EP531 - Dual-Layer Autonomous Surgical Navigation System to Prevent Ureteral Injury: Preclinical Validation Study.

ePoster Presenter(s)

• 

  Hussam Shwaib, MD, MAS

  General Surgery Resident
  Rutgers Robert Wood Johnson, Department of Surgery
  Rutgers Robert Wood Johnson, Department of Surgery
  New Brunswick, NJ, United States

  Disclosure information not submitted.

Author(s)

• SM

  Srivatsan Manikandan, n/a

  Software Engineer
  Rutgers Robert Wood Johnson Medical School, Department of Surgery
  South Brunswick Township, New Jersey, United States

  Disclosure information not submitted.
• JP

  Jenya Pandu, n/a

  Undergraduate Research Assistant
  Department of Surgery, Rutgers Robert Wood Johnson Medical School
  Kendall Park, New Jersey, United States

  Disclosure information not submitted.
• BS

  Brendan Scully, Md

  Assistant professor
  Rutgers Robert Wood Johnson
  New Brunswick, New Jersey, United States

  Disclosure information not submitted.
• 

  Kristen Donohue, MD, FACS, FASCRS

  Assistant Professor of Surgery
  Rutgers Robert Wood Johnson Medical School
  Atlantic Highlands, New Jersey, United States

  Disclosure information not submitted.
• TE

  Theodore Eisenstat, MD FACS FASCRS

  Professor of Surgery
  Robert Wood Johnson Medical School
  New Brunswick, New Jersey, United States

  Disclosure information not submitted.
• 

  Daniel L. Feingold, MD, FACS, FASCRS

  Professor of Surgery
  Rutgers University
  New Brunswick, New Jersey, United States

  Disclosure information not submitted.

Senior Author(s)

• NM

  Nell Maloney Patel, MD

  Professor of Surgery; Vice Chair of Education
  Rutgers Robert Wood Johnson Medical School
  Westfield, New Jersey, United States

  Disclosure information not submitted.

Purpose/Background: Ureteral injury remains one of the most costly and avoidable complications in abdominal and pelvic surgery, occurring in 0.3-2% of cases. Current fluorescence systems require conscious activation and provide only momentary visualization. Our platform creates a dual-layer autonomous navigation system capable of continuous ureter identification without user activation. The first layer is a smart ureteral stent, and the second is a real-time computer vision pipeline that detects and tracks the stent automatically. The goal of this study was to establish proof of function and preclinical validation under simulated surgical conditions.

Methods/Interventions: Bench experiments were conducted using a dual-arm robotic phantom replicating laparoscopic geometry and illumination. A custom real-time image-processing pipeline was developed for contour extraction and distance tracking. The "smart" ureteral stent marker was optimized for continuous visibility. Testing was performed under simulated operating room conditions to assess dual-mode detection stability. System performance was evaluated across key metrics: frame rate ≥ 25 frames per second (fps), latency ≤ 75 milliseconds (ms), tracking error ≤ 1.5 millimeters (mm), and continuity ≥ 98%.

Results/Outcomes: The system continuously identified and tracked the stent without manual activation. Sustained stent detection was achieved in 10 of 12 dynamic-illumination trials (83%), with transient loss only during rapid high-intensity transitions and automatic reacquisition within < 1 second. Across all runs, per-frame color-marker detection averaged 97 ± 2%, and mean tracking continuity was 1,785 ± 65 seconds per 1,800 seconds (≈99%). Real-time performance was maintained at 30 ± 2 fps with a latency of 47 ± 8 ms, and distance estimation error remained ≤ 1.5 mm. The marker was consistently detected at ~20 mm.

Conclusion/Discussion: This platform demonstrates an activation-free, continuously visible ureteral stent with autonomous tracking capability. The dual-layer optical navigation approach enables sustained detection under surgical lighting and motion, addressing a major unmet need in intraoperative safety. These findings establish a technical foundation for future navigation systems designed to prevent ureteral injury. Next-phase work includes quantitative validation, porcine model studies, and regulatory preparation for first-in-human trials.

Experimental Bench Setup for Dual-Robot Phantom Validation
Custom robotic phantom platform used for preclinical validation of the navigation system. Two robotic arms simulate laparoscopic geometry and instrument motion for real-time fluorescent and visible tracking of the smart ureteric stent under controlled illumination conditions.
Dual-Layer Detection Interface for Ureteric Stent Tracking
(A) Real-time color marker detection demonstrating near-field tracking at 20 mm under visible illumination.

(B) Fluorescent stent contour detection at operative working distance (~380 mm) using near-infrared imaging. Both confirm continuous, activation-free object recognition by the custom software.