Frame-based or frameless approaches to deep brain targets require planning of insertion trajectories that mitigate hemorrhagic risk and loss of function. Currently, this is done empirically by manual inspection of imaging data.

     We propose an intuitive software framework for computer-assisted trajectory planning, applied initially to insertion of deep brain stimulator (DBS) electrodes.

      

     Our framework accepts the following user-inputs: a target point (e.g. subthalamic nucleus), a set of surgical constraints (e.g. avoidance of blood vessels, cerebral sulci, ventricles), and multi-modal patient datasets (T1-weighted, T2-weighted, susceptibility-weighted and time-of-flight MR images). These datasets are automatically processed to extract a list of entry points after definition of a broad search-space avoiding critical brain areas (e.g. primary sensorimotor and speech areas). An automatic algorithm aggregates the many requirements into a meaningful ranking of optimized trajectories.

     Thousands of trajectories are processed in <4 minutes. The problem is reduced to well-defined patches of best-ranked trajectories, presented to the surgeon for evaluation using an intuitive color-scale overlaid onto a 3D reconstructed cortex. Analysis of 8 DBS cases reveals that automatic planning can provide alternative trajectories further away from vessels or sulci compared to manual planning alone. This novel method may improve insertion safety, and reduce surgical time.

     This is a retrospective study.

     Using high computational power, our framework provides rapid, objective optimization of many customizable constraints across dense, multi-modal, datasets.

     The tool allows efficient optimization of patient-specific DBS lead trajectories, and can be generalized to trajectory planning for biopsies, endoscopy, insertion of depth electrodes, and approach corridors to deep-seated tumors.