As Clinical-diffusion mismatch (CDM) well reflects true penumbra, we hypothesized that patients with CDM may better respond to endovascular thrombectomy (EVT) than those without. 

Clinical-diffusion mismatch (CDM) and diffusion-perfusion mismatch (DPM) is used to select patients eligible for endovascular thrombectomy (EVT) in the late-time window. 

Acute ischemic stroke patients who received EVT between 6 to 24 hours from stroke onset were enrolled. All patients showed DPM (ratio of penumbra/ischemic core>1.8). CDM was defined according to the DAWN criteria; 1) age≥80, National Institute of Health Stroke Scale (NIHSS) score≥10 and DWI lesion≤21 ml, 2) age<80, NIHSS score≥10 and DWI lesion≤31 ml and 3) age<80, NIHSS ≥ 20 and 31<DWI lesion≤51 ml. RAPID software was used to measure lesion volume. Characteristics and response to EVT was compared between those with and without CDM. Early neurological improvement (ENI) was defined as improvement of more than 4 points of NIHSS score from baseline 24 hours after EVT. Multivariable analysis was performed to find independent factors associated with ENI.

Among 94 patients enrolled, 44 (46.3%) patients showed CDM. ENI was observed from 48 (51.1%) of patients. Patients with CDM showed higher prevalence of hypertension (p=0.047) initial NIHSS score (14±4 vs. 8±4; p<0.001), more improvement of NIHSS after EVT (6±6 vs. 1±5; p<0.001) and a higher prevalence of ENI. From the multivariable analysis, ENI was associated with onset-to-door time (OR=0.998, 95% CI 0.997–1.000; p=0.042), complete recanalization (OR=23.912, 95 % CI 2.238–255.489; p=0.009), NIHSS score (OR=1.180, 95% CI 1.012–1.377; p=0.034) and CDM (OR=5.160, 95% CI 1.448–18.386; p=0.011). Interestingly, the correlation between DWI lesion volume and NIHSS score was strong in those without CDM (r=0.731), but only moderate in those with CDM (r=0.355).

Patients with both CDM and DPM showed a better response to EVT in late-window than those with DPM only.