Combined Gaze Metrics as Stress-Sensitive Indicators of Microsurgical Proficiency

Abstract

Background. Evaluation of microsurgical proficiency is conventionally subjective, time consuming, and unreliable. Eye movement–based metrics have been promising not only in detection of surgical expertise but also in identifying actual cognitive stress and workload. We investigated if pupil dilations and blinks could be utilized in parallel to accurately classify microsurgical proficiency and its moderating features, especially task-related stress. Methods. Participants (n = 11) were divided into groups based on prior experience in microsurgery: novices (n = 6) with no experience and trained microsurgeons (n = 5). All participants conducted standardized suturing tasks with authentic instruments and a surgical microscope. A support vector machine classifier was used to classify features of microsurgical expertise based on percentage changes in pupil size. Results. A total of 109 successful sutures with 1090 segments were recorded. Classification of expertise from sutures achieved accuracies between 74.3% and 76.0%. Classification from individual segments based on these same features was not feasible. Conclusions. Combined gaze metrics are applicable for classifying surgical proficiency during a defined task. Pupil dilation is also sensitive to external stress factors; however, the usefulness of blinks is impaired by low blink rates. The results can be translated to surgical education to improve feedback and should be investigated individually in the context of actual performance and in real patient operations. Combined gaze metrics may be ultimately utilized to help microsurgeons monitor their performance and workload in real time—which may lead to prevention of errors.

Methods and Measures for Mental Stress Assessment in Surgery: A Systematic Review of 20 Years of Literature

“…Furthermore, experienced surgeons showed smaller changes in pupil dilation than less experienced operators [63]. Finally, average and SD of pupil size changes during suturing successfully classified surgical expertise while fatigue, caffeine, and illumination changes were possible reasons for no classification improvement from blink rate features [62].…”

“…Intraoperative stress across surgical roles and expertise. Seventeen papers (23.94%) compared surgical performance and stress markers between individuals from two expertise levels -novice versus expert or junior versus senior residents and surgeons [38], [62], [63], [66], [78], [86], [90], [96], [105], [109], [117], [121], or among surgical professionals with diverse professional backgrounds [6], [44], [64], [84], [118]. Six studies compared stress and technical performance when participants switched their roles between primary and assistant surgeons [4], [77], [86], [92], [108], [122], and two papers measured stress, anxiety, and workload from teams of surgeons, nurses, technicians, and assistants [66], [92].…”

“…The highest cortisol levels were observed 20-30 minutes after task completion in a high-stress scenario [114]. c) Ocular activity: Eye activities were recorded using surface EOG [98], eye-tracking glasses [38], [63], [106], [107], [115], [121], or eye trackers connected to microscope oculars [62]. Gaze and fixation metrics were computed from image frames using commercial or custom scripts [38], [107], [115], [121].…”

“…Synchronization of multiple sources such as physiological signals, motion data, speech, and operating scene videos with each other and with ongoing surgical phases is a critical process in data acquisition and processing [67]. Among collected studies, 16 papers (22.53%) clearly mentioned marking and annotating surgical phases and major time-locked events [6], [39], [62], [66], [77], [82], [86], [91], [96]- [99], [111], [121], [123], [125] and one study synchronized physiological data with an endoscopic video stream [97].…”

“…4) Baseline and Multi-Point Data Acquisition: Humans demonstrate large inter-subject variability in their sympathetic, parasympathetic, and cortical activations [5], [52], [62], [66], [77], [139], and it is beneficial to interpret intraoperative measurements with respect to individual baseline values to reliably identify the effects of stress and workload demands [106]. For example, surface EMG amplitudes can be normalized with respect to maximum amplitudes from voluntary contractions [89], [95].…”