A Model-Based Framework for Assessing Operator Performance in Navigational Bronchoscopy
Zhaoxing Deng, David Hanley, Francis Xiatian Zhang, Kev Dhaliwal, Mohsen Khadem
AI summary
Problem
Current bronchoscopy skill assessment relies heavily on subjective expert judgment and lacks objective, quantitative metrics, hindering consistent training and skill transfer.
Approach
The framework uses electromagnetic pose data to generate optimal reference trajectories via a nonholonomic kinematic model and MPPI control, then quantifies operator deviations from these references to assess skill.
Key results
- Nonholonomic kinematic model encoding expert steering constraints
- Model-based semi-supervised classifier using trajectory deviations
- Significant separation of expert and novice performance via positional and orientation metrics
- Identification of local variance and high-error events as key discriminators
Why it matters
Provides an interpretable, data-driven alternative to subjective supervision, enabling scalable objective skill evaluation for bronchoscopy training and robotic platforms.
Abstract
Bronchoscopy is a critical procedure for diagnos- ing and treating pulmonary diseases, but its safe and effective execution demands substantial operator training. Insufficient experience is associated with higher complication rates, in- cluding bleeding, pneumothorax, and bronchospasm. Existing assessment tools provide structured evaluations, yet they remain heavily reliant on subjective expert judgment and limited sensory feedback. To address this limitation, we propose a model-based framework for objective performance evaluation in navigational bronchoscopy. Our approach leverages pose data from electromagnetic (EM) trackers, routinely used in clinical navigation, and embeds nonholonomic kinematic con- straints that characterize expert-like trajectories. Using the model and a Model Predictive Path Integral (MPPI) control, we generate optimal reference trajectories and define error metrics that quantify deviations between operator-executed and model- predicted motions. We hypothesize that these deviations provide robust discriminative features for distinguishing between expert and novice performance. Experiments on a phantom lung dataset comprising 11 operators and 98 procedures demonstrate that the proposed metrics significantly separate skill levels, enabling the construction of an effective classifier for operator proficiency. This framework offers an interpretable, data-driven alternative to supervisor-dependent assessments and represents a step toward scalable, objective skill evaluation and transfer in bronchoscopy training and robotic platforms.