Skip to main content

Table 1 Comparison of two analysis pipelines used to estimate metabolic rate of 18F-FDG to assess pulmonary inflammation

From: Reproducibility of compartmental modelling of 18F-FDG PET/CT to evaluate lung inflammation


Pipeline A

Pipeline B: initial

Pipeline B: final

Blood ROI size

Circular, 5 pixel diameter, aortic arch to variable

Circular, 8 pixel diameter, 25 slices beginning aortic arch

As pipeline A

Lung ROI closing and erosion operation

5 pixel diameter disc

3 pixel diameter disc

As pipeline B initial

Input function model

Exponential basis functions

Tri- or biexponential fits

As pipeline A

Time delay estimation

Inside compartmental model optimisation

Outside compartmental model optimisation

As pipeline A

Time delay fitting

Delays spanning −50 to 50 s using 1-compartmental model fitted for first 5 min—lowest residual sum of squares

Additional parameter within estimation of rate constants

As pipeline A


Local optimum

Global optimum

As pipeline A

Start point of optimisation

K1 = 0.5, k2 = 0.2, k3 = 0.3, Vb = 0.1

Multiple start points generated finds best guess (lowest objective function value)

As pipeline A

  1. The columns “Pipeline B: initial” and “Pipeline B: final” describes the parameters which were used in the initial evaluation and the final settings used following adjustments to pipeline B respectively. The table highlights the key differences between the implementations of the compartmental model. Parameters not included below were identical between the analysis pipelines. WL = whole lung, DA = descending aorta, ROI = region of interest