Long-term radiographic outcomes of microemboli following carotid interventions

Subclinical microemboli on diffusion-weighted magnetic resonance imaging (DWI) have been identified immediately following carotid revascularization procedures, but the clinical significance and long-term effects are largely unknown. The purpose of this study was to evaluate long-term radiographic outcomes of these DWI lesions.

Patients who underwent perioperative magnetic resonance imaging (MRI) evaluations for carotid interventions at a single institution from July 2004 to December 2008 were evaluated, particularly those who had additional follow-up MRI. DWI with apparent diffusion coefficient (ADC), fluid-attenuated inversion recovery (FLAIR), and T2-weighted MRI images were compared to determine long-term effect of microemboli.

One-hundred sixty-eight consecutive patients (68 carotid artery stenting [CAS] and 100 carotid endarterectomy [CEA]) who received perioperative MRI were included. All CAS were performed with an embolic protection device. The incidence of microemboli was significantly higher in the CAS group than the CEA group (46.3% and 12%, respectively, P < .05) despite a relative low incidence of procedure-associated neurologic symptoms in both groups (2.9% vs 2%). Thirty patients (16 CAS and 14 CEA) who had follow-up MRI were further analyzed and a total of 50 postoperative DWI lesions (mean size 46.57 mm(2); range 16 to 128 mm(2)) were identified among them. During a mean MRI follow-up of 10 months (range, 2 to 23 months), residual MRI abnormalities were only identified in DWI lesions larger than 60 mm(2) on postoperative MRI and on postoperative FLAIR images (n = 5, P < .001). The CEA group had fewer but larger ipsilateral distributed emboli (total 12 lesions, mean 79 mm(2)) compared with the CAS group (total 38 lesions, mean 27.5 mm(2), P < .05).

The majority of microemboli do not have long-term radiographic sequelae. Size and hyperintensity on postoperative FLAIR are predictive of residual brain structure abnormality, and further neurocognitive evaluations are warranted.