ESOC 2022 | Predicting poor outcomes despite thrombolysis and thrombectomy in ischemic stroke

A notable proportion of patients with ischemic stroke suffer very poor outcomes despite receiving intravenous thrombolysis and/or mechanical thrombectomy. As these treatments are associated with a high economic and logistical burden, it would be beneficial to be able to predict who will have poor outcomes despite maximal treatment. Thomas Meinel, MD, Bern University Hospital, and University of Bern, Bern, Switzerland, talks on efforts to develop a multivariable prognostic model able to identify futile acute revascularizations (FAR) in patients undergoing thrombolysis and thrombectomy. Machine learning was used to identify the most important predictors of FAR, defined as a modified Rankin scale score of 5-6 at 3 months. Clinical variables, laboratory values, imaging biomarkers, and time-to-treatment were all included in the model. The final model was discriminatory for predicting 3-month FAR, with an AUC of 0.87. However, when assessing the clinical value of the model by decision curve analysis, it had no added benefit at the thresholds required. This interview took place at European Stroke Organisation Conference 2022 in Lyon, France.