JMIR Publications

ABSTRACT

Background:

Laparoscopic surgical performance relies on accurate visual perception and attention, yet visual misperception is a leading cause of error in laparoscopic cholecystectomy. Eye-tracking research has shown differences in visual strategies across expertise levels. However, there is a need to better understand how surgeons of varying expertise perceive anatomy and how indocyanine green (ICG) fluorescence influences attention and perception. This study uses a mixed-methods approach to examine gaze behavior and perceptual accuracy during surgical viewing.

Objective:

This study explores differences in visual perception and attention when viewing laparoscopic cholecystectomy procedures under white light and ICG fluorescence imaging conditions across surgeons of varying levels of expertise.

Methods:

Sixty-one surgeons (20 experts, 16 intermediates, 25 novices) participated in this study conducted at the 2025 Society of American Gastrointestinal and Endoscopic Surgeons Meeting. Surgeons watched a series of laparoscopic cholecystectomy video clips while their eye movements were recorded using a Tobii Pro Fusion device. The videos included both continuous viewing and paused scene conditions which required that they identify anatomical structures. Eye-tracking metrics and anatomical identification accuracy were analyzed across expertise levels and imaging conditions. Identification accuracy for each paused scene was scored as correct, suggestive, or incorrect.

Results:

Experts demonstrated higher anatomical identification accuracy, particularly in ICG fluorescence scenes with complex or abnormal anatomy. Significant associations between expertise level and anatomical identification accuracy were found for identifying the cystic artery (p=0.018) and duct (p=0.046) in the ICG scene. Level of expertise had a significant effect on average fixation (p=0.034), with experts exhibiting higher saccade counts than novices (p=0.032). Heatmaps visualizing the spatial distribution of fixations showed that differences in the spatial distribution of attention across expertise levels were more pronounced in mixed lighting conditions during dissection or confirmation of the CVS. Across imaging conditions, differences in gaze behavior were more evident in complex scenes, although differences in many eye-tracking metrics did not reach statistical significance.

Conclusions:

The integration of eye-tracking measures with perceptual accuracy scoring provided insight into how surgeons perceive and attend to visual cues across surgical scenes. Differences in perceptual accuracy and gaze patterns across levels of expertise were more pronounced in perceptually complex scenes under ICG fluorescence.