Cerebral arteriovenous malformation ai intelligent embolization auxiliary system

Abstract

This invention discloses an AI-powered intelligent embolization assistance system for arteriovenous malformations of the brain and spinal cord. By integrating an AI-powered intelligent embolization assistance subsystem and an image-guided platform, it matches 3D-DSA angiography images with internally stored hydrodynamic models of different embolization materials to obtain a three-dimensional vascular architecture map of the lesion and surrounding vessels. Through AI calculation, it completes virtual path planning, simulating the embolization path, three-dimensional course, and embolization range. Based on this, it provides real-time navigation of the microcatheter during EVT, intelligently following its path. After the procedure, it re-acquires 3D-DSA angiography images, performs multimodal image fusion, and assesses the embolization status, the three-dimensional relationship with surrounding structures, and potential related complications. This system enables preoperative planning and simulation of the interventional embolization path and range, and real-time navigation, tracking, and assessment of the lesion during the procedure, effectively improving the efficacy of EVT.

Description

Technical Field

[0001] This invention belongs to the field of medical imaging technology, specifically relating to an AI-powered intelligent embolization assist system for arteriovenous malformations of the brain and spinal cord. Background Technology

[0002] Arteriovenous malformations (AVMs) of the cerebral and spinal cord are relatively common vascular diseases of the central nervous system. Cerebral and spinal AVMs are known to be non-self-limiting and progressive, with an annual hemorrhage risk of approximately 1%-3%. The mortality rate from the first hemorrhage can reach 10%-50%. Among survivors, about 40% may experience severe neurological dysfunction, such as hemiplegia, aphasia, and cognitive impairment. Approximately 20%-30% of patients will experience seizures, and about 10%-20% will develop neurological dysfunction. Therefore, the treatment principle should be "early detection and early treatment."

[0003] Treatment options include surgical resection, endovascular treatment (EVT), stereotactic radiotherapy, and combinations thereof. Surgical resection is suitable for superficial, small AVMs, but it is highly invasive and may cause complications such as intraoperative bleeding, neurological damage (e.g., motor or speech impairment), and infection. The recurrence rate is approximately 5%. Endovascular treatment (EVT) is minimally invasive, but the complete occlusion rate is low (approximately 50%). Incomplete embolization or vascular damage may lead to bleeding, ischemic stroke, or permanent neurological deficits. Stereotactic radiotherapy has a slow onset of action (requiring 6-24 months), with a continued risk of bleeding during this period, and may cause complications such as radiation-induced brain injury, headache, seizures, or cyst formation.

[0004] With the rapid development of materials and technology, minimally invasive endovascular transluminal thoracotomy (EVT) has gradually become the preferred treatment for cerebrospinal vascular diseases. Therefore, a combination of EVT and stereotactic radiotherapy is generally used nowadays. However, the complete occlusion rate of EVT is relatively low (about 50%), especially for spinal AVMs. This is because spinal cord vessels are delicate and tortuous, making it difficult to insert microcatheters. Furthermore, the spinal cord lacks the "quiet zone" (non-functional area) of the brain, resulting in generally inferior efficacy of EVT for SCVMs compared to brain lesions, and a higher complication rate. Summary of the Invention

[0005] The purpose of this invention is to address the challenges of EVT treatment for brain and spinal cord AVMs. If an AI-powered intelligent embolization-assisted system can be developed for clinical applications, utilizing existing image segmentation technology platforms, it can achieve: preoperative planning and simulation of the interventional embolization path and embolization range; and real-time navigation, tracing, and lesion assessment during the procedure, effectively improving the efficacy of EVT.

[0006] To achieve the above-mentioned objectives, the present invention provides an AI-assisted intelligent embolization system for arteriovenous malformations of the brain and spinal cord, comprising:

[0007] The AI-powered intelligent embolization assistance subsystem matches 3D-DSA angiography images acquired by DSA with internally stored hydrodynamic models of different embolization materials to obtain a three-dimensional vascular architecture map of the lesion and surrounding blood vessels. The image-guided platform uses AI calculations to plan a virtual path based on a three-dimensional vascular architecture map of the lesion and surrounding blood vessels, simulating the embolization path, three-dimensional course, and embolization range. Based on this, it can navigate the microcatheter in real time during EVT, intelligently following its course. After the procedure, 3D-DSA angiography images are acquired again, and after multimodal image fusion, the embolization status, the three-dimensional relationship with surrounding structures, and possible related complications are assessed.

[0008] Preferably, the AI ​​intelligent embolization assistance subsystem, based on the Monte Carlo principle, constructs different vascular structures for different locations and different vascular diseases, and builds fluid dynamic models for different embolization materials.

[0009] Preferably, the AI ​​intelligent embolization auxiliary subsystem is implemented based on the AW4.7 intelligent post-processing workstation platform.

[0010] Preferably, the AI ​​intelligent embolization assistance subsystem is built upon the underlying R&D framework of GE's global database and AI vision technology.

[0011] Preferably, the different embolization materials include colloids and microspheres.

[0012] Preferably, the image-guided platform is built based on AI autonomous deep learning, Two-Click automatic vascular tracing technology, and virtual embolization technology.

[0013] This invention provides an AI-powered intelligent embolization assistance system for arteriovenous malformations of the brain and spinal cord. By integrating an AI-powered intelligent embolization assistance subsystem and an image-guided platform, it matches 3D-DSA angiography images with internally stored hydrodynamic models of different embolization materials to obtain a three-dimensional vascular architecture map of the lesion and surrounding vessels. Through AI calculation, it completes virtual path planning, simulating the embolization path, three-dimensional course, and embolization range. Based on this, it provides real-time navigation of the microcatheter during EVT, intelligently following its path. Postoperatively, it acquires 3D-DSA angiography images again, performs multimodal image fusion, and assesses the embolization status, the three-dimensional relationship with surrounding structures, and potential related complications. This system enables preoperative planning and simulation of the interventional embolization path and range, and real-time navigation, tracking, and assessment of the lesion during the procedure, effectively improving the efficacy of EVT. Attached Figure Description

[0014] Figure 1 This is a schematic diagram illustrating the application of the AI-powered intelligent embolization assistance system (Embo ASSIST) in interventional treatment of arteriovenous malformations. Detailed Implementation

[0015] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0016] This invention provides an AI-powered intelligent embolization assistance system for arteriovenous malformations of the brain and spinal cord, comprising: ① AI Intelligent Embolization Assist System (Embo ASSIST): Based on the AW4.7 intelligent post-processing workstation platform. Relying on the database of 6000+ vascular disease EVT cases in GE's global database and the underlying R&D framework of AI vision technology (such as ConDSeg and MIScnn medical image segmentation framework), under the Monte Carlo principle, different vascular structures were constructed for different locations and different vascular diseases, and fluid dynamic models of different embolization materials (colloids, microspheres, etc.) were also constructed.

[0017] ② Image-guided platforms (such as Embo ASSIST—GE Healthcare): Developing clinical application plugins based on this research using frameworks such as MIScnn (an open-source Python library providing an intuitive API for medical image segmentation). This includes AI-driven deep learning, Two-Click automated vascular tracing technology, virtual injection technology, and continuous software updates and dataset accumulation. Before interventional procedures, AI-suggested clinical decisions are made, enabling preoperative planning and simulation of the embolization path and extent, as well as real-time intelligent navigation, tracing, and assessment during the procedure.

[0018] ③ Procedure: First, 3D-DSA angiography images are acquired via DSA. Then, Embo ASSIST software is used to obtain a three-dimensional vascular architecture map of the lesion and surrounding vessels. In this map, AI calculations are used to complete virtual path planning, simulating the embolization path (related vessels), three-dimensional course, and embolization range. During EVT, the microcatheter is guided in real-time and intelligently followed. After interventional embolization, a second 3D-DSA scan can be used to assess the embolization status, the three-dimensional relationship with surrounding structures, and potential complications (such as intracranial hemorrhage) on multimodal image fusion.

[0019] The technical effects of the embodiments of the present invention are as follows: Since 2023, the above-mentioned technology has been applied in multiple centers. Its technical effects include: preoperative planning and simulation of interventional embolization path and embolization range, and real-time intelligent navigation, tracking and evaluation during the operation, which can effectively improve the effectiveness and safety of endovascular interventional treatment (EVT) of SCVM.

[0020] An analysis was conducted on 12 cases of arteriovenous malformation (AVM) of the brain and spinal cord treated at three centers (Huashan Hospital affiliated with Fudan University, Putuo District People's Hospital affiliated with Tongji University, and Huadong Hospital affiliated with Fudan University) over the past three years (2023 to present). The complete occlusion rate of the EVT increased by approximately 12.78% compared to previous data (reaching approximately 74.56%), with no significant complications, such as... Figure 1 As shown.

Claims

1. An AI-powered intelligent embolization-assisted system for arteriovenous malformations of the brain and spinal cord, characterized in that, include: The AI-powered intelligent embolization assistance subsystem matches 3D-DSA angiography images acquired by DSA with internally stored hydrodynamic models of different embolization materials to obtain a three-dimensional vascular architecture map of the lesion and surrounding blood vessels. The image-guided platform uses AI calculations to plan a virtual path based on a three-dimensional vascular architecture map of the lesion and surrounding blood vessels, simulating the embolization path, three-dimensional course, and embolization range. Based on this, it can navigate the microcatheter in real time during EVT, intelligently following its course. After the procedure, 3D-DSA angiography images are acquired again, and after multimodal image fusion, the embolization status, the three-dimensional relationship with surrounding structures, and possible related complications are assessed.

2. The AI-powered intelligent embolization assistive system for arteriovenous malformations of the brain and spinal cord as described in claim 1, characterized in that, Based on the Monte Carlo principle, the AI-powered intelligent embolization assistance subsystem constructs different vascular structures for different locations and different vascular diseases, and also builds fluid dynamic models for different embolization materials.

3. The AI-powered intelligent embolization assistive system for arteriovenous malformations of the brain and spinal cord as described in claim 1, characterized in that, The AI-powered intelligent embolization assistance subsystem is based on the AW4.7 intelligent post-processing workstation platform.

4. The AI-powered intelligent embolization assistive system for arteriovenous malformations of the brain and spinal cord as described in claim 1, characterized in that, The AI-powered intelligent embolism assistance subsystem is built upon the underlying R&D framework of GE's global database and AI vision technology.

5. The AI-powered intelligent embolization assistive system for arteriovenous malformations of the brain and spinal cord as described in claim 1, characterized in that, The different embolization materials include colloids and microspheres.

6. The AI-powered intelligent embolization assistive system for arteriovenous malformations of the brain and spinal cord as described in claim 1, characterized in that, The image-guided platform is built upon AI autonomous deep learning, Two-Click automatic vascular tracing technology, and virtual embolization technology.

Images