LAUSR

Despite advancement in mechanical thrombectomy (MT) techniques, 10–30% of MT for large vessel occlusions (LVO) are unsuccessful. Current prediction models fail to address the association between patient-specific factors and reperfusion. We aimed to evaluate objective, easily reproducible, admission clinical and radiological biomarkers that predict unsuccessful MT. We analyzed consecutive anterior LVO MT patients at two comprehensive stroke centers. The primary outcome was unsuccessful reperfusion defined by a modified thrombolysis in cerebral infarction (mTICI) score of 0–2a. We quantitatively assessed the hyperdense vessel sign by measuring Hounsfield units (HU) on admission computed tomography (CT). Receiver operating characteristic (ROC) curves were plotted to estimate the predictive value of quantitative hyperdense middle cerebral artery (MCA) measurements (delta and ratio) and of the final model for mTICI scores. We performed multivariable logistic regression to analyze associations with outcomes. Out of 348 patients 87 had unsuccessful MT. Smoking, difficult arch, vessel tortuosity, vessel calcification, diminutive vessels, truncal M1 occlusion, delta HU and HU ratio were significantly associated with unsuccessful MT in the univariate analysis. When we fitted two separate multivariate models including all significant variables and a HU measurement; delta HU <6 (odds ratio, OR = 2.07, 95% confidence intervals, CI 1.09–3.92) and HU ratio ≤1.1 (OR = 2.003, 95% CI 1.05–3.81) were independently associated with failed MT after adjustment for smoking, diminutive vessels, vessel tortuosity, and difficult arch. The area under the curve AUC<9 of the final model was 0.717. Novel radiological biomarkers on CT, CT angiography (CTA) and digital subtraction angiography (DSA) may help identify patients refractory to standard MT and prepare interventionalists for using additional alternative methods. Quantitative assessment of HU (delta and ratio) may be important in developing objective prediction tools for unsuccessful MT.