Abstract
Background A significant proportion of ischemic strokes are caused by emboli from unstable atherosclerotic carotid artery plaques with inflammation being a key feature of plaque instability and stroke risk. Positron emission tomography (PET) depicting the uptake of 2-deoxy-2-(18F)-fluoro-D-glucose (18F-FDG) in carotid artery plaques is a promising technique to quantify plaque inflammation. A consensus on the methodology for plaque localization and quantification of inflammation by 18F-FDG PET/computed tomography (CT) in atherosclerosis has not been established. High inter-reader agreement is essential if 18F-FDG PET/CT is to be used as a clinical tool for the assessment of unstable plaques and stroke risk. The aim of our study was to assess the inter-reader variability of different methods for quantification of 18F-FDG uptake in carotid atherosclerotic plaques with a separate CT angiography (CTA) providing anatomical guidance.
Methods and results Forty-three patients with carotid artery stenosis ≥70% underwent 18F-FDG PET/CT. Two independent readers separately delineated the plaque in all axial PET slices containing the atherosclerotic plaque and the maximum standardized uptake value (SUVmax) from each slice was measured. Uptake values with and without background correction were calculated. Intraclass correlation coefficients were highest for uncorrected uptake values (0.97-0.98) followed by those background corrected by subtraction (0.89-0.94) and lowest for those background corrected by division (0.74-0.79). There was a significant difference between the two readers definition of plaque extension, but this did not affect the inter-reader agreement of the uptake parameters.
Conclusions Quantification methods without background correction have the highest inter-reader agreement for 18F-FDG PET of carotid artery plaque inflammation. The use of the single highest uptake value (max SUVmax) from the plaque will facilitate the method’s clinical utility in stroke prevention.