Advances in thrombus imaging and retrieval in acute ischemic stroke

This thesis focused on advances in thrombus imaging and retrieval. We found that AI-based automated LVO detection is accurate and fast and therefore can be valuable in clinical practice. Furthermore, we showed that the use of thrombus imaging characteristics (thrombus length and distal location) aids in identifying patients who are most likely to benefit from IVT. However, thrombus imaging characteristics were of limited use for outcome prediction in patients who underwent EVT or for selection of first-line EVT device. While thrombus radiomics features describing volume and texture were associated with an increasing number of passes to achieve reperfusion, thrombus radiomics features had no added value for predicting procedural and functional outcome when compared to clinical variables and conventional thrombus measurements. Additionally, we investigated two topics that could possibly contribute to advancement in thrombus retrieval. First, we investigated the effect of number of passes in relation to final reperfusion score and functional outcome in patients with anterior circulation stroke. We found similar outcomes in patients with single pass eTICI 2B and multiple pass eTICI 2C/3, indicating that the advantage of eTICI 2C/3 over eTICI 2B diminishes with an increasing number of passes. Second, we investigated the effect of first-line EVT approach on functional outcome in patients with posterior circulation stroke and found better patient outcomes with first-line aspiration EVT. Our findings suggested that first-line aspiration EVT is a better first-line EVT approach in posterior circulation stroke patients.