Nonrigid liver registration for image-guided surgery using partial surface data: a novel iterative approach

Abstract

In the context of open abdominal image-guided liver surgery, the efficacy of an image-guidance system relies on its ability to (1) accurately depict tool locations with respect to the anatomy, and (2) maintain the workflow of the surgical team. Laser-range scanned (LRS) partial surface measurements can be taken intraoperatively with relatively little impact on the surgical workflow, as opposed to other intraoperative imaging modalities. Previous research has demonstrated that this kind of partial surface data may be (1) used to drive a rigid registration of the preoperative CT image volume to intraoperative patient space, and (2) extrapolated and combined with a tissuemechanics-based organ model to drive a non-rigid registration, thus compensating for organ deformations. In this paper we present a novel approach for intraoperative nonrigid liver registration which iteratively reconstructs a displacement field on the posterior side of the organ in order to minimize the error between the deformed model and the intraopreative surface data. Experimental results with a phantom liver undergoing large deformations demonstrate that this method achieves target registration errors (TRE) with a mean of 4.0 mm in the prediction of a set of 58 locations inside the phantom, which represents a 50% improvement over rigid registration alone, and a 44% improvement over the prior non-iterative single-solve method of extrapolating boundary conditions via a surface Laplacian. PURPOSELiver resection surgery in the open abdomen is a challenging setting for the application of image-guided surgical techniques which have been largely limited to procedures involving the cranium in the past. Surgical liver presentation typically begins with mobilization from the surrounding anatomy, followed by stabilization by packing support material underneath and around the organ. Thus large deformations (on the order of several centimeters) often occur between the preoperative (when CT imaging was performed) and intraoperative organ states.While intraoperative imaging has been used to document the extent of deformation, 1 and guidance solutions using intraoperative imaging have been proposed, 1-5 the workflow requirements and the challenges of integrating preoperative imaging data continue to be a hindrance. As a result, there remains a clinical need to efficiently align preoperative data to the intraoperative patient state, in order to leverage the wealth of preoperative image data that can be collected without incurring the encumbrance of many imaging systems. In prior work, Clements, et al. proposed a robust weighted-patch iterative-closest-point algorithm to perform rigid registration using a surface point cloud obtained from a laser range scan (LRS) combined with salient feature data patches from tooltip swabbing. 6 Subsequently, Dumpuri et al. 7 and Clements et al. 8 investigated methods for an additional nonrigid registration step. Using a linear elastic finite element model generated from the patient's CT data, these methods imposed Dirichlet b...

A Mechanics-Based Nonrigid Registration Method for Liver Surgery Using Sparse Intraoperative Data

“…Some of our work on this subject was presented in preliminary form in [26]. In this paper, we offer several significant contributions beyond our early work including 1) generalization of our basic approach and extensive description of the methods involved in Section II, 2) numerous additional experimental investigations in liver phantoms and extensive experimental analysis of our approach in Section IV, and 3) feasibility studies using clinical case data.…”

Registration of liver images to minimally invasive intraoperative surface and subsurface data

“…Rucker (2013) investigated using a posterior displacement to characterize intraoperative deformation. The displacement field on the posterior side was iteratively solved for to obtain boundary conditions that optimized a metric measuring the partial surface fit between the model and the intraoperative LRS data.…”

“…The phantom was made similar to the one described in Rucker (2013). Briefly, an Ecoflex(Smooth-On, Easton PA) silicone phantom in the shape of a human liver was cast with embedded surface and subsurface targets.…”

“…1) A preoperative tomographic (MR or CT) imaging scan of the relevant organ is segmented and digitized into a soft tissue biomechanical Finite Element (FE) model, 2) a set of sparse intraoperative data is collected from the deformed state and finally, 3) the registration algorithm is used to calculate the displacement field describing how the preoperative model should be deformed to match the intraoperative data. The registration framework used was based off the work of Rucker (2013). Fig.…”

Biomechanical kidney model for predicting tumor displacement in the presence of external pressure load

“…By using a biomechanical model based simulation, Haouchine et al achieved 4 mm position differences between a simulated and scanned tumor inside a liver phantom under known external loads [9]. Rucker et al also achieved average phantom tumor registration errors of 4 mm using a combination of rigid transformations and model-based non-rigid transformations [10].…”