Minimally Invasive Surgery (mis) Assessment Device: Variance Analysis

Sami Abusneineh, PhD, Brent Seales, PhD

University of Kentucky

The objective of this study is to analyze the variance of the skill level of MIS trainees using a multi-sensor camera system. We built multiple sensors system in order to extract a large number (more than 50) of assessment metrics and their relationships to the surrounding environment. The system is designed to synchronize non-invasive, real-time sensors and coordinate them over many cues (eyes, external shots of body and instruments, internal shots of operative field). The system analyzes the set of metrics and chooses the set of metrics that has high correlation with the skill level.
The designed system contains a number of camera sensors to capture the eyes, external snapshots of body and instruments, and internal shots of operative field synchronously, in addition to a heart rate monitor. Eight cameras are dedicated to tracking the surgery tools and the trainee’s arms and head; two cameras attached to the surgical display track the trainee’s eyes; a laparoscope camera is used to detect whether the tools are moving within the field of view or not; and a heartbeat rate monitor attached to the trainee’s body tracks the heart rate. Figure 1 shows the setup of the system including the sensors.

To validate the system we captured 55 metrics for 58 subjects in three different skill levels. The pegboard transfer task is used in the experiment. The captured metrics can be categorized into time, Kinematics, and stress and fatigue metrics. From the 55 metrics the system showed high correlation with16 metrics not including completion time and completion status. A variance analysis is applied on subjects from each skill level. The result showed that the variance of the 16 metrics in the novice class is larger than the intermediate. And the variance of the intermediate is larger than the expert. Figure 2 shows the 16 metrics and their variance for each level of experience.
In this research we studied the variance of the metrics produced by system we built that integrates multiple sensors to observe and extract data from the training environment. The variance of the metrics shows significant difference between the skill levels. The future work of this study is to perform multivariate data analysis on the data in order to detect skills patterns using the composite metrics.