Patient-Specific Simulation of Pneumoperitoneum for Laparoscopic Surgical Planning

Abstract

Gas insufflation in laparoscopy deforms the abdomen and stretches the overlying skin. This limits the use of surgical image-guidance technologies and challenges the appropriate placement of trocars, which influences the operative ease and potential quality of laparoscopic surgery. This work describes the development of a platform that simulates pneumoperitoneum in a patient-specific manner, using preoperative CT scans as input data. This aims to provide a more realistic representation of the intraoperative scenario and guide trocar positioning to optimize the ergonomics of laparoscopic instrumentation. The simulation was developed by generating 3D reconstructions of insufflated and deflated porcine CT scans and simulating an artificial pneumoperitoneum on the deflated model. Simulation parameters were optimized by minimizing the discrepancy between the simulated pneumoperitoneum and the ground truth model extracted from insufflated porcine scans. Insufflation modeling in humans was investigated by correlating the simulation's output to real post-insufflation measurements obtained from patients in theatre. The simulation returned an average error of 7.26 mm and 10.5 mm in the most and least accurate datasets respectively. In context of the initial discrepancy without simulation (23.8 mm and 19.6 mm), the methods proposed here provide a significantly improved picture of the intraoperative scenario. The framework was also demonstrated capable of simulating pneumoperitoneum in humans. This study proposes a method for realistically simulating pneumoperitoneum to achieve optimal ergonomics during laparoscopy. Although further studies to validate the simulation in humans are needed, there is the opportunity to provide a more realistic, interactive simulation platform for future image-guided minimally invasive surgery.

Keywords: Laparoscopy; Patient-specific; Pneumoperitoneum; Simulation; Surgical planning.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Fig. 1

Slices from insufflated and deflated porcine CT scans demonstrating the segmentation of different regions: gas (red), abdominal viscera (green), lungs (light blue) and abdominal wall (dark blue). The red line in the deflated scans indicates the boundary between the peritoneal cavity and the abdominal wall

Fig. 2

3D reconstructions of an insufflated and deflated porcine scan, produced by interpolating individually segmented axial slices in Fig. 1a

Fig. 3

Triangular surface meshes of insufflated porcine (abdominal wall in blue, organs in green, pneumoperitoneum in red)

Fig. 4

Particle density of insufflated mesh, separated into the inflatable structure (blue), skin (peach), viscera (red). Green points indicate fixed regions in the back

Fig. 5

Pneumoperitoneum before and after simulation, showing increased volume of the pneumoperitoneum (10x) and resultant organ compression and abdominal wall deformation

Fig. 6

Three measurements were taken from landmarks on the abdominal surface: umbilicus to right and left anterior-superior iliac spines (ASIS), xiphisternum (XS) to pubic symphysis (PS)

Fig. 7

Mean overall error of simulated meshes across simulation pressure, as average distance to corresponding vertices on ground truth meshes from insufflated porcine scans

Fig. 8

Color maps displaying map illustrating the overall error (mm) of the simulated pneumoperitoneum of the 2nd and 7th porcine datasets at their optimal pressure parameters. Colors correspond to the overall error in millimeters (as on the color bar). Warmer colors indicate a higher degree of misalignment with the ground truth mesh, implying greater overall error