Skip to main content

Table 1 Goodness of fits and Akaike weights for the PBMS method

From: A population-based method to determine the time-integrated activity in molecular radiotherapy

Equation number Function name K Coefficient of Variation CV (max) d Off-diagonal values of the correlation matrix (maxc) Akaike weight (%) Jackknife Akaike weights (% median [min,max])
1 \(f_{2a}\) 26 0.04 0.92; 0.95; 0.99b
2 \(f_{2b}\) 26 0.31 0.81; 0.82; 0.86 0.03 0 [0,50]
3 \(f_{2c}\)a 26
4 \(f_{2d}\)a 26
5 \(f_{2e}\)a 26
6 \(f_{3a}\) 39 1.96E5b 0.93; 0.98; 0.99b
7 \(f_{3b}\)a 39
8 \(f_{3c}\)a 39
9 \(f_{3d}\) 39 1.55E + 6b 0.95; 0.98; 0.99b
10 \(f_{2a,3d}\)a 26
11 \(f_{2b, 3d}\)a 26
12 \(f_{2c,3d}\)a 26
13 \(f_{3aS1}\) 26 + 1 4.27E5b 0.95; 0.98; 0.99b
14 \(f_{3aS2}\) 26 + 1 0.32 0.57; 0.59; 0.64 0.04 0 [0,66]
15 \(f_{3aS3}\) 26 + 1 0.31 0.72; 0.73; 0.76 2.49 3 [0,58]
16 \(f_{3aS4}\) 26 + 1 0.37 0.79; 0.81; 0.84 97.40 97 [33,100]
17 \(f_{3aS5}\) 26 + 1 8.64b 0.98; 0.99; 1.00b
18 \(f_{3aS6}\) 26 + 1 0.14 0.64; 0.67; 0.72 0.04 0 [0,3]
19 \(f_{1}\)a 13
20 \(f_{4}\)e 52
  1. The total number of biokinetic data N used in this retrospective analysis is 46, the numbers of parameters of the functions are given in column K
  2. aThe fitting failed based on the visual inspection of the fitted graph
  3. bInadequate goodness of fit (these functions should not be used for model selection)
  4. cThree largest (absolute) values of K* (K − 1)/2 lower-off-diagonal elements. Note that a low percentage of elements only slightly higher than 0.8 is acceptable
  5. dCV for the fit parameters calculated as SD divided by the mean
  6. eThe fitting failed as the number of parameters is larger than the number of data N = 46