# Making the invisible visible: integrated visualization and automated quantification of thrombus deformation during mechanical thrombectomy

## TL;DR

This study introduces a new method to visualize and measure how stent retrievers deform blood clots during stroke treatment, aiming to improve device selection and treatment outcomes.

## Contribution

A novel algorithm and experimental framework were developed to objectively quantify thrombus deformation during mechanical thrombectomy.

## Key findings

- Both Embotrap III and NeVa stent retrievers achieved high first-pass reperfusion rates.
- Embotrap III caused greater thrombus deformation and contour change compared to NeVa.
- NeVa showed no thrombus fragmentation but had more distal migration in some cases.

## Abstract

Mechanical thrombectomy using stent retrievers is a proven treatment for large vessel occlusion stroke, but quantitative and objective evaluation of device-thrombus interaction remains limited.

This study introduces a novel algorithm and experimental framework to visualize and quantify thrombus deformation during retrieval under standardized in vitro conditions, with the long-term goal of informing future clinically applicable analysis tools.

An in vitro model simulating large vessel occlusion was developed using organized human thrombi. Two stent retrievers - the Embotrap III (6.5 × 45 mm) and NeVa (4.5x 29 mm) - were studied in a vascular model under fluoroscopic guidance. An algorithm was developed to automatically segment and quantify thrombus deformation and analyze contour changes during the retrieval.

In total, 36 and 47 thrombectomies were performed with the Embotrap III and NeVa, respectively. Both devices achieved high first-pass reperfusion rates (Embotrap III: 88.9%, NeVa: 87.2%). However, thrombus deformation (mean: 14.7 × 10−3 vs. 8.9 × 10−3) and contour change (mean: 0.52 vs. 0.40) were significantly higher with Embotrap III, suggesting greater mechanical engagement. Thrombus fragmentation occurred in 5.6% of Embotrap III cases, but in none with NeVa. Moderate to marked distal thrombus migration occurred in over half of cases in both groups. Extensive migration occurred exclusively with NeVa in 4 cases (13%). Interrater and intrarater reliability of the automatic segmentation were excellent.

Our new imaging and analysis framework allows for objective, reproducible measurements of thrombus-stent interaction over time. Our findings support the high efficacy of both stent retrievers. While the Embotrap III exerted stronger mechanical deformation effects on the clots, initial distal thrombus migration occurred more often with the NeVa device. This approach may support more informed device selection and optimization of thrombectomy strategies to enhance reperfusion success and reduce distal embolization in acute ischemic stroke.