Introduction: Eloquent areas of brain have been identified more precisely using functional MRI but not often applied for stroke patients. Furthermore, it is widely accepted that stoke outcome depends on a number of factor including age, stroke lateralization, baseline NIHSS and brain imaging. This study aims to develop a new deconvolution algorithm for stroke outcome incorporating the concept of brain eloquent areas. We assessed the effectiveness of a new model, the Tours Stroke Imaging Project (TOSIP), to predict the outcome at three months compared to the conventional ASPECT.
Methods: Patients (n=130) with cerebral infarction on the carotid territory were included. A map of the brain eloquent area was created based on previous MRI mapping. 24 h CT-scan were analyzed by 2 senior neuroradiologists that noted ASPECT score and each eloquent area. Independently, demographic data and outcome (score on the modified Rankin scale at 3 month) were collected by 2 vascular neurologist. A Bayesian estimation algorithm were use to create a prognostic model. Spearman’s rank correlation coefficient was used to test the association between this new model, ASPECT score and the functional outcome.
Results: We observed a strong correlation between TOSIP and outcome at 3 month. (OR = 3.47 [2.96; 4.07], P<0.0001). For the same patients, delay ASPECTS value is not as well correlated (OR = 2.8 [2.23; 3.39], P<0.0001).
Conclusions: TOSIP seems to be a promising new tool to predict more precisely the clinical outcome at 3-month after ischemic stroke. A more important sample is necessary for further validation.