Deep Learning�Driven Tumor Boundary Estimation Using Robotic Palpation in Minimally Invasive Surgery
Youngjun Ryu, Jeongbin Hong, Hyeonwoo Kee, Sukho Park
AI summary
Problem
Intraoperative tumor boundary detection relies on static preoperative imaging that fails to account for tissue deformation, while existing robotic palpation techniques require complex auxiliary sensors or struggle to measure displacement during sweeping.
Approach
The method reconstructs the tissue surface via deep learning to estimate displacement during sweeping palpation, computes a stiffness map from a single force/torque sensor, and applies a sparse autoencoder with clustering to segment tumors.
Key results
- Novel two-step palpation workflow using 3D surface reconstruction for accurate stiffness estimation
- High-resolution stiffness mapping achieved with only a single F/T sensor, removing auxiliary hardware needs
- Sparse autoencoder latent features significantly improve clustering robustness and boundary accuracy over raw data
- Consistently low Hausdorff distances under clinically relevant surgical margins across phantom and ex vivo models
Why it matters
Enables precise, real-time tumor localization during robot-assisted surgery without extra imaging devices, supporting safer margin control and organ preservation.
Abstract
Accurate estimation of tumor boundaries is critical for ensuring adequate surgical margins in robot-assisted minimally invasive surgery (RMIS). In this study, we present a method that estimates tumor boundaries in RMIS using sweeping palpation data acquired with a single force/torque (F/T) sensor. From the reconstructed surface, tissue displacement and normal force were derived to calculate stiffness, which was then used to construct a stiffness map. To reduce noise and enhance feature representations, we employed a sparse autoencoder (SAE). The SAE outputs were subsequently clustered with a Gaussian mixture model (GMM) and K-means to segment the tumor from normal tissue. Experiments with phantom models and an ex vivo model demonstrated that the SAE-based approach significantly improved the Dice similarity coefficient (DSC) and sensitivity while maintaining specificity, and reduced the Hausdorff distances (HD and HD95) and average symmetric surface distance (ASSD), compared with results from raw data. Importantly, when evaluated under clinically relevant surgical margin conditions, the estimated HD consistently remained below threshold across all models. These results indicate that the proposed method achieves both high accuracy and clinical feasibility without additional imaging devices or displacement sensors, highlighting its potential to support margin minimization and organ function preservation in RMIS.