a new window] [Download PPT slide]   Figure 4a: Statistical analysis was performed with Bland-Altman analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. The y-axis is the difference between GFR values obtained by using the CT algorithm and inulin clearance; the x-axis is their average. Bland-Altman graphs show agreement between inulin clearance and GFR derived with (a) the original Patlak method, (b) the modified Patlak method, and (c) EB CT modeling.

View larger version (8K): [in this window] [in a new window] [Download PPT slide]   Figure 4b: Statistical analysis was performed with Bland-Altman analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. The y-axis is the difference between GFR values obtained by using the CT algorithm and inulin clearance; the x-axis is their average. Bland-Altman graphs show agreement between inulin clearance and GFR derived with (a) the original Patlak method, (b) the modified Patlak method, and (c) EB CT modeling.

View larger version (9K): [in this window] [in a new window] [Download PPT slide]   Figure 4c: Statistical analysis was performed with Bland-Altman analysis of GFR values obtained with the original and modified Patlak methods and the gamma variate method; these values were compared with GFR values obtained by means of inulin clearance. The y-axis is the difference between GFR values obtained by using the CT algorithm and inulin clearance; the x-axis is their average. Bland-Altman graphs show agreement between inulin clearance and GFR derived with (a) the original Patlak method, (b) the modified Patlak method, and (c) EB CT modeling.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 
Our findings show that modifications that account for renal anatomy and physiology can improve the utility of CT techniques for minimally invasive estimation of the single-kidney GFR.

With the Patlak graphical method, which was originally developed to measure blood-to-brain transfer constants (22), several assumptions that needed to be met for valid implementation of this method were outlined. Patlak and Blasberg (5) subsequently generalized this method and extended its applicability to other organs. Nevertheless, if the Patlak method is to be acceptable for measurement of GFR, special consideration should be given to regional renal structure and function and to flow dynamics within the kidney, as pointed out by Wolf (23).

Implicit in the application of the Patlak method is the assumption that the indicator (contrast agent or radionuclide) is irreversibly trapped in the system, at least during the time of analysis (5). This assumption is probably met when the entire kidney is scanned for the 2–3 minutes required for renal transit of a bolus of contrast material (24), as with the original Patlak method. Nevertheless, this approach has several limitations. The complex three-dimensional tubular anatomy could lead to the contrast agent leaving the CT plane or entering it after filtration in adjacent sections (23). Moreover, concentration or dilution processes, as well as medullary geometric architecture, could modulate contrast medium concentration and thereby lead to over- or underestimation of the glomerular filtration. This notion is underscored by the demonstration of Patlak slopes in the medulla (7), which is not involved in glomerular filtration. Therefore, inclusion of the entire kidney in a Patlak plot may lead to sampling of unrelated dynamic processes. Hence, Wolf (23) proposed to circumvent this problem by using only a cortical ROI in conjunction with a higher sampling rate.

It is important to note that because the Patlak method is a two-compartment technique, the third space—which is represented in the kidney by the interstitium—is neglected. The amount of contrast material that leaks into the interstial space is indistinguishable from the amount that is still in the vascular space and usually is included in the calculation of the GFR (9). However, the interstitial space in the cortex is smaller than that in the medulla. Moreover, the flow of contrast material to the interstitium occurs particularly during bolus arrival, when the interstitial space is relatively devoid of contrast material. In later phases (such as those sampled with the original Patlak method), backflow from the third compartment may be detectable when the concentration of contrast material tends to decline in the intravascular compartment, thus contributing to an artifactual decrease in the estimated GFR (25). On the other hand, it is possible that in the early phase (sampled when the cortical ROI was used) this phenomenon contributes to the slight overestimation of GFR, which is observed with both the modified Patlak method and the gamma variate model as compared with the estimation of GFR with the reference standard.

Moreover, because one of the central assumptions of the Patlak method pertains to the unidirectional flow of contrast material that remains in the ROI (10), in this study the time of analysis was limited to that corresponding to filtration of contrast material from the glomerulus to the proximal tubule (ie, less than 40 seconds) (3,26). Longer cortical sampling would oppose the requirement for unidirectional flow and retention as the contrast material bolus starts its tubular flow.

Rapid sampling is essential to portray the glomerular filtration of the contrast material prior to contrast agent outflow from the cortex, which invalidates this approach. In particular, although a high injection rate is essential to temporarily trapping the contrast agent in the ROI, depicting the cortical TAC, and selecting the sampling period, it is also associated with rapid (less than half a second) vascular attenuation changes, which might be averaged and missed by using long scanning times. The short acquisition time (50 msec per scan) available with EB CT ensures adequate depiction of both the TAC and the bolus dynamics in the aorta, cortical vasculature, and tubules. Nevertheless, acquisition times of less than 1 second are possible with newer multidetector CT scanners and may be adequate for this purpose.

Sampling renal attenuation for periods of time corresponding to tubular flow beyond the proximal tubule also may have been partly responsible for the poor correlation between inulin clearance and GFR obtained with the original Patlak method in our study. Inclusion of late points corresponding to contrast material located predominantly in the distal nephron tends to decrease the Patlak slope and estimated GFR. Moreover, for validation, we extended the range of GFR values by using different-sized kidneys and vasodilation. However, acetylcholine is a diuretic and dilutes the distal nephron fluid concentration, thereby decreasing cortical attenuation at the corresponding time points, which might further reduce the Patlak slope and estimated GFR. Although the use of an acute challenge provides robust validation, it may also account for the lower rate of correlation between GFR calculated with the original Patlak method and renal clearance observed in the current study as compared with the rate of correlation in previous studies (8,12) in which the kidney was not acutely challenged and creatinine clearance was the reference standard.

In contrast to the GFR calculated with the original Patlak method, GFR measurements obtained in the cortex with the modified Patlak method and the gamma variate model correlated significantly with inulin clearance, and a physiologic increase in GFR was detected. Nonetheless, the Bland-Atlman plot (27), which was used to examine concordance between two measurements of the same physiologic parameter, showed some overestimation of the GFR at high filtration rates with both CT methods when compared with inulin clearance. This might have been due to a concurrent increase in tubular contrast agent concentration secondary to filtration and proximal fluid reabsorption. Overestimation was greater with the modified Patlak method, possibly because it samples intravascular contrast material washing out of the cortex; this is mathematically excluded with gamma variate modeling. Alternatively, the different time periods assayed with inulin clearance as compared with CT assessment (10 minutes vs 10–20 seconds) may unveil some intrinsic temporal variability in GFR (28) or inaccuracy in inulin clearance methods (29).

In our study, only two levels were selected to represent each kidney because of the relative uniformity of pig kidneys. However, multidetector EB CT can be used to scan up to eight CT levels (approximately 8 cm of renal tissue) almost simultaneously; thus, evaluation of segmental renal disease may be feasible. EB CT measurements of renal volumes have been shown to agree with volume displacement (16), and to minimize volume-averaging errors, the 6-mmthick sections were selected for analysis at the hilar region (21), where renal curvature is the lowest. Furthermore, single-kidney GFR was calculated for individual kidneys by estimating their fractional volumes, an assumption that is generally acceptable (3,4,12) but not invariably valid. Nevertheless, although single-kidney GFR (measured in milliliters per minute) allows comparison with reference standards, the value of regional GFR (measured in milliliters per minute per unit of tissue, as derived directly from the TAC) is often used and might be useful for monitoring physiologic or clinical conditions.

In conclusion, by taking into account renal physiologic characteristics, the modifications proposed in this study may enhance the reliability of CT methods for assessment of GFR. Despite greater overestimation of GFR, the relative simplicity and potential clinical usefulness of the modified Patlak method may offer some advantage over extended gamma variate modeling, which may be more accurate but is also more time consuming. Further studies are needed to test this method in humans and validate the utility of alternative modalities (eg, helical CT, magnetic resonance imaging, and positron emission tomography) for this approach in the assessment of single-kidney GFR. On the other hand, limitations of the CT methods include overestimation of GFR, radiation exposure, and potential contrast material nephrotoxicity. However, advances in CT technology may permit reduction of the radiation dose, and use of alternative contrast agents (eg, particulate) may decrease the risk of nephrotoxicity.

Practical application: CT techniques could be an important clinical tool for estimation of single-kidney GFR. Despite some overestimation of GFR, extended gamma variate modeling provided the strongest agreement with inulin clearance. The implementation of this method is more complex and requires a longer data collection period, but it is capable of providing simultaneous additional important assessments, such as those of renal perfusion and tubular fluid concentration (3,21,26,30,31). The Patlak method is relatively simple: It may require less contrast medium and fewer scans, and it could theoretically be integrated into clinical practice in conjunction with conventional CT. Moreover, although central venous bolus injections of contrast material similar to those used in the current study have been used in humans (30–32), the Patlak method may be compatible with peripheral injections and therefore less invasive. Of note, arterial catheterization was performed in this study only for the purpose of blood pressure measurement and acetylcholine infusion, which would not be required in clinical practice.

	ADVANCE IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 

• Our study results show that modifications that account for renal anatomy and physiology can improve the utility of CT techniques for minimally invasive estimation of single-kidney glomerular filtration rate.
  

	FOOTNOTES

 

Abbreviations: EB = electron beam • GFR = glomerular filtration rate • ROI = region of interest • TAC = time-attenuation curve

² Current address: Hospices Civils de Lyon et Université Claude Bernard Lyon 1, Hôpital Edouard Herriot, Lyon, France

Authors stated no financial relationship to disclose.

See also Science to Practice in this issue.

Author contributions: Guarantor of integrity of entire study, L.O.L.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, E.D., L.J., L.O.L.; experimental studies, E.D., L.J., J.D.K., J.C.R., L.O.L.; statistical analysis, E.D., L.J., J.A.H., J.D.K.; and manuscript editing, all authors

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCE IN KNOWLEDGE References

 

1. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39(2 suppl 1):S1–S266.[CrossRef][Medline]
2. Peters AM. The kinetic basis of glomerular filtration rate measurement and new concepts of indexation to body size. Eur J Nucl Med Mol Imaging 2004;31:137–149.[CrossRef][Medline]
3. Krier JD, Ritman EL, Bajzer Z, Romero JC, Lerman A, Lerman LO. Noninvasive measurement of concurrent single-kidney perfusion, glomerular filtration, and tubular function. Am J Physiol Renal Physiol 2001;281:F630–F638.[Abstract/Free Full Text]
4. Lerman LO, Ritman EL, Pelaez LI, Sheedy P II, Krier JD. Quantification of single-kidney glomerular filtration rate with electron beam computed tomography. In: Chen CT, Clough AV, eds. Proceedings of SPIE: medical imaging 2000—physiology and function from multidimensional images. Vol 3978. Bellingham, Wash: International Society for Optical Engineering, 2000;1(23):539–546.
5. Patlak CS, Blasberg RG. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data: generalizations. J Cereb Blood Flow Metab 1985;5:584–590.[Medline]
6. Dawson P, Peters AM. Functional imaging in computed tomography: the use of contrast-enhanced computed tomography for the study of renal function and physiology. Invest Radiol 1993;28(suppl 5):S79–S84.
7. Miles KA, Leggett DA, Bennett GA. CT derived Patlak images of the human kidney. Br J Radiol 1999;72:153–158.[Abstract]
8. Tsushima Y, Blomley MJ, Kusano S, Endo K. Use of contrast-enhanced computed tomography to measure clearance per unit renal volume: a novel measurement of renal function and fractional vascular volume. Am J Kidney Dis 1999;33:754–760.[Medline]
9. Hackstein N, Puille MF, Bak BH, Scharwat O, Rau WS. Measurement of single kidney contrast media clearance by multiphasic spiral computed tomography: preliminary results. Eur J Radiol 2001;39:201–208.[CrossRef][Medline]
10. Hackstein N, Bauer J, Hauck EW, Ludwig M, Kramer HJ, Rau WS. Measuring single-kidney glomerular filtration rate on single-detector helical CT using a two-point Patlak plot technique in patients with increased interstitial space. AJR Am J Roentgenol 2003;181:147–156.[Abstract/Free Full Text]
11. Hackstein N, Wiegand C, Rau WS, Langheinrich AC. Glomerular filtration rate measured by using triphasic helical CT with a two-point Patlak plot technique. Radiology 2004;230:221–226.[Abstract/Free Full Text]
12. Tsushima Y, Blomley MJ, Okabe K, Tsuchiya K, Aoki J, Endo K. Determination of glomerular filtration rate per unit renal volume using computerized tomography: correlation with conventional measures of total and divided renal function. J Urol 2001;165:382–385.[CrossRef][Medline]
13. Achar E, Achar RA, Paiva TB, Campos AH, Schor N. Amitriptyline eliminates calculi through urinary tract smooth muscle relaxation. Kidney Int 2003;64:1356–1364.[CrossRef][Medline]
14. Roshani H, Dabhoiwala NF, Dijkhuis T, Pfaffendorf M, Boon TA, Lamers WH. Pharmacological modulation of ureteral peristalsis in a chronically instrumented conscious pig model. I. Effect of cholinergic stimulation and inhibition. J Urol 2003;170:264–267.
15. Daghini E, Chade AR, Krier JD, Versari D, Lerman A, Lerman LO. Acute inhibition of the endogenous xanthine oxidase improves renal hemodynamics in hypercholesterolemic pigs. Am J Physiol Regul Integr Comp Physiol 2006;290:R609–R615.[Abstract/Free Full Text]
16. Lerman LO, Bentley MD, Bell MR, Rumberger JA, Romero JC. Quantitation of the in vivo kidney volume with cine computed tomography. Invest Radiol 1990;25:1206–1211.[CrossRef][Medline]
17. Lerman LO, Bentley MD, Bell MR, Rumberger JA, Sheedy PF 2nd, Romero JC. The effect of a low-osmolar radiographic contrast medium on in vivo and postmortem renal size. Invest Radiol 1991;26:992–997.[Medline]
18. Lee VS, Rusinek H, Noz ME, Lee P, Raghavan M, Kramer EL. Dynamic three-dimensional MR renography for the measurement of single kidney function: initial experience. Radiology 2003;227:289–294.[Abstract/Free Full Text]
19. Thoeny HC, De Keyzer F, Oyen RH, Peeters RR. Diffusion-weighted MR imaging of kidneys in healthy volunteers and patients with parenchymal diseases: initial experience. Radiology 2005;235:911–917.[Abstract/Free Full Text]
20. van den Dool SW, Wasser MN, de Fijter JW, Hoekstra J, van der Geest RJ. Functional renal volume: quantitative analysis at gadolinium-enhanced MR angiography—feasibility study in healthy potential kidney donors. Radiology 2005;236:189–195.[Abstract/Free Full Text]
21. Lerman LO, Bell MR, Lahera V, et al. Quantification of global and regional renal blood flow with electron beam computed tomography. Am J Hypertens 1994;7(9 pt 1):829–837.[Medline]
22. Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 1983;3:1–7.[Medline]
23. Wolf GL. Using enhanced computed tomography to measure renal function and fractional vascular volume. Am J Kidney Dis 1999;33:804–806.[Medline]
24. Lerman LO, Rodriguez-Porcel M, Sheedy PF 2nd, Romero JC. Renal tubular dynamics in the intact canine kidney. Kidney Int 1996;50:1358–1362.[Medline]
25. Hackstein N, Cengiz H, Rau WS. Contrast media clearance in a single kidney measured on multiphasic helical CT: results in 50 patients without acute renal disorder. AJR Am J Roentgenol 2002;178:111–118.[Abstract/Free Full Text]
26. Chade AR, Rodriguez-Porcel M, Grande JP, et al. Distinct renal injury in early atherosclerosis and renovascular disease. Circulation 2002;106:1165–1171.[Abstract/Free Full Text]
27. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1(8476):307–310.[CrossRef][Medline]
28. Shipley RE, Study RS. Changes in renal blood flow, extraction of inulin, glomerular filtration rate, tissue pressure and urine flow with acute alterations of renal artery blood pressure. Am J Physiol 1951;167:676–688.[Free Full Text]
29. Gaspari F, Perico N, Remuzzi G. Measurement of glomerular filtration rate. Kidney Int Suppl 1997;63:S151–S154.[Medline]
30. Lerman LO, Flickinger AL, Sheedy PF 2nd, Turner ST. Reproducibility of human kidney perfusion and volume determinations with electron beam computed tomography. Invest Radiol 1996;31:204–210.[CrossRef][Medline]
31. Lerman LO, Taler SJ, Textor SC, Sheedy PF 2nd, Stanson AW, Romero JC. Computed tomography-derived intrarenal blood flow in renovascular and essential hypertension. Kidney Int 1996;49:846–854.[Medline]
32. Flickinger AL, Lerman LO, Sheedy PF 2nd, Turner ST. The relationship between renal cortical volume and predisposition to hypertension. Am J Hypertens 1996;9:779–786.[CrossRef][Medline]

This article has been cited by other articles:

				Y. Tsushima, A. Taketomi-Takahashi, K. Endo, E. Daghini, and L. O. Lerman Patlak Plot Analysis for Assessment of Single-Kidney Glomerular Filtration Rate with Dynamic CT Radiology, January 1, 2008; 246(1): 336 - 338. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Daghini, E. Articles by Lerman, L. O. Search for Related Content PubMed PubMed Citation Articles by Daghini, E. Articles by Lerman, L. O. Related Collections Related Article