J. Kissel et al., Noninvasive determination of the arterial input function of an anticancer drug from dynamic PET scans using the population approach, MED PHYS, 26(4), 1999, pp. 609-615
Citations number
35
Categorie Soggetti
Radiology ,Nuclear Medicine & Imaging","Medical Research Diagnosis & Treatment
For the application of a kinetic model to PET data, it is generally necessa
ry to obtain the arterial input function (AIF). It was the aim of the prese
nt study to introduce a method suitable for the determination of the AIF of
a substance that undergoes biochemical transformation from noisy PET data:
the population approach. F-18 labeled 5-fluorouracil (5-[F-18]FU) was admi
nistered i.v. to eight patients suffering from liver metastases of colorect
al carcinoma. Radioactivity concentrations in liver and aorta were dynamica
lly measured with PET over 120 min. Pharmacokinetic analysis was carried ou
t by applying a five-compartment model to individual activity-time data for
the eight patients or to the mean activity-time data among the eight patie
nts. The mean values of all parameters describing 5-FU transport and catabo
lism, i.e., volumes of distribution and clearances, as well as interindivid
ual coefficients of variation (CV) were calculated according to both approa
ches. With our model, we were able to separate the concentration-time cours
e of 5-FU in plasma, i.e., the AIF, from that of its major catabolite alpha
-fluoro-beta-alanine (FBAL). As far as the mean parameter estimates are con
cerned, the differences between both approaches are not significant. For th
e liver data, the CV's are almost the same for both approaches. For the par
ameters concerning the aorta, however, there is a decrease in the CV's by u
sing the population approach. For example, the CV of the central distributi
on volume of 5-FU was 30% for the individual approach and 18% for the popul
ation approach. With the population approach, it is possible to determine t
he AIF of drugs that undergo metabolic conversion, such as anticancer drugs
, from the abdominal aorta visualized on PET images. The population approac
h helps to overcome noise in individual data. Since no measurements are nee
ded in addition to the PET examination, the suggested method helps to reduc
e risk and pain for the patients as well as cost and thus facilitates large
scale patient studies. (C) 1999 American Association of Physicists in Medi
cine. [S0094-2405(99)00904-9].