Abstract
174
Objectives: The Siemens Biograph Vision 600 is a modern PET/CT with capabilities of great resolution. Head and neck 18F-FDG PET/CT images at Northwestern Memorial Hospital are acquired with a zoom (unlike general skull to thigh imaging) to greater magnify smaller lesions in this area. Applying a zoom will create a reconstructed image that will possess smaller image voxel size, potentially allowing for more precise measurements on sub-centimeter lesions. Alternatively, the reduction in pixel size will lead to an unsurmountable amount of noise in PET images which could lead to qualitative and quantitative diagnostic uncertainty. The objective of this study was to see if there are statistically significant differences in Standard Uptake Values (SUV) measurements between reconstructions with different applied zoom factors.
Methods: Seventeen head and neck 18F-FDG PET/CT examinations were included in this study. Of these 17 examinations, 10 presented with metastatic disease within the head and neck region. All 10 patients had their head and neck PET/CT images retro-reconstructed and examined with varying levels of zoom including 1.5 (institution standard), 2.0 and 2.5. Images were reviewed in a SyngoVia viewing software. Lesions were evaluated by a Volume of Interest (VOI). The SUVmax of each lesion in all 3 zoomed reconstructed images were recorded. The SUVmax and SUVmean for the sternocleidomastoid muscle (SCM) were also generated. These values were statistically analyzed using a one sample t-test. Results were considered statistically insignificant with a p-value greater than 0.05.
Results: One sample t-tests were completed comparing the SUVmax of the head and neck lesions for the 1.5 zoom factor (the standard) versus the 2.0 and 2.5 zoom factors, which were statistically insignificant (p > 0.05). The SUVmean values for the SCM were also statistically insignificant when comparing the 1.5 to the 2.0 and 2.5 zoom factors.
Conclusions: Head and neck 18F-FDG PET/CT imaging with reconstructed zooms of 1.5, 2.0 and 2.5 showed no statistical difference in SUV parameter measurements. These findings imply that our clinical standard head and neck imaging could be zoomed further without statistical difference in SUV measurements.