Last updated: 2023.5.11

Design and develop image-free tracking to guide endovascular neurosurgical procedures.

• Students: Huilin (Lin) Xu, Fangjie (FJ) Li, Shanelle Cao

Project Mentor(s):

Dr. Ali Uneri

Dr. Fernando Gonzalez

Brain aneurysms are bulges in blood vessels in the brain primarily due to thinning artery walls. They often occur where blood vessels branch, as vessel walls are usually thinner at these locations. Brain aneurysms can either leak or rupture, which will then quickly develop into life-threatening conditions. Leaking or ruptured brain aneurysms may cause severe headaches, stroke, permanent neurological deficits, or death. In the US, around 6.7 million people live with brain aneurysms. Out of the 6.7 million, around 30,000 patients experience ruptured aneurysms annually, with a fatality rate of up to 50%.

With such a high fatality rate, treatments are mainly concerned with stopping further blood flow into the aneurysm to prevent or stop an ongoing rupture. Two different procedures are available for embolizing aneurysms. Microsurgical clipping, as shown below, is where a small opening is made on the skull near the site of the aneurysm, and a clip is inserted to pinch off further blood flow into the aneurysm. This method is highly invasive, taking patients long periods of time to heal after the operation.

More recently, the endovascular neurosurgical method has been developed, where a catheter is inserted from the femoral artery near the thigh of the patient and navigated to the site of the aneurysm in the brain. Once the catheter reaches the aneurysm, there are two different methods to embolize the aneurysm. One such method is coiling, where a coil of wire is inserted into the aneurysm through the catheter, causing blood clots. The other method is flow diversion, where a mesh tube, acting like a stent, diverts the blood flow away from the aneurysm and along the vessel. With no new blood flow into the aneurysm, existing blood inside the aneurysm will clot. Blood clotting will block further blood flow into the aneurysm, preventing or stopping an ongoing rupture. Compared to microsurgical clipping, the endovascular neurosurgical method is minimally invasive.

Although the endovascular procedure is minimally invasive, it is hard for surgeons to navigate the catheter to the site of the aneurysm smoothly. Pre-operative CT angiography and fluoroscopy is required to help visualize blood vessels and tissues and locate the catheter in order to navigate the catheter to the site of the aneurysm.

As mentioned before, for endovascular techniques, fluoroscopy and CT angiograms are used to help surgeons visualize the position of their catheter in order to navigate the catheter to the site of the aneurysm. However, this process exposes both the patient and surgeon to hundreds of mGy of X-ray radiation. At this level of radiation exposure, there are increased health risks of diseases such as cancer, cataract, non-malignant skin damage, and impaired fertility. Especially, surgeons are at increased risk since they perform this procedure on a regular basis. Thus, there is currently a need for methods of detecting catheter position inside the patient without relying on X-ray imaging techniques, which will greatly reduce the amount of radiation that patients and surgeons are exposed to.

To address this issue, we aim to develop a system where the catheter can be tracked using EM. This way, we can eliminate the need of fluoroscopy during endovascular procedures. Specifically, we hope to achieve:

1. Manufacture a catheter that can be tracked using EM
2. Develop a software where the tracked catheter can be visualized with patient image to provide image guidance

Aurora Field Generator generates an electromagnetic field that causes current to run through the sensor attached to the catheter. By measuring the voltage on the sensor, the catheter's position and orientation can be determined.

Directly stick the EM 5-DOF sensor inside the catheter. (left: 5-DOF sensor; right: catheter prototype)

• Pros: Easy to implement. Provides enough functionality for data collection and algorithm verification.
• Cons: The sensor is not highly securely attached to the catheter. Leaves no space for other tools inside the catheter.
  

Put the EM 5-DOF sensor outside the catheter and use heat shrink tubing to secure the whole system.

• Pros: More suitable for clinical use, more robust compared to iteration 1
• Cons: Added extra thickness to the catheter
  

* Fiducial points registration:

1. Manually mark several fiducial points on CT scan of the phantom
2. Use catheter tip to perform points cloud registration

* Path-based registration:

1. Extract vessel centerlines to represent the essential shape
2. Perform gradient descent to find the transformation between the tracker path and the centerline

The 5 DoF sensor, which is aligned with the tip, has sufficient information to display the tip orientation with a high degree of accuracy. However, the challenge is to estimate the pose of the body, which we do not have direct sensor data of. Therefore, we relied on an assumption for body pose estimation: the catheter is sufficiently rigid for a short span of length. As a result, the catheter pose will resemble the trajectory composed of recent sensor readings. Therefore, we fitted a linear vector to the sensor readings closest to the current sensor position, (1 cm within current location) using PCA based computations. Then, this orientation is attached to the current sensor tip transform, with a small translation offset that is predefined by the user to best reflect the actual geometry of the catheter tip shape.

1. Prototype Design
   • Planned Date: Feb 28th, 2023
   • Expected Date: Feb 28th, 2023
   • Status: Deicide to build two iterations. One for data collection inside the phantom and another for medical surgery use. Details information can be found in Technique Approach section.
2. Prototype Manufacture
   • Planned Date: March 15th, 2023
   • Expected Date: March 15th, 2023
   • Status:
     • Version #1 Prototype: Simple catheter with sensor inserted for data collection purposes, unable to meet clinical standards
       • Status: Done
     • Version #2 Prototype: Sensor attached to the outside of the catheter with heat shrink tubing protection, leaving the working lumen open for clinical use.
       • Status: Done
3. Path Visualization Algorithm
   • Planned Date: Feb 16th, 2023
   • Expected Date: Feb 16th, 2023
   • Status: Done
4. Visualization Algorithm Improvement
   • Planned Date: March 15th, 2023
   • Expected Date: March 4th, 2023
   • Status: Done
5. Fiducial Point Registration Algorithm to Phantom
   • Planned Date: March 22nd, 2023
   • Expected Date: March 15th, 2023
   • Status: Done.
6. Path-based Registration Algorithm to Aortic Arch CT
   • Planned Date: April 21st, 2023
   • Expected Date: April 21st, 2023
   • Status: Done

• Project Plan
  • Plan Presentation
  • Proposal
• Project Background Reading
  • Background Reading
• Project Checkpoint
  • Checkpoint
• Project Final Poster
  • Final Poster
• Project Final Report
  • Final Report

Note: Documentations are in Other Resources and Project Files section.

* here list references and reading material

1. Simultaneous Tracking of Catheters and Guidewires: Comparison to Standard Fluoroscopic Guidance for Arterial Cannulation
2. Learning-based endovascular navigation through the use of non-rigid registration for collaborative robotic catheterization
3. Electromagnetic Tracking for Registration and Navigation in Endovascular Aneurysm Repair: A Phantom Study

• Catheter Prototype Design Documentation: catheter_prototype_design_specification.pdf
• Software Design Documentation: software_design_specification.pdf
• Product Requirements: product_requirements_.pdf
• Link to Github: https://github.com/LinXu12/EM_Catheter_Tracking