Causal role of non-stenotic hypodense plaques with high FDG uptake detected with PET-CT angiography in patients with unknown cause of stroke?

Objectives Ruptured complicated non-stenotic carotid plaques are a well-described cause of ischemic stroke. Accurate identification of complicated atherosclerotic plaques associated with ischemic stroke is crucial because the prevalence of carotid stenosis < 50 % is high in the population. CT angiography (CTA) and 18F-fluoro-deoxyglucose positron emission tomography (18F-FDG PET) imaging allows for the detection of certain morphological and functional aspects characteristics of complicated carotid plaques. The aim of this study was to investigate in patients with ischemic stroke of unknown origin the value of FDG-PET-CTA for the identification of complicated atherosclerotic plaques.

Methods A total of 17 patients with ischemic stroke of unknown origin in the territory of the anterior or middle cerebral artery < 14 days were evaluated in this study. PET imaging of carotid arteries was acquired 2 hours after injection of 4 MBq/kg of FDG and followed by CTA of the supra-aortic trunks. Plaque localization and maximal thickness as well as the presence of hypodense areas (< 30 Hounsfield units) and its extent (maximal hypodense area measured on cross-sections) were identified on CTA. The intensity of FDG uptake was quantified using tissue to background ratio (TBR) with PET on 3 adjacent slices centered on each plaque (TBR max.) and on 12 axial slices along each carotid artery (TBR mean). In addition, hypodense plaques with high FDG uptake on PET-CTA were further imaged in 4 patients with high-resolution MRI to search for the presence of complicated plaques in this location.

Results Patients (mean age: 63 ± 27; 41% female) were imaged 7 days (1-12 days) after ischaemic stroke. A total of 16 non-stenotic plaques were identified in the carotid artery ipsilateral to stroke and 8 contralateral to the stroke. The prevalence of hypodense plaques was significantly higher in the carotid artery ipsilateral vs. contralateral to stroke (50 % vs. 22 %; respectively, p < 0.05). Hypodense area in plaques did, however, not significantly differ between the carotid artery ipsilateral vs contralateral to stroke (2.3 ± 1.8 mm2 vs. 1.5 ± 0.8 mm2; p = 0.50). FDG uptake was not significatively higher in non-stenotic hypodense vs non-hypodense plaques (TBR max. = 2.12 ± 0.97 vs. 1.87 ± 0.53; p = 0.29). Mean and maximal TBR were significantly higher in ipsilateral vs. contralateral carotid artery (1.75 ± 0.73 vs 1.64 ± 0.76 and 2.10 ± 0.91 vs. 1.81 ± 0.55; p= 0.003 and p= 0.006; respectively). The prevalence of non-stenotic hypodense plaques with TBR max. > 1.80 was significantly higher in the ipsilateral carotid artery vs contralateral to stroke (44% vs. 11%, p < 0.05, respectively) and detected in various locations along the carotid arteries (1 in the common carotid artery, 10 in the carotid bulb and 7 in the origin of the internal carotid artery). In 4 patients, high-resolution of the vascular wall was performed at the level of the hypodense plaque with high FDG uptake and confirmed the presence of complicated plaques in all 4 patients (plaque rupture (n = 2), thrombus (n = 1), intra-plaque hemorrhage (n= 1)).

Conclusions In this exploratory study, we found a high prevalence of non-stenotic hypodense plaques with high FDG uptake (TBR max. > 1.80) in carotid arteries ipsilateral to cerebral ischemia suggesting a causal role for these plaque in patients with unknown cause of stroke. In comparison to high-resolution MRI, the imaging criteria used in this study for the identification of complicated plaques were very simple and easy to implement. In addition, FDG-PET-CTA allows for the evaluation of non-stenotic plaques along the whole length of supra-aortic trunks. Further studies with a larger number of patients are requested to confirm the potential of FDG-PET-CTA for the evaluation of carotid arteries in patients with ischemic stroke of unknown origin.