A deep learning-based hybrid approach for the solution of multiphysics problems in electrosurgery.

Multiphysics modeling of evolving topology in the electrosurgical dissection of soft hydrated tissues is a challenging problem, requiring heavy computational resources. In this paper, we propose a hybrid approach that leverages the regressive capabilities of deep convolutional neural networks (CNN) with the precision of conventional solvers to accelerate Multiphysics computations. The electro-thermal problem is solved using a finite element method (FEM) with a Krylov subspace-based iterative solver and a deflation-based block preconditioner. The mechanical deformation induced by evaporation of intra- and extracellular water is obtained using a CNN model. The CNN is trained using a supervised learning framework that maps the nonlinear relationship between the micropore pressure and deformation field for a given tissue topology. The simulation results show that the hybrid approach is significantly more computationally efficient than a FEM-based solution approach using a block-preconditioned Krylov subspace solver and a parametric solution approach using a proper generalized decomposition (PGD) based reduced order model. The accuracy of the hybrid approach is comparable to the ground truth obtained using a standard multiphysics solver. The hydrid approach overcomes the limitations of end-to-end learning including the need for massive datasets for training the network.

A Multiphysics Model for Radiofrequency Activation of Soft Hydrated Tissues.

A multi-physics model has been developed to investigate the effects of cellular level mechanisms on the electro-thermo-mechanical response of hydrated soft tissues with radiofrequency (RF) activation. A micromechanical model generates an equation of state (EOS) that provides the additional pressure arising from evaporation of intra- and extracellular water as well as temperature to the continuum level thermo-mechanical model. A level set method is used to capture the interfacial evolution of tissue damage with the level set evolution equation derived from the second law of thermodynamics, which is consistent with Griffith's fracture evolution criterion. The discretized equations are solved simultaneously using a Krylov subspace based iterative solver (GMRES) with block preconditioning that effectively deflates the spectrum of the system matrix, resulting in exponential convergence of the Arnoldi iterations. Example problems, including experimental validation, illustrate the computational accuracy and efficiency of the technique.

FUSE certification enhances performance on a virtual computer based simulator for dispersive electrode placement.

BACKGROUND: The Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) has developed the fundamental use of surgical energy (FUSE) didactic curriculum in order to further understanding of the safe use of surgical energy. The virtual electrosurgical skill trainer (VEST) is being developed as a complementary simulation-based curriculum, with several modules already existing. Subsequently, a new VEST module has been developed about dispersive electrode placement. The purpose of this study is to assess knowledge about dispersive electrode placement in surgeons and surgical trainees in addition to describing a new VEST module. METHODS: Forty-six subjects (n = 46) were recruited for participation at the 2016 SAGES conference Learning Center. Subjects were asked to complete demographic surveys, a five-question pre-test, and a five-question post-test after completing the VEST dispersive electrode module. Subjects were then asked to rate different aspects of the module using a five-point Likert scale questionnaire. RESULTS: Mean pre-simulator and post-simulator assessment scores were 1.5 and 3.4, respectively, with Wilcoxon signed rank analysis showing a significant difference in the means (p < 0.05). Subjects were grouped by the presence (n = 12) or absence (n = 31) of prior FUSE experience and by training level. Mann-Whitney U testing showed no significant difference in pre-simulator assessment scores between attending surgeons and trainees (p > 0.05). In those with and without FUSE exposure, a significant difference (p < 0.05) was seen in pre-simulator assessment scores, and no significant difference in Likert scale assessment scores was seen. CONCLUSIONS: This study demonstrated a new VEST educational module. Consistently high Likert assessment scores showed that users felt that the VEST module helped their understanding of dispersive electrode placement. Additionally, the study reflected a potential knowledge deficit in the safe use of dispersive electrodes in the surgical community, also demonstrating that even some exposure to the FUSE curriculum developed by SAGES provides increased awareness about dispersive electrode use.

A physics-based algorithm for real-time simulation of electrosurgery procedures in minimally invasive surgery.

BACKGROUND: High-frequency electricity is used in the majority of surgical interventions. However, modern computer-based training and simulation systems rely on physically unrealistic models that fail to capture the interplay of the electrical, mechanical and thermal properties of biological tissue. METHODS: We present a real-time and physically realistic simulation of electrosurgery by modelling the electrical, thermal and mechanical properties as three iteratively solved finite element models. To provide subfinite-element graphical rendering of vaporized tissue, a dual-mesh dynamic triangulation algorithm based on isotherms is proposed. The block compressed row storage (BCRS) structure is shown to be critical in allowing computationally efficient changes in the tissue topology due to vaporization. RESULTS: We have demonstrated our physics-based electrosurgery cutting algorithm through various examples. Our matrix manipulation algorithms designed for topology changes have shown low computational cost. CONCLUSIONS: Our simulator offers substantially greater physical fidelity compared to previous simulators that use simple geometry-based heat characterization.