The Development of Spinal Endoscopic Ultrasonic Imaging System With an Automated Tissue Recognition Algorithm

Area of Science:

• Neurosurgery
• Medical Imaging
• Surgical Technology

Background:

• Endoscopic spinal surgery (ESS) presents challenges due to limited visualization.
• High surgeon proficiency is required for procedures like discectomy and decompression.
• Risk of damaging neural structures exists due to the narrow surgical field.

Purpose of the Study:

• To develop an intraoperative ultrasound imaging device for ESS.
• To assist surgeons in differentiating tissue types during endoscopic spinal procedures.
• To enhance visualization and safety within the confined surgical site.

Main Methods:

• A 4-mm custom ultrasound probe was integrated into an endoscopic spinal surgery (ESS) working channel.
• Ultrasound imaging was performed on ovine lumbar spine samples.
• A 2-stage classification algorithm using DenseNet architecture was developed for automated tissue recognition.

Main Results:

• The ultrasound probe provided clear images within the ESS working channel.
• The automated recognition algorithm achieved over 90% accuracy in identifying 6 spinal tissue types (spinal cord, nucleus pulposus, adipose tissue, bone, annulus fibrosis, nerve roots).
• Recognition time was under 0.1 seconds with a Kappa value of 0.875.

Conclusions:

• The developed intraoperative ultrasound system is compatible with existing ESS channels.
• The system accurately and rapidly identifies critical intraspinal structures.
• This technology has the potential to reduce the learning curve and improve efficiency and safety in ESS.