Image-based laparoscopic bowel measurement.

Area of Science:

• Medical Imaging
• Surgical Technology
• Computer Vision

Background:

• Minimally invasive surgery (MIS) presents challenges for surgeons, notably diminished depth perception.
• Accurate distance estimation is crucial in MIS, particularly for procedures like gastric bypass surgery.
• Current methods lack real-time surface measurement capabilities during interventions.

Purpose of the Study:

• To develop and evaluate an on-the-fly surface measurement approach using stereo endoscopy for MIS.
• To address the challenge of lost depth perception in MIS by providing accurate distance estimations.
• To enable precise measurements of anatomical structures, such as bowel segments, during surgery.

Main Methods:

• Utilized calibrated stereo endoscope image pairs as input.
• Employed random forest segmentation for on-the-fly organ and instrument localization.
• Implemented three surface distance measurement techniques: direct, shortest path, and spline-interpolated shortest path.

Main Results:

• Evaluated the approach on phantom and porcine datasets during simulated minimally invasive bowel measurements.
• Achieved an average error of [Formula: see text] for shorter phantom bowel segments (<5 cm) and [Formula: see text] for longer segments (>10 cm).
• Demonstrated an average error of [Formula: see text] on porcine datasets.

Conclusions:

• Presented a novel stereo endoscopy-based approach for on-the-fly surface measurements in MIS.
• Successfully compared different surface length determination methods.
• The method requires only a calibrated stereo endoscope, minimizing disruption to surgical workflows.