The efficacy of endovascular interventions based on magnetic resonance (MR) imaging for acute ischemic stroke caused by embolic carotid T occlusion was assessed. Endovascular intervention was performed in 10 of 751 consecutive acute ischemic stroke patients admitted to our institution between April 2001 and July 2003. Indications were: presentation within 6 hours of onset of cardioembolic stroke; National Institute of Health Stroke Scale (NIHSS) score ≥10 on admission; absence of extensive signal hyperintensity on diffusion-weighted imaging; and carotid T occlusion on angiography. Localized intra-arterial thrombolysis using urokinase and/or mechanical clot disruption using micro-balloon catheter were performed. Radiographic findings, ΔNIHSS defined as NIHSS on day 7 − NIHSS on admission, 3-month modified Rankin scale (mRS), 3-month mortality, and symptomatic intracranial hemorrhage (ICH) were evaluated. Partial recanalization was achieved in 5 of 10 patients, but no recanalization was obtained in 5 patients. No significant differences in 3-month mRS (0-2) or symptomatic ICH were identified between these groups. However, the partial recanalization group showed significantly better clinical outcomes (median 3-month mRS 3 vs. 6, p = 0.009) and lower mortality rate at 90 days (0% vs. 80%, p = 0.047) with significantly greater ΔNIHSS (−8 vs. +6, p = 0.009). Endovascular intervention based on MR imaging resulting in partial recanalization may reduce poor outcomes and mortality in patients with embolic carotid T occlusion. Improvement of the outcomes of such patients requires more efficient methods such as clot retrieval therapy.