Tactile sensor-based real-time clustering for tissue differentiation.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Surgical Technology

Background:

• Intraoperative tumor delineation relies on surgeon's visual and tactile cues, which is challenging due to low brain tissue contrast.
• Neuroendoscopy and robotic surgery often lack or have diminished tactile feedback, increasing surgical risks.
• Development of artificial tactile perception is crucial for improving surgical safety and accuracy.

Purpose of the Study:

• To enhance and evaluate a novel tactile sensor for improved intraoperative brain tissue differentiation.
• To develop an intelligent artificial tactile perception system for neurosurgery.
• To assess the sensor's capability in distinguishing tumor from healthy brain tissue.

Main Methods:

• A robotic-driven piezoelectric bimorph sensor was utilized for tactile sensing.
• Multisine excitation was applied to obtain the frequency response function of tissue-sensor contact.
• Viscoelastic parameters (elastic moduli E1, E2, viscosity coefficient η) were derived from load-depth, relaxation, and creep tests for tissue classification using a multivariate cluster algorithm.

Main Results:

• The cluster algorithm successfully assigned five distinct clusters for white matter, basal ganglia, and thalamus.
• Thalamus was exclusively delineated, while basal ganglia and white matter were grouped into a common cluster, indicating lower discrimination for these types.
• Gray matter exhibited potential for further sub-clustering, demonstrating high sensitivity.

Conclusions:

• Piezoelectric bimorph-based tactile sensors, excited by multisine, show high sensitivity for ex vivo brain tissue differentiation.
• The developed sensor principle holds promise for enhancing intraoperative surgical guidance.
• Further evaluation is needed, but the results are encouraging for clinical application.