Incision trajectory tracing for electrosurgical navigation by CNN-based knife contacting frames extraction method

Image-guided surgical navigation has been actively studied because of its advantage in identifying subsurface targets and critical structures, while it requires incision trajectories to update the preoperative three-dimensional model dynamically during the surgery. The novelty of this study is the thermal feature distinction regarding whether the energy-based tools are contacting the tissue using a Convolutional Neural Network (CNN), and the extraction of the knife contacting frames, to form incision trajectories that can meet the requirements during the surgery.

This study first verified that a CNN can classify the thermal images of electric knife and ultrasonic cutter operations separately, and can improve the accuracy of the incision trajectories derived from the connection of the thermal intensity centroids of the frames predicted by the CNN as "Contacting".

Our results obtained by employing the electric knife not only reveal a remarkably high accuracy of 97.2% in the CNN's identification, but also achieved an error reduction of more than 2.5 times in the incision trajectory prediction as compared to that obtained by the conventional method. Besides the electric knife, the results obtained by employing another electric tool, the ultrasonic cutter reveal a high accuracy up to 93.7%.

In this study, we confirmed the feasibility of using a CNN for the thermal feature distinction of electric tools contacting the tissue, and confirmed that the proposed method has achieved millimeter-level accuracy under controlled ex vivo semi-rigid and constrained soft-tissue conditions.