Automatic Incision Point and Path Selection for Thoracoscopic Surgery
Minimally invasive thoracic surgery represents one of the most technically demanding fields in modern medicine. Each incision, each instrument trajectory, must be precisely planned to ensure safety, minimize trauma, and maintain ergonomic access for the surgeon. Despite advances in imaging and robotics, identifying optimal incision points and tool paths that avoid vital structures while meeting the Remote Center of Motion (RCM) constraint remains a major challenge.
Our recent research, “Automatic Incision Point and Path Selection for Thoracoscopic Surgery” — available as a preprint on SSRN — introduces a novel software system designed to support these critical preoperative decisions.
From Imaging to Surgical Design
The system starts from CT-based anatomical segmentation, extracting the patient’s organs and critical structures such as the heart, aorta, bronchus, diaphragm, and rib cage. Using this detailed voxelized representation, the software automatically evaluates feasible incision points on the skin and calculates valid tool trajectories that enter through a trocar, navigate within the thoracic cavity, and reach the target vessels safely.
To explore this vast and complex configuration space, the algorithm employs Monte Carlo sampling with bidirectional search strategies:
Target-to-incision: starting from the surgical target and working outward to find feasible entry points.
Incision-to-target: beginning from potential incision areas and testing which trajectories can safely reach the target.
Each candidate trajectory is validated for collision avoidance, joint angle limits, and RCM preservation — ensuring the tool maintains a fixed pivot point during movement, as required in minimally invasive procedures.
Augmented Reality Visualization
Once valid paths are identified, they are ranked by a multi-parameter quality function, learned from expert surgeon evaluations. The best candidates are visualized in 3D or via Augmented Reality (AR) headsets, allowing surgeons to inspect and select their preferred incision and trajectory interactively.
In experimental validation, the calculated tool paths were projected onto porcine thoracic models using AR visualization. This enabled a direct comparison between virtual planning and real anatomical execution — demonstrating that the computed trajectories were not only geometrically valid but also clinically feasible.
Bridging Medicine and Engineering
As both a cardiac surgeon and software engineer, I experienced first-hand how crucial interdisciplinary understanding is in projects like this. Translating clinical requirements into mathematical constraints and computational algorithms requires a dual perspective:
– the precision and intuition of surgery, and
– the rigor and abstraction of engineering.
This dual background helped streamline communication across disciplines and accelerate the development cycle, from concept to prototype.
Collaborative Development and Support
This work was the result of a close collaboration between:
Budapest University of Technology and Economics (BME)
Semmelweis University
National Institute of Oncology
and Gamma Digital Ltd.
We gratefully acknowledge Gamma Digital’s technological and financial support, which enabled the development of the GPU-accelerated simulation engine and AR visualization platform.
Looking Ahead
The current system serves as a pre-clinical surgical design and educational platform, enabling surgeons to plan and visualize minimally invasive interventions in real time.
Future directions include extending the method to other surgical domains — such as liver and cardiac procedures — where constrained workspaces and critical anatomy demand precise path planning.
This project exemplifies how collaboration between medical professionals and software engineers can transform surgical planning, bringing us one step closer to data-driven, personalized surgery.