"Toward tissue interface detection with an electrical impedance sensing surgical drill"

Abstract

Dental implantation is an increasingly common procedure used to treat missing teeth. However, surgical drilling to place the implant poses a high risk of injury to critical anatomy, such as inferior alveolar nerve injury or maxillary sinus perforation. A real-time surgical feedback system sensing proximity of these critical anatomy could reduce injury risks. This dissertation investigates the development of such a system that incorporates an electrical impedance sensor into the tip of a surgical drill.

A simulation framework based on finite element method (FEM) was developed to optimize the sensor as it approached a high impedance boundary. The accuracy of the FEM framework was improved through use of a novel empirical contact impedance model with adaptive mesh refinement strategies. The FEM model closely matched benchtop measurements with an overall mean relative error of +1.7%. Using the simulation platform, the optimal sensing geometry for a 2mm diameter dental-drill was determined to include a 1.6mm exposed length at the drill bit tip.

An in-vivo animal study protocol was developed and followed using 14 adult pigs. 146 holes were drilled, and local impedance data was intermittently recorded at 6 locations on average for each hole. Using intraoperative optical 3D tracking, impedance measurements were located in CT and micro-CT scans. Bone densities were estimated from these scans and compared to impedance measurements.

The cortical boundary thickness when approaching the mandibular canal was 1.29±0.31mm, and when approaching the maxillary sinus was 1.41±0.35 mm. Phase (251 Hz) showed a weak correlation of r = 0.27 (p<0.01) with bone density. In homogenous regions, the correlation with phase (158Hz) improved to r = 0.34 (p<0.01). Measured resistance (10kHz) showed a maximum correlation of r = 0.30 (p<0.01). Measuring phase at 251Hz while drilling showed significant differences (p<0.05) between the boundary and the layer before boundary, and an area under curve (AUC) of 0.71 was determined to detect critical anatomy. Although these results show smaller differences between tissue type and a smaller ability to detect critical anatomy than expected, this work achieved moderate success (results noted above) and uncovered multiple data-collection challenges that may improve performance in subsequent studies.

Thesis Committee

• Prof. Ryan Halter (chair)
• Prof. Ethan Murphy
• Prof. Xiaoyao Fan
• Prof. Kofi Odame
• Prof. Ørjan G. Martinsen (Department of Physics, University of Oslo, Norway)

Contact

For more information, contact Thayer Registrar at thayer.registrar@dartmouth.edu.