UPmémoires

English

Adaptation of MRI-DRAGON score for the early prediction of stroke outcome after mechanical thrombectomy

Background and Purpose: The MRI-DRAGON score has shown good performance to predict 3-month poor outcome in anterior circulation acute ischemic stroke (AIS) patients treated by intravenous thrombolysis (IV t-PA), with MRI as first-line imaging. It contains simple clinical, biological and imaging parameters shortly available at patient admission. The aim of our study is to update this score at the era of endovascular therapy (EVT).

Methods: All consecutive proximal anterior circulation AIS patients treated by EVT, with or without IV t-PA, between January 2015 and December 2017 in the Lille University Hospital were reviewed (derivation cohort). Baseline MRI-DRAGON score parameters were collected and the score was adapted after multivariable logistic regression modelling assuming proportional odds. We considered long-term poor outcome as 3-month modified Rankin Scale >2, and calculated area under the curve (AUC) of the new score: the MT-DRAGON score, to assess the ability of poor outcome prediction. Internal validation was performed with boostrapping, and this new score was externally validated in a population (validation cohort) issued from two French University Hospitals (Centre Hospitalier Sainte-Anne and Centre Hospitalier de Reims).

Results: We included 497 patients in the derivation cohort. MRI-DRAGON score was adapted, by keeping or adjusting all clinical and biological variables (age, National Institutes of Health Stroke Scale at admission (NIHSS), serum glucose level, pre-stroke mRS, DWI-ASPECTS, occlusion localization, onset-to-treat time), and adding the administration or not of IV t-PA. All parameters of the MT-DRAGON score were statistically associated with long-term poor outcome in multivariate analysis. The c-statistic of the new score for poor functional outcome was 0,84 (95% IC [0,80 - 0,87]). The Hosmer-Lemeshow test showed no significant difference between the observed and predicted risks of poor outcome (x2= 9.2, p =0.35). In the validation cohort (n=219), the c-statistic of the MT-DRAGON score was 0.77 (95% CI, [0.70, 0.83]), slightly inferior than the c-statistic of the development cohort (P = 0.21).

Conclusion: The MT-DRAGON score, adapted from the MRI-DRAGON score can reliably predict long-term poor outcome in proximal anterior circulation AIS patients treated by EVT with or without IV t-PA, with simple parameters, shortly available at patient admission.