Device-anatomy interactions in the dural venous sinuses: angioscopic insights from a perfused human cadaveric model

Neuroendovascular venous interventions are increasingly performed using technologies originally developed for arterial procedures and indications. However, the major dural venous sinuses possess a unique intraluminal anatomy that is not present in arteries, raising concerns about device-anatomy interactions that may affect procedural performance. We used a perfused human cadaveric model with direct intraluminal angioscopic visualization to evaluate currently available endovascular devices within the dural venous sinuses and to characterize mechanisms of device-anatomy interactions associated with technical difficulty and failure.

Six fresh human head-and-neck cadaveric specimens were perfused with 0.9% saline solution via bilateral internal jugular vein catheterization using a peristaltic pump. Direct intraluminal angioscopic visualization was achieved through transcranial access to the major dural venous sinuses, allowing real-time observation of target segments during device manipulation. Standard endovascular maneuvers were performed within the dural venous sinuses, including guidewire and microcatheter navigation, catheter advancement, venous stent deployment, stent retriever deployment, aspiration thrombectomy, and balloon angioplasty. Angioscopic and fluoroscopic recordings were independently reviewed by experienced neurointerventionists to identify and categorize technical challenges and failure mechanisms.

Angioscopy revealed multiple device-intraluminal interactions that were not fully appreciated on fluoroscopy alone. Several representative technical challenge and failure scenarios were identified and grouped into four principal mechanisms: (1) catheterization of venous channels parallel to the main sinus lumen, resulting in catheter entrapment and incomplete expansion of venous stents and stent retrievers; (2) device deformation or incomplete expansion due to intraluminal bands, including stent deformation, malposition, and constrained balloon angioplasty; (3) arrested or impaired device advancement caused by intraluminal bands, frequently necessitating microcatheter-assisted support to overcome ledge effects; and (4) interaction with arachnoid granulations leading to occlusion of aspiration catheter inlets and impeded intraluminal navigation.

The venous system differs fundamentally from arteries in luminal geometry and internal architecture. Our findings demonstrate that arterial-derived devices incompletely accommodate these differences, resulting in parallel channel navigation, constrained expansion and deformation of stents, and occlusion of suction catheters. These findings highlight the fact that veins are not arteries and underscore the need for venous-specific techniques and technologies.