ABSTRACT NUMBER: ESOC2026YS86 RELATIVE NCCT TO PREDICT EVT RESPONSE IN LOW ASPECTS PATIENTS

Abstract Background and aims Endovascular thrombectomy (EVT) improves functional outcome in patients with large ischemic lesions. However, EVT benefit was recently demonstrated to be modified by the degree of infarct hypoattenuation, as measured by net water uptake (NWU). Relative non-contrast computed tomography (rNCCT) offers an alternative to NWU for the assessment of tissue status without the need for manual segmentation or perfusion imaging. Building on recent validation and established diagnostic thresholds for rNCCT, we aim to investigate whether rNCCT hypodensity severity modifies the treatment effect of EVT and can identify a cutoff for EVT futility. Methods We will perform a secondary analysis of the Efficacy and Safety of Thrombectomy in Stroke with Extended Lesion and Extended Time Window (TENSION) trial, a multicenter, randomized trial of 253 patients with anterior circulation stroke and an Alberta Stroke Program CT score (ASPECTS) of 3-5 treated with EVT versus best medical treatment. Patients lacking admission NCCT, significant imaging artefacts or prior lesions in the analysis region will be excluded. rNCCT-derived core lesion maps using our previously established thresholds will be automatically generated, in which we will calculate the median (IQR) hypodensity. Modification of EVT by hypodensity severity will then be investigated using interaction analysis and further investigated using receiver operating characteristics (ROC). Results Imaging and clinical data have been obtained and are currently being processed for analysis. Conflict of interest “Laurens Van Herck: nothing to disclose. ”, “Lieselotte Vandewalle: received financial support from the Research Foundation Flanders (Fonds Wetenschappelijk Onderzoek, FWO), PhD fellowship fundamental research 1113925N ”, “Martin Bendzus: received funding for TENSION trial by EU ”, “Bastian Cheng: nothing to disclose ”, “Soren Christensen: nothing to disclose ”, “Maarten Lansberg: received funding from NIH ”, “Robin Lemmens: nothing to disclose ”, “Anke Wouters: consultancy for Bayer ”, “Jelle Demeestere: nothing to disclose ”. Figure 1 - belongs to Background and aims