Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Supervised Autonomous Resection and Retraction Framework for Transurethral Enucleation of the Prostatic Median Lobe

Mariana Smith, Tanner Watts, Susheela Sharma Stern, Brendan Burkhart, Hao Li, Alejandro O. Chara, Nithesh Kumar, James Ferguson, Ayberk Acar, Jesse F. d'Almeida, Lauren Branscombe, Lauren Shepard, Ahmed Ghazi, Ipek Oguz, Jie Ying Wu, Robert J. Webster, Axel Krieger, Alan Kuntz

TL;DR

The paper addresses automated transurethral median lobe enucleation for BPH by merging a CT-guided model-based resection planner with PushCVAE, a learning-based retraction network, on the Virtuoso Endoscopic System. It introduces a three-phase median lobe workflow and a phase-specific retraction policy trained from surgeon demonstrations, enabling Level-3 supervised autonomy on hydrogel prostate phantoms. Key results show 97.1 percent removal of the targeted median lobe volume, with preoperative and postoperative CT measurements supporting accuracy relative to the planned margins. This work establishes image-guided autonomy in MIS prostate surgery and provides a foundation for future perception-driven, closed-loop automated enucleation.

Abstract

Concentric tube robots (CTRs) offer dexterous motion at millimeter scales, enabling minimally invasive procedures through natural orifices. This work presents a coordinated model-based resection planner and learning-based retraction network that work together to enable semi-autonomous tissue resection using a dual-arm transurethral concentric tube robot (the Virtuoso). The resection planner operates directly on segmented CT volumes of prostate phantoms, automatically generating tool trajectories for a three-phase median lobe resection workflow: left/median trough resection, right/median trough resection, and median blunt dissection. The retraction network, PushCVAE, trained on surgeon demonstrations, generates retractions according to the procedural phase. The procedure is executed under Level-3 (supervised) autonomy on a prostate phantom composed of hydrogel materials that replicate the mechanical and cutting properties of tissue. As a feasibility study, we demonstrate that our combined autonomous system achieves a 97.1% resection of the targeted volume of the median lobe. Our study establishes a foundation for image-guided autonomy in transurethral robotic surgery and represents a first step toward fully automated minimally-invasive prostate enucleation.

A Supervised Autonomous Resection and Retraction Framework for Transurethral Enucleation of the Prostatic Median Lobe

TL;DR

The paper addresses automated transurethral median lobe enucleation for BPH by merging a CT-guided model-based resection planner with PushCVAE, a learning-based retraction network, on the Virtuoso Endoscopic System. It introduces a three-phase median lobe workflow and a phase-specific retraction policy trained from surgeon demonstrations, enabling Level-3 supervised autonomy on hydrogel prostate phantoms. Key results show 97.1 percent removal of the targeted median lobe volume, with preoperative and postoperative CT measurements supporting accuracy relative to the planned margins. This work establishes image-guided autonomy in MIS prostate surgery and provides a foundation for future perception-driven, closed-loop automated enucleation.

Abstract

Concentric tube robots (CTRs) offer dexterous motion at millimeter scales, enabling minimally invasive procedures through natural orifices. This work presents a coordinated model-based resection planner and learning-based retraction network that work together to enable semi-autonomous tissue resection using a dual-arm transurethral concentric tube robot (the Virtuoso). The resection planner operates directly on segmented CT volumes of prostate phantoms, automatically generating tool trajectories for a three-phase median lobe resection workflow: left/median trough resection, right/median trough resection, and median blunt dissection. The retraction network, PushCVAE, trained on surgeon demonstrations, generates retractions according to the procedural phase. The procedure is executed under Level-3 (supervised) autonomy on a prostate phantom composed of hydrogel materials that replicate the mechanical and cutting properties of tissue. As a feasibility study, we demonstrate that our combined autonomous system achieves a 97.1% resection of the targeted volume of the median lobe. Our study establishes a foundation for image-guided autonomy in transurethral robotic surgery and represents a first step toward fully automated minimally-invasive prostate enucleation.

Paper Structure

This paper contains 10 sections, 8 figures.

Table of Contents

  1. Introduction
  2. Methods
  3. System Hardware
  4. Phantom Setup
  5. Model-Based Resection Planner
  6. Learning Based Retraction
  7. Experiments and Results
  8. Feasibility Study
  9. Postoperative CT Results
  10. Discussion and Conclusion

Figures (8)

  • Figure 1: Robotic system for experimentation, comprised of a Kuka LBR Med 14 for gross positioning and a VES mounted to its end-effector. The VES consists of a monocular endoscope with two CTR tools: one spatula retractor and one monopolar electrosurgery probe.
  • Figure 2: A CT scan is taken of the hydrogel prostate phantom, which allows segmentation of the capsule and lobes using embedded contrast agent. The robot enters the model through the urethra, where it can visualize the left, right, and median lobes of the prostate in its monocular endoscope.
  • Figure 3: Minimally-invasive robotic procedural workflow for resection of the median prostate lobe using the VES, demonstrated here in teleoperation by a human operator.
  • Figure 4: Process diagram of the model-based resection planner.
  • Figure 5: Groupings of planned cuts to execute at discrete global positions of the endoscope, necessary to ensure reachability. Global movements are executed manually with the Kuka robot, with an RCM constraint at the entry.
  • ...and 3 more figures