Robotic Neurosurgery | Prateek Bhargava

Path Planning for Semi-Automated

Simulated Robotic Neurosurgery

Introduction

The advancements and the continuous improvement on the technological front has allowed the scientific community all around the world to try and incorporate technology into the field of medicine to improve patient care. Likewise, since the advent of robots in the past two decades, robotic surgery has taken leaps and bounds due to the continuous development being made by researchers throughout the globe. It has been proved that the use of robotic tools in a brain biopsy can increase the chances of success and the survival percentage of the patient exponentially. Neurosurgery requiring high accuracy and precision is considered the most demanding area for computer-based surgery. Thus the integration of robots widens the scope by further making microsurgery a reality, given their ability to perform surgeries in places out of reach of human hands.

The Coverage Path Planning for Semi-Automated Simulated Robotic Neurosurgery is an example of the incorporation of technical aspects like path planning of a robot to perform neurosurgery to remove cancerous margins (brain tumor ablation surgery). This task is performed under the supervision of a surgeon.

Here intermediate steps taken towards planned intelligent robotic surgical systems using path coverage algorithms that have the ability to scan for fluorescent(tumor paint) exposed by bulk tumor removal, and automatically treat the material is planned, built, and discussed upon. It is ensured in this algorithm that the maximum area is covered by the robot removing all the tumor marginal cells and that there is minimal to zero damage done to the nearby healthy brain cells.

Inputs have been taken from computer vision systems consisting of camera sensors and various other detection devices to graphically compute a 2-D generated map on a system. The 2-D generated map is further computationally worked upon and a 3-D generated brain picture is simulated using image reconstruction processes. We would be working on a single 2-D generated layer and perform our path coverage algorithm technique on the same using some of the guidelines given below:

• Completely removing the marked tissues.

• Ensuring healthy cells and tissues near the brain tumor cavity are not affected

• Performing iterations and path coverage algorithm on different brain tumor shapes to ensure the productivity of the algorithm.

Methodology

The workspace for the surgical robot is the surgical cavity in the brain after most of the brain tumor is surgically ablated by the surgeon. The figure below is the simulated surgical space for the testing of the path coverage algorithm.

The black contours/regions in the cavity represent the regions from where the path coverage algorithm has to be implemented. The contours (boundary regions) have been considered as the workspace for the algorithms and the rest of the cavity has been ignored. Multiple contours have been considered to test the effectiveness of our algorithms. All the contours are defined using python and are visualized using matplotlib.

Implementation

A. Grid Generation

Since the shape of the contours are irregular, a set of random points that generate an irregular contour shape and is to be used as the workspace for the algorithm. The map is defined in such a way that the total number of Contour space nodes for exploration inside the contour are obtained as the product of width and height of the contour. The robot doesn’t need to explore all the nodes to be effective, but a threshold/resolution is set for that to adjust the fine and coarseness of the path coverage.

B. Sweep Motion and Moving Directions

The algorithm defines two types of motions. One is the sweep motion and the other is the lateral movement of the robot to cover the area in that region. Since it is an irregular contour, the start points and sweep directions have to be determined. This is achieved by mapping the slope distances between points of the input of the contour and selecting the maximum slope distance. Once the sweep direction has been determined it remains the same throughout the entire algorithm. The sweep direction is also obtained from this custom function. This is done to ensure that the robot moves only once over the particular region to ensure safe removal of cancer boundaries without touching healthy brain tissue.

Path Coverage Algorithm

The zigzag path planning algorithm which is computationally the foremost effective algorithm over other algorithms is implemented.

In the algorithm the nodes present inside the defined contour shape and compute n and n-1 number of points, thus creating points on each parallel line. This thus determines if the node lies either inside or outside the contour shape. The algorithm also incorporates the robot radius ensuring that there is a specific threshold maintained and the suction of the brain tumor cells takes places only inside the contour shape generated. This algorithm then ensures that no healthy tissues are damaged in the robotic surgery procedure.

The nodes are then coherently divided into two parts which are as follows.:

i) ”in” nodes.

ii) ”out” nodes.

The nodes are sorted in the following way and are depicted as follows:

Step 1: The in node of the starting line is discovered.

Step 2: If the starting node or point is not found the program is terminated.

Step 3: A Zig-zag line or pattern is created based on the in and out nodes detected.

Step 4: The While loop continuously checks for the out nodes or edges and continues to move in the same direction until the edge is reached.

Step 5: The closest group point of the following row is found and then step 3 is repeated.

The tumor margins are described by the random selection of points forming defined contours. The path coverage algorithm will traverse the whole area inside the contour shape thus marking out the areas of the residual tumor cells to be removed by the surgical robot in a zigzag manner.

The algorithm will produce a zigzag line which will be represented by arrows or a line. The zigzag line would comprise of evenly spaced segments based depending on the suction or ablation area of the robot end-effector so as to cover the whole area present inside the tumor cavity.

Effective implementation of the path coverage can be seen below that worked relatively well with multiple irregular contours.