Background and purpose: Whether to withhold mechanical thrombectomy when the diffusion-weighted imaging (DWI) lesion exceeds a given volume is undetermined. Our aim was to identify markers that will help to select patients with large DWI lesions [DWIÀAlberta Stroke Program Early Computed Tomography Score (DWI-ASPECTS) ≤ 5] that may benefit from thrombectomy. Methods: From May 2010 to November 2016, 82 acute ischaemic stroke patients with DWI-ASPECTS ≤5 (43 men, 64.6 AE 14.4 years, National Institutes of Health Stroke Scale 18.4 AE 5.4) treated with state-of-the-art mechanical thrombectomy were studied. Thrombectomy alone was performed in 28 (34%) and bridging therapy in 54 (66%) patients. Recanalization was defined as a thrombolysis in cerebral infarction score 2B-3 and significant hemorrhagic transformation as parenchymal haematoma type 2 (European Cooperative Acute Stroke Study 3 classification). Pretreatment variables were compared between patients with a good (modified Rankin Scale 0À2) and a poor (modified Rankin Scale 3À6) neurological outcome at 3 months. Results: Overall, 28 patients (34%) achieved good neurological outcome at 3 months. Recanalizers were significantly more likely to achieve good outcome (61% vs. 7.3%, P < 0.0001), had lower mortality (24% vs. 49%, P = 0.03) and similar rates of parenchymal haematoma type 2 (9.8% vs. 7.3%, P = 1) compared to non-recanalizers. Regression modelling identified DWI-ASPECTS >2 [odds ratio (OR) 6.93; 95% confidence interval (CI) 1.05-45.76, P = 0.04), glycaemia ≤6.8 mmol/l (OR 4.05; 95% CI 1.09-15.0, P = 0.03) and thrombolysis (OR 3.67; 95% CI 1.04-12.9, P = 0.04) as independent predictors of good neurological outcome. Conclusions: In patients with DWI-ASPECTS ≤5, two-thirds of patients experienced good neurological outcome when recanalized by state-of-the-art thrombectomy, whilst only one in 14 non-recanalizers achieved similar outcomes. Pretreatment markers of good neurological outcomes were DWI-ASPECTS >2, intravenous thrombolysis and glycaemia ≤6.8 mmol/l.