Value of diffusion-weighted MRI during carotid angioplasty and stenting

Introduction. The incidence of neurological injury following carotid angioplasty and stenting is of great interest to those advocating it as an alternative to endarterectomy in the management of critical carotid stenosis. A significant inter-observer variation exists in determining the presence or absence of a neurological deficit following the procedure objective imaging would be advantageous. In this study, we sought to assess diffusion weighted MRI as a diagnostic tool in evaluating the incidence of neurological injury following carotid angioplasty and stenting (CAS). Patients and methods. The first 110 cases of CAS in our unit were included in this series. The procedure was abandoned in three patients. Patients underwent intracranial and extracranial MR angiography, together with diffusion-weighted MRI (DWI) prior to and following CAS and had a formal neurological assessment in the intensive care unit after the procedure. Results. One hundred and ten Procedures were attempted in 98 patients. Twenty-eight percent were asymptomatic. Following CAS, 7.2% of patients had a positive neurological exam (two major strokes with one fatality) and 21% had positive DWI scans, equating to a sensitivity of 86% and a specificity of 85% for DWI in detecting cerebral infarction following CAS. The positive predictive value of the test was 0.3 and negative predictive value 0.99. The major stroke and death rate was 1.8%. While the use of a cerebral protection device appeared to significantly reduce the incidence of cerebral infarction (5% vs. 25%, p = 0.031) this may be a reflection of the learning curve encountered during the study. Conclusion. The incidence of subclinical DWI detected neurological injury was significantly higher than clinical neurological deficit following CAS. Conventional methods of neurological assessment of patients undergoing CAS may be too crude to detect subtle changes and more sensitive tests of cerebral function are required to establish whether these subclinical lesions are relevant.