Can intravenous contrast enhanced (CE) PET/CT be used for SUV measurements in the detection of response to treatment?

Objectives The bilinear conversion method is commonly used to translate CT Hounsfield Units (HU) to 511 keV attenuation values (µ). We investigated if µ derived from CECT, differ from non-CE µ and might affect the SUV. A phantom study to simulate the effects of different IV contrast concentrations on SUV was added.

Methods PART I: PETCT images were acquired in 21 pts (Philips Gemini TF64). CT was acquired using a low dose protocol (120KeV, 50mAs) and after contrast administration during venous phase (90s post injection, 120KeV, 250mAs). SUVmax was measured in 32 regions of interest, drawn over normal tissues with different densities (lung, liver, bone, muscle and fat) and in 28 pathological sites. Paired 1-tailed student t-test was used. PART II: Physiological contrast concentrations were determined by measuring HU in the left ventricle during arterial and venous phase in 14 individuals. Consecutively a phantom study was performed in which SUVmax was measured after filling 7 vials with identical activities (0,18MBq/ml FDG) and increasing iodine contrast concentrations (0-3%).

Results PART I: The largest increase in SUVmax was measured in the liver: 2.86 for CECT vs 2.74 (or 4,2%) for non-CECT (p<0.01). No significant SUVmax difference was found in muscle, lung, bone and fat. Within the sites of pathology, the SUVmax difference was significant 10.90 vs 10.59 for non-CT (or 2.8%)(p<0.05). PART II: The iodine contrast agent concentration in the arterial and venous phase was resp 2.78 and 1.93%. In the vials containing 0-3% concentrates, the most important increase in SUVmax was 0.055 (or 5.77%) and seen at 1.5% .

Conclusions Quantification of the reconstructed PET image, obtained from non-CECT vs CECT in the venous phase, is significantly different in normal liver tissue and pathological sites. However the difference is small and will unlikely be of clinical relevance when using quantification in the determination of response to treatment. This conclusion was confirmed with the findings of our phantom study