Vol 12, No 1 (2009)
Research paper
Published online: 2009-04-21
Assessment of accuracy and precision of 99mTc-HEPIDA clearances determined by means of a simplified method
Nucl. Med. Rev 2009;12(1):14-22.
Abstract
BACKGROUND: The aim of the present study was the assessment
of the accuracy and precision of our own simplified method
for the determination of 99mTc-HEPIDA liver clearance.
MATERIAL AND METHODS: It has been assumed that archived results of plasma clearance (ClPl) and hepatic (ClHp), determined by means of multisample methods, could be legitimately used as a reference standard. The accuracy and precision of the simplified method was assessed by means of a Monte Carlo method alternatively utilizing three blood sampling times (T) of 68, 75 and 83 minutes post i.v. administration of 99mTc-HEPIDA. The corresponding alternative three urine voiding times (Y) were: 75, 80, and 95 min p.i. The analysed model was created accepting values of ClPl and ClHp, of administered activity Ap and parameters of biexponential function, describing the concentration C(t) decrease of the radiopharmaceutical (RF) in plasma during time as real values. Using the function C(t) for each individual, the plasma concentrations of RF at three sampling times, urinary clearance (ClPl - ClHp), and voided activity (AUr(Y)) were calculated. Simulated random errors were added to the assumed blood sampling times T and to voiding time Y. To the activity Ap and AUr(Y), and RF plasma concentrations random errors were added, assuming normal distribution with relative SD from 0 to 5% and then clearance values were computed. For each process there were 5000 repeated simulated determinations. The accuracy of the simplified methods was assessed by comparing mean values of simulated clearance computations with the reference. Comparison of standard deviations with mean uncertainties enabled us to gain insight into the degree of agreement of the estimator of relative uncertainty with the coefficient of variation as a measure of precision.
RESULTS: There were strong correlations between the reference clearance values and the mean values of determinations by means of the simplified procedure (r > 0.93). The correlations were practically insensitive to the uncertainty of pipetting. The lines of regression differed slightly from the lines of identity, giving an indication that there was a systematic error involved; it amounted to +4 ml/min at ClPl = 60 ml/min and to -7 ml/min for ClPl of 370 ml/min. For ClPl a bias of +6 ml/min was found for a clearance value of 16 ml/min and –13 ml/min at ClPl > 300 ml/min. At uncertainty of pipetting of 2%, a precision of 6-7% was found for ClPl of 300 ml/min. For ClPl of 200 and 150 ml/min the corresponding precisions were 7–8% and 10%, respectively. For ClPl of 200, 150 and 100 ml/min the corresponding precisions were 10, 12 and 17%, respectively. These precisions are 5 percent worse than those that were obtained from determinations by means of multisampling procedures.
MATERIAL AND METHODS: It has been assumed that archived results of plasma clearance (ClPl) and hepatic (ClHp), determined by means of multisample methods, could be legitimately used as a reference standard. The accuracy and precision of the simplified method was assessed by means of a Monte Carlo method alternatively utilizing three blood sampling times (T) of 68, 75 and 83 minutes post i.v. administration of 99mTc-HEPIDA. The corresponding alternative three urine voiding times (Y) were: 75, 80, and 95 min p.i. The analysed model was created accepting values of ClPl and ClHp, of administered activity Ap and parameters of biexponential function, describing the concentration C(t) decrease of the radiopharmaceutical (RF) in plasma during time as real values. Using the function C(t) for each individual, the plasma concentrations of RF at three sampling times, urinary clearance (ClPl - ClHp), and voided activity (AUr(Y)) were calculated. Simulated random errors were added to the assumed blood sampling times T and to voiding time Y. To the activity Ap and AUr(Y), and RF plasma concentrations random errors were added, assuming normal distribution with relative SD from 0 to 5% and then clearance values were computed. For each process there were 5000 repeated simulated determinations. The accuracy of the simplified methods was assessed by comparing mean values of simulated clearance computations with the reference. Comparison of standard deviations with mean uncertainties enabled us to gain insight into the degree of agreement of the estimator of relative uncertainty with the coefficient of variation as a measure of precision.
RESULTS: There were strong correlations between the reference clearance values and the mean values of determinations by means of the simplified procedure (r > 0.93). The correlations were practically insensitive to the uncertainty of pipetting. The lines of regression differed slightly from the lines of identity, giving an indication that there was a systematic error involved; it amounted to +4 ml/min at ClPl = 60 ml/min and to -7 ml/min for ClPl of 370 ml/min. For ClPl a bias of +6 ml/min was found for a clearance value of 16 ml/min and –13 ml/min at ClPl > 300 ml/min. At uncertainty of pipetting of 2%, a precision of 6-7% was found for ClPl of 300 ml/min. For ClPl of 200 and 150 ml/min the corresponding precisions were 7–8% and 10%, respectively. For ClPl of 200, 150 and 100 ml/min the corresponding precisions were 10, 12 and 17%, respectively. These precisions are 5 percent worse than those that were obtained from determinations by means of multisampling procedures.
Keywords: hepatic 99mTc-HEPIDA clearance errorhepatic 99mTc-HEPIDA clearance determination
