Characterization and correction of intraoperative soft tissue deformation in image-guided laparoscopic liver surgery

Abstract

Laparoscopic liver surgery is challenging to perform due to a compromised ability of the surgeon to localize subsurface anatomy in the constrained environment. While image guidance has the potential to address this barrier, intraoperative factors, such as insufflation and variable degrees of organ mobilization from supporting ligaments, may generate substantial deformation. The severity of laparoscopic deformation in humans has not been characterized, and current laparoscopic correction methods do not account for the mechanics of how intraoperative deformation is applied to the liver. We first measure the degree of laparoscopic deformation at two insufflation pressures over the course of laparoscopic-to-open conversion in 25 patients. With this clinical data alongside a mock laparoscopic phantom setup, we report a biomechanical correction approach that leverages anatomically load-bearing support surfaces from ligament attachments to iteratively reconstruct and account for intraoperative deformations. Laparoscopic deformations were significantly larger than deformations associated with open surgery, and our correction approach yielded subsurface target error of [Formula: see text] and surface error of [Formula: see text] using only sparse surface data with realistic surgical extent. Laparoscopic surface data extents were examined and found to impact registration accuracy. Finally, we demonstrate viability of the correction method with clinical data.

Keywords: biomechanical modeling; image-guided surgery; laparoscopy; liver; registration; soft tissue deformation.

Fig. 1

(a) Anatomy of the liver, adapted from Kingham et al. The falciform and left and right triangular ligament attachments shown in red are put in tension during insufflation due to expansion of the abdominal cavity. Two salient anatomical features of the liver are shown in blue. (b) The liver phantom is suspended in an insufflated mock abdomen without mobilization from its ligaments. (c) Positions of the 147 subsurface targets distributed throughout the volume of the phantom. (d) Segmented preoperative and intraoperative phantom volumes are shown in blue and red, respectively. The difference between surfaces demonstrates the deformation reproduced in the laparoscopic phantom simulator.

Fig. 2

Overview of intraoperative organ shape comparison. Sparse point clouds of the intraoperative organ shape under two distinct operative conditions are coregistered to the preoperative liver surface and resampled into full reconstructed surfaces. Distance measures of shape dissimilarity are computed for only the resampled points that are enclosed by the extents of both data sources (purple region).

Fig. 3

Overview of deformation correction algorithm. (a) Model solutions are computed for perturbations of a choice of control points. (b) Nonrigid correction is performed by iteratively updating a set of parameters that are used to reconstruct the intraoperative organ shape from precomputed modes of expected deformation.

Fig. 4

Reconstructed closest point distance error from three representative cases among the $n=25$ between (a) preoperative and laparoscopic surfaces, (b) preoperative and open surfaces, and (c) laparoscopic and open surfaces.

Fig. 5

Registered preoperative liver (blue) in comparison with the ground-truth intraoperative organ shape (red) for each organ deformation and registration technique. The sparse intraoperative data used to perform the registrations are overlaid in black. Attaining perfect alignment is challenging due to incomplete coverage of the intraoperative surface data.

Fig. 6

Surface correction quartiles are shown for rigid and nonrigid registrations to each series of laparoscopic and open organ configurations. The gray panel displays the distribution of surface correction among all intraoperative phantom mobilizations and surface data extents.

Fig. 7

(a) Variation in available surface data extents from clinical data: 31% (top), 20% (center), and 11% (bottom). (b) Average surface data extents through each of the nine ports of the phantom, standard deviation in parentheses. Lateral ports colored in red provide average extents of $<15\%$ of the organ surface. Periumbilical ports in yellow offer moderate extents between 20% and 25%, and ports placed in the medial right upper quadrant yield the best available surface extents, which exceed 25% on average.

Fig. 8

Distributions of (a) TRE and (b) target correction with respect to the extent of intraoperative surface data. The box and whiskers represent the median, upper and lower quartiles, maximum, and minimum of TRE. Our nonrigid correction contributes little improvement over rigid registration at extents smaller than 10%. However, at extents $>22\%$, the nonrigid correction algorithm offers a substantial improvement in TRE.