Ultrasound Aided Vertebral Level Localization for Lumbar Surgery

Area of Science:

• Medical Imaging
• Surgical Navigation
• Spinal Surgery

Background:

• Accurate localization of vertebral levels is critical for safe and effective lumbar hernia surgery.
• Current methods for determining surgical entry points can be complex and lack precision.

Purpose of the Study:

• To develop and evaluate a novel system for intraoperative vertebral level localization using free-hand 2-D ultrasound (US) imaging.
• To improve the accuracy and efficiency of surgical entry point detection in lumbar procedures.

Main Methods:

• A deep convolutional neural network was developed for automatic segmentation of bone structures from US images.
• A registration strategy was implemented to match pre-operative X-ray outlines of spinous processes with intraoperative US images.
• The system leverages differences in spinous process shapes for robust identification and labeling of vertebral levels.

Main Results:

• The proposed deep learning-based bone segmentation algorithm demonstrated superior performance and speed compared to existing methods.
• The complete US-based localization procedure achieved a correct labeling rate of 92% across 19 clinical datasets.
• Evaluation involved two independent observers, confirming the system's reliability.

Conclusions:

• The developed ultrasound-based system offers a feasible and accurate solution for intraoperative surgical entry point detection in spinal surgeries.
• This technology has the potential to enhance surgical navigation and patient safety during lumbar procedures.