- Email: [email protected]
Estimating Relative Renal Function
THE JOURNAL OF UROLOGY Copyright© 1976 by The Williams & Wilkins Co.
Vol. 115, June
Printed in U.S.A.
Original Articles ESTIMATING RELATIVE RENAL FUNCTION R. H. SECKER-WALKER
AND
R. E. COLEMAN
From the Edward Mallinckrodt Institute of Radiology, Washington University School of Medicine and the Department of Internal Medicine, St. Louis University School of Medicine, St. Louis, Missouri
ABSTRACT
Radionuclide measurements of relative .renal function have been made using conventional dual-probe renography, computer assisted triple-probe renography, rectilinear kidney scanning and computer assisted gamma camera renography and kidney scanning. The relative area of each kidney from the excretory urogram also has been used to measure relative function. The best correlation between renography and kidney scanning was obtained with the gamma camera computer system. The relative function of each kidney was obtained by outlining regions of interest that corresponded closely to the renal images and subtracting the blood and tissue background activity, using a special background region. Triple-probe renography, using computer assisted blood background subtraction, provides another reliable indicator of relative function, which is considerably more accurate than conventional dual-probe renography. Relative renal function also may be determined, but with less precision, from the relative counts in a rectilinear kidney scan, from the relative scan areas and from the relative areas found on excretory urography. Both relative area methods are less reliable in the presence of outflow tract obstruction, renal tumors and renal cysts. There are a number of situations that arise in the management of patients with unilateral renal disease when it is useful to know how renal function is proportioned between each kidney. The blood urea nitrogen, plasma creatinine and creatinine clearance provide clinically useful indications of over-all renal function. Excretory urograms (IVPs) will shoW, whether each kidney is functioning and some qualitative judgment can be made as to how well each is working. Kidney length and cortical thickness provide additional information but neither can be used to determine relative function. Herein we review several different methods of estimating relative renal function based on radioisotope renography, kidney scanning and the IVP. METHODS
The patients were studied in an 8-year period, during which time the available apparatus became more sophisticated. Early studies were done with a dual-probe renography system and a rectilinear scanner. Later a third probe was added to the renography system and the output from all 3 probes was recorded on paper tape and processed by digital computer. More recently, studies have been performed with a gamma scintillation camera interfaced to a small digital computer. For renography the patients were given 1 µCi. 1311 hippuran per 7 kg. body weight while lying supine. The renography apparatus consisted of 3, 1 ¾-inch Nal(Tl) detectors with wide angle collimators whose output, after suitable pulse height analysis and amplification, was fed into a Honeywell chart recorder. Two detectors viewed the kidneys, while the third viewed the upper mediastinum. Accepted for publication September 12, 1975.
When facilities for computer processing of the data became available the output from the amplifiers also was fed into scalers that were read every 10 seconds by a paper tape machine. 1 Rectilinear scans were made with either a 3 or 5-inch Picker Magna Scanner. Scans were performed 2 to 4 hours after the intravenous injection of 1 • 1 Hg chlormerodriri (1 µCi. per kg. body weight) using 35 per cent background cutoff. Color and photoscans were produced. After the scan was complete a slit collimator was used instead of the 19-hole collimator and the proportion of activity in each kidney was determined from the average number of counts recorded as the slit passed over each kidney 3 times. Studies with the gamma scintillation camera used a Phogamma 111 interfaced to a PDP-12. Details of the interface have been described elsewhere. 2 For these studies each patient was given 200 to 250 µCi. 1 • 1 Hg chlormerodrin 1 to 2 hours before the study. Then the patients sat upright with their backs to the gamma camera. Static images of the kidneys were obtained during 4 to 8 minutes and recorded on magnetic tape. With the subject still seated 200 to 250 µCi. 1311 hippuran was given and images were collected at 20-second intervals for 20 minutes. DATA PROCESSING
The dual-probe renograms were analyzed by assessing the relative rise of the second phase for each kidney after the method described by Tauxe and associates. 3 A straight line was drawn between the 1 and 2 ½-minute points on the second phase of each curve (phase ID and the height that each curve had risen during this time was measured and expressed as a percentage of the total height that each had risen. The renograms processed by the digital computer were 621
622
SECKER-WALKER AND COLEMAN
analyzed using the computer assisted blood background subtraction technique of Britton and Brown• and described previously.' The kidney scans were quantified in 2 ways: 1) By expressing the slit collimator counts from each kidney as a percentage of the total counts. Background subtraction was incorporated into the slit counts. In the results obtained from 30 patients slit counts were compared to the actual counts in each kidney after these had been counted manually on the color scans. There was a highly significant correlation between these 2 methods, r equals 0.99, p less than 0.0001, so that the slit collimator technique was used for all subsequent scans. 2) By drawing a smooth line around the image of each kidney on the color scan and then measuring the enclosed areas by planimetry. The relative function of each kidney was determined from the fraction that each kidney was of the total area of both kidneys. In the gamma camera studies the static scan was presented to the operator on the computer oscilloscope and each kidney and a tissue background area were outlined using a joy-stick control (fig. 1). The counts per cell in the background region were scaled in proportion to the number of cells outlined for each kidney and then subtracted from the counts within the kidney region. With the same areas of interest the counts under the first 2 ½ to 3 minutes of each renogram were integrated and the tissue background was subtracted. The time for integration was the same for each kidney in a particular patient and included the frame that was 20 seconds before the first peak. If the peak occurred after 3 minutes, the counts up to the 3-minute frame were used. The relative area of each kidney was derived from the cells outlined for each kidney. In 25 patients with evidence of unilateral renal disease the outline of the kidneys on the IVP was traced onto transparent paper and then this area was measured by planimetry. The area of the right kidney was expressed as a percentage of the total area of both kidneys. For each of these methods the relative function of the right kidney is considered in the results.
RESULTS
The results obtained with the dual-probe system have been divided into several groups, depending on the underlying renal disease: 1) renal artery stenosis, 2) chronic pyelonephritis-using radiological criteria for the diagnosis-a clubbed calix with an overlying cortical scar and 3) back pressure atrophy -uniformly clubbed calices with a smooth cortical outline. 5 When the relative function from the dual-probe renograms was compared to that derived from the counts in the kidney scans, highly significant correlations were obtained in each of the diagnostic categories (table 1). Highly significant correlations also were obtained when the relative area of the scan was compared to the relative counts in that kidney (table 1). Similar correlations were obtained when the relative area of the scan was compared to the relative function derived from the 131 I hippuran renograms. Only in the patients with back pressure atrophy was the correlation coefficient less than 0.9 (table 1). In this group the relative function is overestimated by the area method, with the largest difference being found in patients with pelvioureteral junction obstruction. The standard deviation of the regression equations for each of these major comparisons is shown in table 2. Figure 2 shows the comparison between phase II estimates of relative function from dual-probe renography and simultaneously derived estimates from computer assisted blood background subtraction renography. A highly significant correlation exists but the phase II method tends to underestimate the contribution of the poorly functioning kidneys because of the decreasing tissue and blood background activity. Highly significant correlations were found between the computer as~isted blood background subtraction technique of estimating relative renal function and kidney scans (n equals 53, r equals 0.96, p less than 0.001) as has been reported previously.' Similar correlations were obtained using the gamma camera computer system in a smaller number of patients in whom the tissue background was determined from selected regions above, between and below each kidney (fig. 3 and table 3). In a selected group of 25 patients with unilateral renal disease the relative area of the right kidney on the IVP was compared to the relative area of the right kidney on the scan and also to the relative function determined from the counts in the scan and from conventional dual-probe renography. A highly significant correlation was found between the 2 area methods, r equals 0.95, p less than 0.001 (fig. 4). The correlations between the relative area of the right kidney on IVP and the relative counts in the right kidney scan, r equals 0.88, p less than 0.001, and the relative function from renography, r equals 0.89, p less than 0.001, were less satisfactory but still highly significant. The case report indicates the value of estimating relative function preoperatively and also the accuracy of the prediction of postoperative function in patients whose renal function is already compromised. CASE REPORT
FIG. 1. Areas used for background region: T-shaped area above and between kidneys and 2 square areas below kidneys. In addition regions selected to represent each kidney from kidney scan are shown. Positioning of bright dots for region selection was under user's joy-stick control. Figure was photographed from computer oscilloscope.
A 69-year-old woman had an abdominal mass that had been present for 1 ½ years and was enlarging. There was no history of renal disease but the creatinine clearance was 80 ml. per minute. An IVP revealed a large mass 1':l the lower pole of the left kidney and pyelonephritic changes in the right kidney. An aortogram and bilateral selective renal arteriogram revealed a left lower pole vascular mass with abnormal vessels. Again noted were changes of pyelonephritis of the right kidney. Kidney scanning and renography were undertaken to determine the contribution of each kidney to over-all renal function (fig. 5). The right kidney accounted for 25 per cent of the total renal function. The patient then underwent indirect split renal
623
ESTIMATING RELATIVE RENAL FUNCTION TABLE
1. Dual-probe renography and kidney scans: correlations among phase II, scan area and counts in scan Renography: Counts in Scan
Diagnostic Category
Scan Area: Counts in Scan p
p
0.96 0.97 0.96
Renal artery stenosis (40 pts.) Chronic pyelonephritis (25 pts.) Back pressure atrophy (43 pts.)
0.97 0.97 0.92
<0.001 <0.001 <0.001
TABLE 2. Dual-probe renography and kidney scans: standard deviations of regression lines of the relationships among phase II, scan area and counts in scan Diagnostic Category Renal artery stenosis (40 pts.) Chronic pyelonephritis (25 pts.) Back pressure atrophy (43 pts.)
Renography: Counts in Scan
Scan Area: Counts in Scan
(%)
(%)
100
>--
C
-0
::.:::
......
....c 0)
Ck::
Q_
0
±7.2
...... C
±6.2
±6.5
±9.6
::J
±13.9
I...
0 C Q.)
0
Ck::
u
(/)
C ::J LL
o:::l o:::l
n y s(x.y)
•
96 0.91 p<0.001 0.68 X + 17.7 ± 7.7
•
I
80
+....c
0)
0 ·-
r
y s(x.y)
::.:::
0.94 0.93 0.86
<0.001 <0.001 <0.001
37 0.97 p <0.001 0.98x + 0.65 ± 5.8
•
60
Uo,::
E 0
I...
0
0)
Q.)
0
Ck::
40
C Q.)
Ck::
E
20
0
I...
LL
80
•
60
I 40
100
Relative Counts From Scan, Right Kidney % FIG. 3. Comparison of relative function of right kidney in 37 patients with variety of renal diseases determined from relative counts in 197 Hg chlormerodrin images and relative uptake of 1311 hippuran using gamma camera computer system. Regression line is almost identical to line of identity.
<(
>
u 20
0
~
Q.)
V,
Q.)
0)
C
Q.)
80
C
p
GAMMA CAMERA STUDIES n
-0
......>
r >-....c
>--
Q.)
(%)
CONVENTIONAL RENOGRAPHY VERSUS C.A.B.B.S. RENOGRAPHY
Q.)
~
Scan Area: Renography
±5.5
±10.6
<0.001 <0.001 <0.001
100
±6.9
±7.5
Scan Area: Renography
tumor-bearing kidney was likely to be more accurate than if the function of 1 kidney had been completely normal. DISCUSSION
Ck::
0
60 100 40 80 20 Relative Function Right Kidney % Conventional Renography
FIG. 2. Comparison of relative function of right kidney in 96 patients with variety of renal diseases determined from second phase of conventional dual-probe renogram and by simultaneous computer assisted blood background subtraction. Conventional method underestimates function of poorly functioning kidneys.
function studies by catheterizing the left ureter and bladder. The creatinine clearance from the left kidney was 61 ml. per minute and from the right kidney it was 19 ml. per minute (24 per cent of the total creatinine clearance). It was predicted from the radionuclide studies and the divided clearance studies that the postoperative creatinine clearance would decrease to approximately 20 ml. per minute. The patient underwent a successful left nephrectomy with removal of a renal cell carcinoma. Postoperatively, the creatinine clearance was 18 ml. per minute. The modest reduction in creatinine clearance found preoperatively implied that both kidneys were already functioning maximally. In these circumstances, without any functional reserve, the prediction of relative function after removal of the
These studies show that renography and kidney scanning provide similar estimates of the relative function of each kidney. With the slit collimator the function of the right kidney tends to be overestimated by about 5 per cent because of the 1 " 7 Hg chlormerodrin content of the adjacent liver. This error was smaller with the gamma camera studies, in which the selection of the regions of interest that correspond to the renal outlines helped separate liver activity from renal activity. In comparisons between the relative scan area and the relative counts in the scan or the relative uptake of 131 I hippuran, the correlation coefficient exceeded 0.93, except in patients with back pressure atrophy. In these patients the partially obstructed kidney tended to have a larger relative area than expected from the relative counts of 197Hg chlormerodrin or relative uptake of 131I hippuran, reflecting the distension of the kidney on the side of the obstruction. The correlations between the relative number of outlined cells on the gamma camera images and the other estimates of relative function were less precise (r equals 0.80 for 197 Hg chlormerodrin and 0.79 for 131I hippuran). Several of these patients were scanned because of the presence of outflow tract obstruction, so that the poorer correlation between relative area and relative function may in part be ascribed to this cause. In addition, the coarse matrix used to outline the
624
SECKER-WALKER AND COLEMAN
3. Gamma camera computer studin in 37 patients
TABLE
Renography: Counts in Sc!ln 0.97 p < 0.001 0.98x + 0.65 s(x -y} ~ 5.8
Correlation coefficient Regression equation
Scan Atta: Ceuntl! ill
s.,-
< o.•1
r
~
r - 0.80 p
y
~
y - O.f>h + Hl.3 o{x-y) - 11.2
Scan Area: RenO§"r&phy
r ~ 0.79 p < 0.001 y ~ 0.58x + 20.4 a(x.y) - 11.5
1.V.P. AREA VERSUS SCAN AREA
100 V,
r
0
>(1)
0
C
(1)
-0
....
80
- c½
60
.2
40
C
0
25 0.95 p < 0.001
Y 1.06 s(x. y)
<(
~
....s:::::.
n
±5.4
X -
2.03
•
O> 0
(1)
4: c 0
0
t-
~
20
u
(/)
0
40 60 80 100 I.V.P. Area Right Kidney as % Tota I I.V.P. Area
20
FIG. 4. Comparison of relative area of right kidney measured by planimetry from IVP and from renal scan in 25 patients with unilateral renal disease but with no evidence of back pressure atrophy. Regression line is close to line of identity.
kidneys leads to a less precise delineation of the areas than can be accomplished on a rectilinear scan. The regions chosen for tissue background subtraction were found by trial and error in early studies with the gamma camera in patients who had undergone unilateral nephrectomy or who had 1 completely functionless kidney. By choosing an area wholly above the kidneys it was found that too much background activity was subtracted, giving negative values on the functionless side. When an area below the kidneys or between them was chosen too little background was subtracted and the apparently functionless kidney had positive values. By choosing a combination of areas, as shown in figure 1, a figure could be obtained that gave the most accurate estimate of blood and tissue background activity. On 2 occasions, using an area similar to the background area shown in figure 1, negative values were still obtained over a non-functioning kidney. One patient proved to have a large functionless hydronephrosis and the other one had a non-functioning polycystic kidney. In each of these patients the contents of the non-functioning kidney not only contained no radionuclide but also attenuated some of the gamma rays from the tissues beyond the kidney. Although 197Hg chlormerodrin kidney scanning and 1311 hippuran renography yield similar estimates of relative renal function, the mechanism of accumulation of the radionuclide differs in each case. It is important to realize that only the first 2 to 3 minutes of the renogram can be used for this purpose if a valid comparison is to be made. During this time the 1311 hippuran is still within the kidney and the relative quantity present accurately reflects relative effective renal plasma flow. Once any 1311 hippuran has left the outlined region or the field
FIG. 5. Serial gamma camera images at 2½-minute intervals after intravenous 1311 hippuran in 69-year-old woman with renal cell carcinoma in left kidney and chronic pyelonephritis of right kidney. Activity-time curves from each kidney are shown below and were derived by selecting areas of interest corresponding to each kidney using PDP-12 digital computer. Left kidney is on left of gamma camera images and its activity-time curve is also on left. Right kidney accounted for 25 per cent of total renal function.
of view of an external probe the estimate of relative function is no longer accurate. 197 Hg chlormerodrin was used for all kidney scans. A small proportion of the administered dose is excreted during the first 2 to 4 hours after injection• but the small quantity present in any of the hydronephrotic renal pelves did not reduce the correlations between the relative counts in these scans and the figures derived from renograms. Studies with commercially available ••mtechnetium diethylenetriamine pentaacetic acid suggest that it cannot be used so readily for quantitative comparison (unpublished observations) because the renal pelvis may contain significant quantities of activity at the time the images are collected. Raynaud and associates have used the accumulation of 197 Hg chloride as an estimate of renal function and have shown that the quantity within each kidney at 48 hours can be used as a measurement of renal function. 7 • 8 By establishing normal values for the renal content of 1 • 1 Hg chloride this technique provides figures that can be used to judge the impairment of individual kidneys without reference to blood urea nitrogen or creatinine clearance. Britton and Brown have extended their original observations with computer assisted blood background subtraction renography by more detailed analysis of the kidney trace.•. 10 They have obtained figures for the frequency function of transit times through the kidney and also an index of the efficiency of transport of 1311 hippuran out of the kidneys. Secker-Walker and associates showed that the transport efficiency was closely related to creatinine clearance and inversely related to blood
625 urea. 1 They
that this n12a.surernen_t of transport could be used to indicate how well a was 3.
A considerably more of the renogram has been proposed DeGrazia and associates. 11 It remains to be seen what use can be made of this wealth of new data in the everyday management of with renal disease. In the comparisons reported here the best correlation between renography and kidney scanning was obtained with the gamma camera computer system. The relative function of each kidney was obtained outlining regions of interest that corresponded closely to the renal images and subtracting the blood and tissue background using the special background region. We suggest that either of these methods can be used to provide a clinically reliable indication of relative renal function. There is some experimental evidence to support this view. Hayes and associates have shown an excellent correlation (r equals 0.97) between estimations of relative renal function using 131 I hippuran and a gamma camera system and relative function measured by divided clearance in dogs. 12 Schlegel and associates showed, in a study of 10 patients, that the relative quantity of 203 Hg chlormerodrin in the kidney image from a gamma camera was closely correlated with relative effective renal plasma flow measured during split-renal function tests. 13 In the absence of a gamma camera triple-probe renography with computer assisted blood background subtraction provides another reliable indication of relative renal function and the possibility of deriving other measurements of kidney function. Rectilinear scans 2 ways of estimating relative function-either from the counts within the renal image or from their relative areas. This last technique is easily accomplished planirnetry and is reliable in the absence of obstruction to the outflow urine. Unfortunately, it is often in patients with varying degrees of outflow tract obstruction that information about relative function is required. When none of the radioisotopic methods is available the IVP can be used. In the absence of hydronephrosis, renal cysts or renal tumors the relative area of each kidney traced from the IVP provides a good indication of relative function. REFERENCES
1. Secker-Waiker, R. H., Shepherd, E. P. and Cassell, K. J.: Clini-
cal applications of computer-assisted renography. J. Nu.cl. Med., 13: 235, 1972. 2. Hill, R. L. and Cox, J. R.: Design of Washington University camera-computer interface. In: Computer Processing of namic Images From an Anger Scintillation Camera. Edited
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
K. [( La:tson and J. Cox. l"Jew Yo:rk: The Society vf Nuclear Medicine, Inc., pp. 52-58, 1974. Tauxe, W. N., Hunt, J. C. and Burbank, M. K.: The radioisotope renogram (ortho-iodohippurate-I1' 1 ). Standardization of technic and expression of data. Amer. J. Clin. Path., 37: 567, 1962. Britton, K. E. and Brown, N. J.: The clinical use of C.A.B.B.S. renography. Investigation of the "non-functioning kidney" and renal artery stenosis by the use of 131 1 hippuran renography modified by computer assisted blood background subtraction (C.A.B.B.S.). Brit. J. Radial., 41: 570, 1974. Hodson, C. J., Craven, J. D., Lewis, D. G., Matz, L. R., Clarke, R. J. and Ross, E. J.: Experimental obstructive nephropathy in the pig. Brit. J. Urol., suppl. vol. 41, p. 4, 1969. Reba, R. C., McAfee, J. G. and Wagner, H. N.: Radiomercurylabelled chlormerodrin for in vivo uptake studies and scintillation scanning of unilateral renal lesions associated with hypertension. Medicine, 42: 269, 1963. Raynaud, C., Ricard, S., Karam, Y. and Kellershohn, C.: The use of the renal uptake of 197 Hg as a method for testing the functional value of each kidney. J. Nucl. Med., ll: 125, 1970. Raynaud, C.: A technique for the quantitative measurement of the function of each kidney. Sem. Nucl. Med., 4: 51, 1974. Britton, K. E. and Brown, N. J. G.: The Value in Obstructive Nephropathy of the Hippuran Output Curve Derived by Computer Analysis of the Renogram. Vienna: International Atomic Energy Agency, pp. 263-275, 1971. Britton, K. E. and Brown, N. J. G.: Advances in renography in relation to radiology and radiotherapy. Proc. Roy. Soc. Med., 64: 342, 1971. DeGrazia, J. A., Scheibe, P. 0., Jackson, P. E., Lucas, Z. J., Fair, W.R., Vogel, J. Iv.I. and Blumin, L. J.: Clinical applications of a kinetic model of hippurate distribution and renal clearance. J. Nucl. Med., 15: 102, 1974. Hayes, M., Brosman, S. and Taplin, G. V.: Determination of differential renal function by sequential renal scintigraphy. J. Urol., lll: 556, 1974. Schlegel, J. U., Varela, R. and Stanton, J. J.: Individual renal plasma flow determination without ureteral catheterization. J. Urol., 96: :20, 1966.
COMMENT The authors have presented an assessment of 2 renographic means for determining separate kidney function in various disease states. Urinary tract obstruction may cause some underestimation of function of the ipsilateral kidney but it is clearly evident that renographic evaluation provides an essentially non-invasive means of assessing individual kidney function. Standard pyelographic studies are often of limited value in estimating differential kidney function and the renographic procedures reported in this article may be technically more feasible and clinically more reliable than the classical split-function measurements of inulin or creatinine clearance. W.C.T.
Vol. 115, June
Printed in U.S.A.
Original Articles ESTIMATING RELATIVE RENAL FUNCTION R. H. SECKER-WALKER
AND
R. E. COLEMAN
From the Edward Mallinckrodt Institute of Radiology, Washington University School of Medicine and the Department of Internal Medicine, St. Louis University School of Medicine, St. Louis, Missouri
ABSTRACT
Radionuclide measurements of relative .renal function have been made using conventional dual-probe renography, computer assisted triple-probe renography, rectilinear kidney scanning and computer assisted gamma camera renography and kidney scanning. The relative area of each kidney from the excretory urogram also has been used to measure relative function. The best correlation between renography and kidney scanning was obtained with the gamma camera computer system. The relative function of each kidney was obtained by outlining regions of interest that corresponded closely to the renal images and subtracting the blood and tissue background activity, using a special background region. Triple-probe renography, using computer assisted blood background subtraction, provides another reliable indicator of relative function, which is considerably more accurate than conventional dual-probe renography. Relative renal function also may be determined, but with less precision, from the relative counts in a rectilinear kidney scan, from the relative scan areas and from the relative areas found on excretory urography. Both relative area methods are less reliable in the presence of outflow tract obstruction, renal tumors and renal cysts. There are a number of situations that arise in the management of patients with unilateral renal disease when it is useful to know how renal function is proportioned between each kidney. The blood urea nitrogen, plasma creatinine and creatinine clearance provide clinically useful indications of over-all renal function. Excretory urograms (IVPs) will shoW, whether each kidney is functioning and some qualitative judgment can be made as to how well each is working. Kidney length and cortical thickness provide additional information but neither can be used to determine relative function. Herein we review several different methods of estimating relative renal function based on radioisotope renography, kidney scanning and the IVP. METHODS
The patients were studied in an 8-year period, during which time the available apparatus became more sophisticated. Early studies were done with a dual-probe renography system and a rectilinear scanner. Later a third probe was added to the renography system and the output from all 3 probes was recorded on paper tape and processed by digital computer. More recently, studies have been performed with a gamma scintillation camera interfaced to a small digital computer. For renography the patients were given 1 µCi. 1311 hippuran per 7 kg. body weight while lying supine. The renography apparatus consisted of 3, 1 ¾-inch Nal(Tl) detectors with wide angle collimators whose output, after suitable pulse height analysis and amplification, was fed into a Honeywell chart recorder. Two detectors viewed the kidneys, while the third viewed the upper mediastinum. Accepted for publication September 12, 1975.
When facilities for computer processing of the data became available the output from the amplifiers also was fed into scalers that were read every 10 seconds by a paper tape machine. 1 Rectilinear scans were made with either a 3 or 5-inch Picker Magna Scanner. Scans were performed 2 to 4 hours after the intravenous injection of 1 • 1 Hg chlormerodriri (1 µCi. per kg. body weight) using 35 per cent background cutoff. Color and photoscans were produced. After the scan was complete a slit collimator was used instead of the 19-hole collimator and the proportion of activity in each kidney was determined from the average number of counts recorded as the slit passed over each kidney 3 times. Studies with the gamma scintillation camera used a Phogamma 111 interfaced to a PDP-12. Details of the interface have been described elsewhere. 2 For these studies each patient was given 200 to 250 µCi. 1 • 1 Hg chlormerodrin 1 to 2 hours before the study. Then the patients sat upright with their backs to the gamma camera. Static images of the kidneys were obtained during 4 to 8 minutes and recorded on magnetic tape. With the subject still seated 200 to 250 µCi. 1311 hippuran was given and images were collected at 20-second intervals for 20 minutes. DATA PROCESSING
The dual-probe renograms were analyzed by assessing the relative rise of the second phase for each kidney after the method described by Tauxe and associates. 3 A straight line was drawn between the 1 and 2 ½-minute points on the second phase of each curve (phase ID and the height that each curve had risen during this time was measured and expressed as a percentage of the total height that each had risen. The renograms processed by the digital computer were 621
622
SECKER-WALKER AND COLEMAN
analyzed using the computer assisted blood background subtraction technique of Britton and Brown• and described previously.' The kidney scans were quantified in 2 ways: 1) By expressing the slit collimator counts from each kidney as a percentage of the total counts. Background subtraction was incorporated into the slit counts. In the results obtained from 30 patients slit counts were compared to the actual counts in each kidney after these had been counted manually on the color scans. There was a highly significant correlation between these 2 methods, r equals 0.99, p less than 0.0001, so that the slit collimator technique was used for all subsequent scans. 2) By drawing a smooth line around the image of each kidney on the color scan and then measuring the enclosed areas by planimetry. The relative function of each kidney was determined from the fraction that each kidney was of the total area of both kidneys. In the gamma camera studies the static scan was presented to the operator on the computer oscilloscope and each kidney and a tissue background area were outlined using a joy-stick control (fig. 1). The counts per cell in the background region were scaled in proportion to the number of cells outlined for each kidney and then subtracted from the counts within the kidney region. With the same areas of interest the counts under the first 2 ½ to 3 minutes of each renogram were integrated and the tissue background was subtracted. The time for integration was the same for each kidney in a particular patient and included the frame that was 20 seconds before the first peak. If the peak occurred after 3 minutes, the counts up to the 3-minute frame were used. The relative area of each kidney was derived from the cells outlined for each kidney. In 25 patients with evidence of unilateral renal disease the outline of the kidneys on the IVP was traced onto transparent paper and then this area was measured by planimetry. The area of the right kidney was expressed as a percentage of the total area of both kidneys. For each of these methods the relative function of the right kidney is considered in the results.
RESULTS
The results obtained with the dual-probe system have been divided into several groups, depending on the underlying renal disease: 1) renal artery stenosis, 2) chronic pyelonephritis-using radiological criteria for the diagnosis-a clubbed calix with an overlying cortical scar and 3) back pressure atrophy -uniformly clubbed calices with a smooth cortical outline. 5 When the relative function from the dual-probe renograms was compared to that derived from the counts in the kidney scans, highly significant correlations were obtained in each of the diagnostic categories (table 1). Highly significant correlations also were obtained when the relative area of the scan was compared to the relative counts in that kidney (table 1). Similar correlations were obtained when the relative area of the scan was compared to the relative function derived from the 131 I hippuran renograms. Only in the patients with back pressure atrophy was the correlation coefficient less than 0.9 (table 1). In this group the relative function is overestimated by the area method, with the largest difference being found in patients with pelvioureteral junction obstruction. The standard deviation of the regression equations for each of these major comparisons is shown in table 2. Figure 2 shows the comparison between phase II estimates of relative function from dual-probe renography and simultaneously derived estimates from computer assisted blood background subtraction renography. A highly significant correlation exists but the phase II method tends to underestimate the contribution of the poorly functioning kidneys because of the decreasing tissue and blood background activity. Highly significant correlations were found between the computer as~isted blood background subtraction technique of estimating relative renal function and kidney scans (n equals 53, r equals 0.96, p less than 0.001) as has been reported previously.' Similar correlations were obtained using the gamma camera computer system in a smaller number of patients in whom the tissue background was determined from selected regions above, between and below each kidney (fig. 3 and table 3). In a selected group of 25 patients with unilateral renal disease the relative area of the right kidney on the IVP was compared to the relative area of the right kidney on the scan and also to the relative function determined from the counts in the scan and from conventional dual-probe renography. A highly significant correlation was found between the 2 area methods, r equals 0.95, p less than 0.001 (fig. 4). The correlations between the relative area of the right kidney on IVP and the relative counts in the right kidney scan, r equals 0.88, p less than 0.001, and the relative function from renography, r equals 0.89, p less than 0.001, were less satisfactory but still highly significant. The case report indicates the value of estimating relative function preoperatively and also the accuracy of the prediction of postoperative function in patients whose renal function is already compromised. CASE REPORT
FIG. 1. Areas used for background region: T-shaped area above and between kidneys and 2 square areas below kidneys. In addition regions selected to represent each kidney from kidney scan are shown. Positioning of bright dots for region selection was under user's joy-stick control. Figure was photographed from computer oscilloscope.
A 69-year-old woman had an abdominal mass that had been present for 1 ½ years and was enlarging. There was no history of renal disease but the creatinine clearance was 80 ml. per minute. An IVP revealed a large mass 1':l the lower pole of the left kidney and pyelonephritic changes in the right kidney. An aortogram and bilateral selective renal arteriogram revealed a left lower pole vascular mass with abnormal vessels. Again noted were changes of pyelonephritis of the right kidney. Kidney scanning and renography were undertaken to determine the contribution of each kidney to over-all renal function (fig. 5). The right kidney accounted for 25 per cent of the total renal function. The patient then underwent indirect split renal
623
ESTIMATING RELATIVE RENAL FUNCTION TABLE
1. Dual-probe renography and kidney scans: correlations among phase II, scan area and counts in scan Renography: Counts in Scan
Diagnostic Category
Scan Area: Counts in Scan p
p
0.96 0.97 0.96
Renal artery stenosis (40 pts.) Chronic pyelonephritis (25 pts.) Back pressure atrophy (43 pts.)
0.97 0.97 0.92
<0.001 <0.001 <0.001
TABLE 2. Dual-probe renography and kidney scans: standard deviations of regression lines of the relationships among phase II, scan area and counts in scan Diagnostic Category Renal artery stenosis (40 pts.) Chronic pyelonephritis (25 pts.) Back pressure atrophy (43 pts.)
Renography: Counts in Scan
Scan Area: Counts in Scan
(%)
(%)
100
>--
C
-0
::.:::
......
....c 0)
Ck::
Q_
0
±7.2
...... C
±6.2
±6.5
±9.6
::J
±13.9
I...
0 C Q.)
0
Ck::
u
(/)
C ::J LL
o:::l o:::l
n y s(x.y)
•
96 0.91 p<0.001 0.68 X + 17.7 ± 7.7
•
I
80
+....c
0)
0 ·-
r
y s(x.y)
::.:::
0.94 0.93 0.86
<0.001 <0.001 <0.001
37 0.97 p <0.001 0.98x + 0.65 ± 5.8
•
60
Uo,::
E 0
I...
0
0)
Q.)
0
Ck::
40
C Q.)
Ck::
E
20
0
I...
LL
80
•
60
I 40
100
Relative Counts From Scan, Right Kidney % FIG. 3. Comparison of relative function of right kidney in 37 patients with variety of renal diseases determined from relative counts in 197 Hg chlormerodrin images and relative uptake of 1311 hippuran using gamma camera computer system. Regression line is almost identical to line of identity.
<(
>
u 20
0
~
Q.)
V,
Q.)
0)
C
Q.)
80
C
p
GAMMA CAMERA STUDIES n
-0
......>
r >-....c
>--
Q.)
(%)
CONVENTIONAL RENOGRAPHY VERSUS C.A.B.B.S. RENOGRAPHY
Q.)
~
Scan Area: Renography
±5.5
±10.6
<0.001 <0.001 <0.001
100
±6.9
±7.5
Scan Area: Renography
tumor-bearing kidney was likely to be more accurate than if the function of 1 kidney had been completely normal. DISCUSSION
Ck::
0
60 100 40 80 20 Relative Function Right Kidney % Conventional Renography
FIG. 2. Comparison of relative function of right kidney in 96 patients with variety of renal diseases determined from second phase of conventional dual-probe renogram and by simultaneous computer assisted blood background subtraction. Conventional method underestimates function of poorly functioning kidneys.
function studies by catheterizing the left ureter and bladder. The creatinine clearance from the left kidney was 61 ml. per minute and from the right kidney it was 19 ml. per minute (24 per cent of the total creatinine clearance). It was predicted from the radionuclide studies and the divided clearance studies that the postoperative creatinine clearance would decrease to approximately 20 ml. per minute. The patient underwent a successful left nephrectomy with removal of a renal cell carcinoma. Postoperatively, the creatinine clearance was 18 ml. per minute. The modest reduction in creatinine clearance found preoperatively implied that both kidneys were already functioning maximally. In these circumstances, without any functional reserve, the prediction of relative function after removal of the
These studies show that renography and kidney scanning provide similar estimates of the relative function of each kidney. With the slit collimator the function of the right kidney tends to be overestimated by about 5 per cent because of the 1 " 7 Hg chlormerodrin content of the adjacent liver. This error was smaller with the gamma camera studies, in which the selection of the regions of interest that correspond to the renal outlines helped separate liver activity from renal activity. In comparisons between the relative scan area and the relative counts in the scan or the relative uptake of 131 I hippuran, the correlation coefficient exceeded 0.93, except in patients with back pressure atrophy. In these patients the partially obstructed kidney tended to have a larger relative area than expected from the relative counts of 197Hg chlormerodrin or relative uptake of 131I hippuran, reflecting the distension of the kidney on the side of the obstruction. The correlations between the relative number of outlined cells on the gamma camera images and the other estimates of relative function were less precise (r equals 0.80 for 197 Hg chlormerodrin and 0.79 for 131I hippuran). Several of these patients were scanned because of the presence of outflow tract obstruction, so that the poorer correlation between relative area and relative function may in part be ascribed to this cause. In addition, the coarse matrix used to outline the
624
SECKER-WALKER AND COLEMAN
3. Gamma camera computer studin in 37 patients
TABLE
Renography: Counts in Sc!ln 0.97 p < 0.001 0.98x + 0.65 s(x -y} ~ 5.8
Correlation coefficient Regression equation
Scan Atta: Ceuntl! ill
s.,-
< o.•1
r
~
r - 0.80 p
y
~
y - O.f>h + Hl.3 o{x-y) - 11.2
Scan Area: RenO§"r&phy
r ~ 0.79 p < 0.001 y ~ 0.58x + 20.4 a(x.y) - 11.5
1.V.P. AREA VERSUS SCAN AREA
100 V,
r
0
>(1)
0
C
(1)
-0
....
80
- c½
60
.2
40
C
0
25 0.95 p < 0.001
Y 1.06 s(x. y)
<(
~
....s:::::.
n
±5.4
X -
2.03
•
O> 0
(1)
4: c 0
0
t-
~
20
u
(/)
0
40 60 80 100 I.V.P. Area Right Kidney as % Tota I I.V.P. Area
20
FIG. 4. Comparison of relative area of right kidney measured by planimetry from IVP and from renal scan in 25 patients with unilateral renal disease but with no evidence of back pressure atrophy. Regression line is close to line of identity.
kidneys leads to a less precise delineation of the areas than can be accomplished on a rectilinear scan. The regions chosen for tissue background subtraction were found by trial and error in early studies with the gamma camera in patients who had undergone unilateral nephrectomy or who had 1 completely functionless kidney. By choosing an area wholly above the kidneys it was found that too much background activity was subtracted, giving negative values on the functionless side. When an area below the kidneys or between them was chosen too little background was subtracted and the apparently functionless kidney had positive values. By choosing a combination of areas, as shown in figure 1, a figure could be obtained that gave the most accurate estimate of blood and tissue background activity. On 2 occasions, using an area similar to the background area shown in figure 1, negative values were still obtained over a non-functioning kidney. One patient proved to have a large functionless hydronephrosis and the other one had a non-functioning polycystic kidney. In each of these patients the contents of the non-functioning kidney not only contained no radionuclide but also attenuated some of the gamma rays from the tissues beyond the kidney. Although 197Hg chlormerodrin kidney scanning and 1311 hippuran renography yield similar estimates of relative renal function, the mechanism of accumulation of the radionuclide differs in each case. It is important to realize that only the first 2 to 3 minutes of the renogram can be used for this purpose if a valid comparison is to be made. During this time the 1311 hippuran is still within the kidney and the relative quantity present accurately reflects relative effective renal plasma flow. Once any 1311 hippuran has left the outlined region or the field
FIG. 5. Serial gamma camera images at 2½-minute intervals after intravenous 1311 hippuran in 69-year-old woman with renal cell carcinoma in left kidney and chronic pyelonephritis of right kidney. Activity-time curves from each kidney are shown below and were derived by selecting areas of interest corresponding to each kidney using PDP-12 digital computer. Left kidney is on left of gamma camera images and its activity-time curve is also on left. Right kidney accounted for 25 per cent of total renal function.
of view of an external probe the estimate of relative function is no longer accurate. 197 Hg chlormerodrin was used for all kidney scans. A small proportion of the administered dose is excreted during the first 2 to 4 hours after injection• but the small quantity present in any of the hydronephrotic renal pelves did not reduce the correlations between the relative counts in these scans and the figures derived from renograms. Studies with commercially available ••mtechnetium diethylenetriamine pentaacetic acid suggest that it cannot be used so readily for quantitative comparison (unpublished observations) because the renal pelvis may contain significant quantities of activity at the time the images are collected. Raynaud and associates have used the accumulation of 197 Hg chloride as an estimate of renal function and have shown that the quantity within each kidney at 48 hours can be used as a measurement of renal function. 7 • 8 By establishing normal values for the renal content of 1 • 1 Hg chloride this technique provides figures that can be used to judge the impairment of individual kidneys without reference to blood urea nitrogen or creatinine clearance. Britton and Brown have extended their original observations with computer assisted blood background subtraction renography by more detailed analysis of the kidney trace.•. 10 They have obtained figures for the frequency function of transit times through the kidney and also an index of the efficiency of transport of 1311 hippuran out of the kidneys. Secker-Walker and associates showed that the transport efficiency was closely related to creatinine clearance and inversely related to blood
625 urea. 1 They
that this n12a.surernen_t of transport could be used to indicate how well a was 3.
A considerably more of the renogram has been proposed DeGrazia and associates. 11 It remains to be seen what use can be made of this wealth of new data in the everyday management of with renal disease. In the comparisons reported here the best correlation between renography and kidney scanning was obtained with the gamma camera computer system. The relative function of each kidney was obtained outlining regions of interest that corresponded closely to the renal images and subtracting the blood and tissue background using the special background region. We suggest that either of these methods can be used to provide a clinically reliable indication of relative renal function. There is some experimental evidence to support this view. Hayes and associates have shown an excellent correlation (r equals 0.97) between estimations of relative renal function using 131 I hippuran and a gamma camera system and relative function measured by divided clearance in dogs. 12 Schlegel and associates showed, in a study of 10 patients, that the relative quantity of 203 Hg chlormerodrin in the kidney image from a gamma camera was closely correlated with relative effective renal plasma flow measured during split-renal function tests. 13 In the absence of a gamma camera triple-probe renography with computer assisted blood background subtraction provides another reliable indication of relative renal function and the possibility of deriving other measurements of kidney function. Rectilinear scans 2 ways of estimating relative function-either from the counts within the renal image or from their relative areas. This last technique is easily accomplished planirnetry and is reliable in the absence of obstruction to the outflow urine. Unfortunately, it is often in patients with varying degrees of outflow tract obstruction that information about relative function is required. When none of the radioisotopic methods is available the IVP can be used. In the absence of hydronephrosis, renal cysts or renal tumors the relative area of each kidney traced from the IVP provides a good indication of relative function. REFERENCES
1. Secker-Waiker, R. H., Shepherd, E. P. and Cassell, K. J.: Clini-
cal applications of computer-assisted renography. J. Nu.cl. Med., 13: 235, 1972. 2. Hill, R. L. and Cox, J. R.: Design of Washington University camera-computer interface. In: Computer Processing of namic Images From an Anger Scintillation Camera. Edited
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
K. [( La:tson and J. Cox. l"Jew Yo:rk: The Society vf Nuclear Medicine, Inc., pp. 52-58, 1974. Tauxe, W. N., Hunt, J. C. and Burbank, M. K.: The radioisotope renogram (ortho-iodohippurate-I1' 1 ). Standardization of technic and expression of data. Amer. J. Clin. Path., 37: 567, 1962. Britton, K. E. and Brown, N. J.: The clinical use of C.A.B.B.S. renography. Investigation of the "non-functioning kidney" and renal artery stenosis by the use of 131 1 hippuran renography modified by computer assisted blood background subtraction (C.A.B.B.S.). Brit. J. Radial., 41: 570, 1974. Hodson, C. J., Craven, J. D., Lewis, D. G., Matz, L. R., Clarke, R. J. and Ross, E. J.: Experimental obstructive nephropathy in the pig. Brit. J. Urol., suppl. vol. 41, p. 4, 1969. Reba, R. C., McAfee, J. G. and Wagner, H. N.: Radiomercurylabelled chlormerodrin for in vivo uptake studies and scintillation scanning of unilateral renal lesions associated with hypertension. Medicine, 42: 269, 1963. Raynaud, C., Ricard, S., Karam, Y. and Kellershohn, C.: The use of the renal uptake of 197 Hg as a method for testing the functional value of each kidney. J. Nucl. Med., ll: 125, 1970. Raynaud, C.: A technique for the quantitative measurement of the function of each kidney. Sem. Nucl. Med., 4: 51, 1974. Britton, K. E. and Brown, N. J. G.: The Value in Obstructive Nephropathy of the Hippuran Output Curve Derived by Computer Analysis of the Renogram. Vienna: International Atomic Energy Agency, pp. 263-275, 1971. Britton, K. E. and Brown, N. J. G.: Advances in renography in relation to radiology and radiotherapy. Proc. Roy. Soc. Med., 64: 342, 1971. DeGrazia, J. A., Scheibe, P. 0., Jackson, P. E., Lucas, Z. J., Fair, W.R., Vogel, J. Iv.I. and Blumin, L. J.: Clinical applications of a kinetic model of hippurate distribution and renal clearance. J. Nucl. Med., 15: 102, 1974. Hayes, M., Brosman, S. and Taplin, G. V.: Determination of differential renal function by sequential renal scintigraphy. J. Urol., lll: 556, 1974. Schlegel, J. U., Varela, R. and Stanton, J. J.: Individual renal plasma flow determination without ureteral catheterization. J. Urol., 96: :20, 1966.
COMMENT The authors have presented an assessment of 2 renographic means for determining separate kidney function in various disease states. Urinary tract obstruction may cause some underestimation of function of the ipsilateral kidney but it is clearly evident that renographic evaluation provides an essentially non-invasive means of assessing individual kidney function. Standard pyelographic studies are often of limited value in estimating differential kidney function and the renographic procedures reported in this article may be technically more feasible and clinically more reliable than the classical split-function measurements of inulin or creatinine clearance. W.C.T.