| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Diagnostic Radiology, Justus-Liebig Universität Gießen, Klinikstr. 36, 35385 Gießen, Germany.
Received April 30, 2001;
accepted after revision August 3, 2001.
Address correspondence to N. Hackstein.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Fifty adult patients without acute renal disorder were included in this study. Each patient underwent CT for clinical indications. The CT protocol consisted of an unenhanced scan and three contrast-enhanced scans 45, 75, and 105 sec after starting an injection of 120 mL of iopromide using an injection rate of 3 mL/sec. All scans included both kidneys. As a reference, plasma clearance of contrast media was determined as a slope clearance by measuring iodine concentration in eight blood specimens up to 8 hr postinjection.
RESULTS. CT clearance was calculated three times for each patient, including early CT clearance, 45-75 sec postinjection; late CT clearance, 75-105 sec postinjection; and overall CT clearance, 45-105 sec postinjection. An overall CT clearance yielded the best correlation with plasma clearance with a correlation coefficient of r = 0.84 and a regression line of y = 7.5 + 0.94x. The mean difference was -3 mL/min (95% confidence interval, -35 to 29 mL/min).
CONCLUSION. CT clearance calculated from data acquired with a minimally modified diagnostic abdominal CT protocol was well correlated with the reference method in determining contrast media clearance for patients without acute renal disorders. The presented method can be used to calculate single kidney contrast media clearance in patients receiving contrast-enhanced abdominal CT for clinical indications.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We investigated the possibility of measuring single-kidney contrast media clearance, which equals glomerular filtration rate, by contrast-enhanced CT. The proposed method, hereinafter called CT clearance, is based on a two-compartment model of the kidney known as the Patlak plot or Patlak-Rutland plot.
The method is used in nuclear medicine for background correction in renal scintigraphy and was first described by Rutland [2] in 1979. In 1983, Patlak et al. [3] described a method measuring transfer constants of the blood—brain barrier by obtaining a single location dynamic CT. In 1993, Blomley et al. [4] and Dawson and Peters [5, 6] measured renal contrast media clearance using the Patlak plot by performing dynamic CT for a single location.
We present the results of a modified Patlak plot technique. We extended the single-location dynamic CT to whole-volume scans of the kidneys. This study is a continuation of an earlier study [7] investigating the same technique. Unlike the earlier study, the present study focuses on a larger study population. Only patients without acute renal disorder were examined.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Fifty consecutive patients were included in the study (31 men and 19 women). The age range was 26-86 years (mean age, 62 years). Mean serum creatinine was 1.07 mg/dL (range, 0.7-2.2 mg/dL; reference range, 0.6-1,2 mg/dL). Mean hematocrit was 37% (range, 25-55%; reference range: men, 41-50%; women, 35-45%).
Contrast Media and Injection Techniques
Plasma clearance determination and CT scans were obtained using a single
injection of nonionic contrast media. All patients received iopromide with 300
mg I/mL (Ultravist 300; Schering, Berlin, Germany).
Injection was performed by an EnVision CT power injector (Medrad, Indianola, PA) through a peripheral venous canula placed in the forearm.
The amount of injected contrast media varied from 90 mL (n = 1), 100 mL (n = 4), 110 mL (n = 1), to 120 mL (n = 44), with an injection rate of 3 mL/sec.
Determination of Plasma Clearance
Plasma clearance was measured according to the method described by
Sapirstein et al. [8]. This
method applies to a single-shot or slope model. In this model, the test
substance is applied in a short time (as an injection). The plasma
concentration curve of the test substance follows the sum of two exponential
functions (equation 1).
where C(t) equals the iodine concentration at time t. The parameters A—D have to be individually determined from several plasma concentration measurements to describe the plasma concentration curve.
Plasma clearance of the test substance can be calculated by equation 2 in which Q equals the amount of injected iodine.
The amount of injected iodine is known from the injected volume and the iodine concentration of the contrast media. Eight heparinized blood specimens of 5 mL were drawn after contrast media injection for CT at 10, 20, and 40 min postinjection and 1, 2, 3, 4, and 8 hr postinjection. For blood withdrawals, a canula other than the one used for the injection was placed to prevent iodine contamination. Additionally, a baseline blood specimen was drawn to exclude any plasma iodine concentration before CT. The blood was centrifuged, and the iodine concentration in the plasma was measured using an X-ray fluorescence analyzer (manufactured in part by L. Kaufman, Department of Radiology, California School of Medicine, San Francisco, CA, and, in part, by EG&G Ortec, Munich, Germany). A two-exponential function was fitted on the iodine measurements yielding accurate fits (Figs. 1 and 2).
|
|
CT Clearance
The measured net attenuation of the whole kidney K(t) (H), which
is proportional to the amount of iodine in the kidney, can be expressed as the
sum of net attenuation in the vascular space B(t) (H) and net
attenuation in the nephron O(t) (H):
Net attenuation is calculated by subtracting unenhanced CT numbers from corresponding enhanced CT numbers. K(t) is the net sum of CT numbers of the kidney at time t. Mean density and the area of the left and the right kidneys are measured by drawing a region of interest (ROI) on every CT slice. K(t) is calculated by multiplying mean density (H/mm3) by area (mm2) and thickness (mm) for each kidney slice. It is assumed that the amount of contrast media in the vascular space B(t) (H) is proportional to the concentration of the contrast media in the aorta b(t) (H/mm3):
where b(t) (H/mm3) represents the net density of the aortic lumen at time t after injection; c1 (mm3) represents a constant equivalent to the vascular space. It is further assumed that the amount of contrast media filtered into the nephron is proportional to the integral of the concentration of contrast media in the aorta, which is the generally known definition of clearance.
where c2 (mm3/sec) is equivalent to the clearance from the vascular space into the nephron. Combining equations 3-5 yields equation 6:
When the integral of b(t) and K(t) are known for two time periods after injection, say at t1 and t2, the constant c2 can be calculated from the arising equation system, as shown by equations 7 and 8.
For calculation of CT clearance, a correction for hematocrit has to be made because the density measurement of the iodine concentration in the aorta b(t) (H/mm3) belongs to full blood, whereas the reference method measures plasma clearance. For that correction, hematocrit was set for all patients at 0.36%. Reasons for using that number are given in the discussion section. CT clearance (mL/min) was calculated according to equation 9.
CT Examinations
CT was performed on a Somatom Plus 4 (Siemens, Erlangen, Germany), which is
a one-row CT scanner. Patients were encouraged to drink ad lib. before the CT
examination and for as long as plasma specimens were taken. Oral contrast
material (barium sulphate for CT) was given if necessary, depending on the
clinical question. Both kidneys were scanned unenhanced and after
administration of IV contrast material during the arterial, early parenchymal,
and late parenchymal phases (diagnostic scans).
The arterial spiral was started 45 sec postinjection (range, 30-103 sec; SD, 13.3 sec), early parenchymal spiral was started 75 sec postinjection (range, 53-133 sec; SD, 15.4 sec), and late parenchymal spiral was started 106 sec postinjection (range, 76-158 sec; SD, 14.9 sec). The wide range resulted from the necessity of embedding the study protocol into the clinical CT protocol. In 20 patients, a recommendation for a whole-body CT examination was made. In the remaining 30 patients, who were referred for abdominal CT, bolus triggering was performed with a 3-sec interscan delay and a 25-mAs tube current. The threshold for bolus triggering was chosen between 100 and 200 H. The bolus-triggering scans were all obtained at the same table position and at the height of the upper renal pole to minimize table movement between the end of the bolus triggering and start of the first enhanced scan.
Scanning the whole kidney volume lasted between 10 and 15 sec, depending on the slice thickness and variation in anatomy. No pelvic kidneys were encountered in the current study. Pitch was 1.5, and rotation time, 0.75 sec. The usual abdominal CT protocol was used, with 120 kV and 98 mAs for unenhanced scans and 150 mAs for enhanced scans. Slice thickness was either 5 or 8 mm. In most patients, slice thickness was 8 mm in the unenhanced scan and 5 mm in at least one enhanced scan. However, slice thickness was adapted to the diagnostic needs of the patients.
For evaluation, ROIs were manually drawn around the left and right kidneys and the aortic lumen (Fig. 3) with Magic View software (Siemens). The border of the ROI was drawn as closely as possible around the kidney parenchyma. The renal pelvis was included in the ROI. Pathologic changes inside the kidney, such as cysts or stones, were taken into the ROI. We decided to take these parts into the ROI to simplify the drawing of the ROI because it is sometimes difficult or impossible to differentiate the pathologic processes from healthy kidney parenchyma in every scan. We took care, as was possible, to draw identical ROIs in every scan for each patient. From each kidney ROI, the mean CT number and area in square millimeters were measured. Taking the distance between two slices into account, we calculated whole organ attenuation of the right and left kidneys.
|
Aortic Density Curve
The aortic ROI was circular, and we placed it inside the aortic lumen,
taking care not to include the aortic wall, which was in many cases calcified
because of artherosclerosis. The unenhanced density of the aorta was
determined as the mean of five density measurements in the aorta at the level
of the kidneys.
The integral of the aortic density curve needed to be determined for calculation of CT clearance. The integral was estimated by linear interpolation of the missing parts of the curve (Figs. 4 and 5). Bolus tracking was performed in 30 patients (Fig. 4). In the remaining 20 patients, CT scans of the thorax and the abdomen were obtained (Fig. 5).
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Kidney Volumes
The kidney volumes for the left and right kidneys were calculated from
manually drawn ROIs for the unenhanced scan and the three enhanced scans. Mean
kidney volume was 415 mL for both kidneys.
Aortic Density Curves
In all patients, four groups of aortic density data points could be used to
calculate the integrals needed. In 30 patients, data from bolus triggering
were present; in the remaining patients, data from a thoracic scan could be
used. Additionally, the three enhanced scans provided data for the aortic
density curve. However, missing parts of the curve had to be interpolated.
Mean aortic density in the arterial scan was 334 H (45 sec postinjection); in
the early parenchymal scan, 161 H (75 sec postinjection); and in the late
parenchymal scan, 126 H (106 sec postinjection). The slope of the aortic
density curve was much flatter between the two parenchymal scans than it was
between the arterial and early parenchymal scans.
CT Clearance
CT clearance was calculated three times for one examination, including
early CT clearance from the arterial and early parenchymal scans, late CT
clearance from the early and late parenchymal scans, and overall CT clearance
from the arterial and late parenchymal scans (Figs.
6,7,8).
The highest correlation was found between overall CT clearance and plasma
clearance with Pearson's correlation coefficient (r = 0.84). An
analysis according to the method of Bland and Altman
[9] was performed, showing a
mean difference between the two methods (plasma clearance and overall CT
clearance) of -3 mL/min with a 95% confidence interval of -35 to 29 mL/min
(Fig. 9).
|
|
|
|
The mean of early CT clearance was higher (89 mL/min) than the mean of late CT clearance (74 mL/min) (Figs. 6 and 7). Thirty-six of the 50 patients had higher early CT clearance numbers than late CT clearance numbers (Fig. 10). Using the Student's t test, we found a significant (p > 0.0001) difference in interindividual early and late clearances.
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The reasons for better correlation of overall CT clearance with plasma clearance rather than correlation of early or late CT clearance with plasma clearance are twofold. First, the more contrast media are excreted into the nephron during the two enhanced scans, the less important is the measurement error concerning the measured kidney attentuation. Second, the error in estimating the integral of the aortic density curve is smaller if more portions of the aortic density curve are known, such as for overall CT clearance.
Hematocrit
For calculation of CT clearance, a correction for the hematocrit in the
aorta at the time of the CT scan needed to be made. We decided to set the
hematocrit for CT clearance calculation for all patients at 0.36%. This number
results from the normal hematocrit for adults, which is 0.42%
[10]. We subtracted 0.06%,
which is the estimated decrease of hematocrit in the first 2 min after
injection of contrast media
[11].
Measurement of Attenuation of Entire Kidney
In the present study, the entire kidney, including the renal pelvis and
focal lesions as renal cysts, was included into the kidney ROI. From a
theoretic point of view, any measurement outside the renal parenchyma that
shows enhancement (such as the renal pelvis, an abscess, or a tumor)
introduces additional interstitial and vascular spaces that do not contribute
to the renal function. Therefore, it would be advantageous in future studies
to include only renal parenchyma into the kidney ROIs. As an exception, simple
renal cysts and stones could be included because these structures do not show
enhancement.
The average measured kidney volume was higher than the expected standard value of 300 mL for both kidneys, which is explained by including the renal pelvis in the kidney ROI. The mean coefficient of variance of the measured kidney volumes ranged between 5.0% and 5.2%. The volumes are greater than those reported by Staron and Ford [12], who found a coefficient of variation of 2.05% for the area of an ROI of a single slice, including one kidney (10 measurements; mean, 17.4 cm2).
The observed differences of intraindividual kidney volume found in the present study might be explained partially by unconscious breathing movements, because the scanning of one kidney lasted approximately 10-15 sec, and partially by differences introduced by drawing the ROIs manually. The use of different slice thicknesses (5 or 8 mm) might also have played a role.
Duration of Kidney Scan
CT scans of both kidneys in the present study lasted between 10 and 15 sec.
The time needed for the scan introduces an error because the upper pole of the
kidney is scanned earlier than the lower pole, which means that lower portions
of the kidney had filtered more iodine into the nephron than the upper
portions when that part was scanned. The difference to the middle portion of
the kidney is approximately 7.5 sec. The error of the upper half might be
corrected partially by the lower half because the differences in the densities
of the upper and lower halves offset each other.
Measurement of Aortic Density Curve
No blood flow correction was performed on the aortic density measurements,
although the diagnostic scans were obtained at different table positions
because mean flow velocity of blood in the aorta is rapid (
20 cm/sec
[13]). Therefore, a correction
of the time of the measurement of the aortic density, which is dependent on
the position of the slice, would change the measurements of the thoracic aorta
by approximately 2-3 sec where the distance of blood flow is greatest (with an
estimated distance of 40-60 cm from the aortic arch to the kidneys). Adding
2-3 sec to the time of the thoracic aortic density measurement will result in
only minor changes of the calculated integrals of the aortic density
curve.
The aortic density curve is needed to calculate the integral from t = 0 to the point of kidney measurement. As previously described, the missing portions of the aortic density curves in the present study were interpolated linearly. This introduces an error, which could be minimized in the future by adding (low-dose) measurements between the diagnostic scans.
CT Protocol and Radiation Exposure
To acquire enough data to calculate CT clearance, two diagnostic enhanced
scans in the arterial and later parenchymal phases and one unenhanced scan are
needed for measuring attenuation of the kidneys. In the present study, three
enhanced scans were obtained, allowing the calculation of three CT clearance
values in one examination. As a result, it is possible to calculate CT
clearance from a CT protocol omitting the early parenchymal scan. The
radiation exposure for one CT scan of the kidneys using a collimation of 8 mm,
a pitch of 1.5, and a tube current of 150 mAs is 3.3 mSv for male patients and
3.6 mSv for female patients
[14].
The radiation exposure of renal scintigraphy with 99mTc-MAG3 ([mercaptoacetyltriglycine] 185 MBq = 1.3 mSv) or 99mTc-DMSA ([dimercaptosuccinic acid] 74 MBq = 0.6 mSv) [15] is much lower.
Other Studies Measuring Contrast Media Clearance by CT
To our knowledge, all previous studies, except those by our group
[7], investigated data from a
single-location dynamic CT, indicating repetitive scanning of one slice at the
same table position.
In 1993, Dawson and Peters [5] described a study measuring contrast media clearance per milliliter of renal tissue using single-location dynamic CT. Contrast media clearance was calculated using the Patlak plot method. Three patients with normal kidney function were studied; the calculated results of 0.35, 0.41, and 0.44 mL/min per milliliter of renal tissue were in the physiologic range of the parameter (normal kidney volume for both kidneys was 300 mL [16], and normal glomerular filtration rate for both kidneys was 120 mL/min [17]). In the present study, the ratio of overall CT clearance and kidney volume was 0.2 mL/min per milliliter of renal tissue. The lower ratio compared with the findings of Dawson and Peters resulted from including the renal pelvis into the ROIs for measuring kidney attenuation.
Inspired by the promising results of Dawson and Peters [5], other researchers have investigated single-location dynamic CT as a technique for measuring relative clearance values. In 1999, Miles et al. [18] studied 14 patients with single-location dynamic CT who had lymphoma with no known kidney disease. These researchers found a mean entire kidney permeability of 0.518 mL · min-1 · mL-1, which slightly exceeded that reported by Dawson and Peters. However, because the single-location dynamic CT technique measures renal clearance of only one CT slice, correlation with plasma urea was weak in that study (r = 0.47) [18].
In 1999, Tsushima et al. [19] measured renal clearance with the Patlak plot method by single-location dynamic CT. Multiplying relative clearance (mL/min per milliliter of renal tissue) by the kidney parenchyma volume, which was measured by an additional kidney scan, resulted in an entire kidney clearance. Twenty-four patients with diabetes, with or without renal impairment, were studied. As a reference, creatinine clearance was determined by a 24-hr urine collection and a determination of serum creatine. The correlation between CT contrast media clearance and creatinine clearance was r = 0.87, with a line of regression of y = 29.2 + 0.64x.
Drawbacks of the single-location dynamic CT method described in previous studies include three problems. First, it is difficult to reproduce exactly the same position of the kidneys between inspiration and expiration maneuvers. Second, the results of single-location dynamic CT and the Patlak analysis have to be extrapolated on the whole kidney volume. Third, it is not possible to include single-location dynamic CT into a clinical abdominal CT protocol. This means that single-location dynamic CT has to be performed as an additional examination with additional radiation and additional contrast media injection.
Interstitial Space
The Patlak plot model was developed by Patlak
[3] for measuring the transfer
constant of the blood—brain barrier. Using this two-compartment model
for measuring renal clearance, we found that the renal interstitium was
neglected, which resulted in an error depending on the size of the
interstitial space. According to Blomley and Dawson
[20], the interstitium in the
healthy kidney appears to be small. Bohle et al.
[21] measured interstitial
volume of the kidney cortex and the outer stripe of the outer medulla from
biopsies using light microscopy in 56 patients suffering from acute renal
failure and 21 control patients. The relative interstitial volume in the renal
cortex (results for the outer stripe of the outer medulla) was measured by
8.4% (17.6%) in control patients and by 16.7% (27.2%) in patients with acute
renal failure.
Interestingly, 99mTc-DTPA scintigraphy fights the same interstitium problem, which involves determining relative kidney function in the first minutes of scintigraphy because the pharmacokinetics of DTPA and nonionic low-osmolality contrast media are similar. As shown by Piepsz et al. [22], the Patlak plot, applied to 99mTc-DTPA scintigraphy, does not yield exact straight plots, as the theory predicts. The points of the Patlak plot lie on a somewhat curved line with a steeper initial slope, which equals a higher initial clearance number. The same observation could be made in the present study, where early CT clearance was higher than late CT clearance (Fig. 10). When contrast media initially permeate the kidney, the interstitium is empty so that some of the contrast media diffuse into it. The amount is dependent on the size of the interstitial space and the permeability (transfer constant) between vascular space and interstitium. Later, when the concentration of the contrast media in the vessels is decreasing, backflow of contrast media from the interstitium into the vascular space will occur. Any flow from the vascular space into the interstitium contributes to constant c2 in equation 5 (i.e., increases the calculated clearance). On the contrary, backflow will decrease the calculated clearance. Corresponding to this theoretic consideration, higher early CT clearance rather than late CT clearance values were found, as shown in the present study (Figs. 6 and 7). Acute kidney disease, similar to acute pyelonephritis, increases the interstitial volume. In these cases, the interstitium error measuring contrast media clearance by CT clearance will increase, resulting in false high CT clearance values. This was recently shown by our earlier study that included patients with acute pyelonephritis [7].
In conclusion, CT clearance correlated well with the reference method. CT clearance can be used to measure single kidney contrast media clearance in patients without acute renal disorder. Patients with acute renal disorders, such as acute pyelonephritis or acute obstruction, were excluded to avoid any problems with increased interstitial space that might lead to faulty results. In the present study, the highest accuracy of CT clearance compared with the reference method could be achieved by scanning the kidneys in the arterial (45 sec postinjection) and parenchymal phases (105 sec postinjection).
CT clearance can be calculated with minor modifications of the CT protocol in patients who are scheduled to undergo contrast-enhanced CT for clinical indications and when the determination of single kidney glomerular filtration rate is of clinical interest. Possible indications might be hypernephroma, a one-sided atrophic kidney, or malignancies in patients with potential renal damage who are planning to receive chemotherapy to monitor kidney function together with the follow-up CT.
If there is no indication for contrast-enhanced CT, renal scintigraphy or laboratory methods should be used to determine glomerular filtration rate.
Acknowledgments
We thank Schering, Berlin, Germany, for performing the iodine concentration
measurements.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Arakawa, M.P. Marks, H.M. Do, D.M. Bouley, N. Strobel, T. Moore, and R. Fahrig Experimental Study of Intracranial Hematoma Detection with Flat Panel Detector C-Arm CT AJNR Am. J. Neuroradiol., April 1, 2008; 29(4): 766 - 772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Daghini, L. Juillard, J. A. Haas, J. D. Krier, J. C. Romero, and L. O. Lerman Comparison of Mathematic Models for Assessment of Glomerular Filtration Rate with Electron-Beam CT in Pigs Radiology, February 1, 2007; 242(2): 417 - 424. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Hackstein, C. Wiegand, W. S. Rau, and A. C. Langheinrich Glomerular Filtration Rate Measured by Using Triphasic Helical CT with a Two-Point Patlak Plot Technique Radiology, January 1, 2004; 230(1): 221 - 226. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Hackstein, J. Bauer, E. W. Hauck, M. Ludwig, H.-J. Kramer, and W. S. Rau Measuring Single-Kidney Glomerular Filtration Rate on Single-Detector Helical CT Using a Two-Point Patlak Plot Technique in Patients with Increased Interstitial Space Am. J. Roentgenol., July 1, 2003; 181(1): 147 - 156. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
