Quantifying the Potential Benefit of Applying Remote Magnetic Navigation Systems to Abdominal SBRT

Abstract

Purpose: Radiation induced bowel toxicity is challenging for abdominal SBRT due to proximity to PTV and high dose/fraction. Remote magnetic navigation systems (RMNS) have been developed for magnetic capsule endoscopy, in which the RMNS mounted on a robotic arm controls the location and orientation of a magnetic capsule within the bowel. Our goal is to apply this technology to enable subtle shifts in bowel location relative to PTV to achieve more favorable anatomy and reduce SBRT radiation dose. The purpose of this study is to examine the frequency of cases that would benefit from this application in various SBRT treatments near the GI tract and the possible dosimetric advantage this method of displacement could be via analytical modeling and finite element analysis.Methods: First, prior SBRT cases (≤5 fractions) ranging from 2019-2023 were reviewed; prescriptions labeled as liver, pancreas, adrenal, and abdominal were identified and stratified for analysis. The maximum dose to GI tract (small bowel, large bowel, and stomach) was recorded, along with distance to the closest edge of the Planning Target Volume (PTV) on the simulation CT. Daily CBCT images were also reviewed to quantify interfractional variations. These SBRT cases were then selectively refined based on physician review and recommendation. Secondly, using force simulation data of a prototype RMNS developed by a collaborator, two mathematical models were developed to predict behavior of the capsule in the bowel; both portrayed bowel and surrounding tissue as a double spring-based system with spring constant determined via Young’s Modulus of tissue. The first used springs in series and utilized Euler-Bernoulli Beam Theory; the second used springs in parallel to model a capsule fixed within abdominal cavity. The models were applied to calculate the bowel displacement achievable from a prototype magnetic capsule (neodymium, 9mm-diameter, 18mm-length) and RMNS (4 coils, 571 loops, 10A) mounted on a robotic arm. Potential dose sparing was quantified for prior clinical SBRT cases (n=10) identified as potential candidates for RMNS intervention. Finally, exploratory finite element analysis was performed in FEBio Studio to investigate gastrointestinal deformation under magnetic actuation. Patient-specific STL meshes derived from CT segmentation were imported initially, although simplified geometric models were also developed to reduce computational complexity. Soft tissue behavior was explored using Neo-Hookean, Mooney–Rivlin, and elastic-damage material formulations. Simulations incorporated nodal loading, prescribed displacements, rigid-body capsule representations, sliding contact conditions, and fixed boundary constraints, with solver parameters adjusted iteratively to improve convergence.Results: At simulation, GI tract was within 2cm of the PTV for 16/50, 50/50, 22/28, and 11/14 for liver, pancreas, adrenal, and abdominal SBRT cases, respectively; the GI tract overlapped with the PTV for 7/50, 36/50, 13/28, and 4/14 cases, respectively. Small bowel, large bowel, and stomach were roughly equally likely to be within 2cm of the PTV for liver (11/50, 12/50, and 7/50, within 2cm respectively), and pancreas SBRT cases (49/50, 48/50, and 47/50 within 2cm, respectively); while there was more variation for adrenal (19/28, 11/28, and 14/28, respectively) and abdominal SBRT cases (10/14 and 4/14 within 2cm, for bowel and stomach respectively). Overall, 99/142 (69.7%) of SBRT cases examined exposed healthy GI tissue to high amounts of dose due to proximity (2cm) to the target. The original 99/142 cases that were deemed potentially clinically applicable were cut to a total of 28 favorable cases. Of these 28, 6 were liver cases, 12 were pancreas cases, 8 were adrenal cases, and 2 were abdominal cases. 10 of these physician narrowed cases (n=3 liver, n=2 adrenal, n=3 pancreas, n=2 abdominal) were selected for in-depth dosimetric analysis using mathematical models. Application of the mathematical models to these representative cases demonstrated reductions in dose to gastrointestinal organs-at-risk following simulated displacement of the OAR contours. Across the analyzed cohort, both models predicted decreases in maximum dose, mean dose, and high-dose volumetric metrics (D1cc and D0.5cc). Model 1 generally produced larger reductions compared to Model 2. On average, Model 1 resulted in reductions of approximately 8.8%Rx in maximum dose, 1.2%Rx in mean dose, 6.8%Rx in D1cc, and 7.5%Rx in D0.5cc. Model 2 produced smaller but measurable reductions of approximately 3.8%Rx in maximum dose, 0.5%Rx in mean dose, 3.3%Rx in D1cc, and 3.7%Rx in D0.5cc. Exploratory finite element simulations demonstrated qualitative deformation of simplified GI geometries under applied loading; however, numerical convergence limitations prevented extraction of reliable quantitative displacement metrics.Conclusion: These results demonstrate that GI organs-at-risk are frequently located near abdominal SBRT targets, creating significant challenges for safe dose delivery. The mathematical displacement models developed in this work suggest that magnetically induced GI displacement could provide measurable reductions in dose to critical GI structures in select clinical scenarios. While exploratory finite element simulations highlight the potential for more advanced biomechanical modeling, further development of patient-specific modeling techniques and magnetic actuation systems will be required to fully realize this approach. Overall, this work establishes a preliminary framework for evaluating magnetic GI displacement as a potential strategy for improving organ-at-risk sparing in abdominal SBRT.