BACKGROUND: The International Carotid Stenting Study (ICSS) of stenting and endarterectomy for symptomatic carotid stenosis found a higher incidence of stroke within 30 days of stenting compared with endarterectomy. We aimed to compare the rate of ischaemic brain injury detectable on MRI between the two groups. METHODS: Patients with recently symptomatic carotid artery stenosis enrolled in ICSS were randomly assigned in a 1:1 ratio to receive carotid artery stenting or endarterectomy. Of 50 centres in ICSS, seven took part in the MRI substudy. The protocol specified that MRI was done 1-7 days before treatment, 1-3 days after treatment (post-treatment scan), and 27-33 days after treatment. Scans were analysed by two or three investigators who were masked to treatment. The primary endpoint was the presence of at least one new ischaemic brain lesion on diffusion-weighted imaging (DWI) on the post-treatment scan. Analysis was per protocol. This is a substudy of a registered trial, ISRCTN 25337470. FINDINGS: 231 patients (124 in the stenting group and 107 in the endarterectomy group) had MRI before and after treatment. 62 (50%) of 124 patients in the stenting group and 18 (17%) of 107 patients in the endarterectomy group had at least one new DWI lesion detected on post-treatment scans done a median of 1 day after treatment (adjusted odds ratio [OR] 5.21, 95% CI 2.78-9.79; p<0.0001). At 1 month, there were changes on fluid-attenuated inversion recovery sequences in 28 (33%) of 86 patients in the stenting group and six (8%) of 75 in the endarterectomy group (adjusted OR 5.93, 95% CI 2.25-15.62; p=0.0003). In patients treated at a centre with a policy of using cerebral protection devices, 37 (73%) of 51 in the stenting group and eight (17%) of 46 in the endarterectomy group had at least one new DWI lesion on post-treatment scans (adjusted OR 12.20, 95% CI 4.53-32.84), whereas in those treated at a centre with a policy of unprotected stenting, 25 (34%) of 73 patients in the stenting group and ten (16%) of 61 in the endarterectomy group had new lesions on DWI (adjusted OR 2.70, 1.16-6.24; interaction p=0.019). INTERPRETATION: About three times more patients in the stenting group than in the endarterectomy group had new ischaemic lesions on DWI on post-treatment scans. The difference in clinical stroke risk in ICSS is therefore unlikely to have been caused by ascertainment bias. Protection devices did not seem to be effective in preventing cerebral ischaemia during stenting. DWI might serve as a surrogate outcome measure in future trials of carotid interventions. FUNDING: UK Medical Research Council, the Stroke Association, Sanofi-Synthélabo, European Union, Netherlands Heart Foundation, and Mach-Gaensslen Foundation.