TABLE 1

Kinetic Modeling of Prostate

Parameter	3C-5p	3C-5p	3C-4p	3C-4p	3C-3p	3C-3p	2C-3p	2C-3p	Patlak-2p	Patlak-2p
Metabolite correction	No	Yes	No	Yes	No	Yes	No	Yes	No	Yes
n	22	22	22	22	22	22	22	22	22	22
k₁	0.289 ± 0.130	0.318 ± 0.145	0.301 ± 0.131	0.319 ± 0.142	0.295 ± 0.134	0.316 ± 0.145	0.289 ± 0.132	0.231 ± 0.099
k₂	0.190 ± 0.051	0.184 ± 0.108	0.206 ± 0.054	0.216 ± 0.091	0.190 ± 0.047	0.202 ± 0.100	0.177 ± 0.046	0.048 ± 0.013
k₃	0.006 ± 0.010	0.126 ± 0.078	0.012 ± 0.018	0.141 ± 0.046	0.005 ± 0.008	0.124 ± 0.046
k₄	0.002 ± 0.002	0.008 ± 0.008	0.008 ± 0.017	0.006 ± 0.004
K	0.006 ± 0.010	0.124 ± 0.062	0.012 ± 0.013	0.124 ± 0.049	0.006 ± 0.008	0.118 ± 0.049			0.005 ± 0.011	0.109 ± 0.056
V_b	0.020 ± 0.020	0.024 ± 0.026					0.001 ± 0.003	0.000 ± 0.001
V_d	1.52 ± 0.63	1.50 ± 1.28	1.44 ± 0.67	0.96 ± 0.46	1.53 ± 0.59	1.08 ± 1.03	1.62 ± 0.60	4.87 ± 1.80	1.51 ± 0.61	1.57 ± 1.50
WRSS	397 ± 420	295 ± 278	439 ± 448	376 ± 368	491 ± 478	372 ± 386	491 ± 470	852 ± 722
N_F	0	9	1	6	0	4	Reference M−	Reference M+
AIC	71 ± 28	65 ± 26	72 ± 28	68 ± 28	74 ± 28	66 ± 28	75 ± 27	92 ± 24

2C, 3C = number of compartments; 2p, 3p, 4p, 5p = number of parameters in model. In model 3C-4p, V_b = 0 and in model 3C-3p both V_b and k₄ are fixed to 0. N_F = number of studies with a significantly lower error compared with corresponding 2C-3p (M− and M+—i.e., without and with metabolite correction as reference) determined with F test (P < 0.05). AIC (Akaike Information Criterion) is defined in Equation 3, and WRSS was used to calculate the likelihood function (35).
Estimates of fitted parameters k₁–k₄, as well as influx rate constant K, V_b, and V_d. Results are expressed as mean ± 1 SD of 22 studies (n) in 22 patients. Input function was corrected for partial-volume effects with 0.6 as recovery coefficient and for metabolites as indicated. Error estimate WRSS (Eq. 4) is a measure of “goodness of fit” between measured and calculated output function. Graphical analysis with Patlak plots was applied on images between 7 and 20 min.