Background Risk of a subsequent stroke following an acute transient ischemic attack (TIA) or minor stroke is high. The ABCD² tool was proposed as a method to triage these patients using five clinical factors. Modern imaging of the brain was not included. The present study quantified the added value of magnetic resonance imaging (MRI) factors to the ABCD² tool. Methods Patients with TIA or minor stroke were examined within 12 h and had a brain MRI within 24 h of symptom onset. Primary outcomes were recurrent stroke and functional impairment at 90 days. A new tool, ABCD²+ MRI, was created by adding diffusion-weighted imaging lesion and vessel occlusion status to the ABCD² tool. The predictive accuracy of both tools was quantified by the area under the curve (AUC). Results One hundred and eighty patients were enrolled and 11·1% had a recurrent stroke within 90 days. The predictive accuracy of the ABCD²+MRI was significantly higher than ABCD² (AUC of 0·88 vs. 0·78, P = 0·01). Those with a high score ( 7 – 9 ) had a 90-day recurrent stroke risk of 32·1%, moderate score ( 5 – 6 ) risk of 5·4%, and low score (0-4) risk of 0·0%. The ABCD² tool did not predict risk of functional impairment at 90 days ( P = 0·33), unlike the ABCD²+MRI ( P = 0·02): high score (22·9%), moderate (7·5%), low (7·7%). Conclusions Risk of recurrent stroke and functional impairment after a TIA or minor stroke can be accurately predicted by a scoring system that utilizes both clinical and MRI information. The ABCD²+MRI score is simple and its components are commonly available during the time of admission.