Overview of 5 Methods Used to Determine 18F-FDG Input Function from Vena Cava Time–Activity Curve in Rat
| IDIF extraction approach | Method 1, dispersion-corrected IDIF | Method 2, dispersion-corrected IDIF, a = 1 | Method 3, dispersion-corrected IDIF, 1 blood sample | Method 4, raw IDIF | Method 5, raw IDIF, 1 blood sample |
| Value of parameter a | 1.10 | 1 | 1 | — | — |
| Scaling with 1 manual blood sample | No | No | Yes | No | Yes |
| AUC ratio (-) | 1.02 ± 0.16 | 0.93 ± 0.15 | 1.00 ± 0.06 | 0.93 ± 0.15 | 1.15 ± 0.12 |
| CMRGlc ratio (-) | 0.94 ± 0.16 | 1.03 ± 0.18 | 0.94 ± 0.07 | 0.93 ± 0.16 | 0.75 ± 0.09 |
| CMRGlc, Patlak ratio (-) | 0.91 ± 0.16 | 1.00 ± 0.16 | 0.92 ± 0.07 | 1.00 ± 0.16 | 0.81 ± 0.09 |
For methods 1–3, parameters of generic dispersion function used in dispersion correction (Eq. 1) were set to their average values found on the 4 rats (Table 1). In methods 2 and 3, parameter a was set to 1 in generic dispersion function. In method 3, manual blood sample withdrawn 1.5 min after infusion was used as scaling reference for corrected input function. In method 4, raw IDIF from vena cava was used. In method 5, manual blood sample withdrawn 1.5 min after infusion was used as scaling reference for raw IDIF. Third line summarizes ratio of AUC found with different image-derived input function and AUC of arterial input function measured with external blood counter. Uncertainty in AUC represents SEM. Fourth and fifth lines show results of simulation studies undertaken to test effects of different IDIF approaches on determination of CMRGlc. In fourth line, ratio of CMRGlc determined by compartmental modeling on simulated brain tissue–activity to true CMRGlc used to simulate tissue–activity curve is summarized, whereas last line shows results for CMRGlc_Patlak determined with Patlak approach. Uncertainty in CMRGlc and CMRGlc_Patlak represents SEM.
(-) = values are dimensionless.