Development of an Augmented Reality Guidance System for Head and Neck
Cancer Resection
Abstract
The use of head-mounted augmented reality (AR) for surgeries has grown
rapidly in recent years. AR aids in intraoperative surgical navigation
through overlaying 3D holographic reconstructions of medical data.
However, performing AR surgeries on complex areas poses challenges in
terms of accuracy and speed. This study explores the feasibility of an
AR guidance system for resections of positive tumor margins at the head
and neck region. We present an intraoperative solution that enables
surgeons to access holographic reconstructions of resected cadaver
tissues. The solution involves using 3D scanner to capture detailed
scans of the resected tissue, which are uploaded into our software. It
then converts these scans into holograms that are viewable through a
head-mounted AR display. Surgeons navigates the tumor site by
re-aligning these holograms with cadavers using gestures or voice
commands. This workflow runs concurrently with frozen section analysis.
On average, we achieve an uploading time of 2.98 min, visualization time
of 1.05 min and re-alignment time of 4.39 min, within the 20 - 30 min
window for frozen section analysis. We achieve a mean re-alignment error
of 3.1 mm. Our software provides a foundation for new product
development in using AR to navigate complex anatomy in surgery.