A Pipeline for 3D Multimodality Image Integration and Computer-assisted Planning in Epilepsy Surgery

The overall goal of this description of the software Pipeline is to allow the dissemination of the software itself in other centers, allowing other groups to benefit from 3D multi-modality image integration in epilepsy surgery. The main advantages that all steps in the pre-surgical evaluations of epilepsy including image registration, anatomy segmentation, and computer-assisted planning can be performed on one software platform. The 3D visualization of brain structures, coupled with multi-modality image integration, and the automatic multi-trajectory planner for electrode implantation, are relevant tools to increase the efficacy and safety of epilepsy surgery.

The 3D multi-modal structural and functional brain imaging display, including normal and abnormal structures and functions, allows for more accurate and quicker planning and implantation of depth electrodes. Begin by opening the in-house software on a P.C.and load the data. Note the two by two window display, data manager on the far left, icons on the top representing different image processing tools, and the selected tool on the far right.

Import data by clicking the open"icon. Scroll through different data sets to ensure completeness. Next, co-register single images by selecting the NiftyReg tool from the speed icons.

Select neuronavigation T1 with gadolinium in data manager to use as a reference image. Then, select the floating image to be co-registered to the reference image. Define the name and location of the registered image.

Set optimization parameters to level number four, level to perform as three, iteration number to five, and co-registration type to rigid body. Click on Run to co-register the automated rigid body. Then, check the accuracy of the co-registration by inspecting the registered image over the reference image.

Alter the transparency of the registered image by right-clicking on the image in data manager, and moving the opacity cursor. Begin by selecting the image to be segmented in data manager, and then select the segmentation editor tool from the speed icons. Use advanced segmentation tools to draw the region of interest on several slices of imaging in axial, coronal, and sagittal planes.

Finally, select 3D interpolation to visualize the evolving segmented structure in a 3D window. Confirm segmentation to generate a new Nifty file of the segmented structure. Right-click on the Nifty file in data manager, and select smooth polygon surface.

Begin by selecting the image file from the whole brain parcellation on data manager, and ensure that this image is co-registered with the reference image. Select the basic processing tools from the speed icons. Apply a threshold from one to 5002 to create a binarized mask of the cortex.

Right-click on the Nifty file in data manager, and select smooth polygon surface. Select the vessel extractor tool to extract the surface models of the vessels. Then, choose the vascular image data set.

Specify the name and location of the vessel extraction Nifty file. Click Run"to execute the vessel extractor. Apply an intracranial mask to the result by using the multiply function from the basic image processing tools to remove extracranial vessels.

Finally, to render the regions of interest as 3D surfaces, select the surface extractor icon, and define the threshold for surface extraction. Select apply, and name the surface rendering in data manager. Begin by running a multi-trajectory planner by selecting the trajectory planner icon.

Select the Neuronavigation T1 MRI as the reference image. Then, select target points, such as the amygdala, hippocampus, insula, or cingulate gyrus, by holding down the Shift key and clicking the left mouse button. Or load the target points from the previously saved file.

Next, select entry points, as well as the scalp exclusion mask on the attached drop-down menu in order to restrict the search of possible entry points to only a surgically feasible area. Mark the surfaces of critical structures that the trajectories should avoid from the drop-down list. Select advanced settings, and adjust the user-defined constraints regarding trajectory length, angle of entry, and distance between trajectories.

Then, run a multi-trajectory planner by selecting add new plan, and recompute plan. Next, use the risk visualization"speed icon to assess risk and safety profiles after trajectory planning. Note the metrics for length, angle of entry, cumulative risk, minimum distance to blood vessel, and grey/white matter ratios.

Select risk map"in data manager by clicking on a specific trajectory to show a color-coded contour map overlying the scalp exclusion mask. Note that the potential entry points are color-coded based on level of risk, such that red represents high risk, and green represents low risk for any selected trajectory. Following that, export the plans and models to the operating room.

To do this, first make sure that the reference image was loaded into icon format, and then open the S7 export tool. Define the reference image, plans, trajectories, and models that are to be exported, and specify the destination of the saved archive. Then run the S7 export tool.

Finally, upload the saved archive onto a USB stick for transfer to a neuronavigation system in the operating room. Load the archived folder on the neuronaivgation system for clinical implementation of the planned trajectories. This protocol allows for more streamlined image integration, as well as 3D visualization and planning prior to an epilepsy surgery.

Computer-assisted planning generates safer, more efficient implantations, that can be completed in a time-effective manner compared to manual planning. Here, a typical outcome is shown from the 3D multi-trajectory planner. The critical structures that have been entered are veins, arteries, and surface solci, which allow for more precise electrode implantations.

This video should give you a good understanding of the principles of multi-modal image integration, and the use of computer-assisted planning. Depending on the amount of imaging to be included, the preparation of imaging components is hard to be integrated into the 3D multi-modal display, it can take hours. If there are any errors in the source data, these will remain after the integration.

While attempting this protocol, it is important to remember the quality of the segmented anatomical structures is key to ensuring accurate computer-assisted planning results. This software provides easy to use tools in one single platform, such that it does not require specialist training or expertise, it's cost-effective, and it's easily translated into clinical practice. Additionally, this software could easily be applied to other areas of neurosurgery, such as resection of tumors close to eloquent cortex, focal lesioning, and delivery of target stimulation.

The in-house software platform is in continuous development with new tools and functionality being added to support all stages of pre-surgical management, and surgical evaluation. Don't forget that neurosurgery and the implantation of devices into the brain carries significant risk, and needs to be carefully checked every step of the way. Including the actual execution of the trajectories, which implantation of the electrodes.