Eur J Nucl Med (1985) 11 : 1-6 European

Journal of
Nuclear
Medicine
© Springer-Verlag1985

99mTc-DTPA gamma-camera renography:

Normal values and rapid determination
of single-kidney glomerular filtration rate
Michael Rehling 1, Michael Lehd M¢ller 1, Jens Otto Lund 1, Klaus Bo Jensen 2, Birger Thamdrup a
and Jens Trap-Jensen l
1 Department of Clinical Physiology, 2 Department of Ophthalmology and ~ Department of Nephrology,
Frederiksberg Hospital, DK-2000 Copenhagen F, Denmark

Abstract. A method for 99mTc-diethylenetriaminepen- mainly secreted in the tubules (Smith 1951; Britton and
taacetate (DTPA) gamma-camera renography is pre- Brown 1971), while D T P A is excreted solely by glome-
sented. F r o m each renogram, an uptake index (UI) pro- rular filtration (Rehling et al. 1984). The greater part of
portional to the single-kidney glomerular filtration rate the literature concerning renography is based on stud-
(SKGFR) is defined. If the proportionality factor be- ies with 13 li_hippuran"
tween UI and S K G F R is the same in all patients, UI The purpose of the present paper was, firstly, to
can be used as an accurate measure of S K G F R . In present a method for 99mTc-DTPA gamma-camera re-
order to test this, 99mTc-DTPA renography was per- nography, secondly, to present a technique for the cal-
formed in 101 patients with glomerular filtration rates culation of the single-kidney G F R (SKGFR) from the
(GFR) varying between 4 and 172 ml/min. The sum of renogram not requiring the determination of the in-
the right- and left-kidney UIs correlated well with the jected dose or the collection of urine or blood samples
total G F R calculated from the simultaneously mea- and, thirdly, to present normal values for some parame-
sured plasma clearance of 99mTc-DTPA after a single ters obtained from the renogram.
injection. The correlation coefficient was 0.97. The
method was tested in a prospective study of 57 patients. Theory
The total G F R estimated from the renograms was not
significantly different from the G F R calculated from Glomerular filtration is a simple, first-order process.
the plasma clearance of 99mTc-DTPA. The coefficient Therefore, the amount of tracer filtered by the kidney
of variation a combination of inaccuracy and impre- in a given time interval (t) equals:
cision in the estimates as well as in the reference
values - was 11.8 % at a G F R of 100 ml/min. It is con- Q(t) = S K G F R i Cp(T) d T (1)
cluded that, in adults, the S K G F R can be calculated as 0

part of the clinical routine from 99mTc-DTPA gamma- where Cp(T) is the plasma concentration of tracer at
camera renography without determining the injected time T.
dose or collecting urine or blood samples. Normal val- In what follows, we only consider values for t of
ues for some parameters of the renogram obtained in less than the shortest transit time through the kidney
25 normal subjects are given. after the injection. In this time interval, the activity in
the nephrons equals the total amount of tracer filtered
Key words: 99mTc-DTPA gamma-camera - reno- by the kidney. This activity is registered externally with
graphy - normal values glomerular filtration rate an efficiency designated E k (including tissue attenuation
plasma clearance and gamma-camera efficiency). The net renogram, R(t),
can be expressed as:
t
In daily clinical routine, the glomerular filtration rate
(GFR) is usually calculated from the endogenous crea- R(t) = E k S K G F R ~ Cp(T) d T (2)
0
tinine clearance or from the plasma clearance of 5~Cr-
ethylenediaminetetraacetate (EDTA) after a single in- The externally registered renogram is composed of
jection. The single-kidney fractional share of the total activity within the nephrons and background activity,
G F R can be measured by 131I-hippuran renography B(t), within the vascular and extravascular fluids in the
(Pedersen and Poulsen 1967). In recent years, 99mTc- region of interest (ROI) of the kidney. Therefore, the
diethylenetriaminepentaacetate (DTPA) has gained composite renogram, CR(t), can be expressed as:
ground for gamma-camera renography due to its ra-
diophysical properties. However, hippuran and D T P A
are handled differently by the kidneys. Hippuran is CR(t) = E k S K G F R i Cp(T) d r + B(t). (3)
0

Offprint requests to." Michael Rehling, M.D. (address see It is assumed that an externally registered curve in the
above) ROI over the heart, H(t), is proportional to the plasma
concentration of tracer during the first few minutes Count rate

after the injection:

H(t) = E h Vh Cp(t) (4)

where E h is a counting-efficiency factor, and Vh is the
mean plasma volume within the ROI of the heart.
Also, this curve is assumed to be proportional to the
background activity within the ROI of the kidney:

B(t) = c~H(t) (5)

where ~ is the background-subtraction factor.

Inserting Eqs. (4) and (5) into Eq. (3) gives:

gk t R(t)
CR(t) = E ~ S K G F R o~ H ( T ) d T + ~ H(t). (6)
trnax |0 20 min
Ek
Substituting ~ S K G F R by the uptake index (UI) Fig. 1. R(t): the renogram corrected for background activity.
The uptake curve, U(t), is defined by fitting the integral of the
and dividing by H(t), Eq. (6) can be reduced to: biexponentially fitted heart curve, H(t), to the uptake phase of
the renogram 1-2.5 min after the injection of tracer. For the
calculation of the UI, see text and the appendix
i H(T) dT
CR(t) = UI o ~-c~. (7)
H(t) H(t)
sampling times. The area under the first part of the
This equation, which is a simple linear expression in heart curve (0_<t-<l min) was calculated by totalling
the unknowns ~ and UI, can be solved by linear re- the counts registered in this time interval. To minimize
gression in the interval 1 _<t<2.5 min (the uptake phase the statistical noise in the calculations, the second part
of the renogram) to give e and UI. The UI is pro- of the heart curve (1 < t < 2 1 rain) was fitted with the
portional to the S K G F R with the factor Ek/(E h Vh). If sum of two exponentials using an iterative least-square
this proportionality factor is the same in all patients, method (Kirkegaard 1970). The UIs and subtraction
the UI can be used as an accurate measure of the factors were estimated as described in the Appendix.
SKGFR. Using these subtraction factors and the fitted heart
curve, background subtraction was performed on the
composite renograms to give the net renograms. As
Materials and methods already described, the UI are proportional to the
S K G F R ; therefore, the UIs were used as a measure of
99mTc-DTPA gamma-camera renography
absolute single-kidney function. The right-kidney frac-
Prior to examination, the patient was hydrated with tional share of the total G F R was calculated as the
10 ml water per kilogram body weight. Diuresis during ratio between the right-kidney UI and the sum of the
renography was estimated by voluntary voiding before right- and left-kidney UI.
and just after the study. Renography was carried out Figure 1 shows a background-corrected renogram
with the patient in the supine position. A large-field-of- and an uptake curve calculated to show the expected
view gamma camera (Maxicamera II; General Electric) count rate if removal of tracer with the urine had not
fitted with a high-resolution, low-energy, parallel-hole taken place. The removal function was described by
collimator was placed underneath the patient so that three parameters. The peak time (tmax) was defined as
the kidneys and heart were in the field of view. Fifty- the time which elapsed between the bolus passage
four microcuries (2 MBq) of 99mTc-DTPA (C.I.S., France) through the left ventricle of the heart and the maxi-
per kilogram body weight was rapidly injected into mum count rate of the renogram. The residual fraction
a medial antecubital vein. Frames of 64 x 64 pixels were (Aao) was defined as the ratio between the count rate of
recorded with an on-line computer (CINE 200, Inter- the renogram at t = 2 0 m i n (Fig. 1, a) and at t=tma x
technique), initially at 1-s intervals for 60 s, followed by (Fig. 1, d). The removed fraction (Rzo) was defined as
120 frames at 10-s intervals. An ROI over the left the difference (Fig. 1,b) between the count rate of the
ventricle of the heart was always defined with the same uptake curve (Fig. 1, c) and the renogram (Fig. 1, a) di-
number of pixels. (From previous experience, we chose vided by the count rate of the uptake curve (Fig. 1, c) at
80 pixels which, on our system, corresponds to a sur- t = 20 min.
face area of about 32 cm2.) The ROIs over the two
kidneys were defined on the summed frames obtained
Glomerular filtration rate
between 1 and 3 min postinjection.
Time-activity curves for ROIs over the heart and 99mTc-DTPA was administered as already described.
kidneys were computed. The curves were corrected for The 99mTc-DTPA plasma clearance was measured by
physical decay. Each curve was separated into two taking blood samples 180, 210, 240, 270 and 300 rain
parts corresponding to the two periods with different postinjection (Brochner-Mortensen 1972). The plasma
clearance of 99mTc-DTPA (ClpDTPA) was converted Uptake Index

to renal clearance of inulin (Cl r inulin) according to the

equation: Clrinulin=(C/pDTPA-3.7) 1.1 (Brochner-
Mortensen and Rodbro 1976. This calculation includes
1.50
correction for (1) the assumption of a one-pool system, /
(2) the extrarenal clearance of D T P A and (3) the differ-
ence in renal clearance of D T P A and inulin. This cal-
culated renal clearance of inulin is taken to equal 1.00
GFR. . •

Statistical methods 0.50

Weighted linear regression and Student's t-test for pai-
red observations were applied. A difference was consid-
ered significant when P < 0.05.
50 100 150 ml/min

Patients Fig. 2. The sum of right- and left-kidney UI compared to the

GFR calculated from the plasma clearance of 99~Tc-DTPA
Three groups of patients were investigated. in 101 consecutive patients; y=0.007886x+0.01778 (r=0.97)

Group I. To evaluate the uptake indices as a measure of

the G F R , we compared the sum of the right- and left- estimated GFR
kidney UI obtained from renograms with the G F R ml/min
y:x,/
calculated from the plasma clearance of 99mTc-DTPA 150
measured on the same day. Over 1 year, 101 con-
secutive patients referred for routine renography and
,o
measurement of G F R were studied. This group con-
sisted of 47 females and 54 males aged 24-81 years 100
(median, 58 years) with G F R s varying from 4 to
172 ml/min. ,-Z."."
Group II. In a prospective study we tested the cor- 50
relation between the UIs and G F R s found in group I -./.
in 57 consecutive patients. This group consisted of 34 jr.,.
females and 23 males aged 22-85 years (median, 60
years) with G F R s varying from 26 to 165 ml/min. I I I I I I GFR
5O 100 150 ml/min

Group IlI. In order to obtain normal values for the Fig. 3. The GFR estimated from renograms (using the equa-
parameters obtained from renograms, 99mTc-DTPA re- tion for the line of regression given in Fig. 2) compared to the
nography was performed in 27 healthy subjects. All had GFR calculated from the plasma clearance of 99mTc-DTPA
normal blood pressure, normal serum-creatinine con- in 57 prospective patients
centration, normal urine sediment and no albuminuria
or glucosuria. Intravenous pyelography was not perfor-
med. The scintigrams were described without know- was 0.97. The right-kidney fractional share of the total
ledge of the renograms. Two subjects were excluded G F R varied between 0~o and 100%.
due to a big difference in size of their two kidneys.
Consequently, the subjects studied consisted of 6 fe- Group I1. In Fig. 3, the G F R estimated from renograms
males and 19 males aged 27-79 years (median, 51 years). using the equation for the line of regression found in
The removal function was only analysed in subjects group I is compared to the G F R calculated from the
with a diuresis exceeding 2 ml/min. plasma clearance of 99mTc-DTPA measured on the
Informed consent was obtained from all subjects same day. The line of regression (y = 0.96 x-0.9) did not
before they participated in the study. The study was differ significantly from the line of identity. The mean
approved by the local ethical committee. difference between the G F R estimated from renograms
and that calculated from the plasma clearance of
99mTc-DTPA did not differ from zero.
Results
With increasing GFR, the standard deviation in-
Group I. In Fig. 2 the sum of the right- and left-kidney creased, and the coefficient of variation decreased, e.g.
UIs obtained from renograms is compared to the G F R 8.3ml/min (16.6%) at a G F R of 50ml/min, and
calculated from the plasma clearance of 99mTc-DTPA 11.8 ml/min (11.8 %) at a G F R of 100 ml/min.
measured on the same day. A weighted linear re-
gression was performed, with the weight equal to 1/y. Group III. The subtraction factor c~, which converts the
The equation for the line of regression was y heart curve to the background curve, varied between
=0.007886x+0.01778. The intercept was not signifi- 0.64 and 1.96 for the right renogram, and between 0.56
cantly different from zero. The correlation coefficient and 1.69 for the left renogram. The subtraction factor
Table 1. Range and 95% confidence limits for some para- in different ways. The most widely accepted and
meters obtained from 99~"Tc-DTPA renograms in 21 (25) theoretically the most valid method is the use of an
normal subjects initial injection of 131I human serum albumin (RIHSA;
Britton and Brown 1971). This technique requires two
95 % Confidence Range
injections and increases the radiation dose to the pa-
limits
tient. Calculation of the subtraction factor using the
Single-kidney fractional 0.44-0,56 0.44-0.56 a present single-injection technique has previously been
share of total GFR described for 131I-hippuran renography (Meldolesi
et al. 1973; Makler et al. 1979; Rutland 1979) but not
tmax(min) < 4.8 2.1-6.5 for 99mTc-DTPA gamma-camera renography. Makler
IA tma~[(min) < 1.1 0.0-1.5 et al. (1979) and Rutland (1979) found that the sub-
A2o(~o) <65 11.4-88.9 traction factors calculated using this simplified tech-
nique, when compared to those simultaneously calcu-
IA12o1(%) <15 0.2-19.8 lated with the R I H S A technique, were overestimated by
R2o (%) > 82 77.9-97.6 5 % and 11%, respectively. In 25 normal subjects the
IARzol(%) < 4.4 0.0-6.2 subtraction factor for the right renogram exceeded that
for the left, indicating a greater blood background in the
Denotes 25 patients. For details, see text field of view over the right kidney, probably due to the
contribution from activity in the liver. This is in accor-
dance with the findings obtained using the R I H S A
for the right renogram significantly exceeded that of the technique (Rodbro and Brochner-Mortensen 1976).
left kidney, i.e. on average by 0.21. In contrast to 131I-hippuran, the extraction fraction
The right-kidney fractional share of the total G F R of 99mTc-DTPA in the kidney after a single injection is
was calculated in all 25 normal subjects. The mean constant (McAfee et al. 1981). Consequently, the UI is
value did not differ significantly from 0.50. The range proportional to the S K G F R in each individual patient.
and 95 ~o confidence limits of the single-kidney frac- In order to use the UI as an accurate measure of
tional share of the total G F R are given in Table 1. S K G F R , the proportionality factor between the UI
The removal function was studied in 21 subjects and the S K G F R , Ek/(E h Vh), should be constant within
with a diuresis exceeding 2 ml/min. The distribution of each individual patient as well as between patients. We
t . . . . A20 and R20 was evaluated after display in the tried to keep the plasma volume within the field of view
histogram mode. A reasonable approximation of the over the heart (Vh) constant by always defining this with
normal distribution was ascertained for the logarithm the same number of pixels. For the fraction Ek/E h to be
of all three parameters; the t . . . . A2o and R20 of the constant within the same patient, the depth of the right
right and left kidneys did not differ. Table 1 gives the and left kidney must be the same. In 45 out of 55
95 ~o lower confidence limits of R20 and the 95 % upper patients, the difference in the skin-to-kidney distance
confidence limits of tm,x and A20; the numerical differ- was only 1 cm or less (Tonnesen et al. 1975). The
ence between the two kidneys for t . . . . Aa0 and R2o are efficiency factor for the kidney, Ek, decreases with the
also shown, and the range of all parameters is given in increasing depth of the kidney. However, it might be
the right column. expected that the efficiency factor for the heart, E h
decreases in parallel with Ek, so that the fraction E k / E h
is nearly the same in all individuals. The proportio-
Discussion
nality factor cannot be calculated directly, and the con-
The present technique for 99~"Tc-DTPA renography stancy cannot be evaluated. However, in 101 conse-
using a gamma camera is based on two assumptions. cutive patients (group I) with G F R s varying between 4
Firstly, the count rate in the ROI over the heart is and 172 ml/min, a good linear correlation was found
proportional to the concentration of tracer in arterial between the sum of the right- and left-kidney UIs and
plasma during the first few minutes of the study and, the G F R calculated from the plasma clearance of
secondly, this count rate is representative of the 99mTc-DTPA. The coefficient of variation in the pro-
changes in the background activity in the field of view spective patients (group II) was, for example, 11.8 % at
over the kidneys. During the first few minutes after the a G F R of 100 ml/min. This variation may have been
injection of tracer, the heart curve reflects the time- due to other factors than those already mentioned.
activity curve for the tracer in arterial plasma. Thereaf- Firstly, the UIs were calculated over 1.5 min, whereas
ter, the heart curve decreases more slowly due to the the plasma clearance of 99mTc-DTPA was calculated
contribution from increasing activity in the extravas- from plasma samples taken 3 5 h after the injection.
cular space within the field of view (personal obser- Changes in the G F R might have occurred during the
vation). In order to be valid for the changes in the investigation. Secondly, the G F R calculated from plas-
background activity over the kidneys, the heart curve ma clearance of 99mTc-DTPA was taken to be an ac-
has to represent the activity from the vascular and curate measure of the actual GFR. However, even at
extravascular volumes in the same ratio and at the these values, there is some imprecision that will in-
same depth as that in the kidney regions. crease the variation. Therefore, the error in the estimate
Taking the second assumption to be valid, the heart of the S K G F R from the renogram is somewhat over-
curve is proportional to the background activity within estimated.
the ROI over the kidney. The proportionality factor Other investigators have also been occupied with
(usually called the subtraction factor) can be calculated the problem of estimating the G F R from renograms
(Aurell et al. 1971; Meldolesi et al. 1973; Nielsen et al. 2.5 2.5
1977; Piepsz et al. 1978; Gates 1982). Recently, Gates y ( t ) - U I y, x(t)
t=l t=l
(1982) has described a method for estimating the c~= {A4)
S K G F R from the fractional renal accumulation of 9
99mTc-DTPA within each kidney. The precision of
As already mentioned, the heart curve, H(t), is fitted
Gates' method seems to be equal to that of our method.
However, with the present technique, the S K G F R can with a sum of two exponentials in the interval
be calculated without determining the injected dose. l_<t<21 min. Therefore, the following expressions are
Therefore, should a partly paravenous injection of the used in the calculations:
tracer accidentally occur, the result will not be in-
validated. ca(t) cR(t)
A renogram is a complex time-activity curve com- y(t)= H(t) al e b~(t-1)+a2e-b2(, 11 (A5)
posed of tracer uptake in the kidney and removal of
tracer from the kidney. N o single parameter from a 1 t
renogram describes the removal function. This is often i H(r)dT S U(rldr+ ~u(r)dr
described by the time required to reach the maximum x (t) - o _o i
count rate of the renogram and by the ratio between H(t) H(t)
the count rate at t = 2 0 min and that at t=tma x. Both 1
a 1
these parameters are greatly influenced by the uptake ! H ( t ) & +~-1 (1-e-bl(t-l))+~2(1-e bz(t-1))
function (Bergstron 1971). As a measure of the removal (A6)
function (being less sensitive to differences in the up- al e-bl(~-1) + a2e-b~-(t-1)
take function), we propose the ratio between the ac-
tivity removed with the urine and the total activity
taken up calculated at t = 2 0 rain (Fig. 1). References
In normal subjects, the values for tma x and A2o
obtained from 99mTc-DTPA renograms exceeded those Aurell M, Jonsson S, Nilsson S (1971) On the quantitation of
obtained from 131I-hippuran renograms using single radiorenograms. Scand J Clin Lab Invest 28:291-297
detectors (Mogensen et al. 1975). This was to be expect- Bergstron H (1971) Influence on the radiorenogram of varia-
ed, since the extraction of D T P A in the kidneys is less tion in renal clearance, renal pelvic volume and urinary
than that of hippuran (McAfee et al. 1981). flow rate. Scand J Clin Lab Invest 28:299 311
In conclusion, in adults the S K G F R can be calcu- Britton KE, Brown NJG (1971) Clinical renography. Lloyd-
lated, as part of the clinical routine, from 99mTc-DTPA Luke, London 15-26:2-4
Brochner-Mortensen J, Rodbro P (1972) A simple method for
gamma-camera renography without determining the in-
the determination of glomerular filtration rate. Scan J Clin
jected dose and without collecting urine or blood sam- Lab Invest 30:271-274
ples. Brochner-Mortensen, Rodbro P (1976) Comparison between
total and renal plasma clearance of SlCr-EDTA. Scand J
Clin Lab Invest 36:247 249
Appendix Carlsen JE, Moiler ML, Lund JO, Trap-Jensen J (1980) Tc-
Estimation of the uptake index and background-subtrac- 99m(Sn)DTPA preparations used for the measurement of
tion factor(o:) glomerular filtration rate. J Nucl Med 21 : 126-129
Gates GF (1982) Glomerular filtration rate: Estimation from
As already described, the equation from which UI and fractional renal accumulation of 99mTc-DTPA. Am J
can be estimated is: RoentgenoI Radium Ther Nucl Med 138:565-570
Kirkegaard P (1970) A FORTRAN IV version of the sum-of-
exponential least-squares code EXPOSUM. Report M-
i H(T) dT 1279 of the Danish Atomic Energy Commission (research
CR(t) = UI o ~-c~ (A1) establishment Riso)
H(t) H(t) McAfee JG, Grossman ZD, Gagne G, Zens AL, Subramanian
G, Thomas FD, Fernandez P, Roskopf ML (1981) Com-
i H(T) dT parison of renal extraction efficiencies for radioactive age-
cR(O nts in the normal dog. J Nucl Med 22:333-338
Substituting H ( t ) - as y(t), and 0 H(t) as x(t), the
Makler PT, Charkes ND, Malmud LS (1979) Calculation of
equation can be reduced to: net renal uptake curves without use of a blood pool agent.
Eur J Nucl Med 4:339 342
y(t) = UI x(t) + c~. (A2) Meldolesi, U, Mombelli L, Roncari G, Conte L (1973) A
simple method of estimating renal clearance by renog-
Since Y(0 and x(0 are known, UI and c~ can be esti- raphy. J Nucl Biol Med 17 79-83
mated for each kidney by simple linear regression in Mogensen P, Munck O, Giese J (1975) 131I-Hippuran re-
the interval 1 G t G 2.5 min (nine curve points). nography in normal subjects and in patients with essential
The solutions are: hypertension. Scand J Clin Lab Invest 35:301 306
Nielsen SP, Moller ML, Trap-Jensen J (1977) 99mTc-DTPA
2.5 2.5 2.5 scintillation camera renography: A new method for esti-
9 ~ x(O y(t)- y, x(f) E Yi t) mation of single-kidney function. J Nucl Med 18:112 117
UI= t=l
9 ~ X2(t) -
t=l t=l
x(t
,)2 (A3) Pedersen F, Poulsen PE (1967) Hippuran 1-131 renography
and clearance measurements used for determination of the
individual kidney function. J Urol 97:180 183
1=1 t=l Piepsz A, Denis R, Ham HR, Dobbeleir A, Schulman C,
 Erbsman F (1978) A simple method for measuring sepa- Smith H W (1951) The kidney structure and function in health
rate glomerular filtration rate using a single injection of and disease. Oxford University Press, New York, pp 153-
99~Tc-DTPA and a scintillation camera. J Pediatr 93: 164
769-774 Tonnesen KH, Munck O, Hald T, Mogensen P, Wolf H
Rehling M, Moller ML, Thamdrup B, Lund JO, Trap-Jensen J (1975) Influence on the radiorenogram of variation in skin
(1984) Simultaneous measurement of renal clearance and to kidney distance and the clinical importance hereof. In:
plasma clearance of 99~Tc-DTPA, 51Cr-EDTA and inulin zum Winkel K, Blaufox MD, Funck-Brentano J-L (eds)
in man. Clin Sci 66:613-619 Radionuclides in nephrology. Georg Thieme, Stuttgart, pp
Rodbro P, Brochner-Mortensen J (1976) Systematic difference 79-86
of extrarenal background in left and right kidney regions
at renography. Scand J Clin Lab Invest 36:189-192
Rutland M D (1979) A single injection technique for sub-
traction of blood background in 131i_hippuran renograms.
Br J Radiol 52:134-137 Received January 12, 1985