Abstract
Abstract
Background:
Three-dimensional (3D) laparoscopes have been developed to maintain the perception of depth in the operating field. Two-dimensional (2D) imaging relies on tactile feedback, visual cues, and previous experience. The aim of this study was to test if 3D laparoscopic vision is superior to 2D laparoscopic vision in training novice operators in completing set laparoscopic tasks.
Materials and Methods:
The study population consisted of 20 interns and medical students. The participants were randomized to completing tasks using a 2D or 3D system. These included pegboard transfer (PT), continuous suturing (CS), and intracorporeal knot-tying (IK). The time to complete the task and number of errors made were recorded.
Results:
Following adjustment for potential confounders, time to complete CS and IK was significantly longer among participants who used the 2D laparoscope compared with those who used the 3D laparoscope (CS, P<.0001; IK, P<.0001). This same effect was not demonstrated in time to perform PT (PT, P=.04). The 2D laparoscope was associated with a significant increase in the number of errors on the IK task (P<.0001) but not on the PT or CS tasks (PT, P=.35; CS, P=.26).
Conclusions:
The 3D system assists novice operators perform more complex laparoscopic tasks in a decreased amount of time and with fewer errors.
Introduction
L
Materials and Methods
Study population
Ethics approval for this study was obtained from the Hospital Research and Ethics Committee of The Queen Elizabeth Hospital, Adelaide, SA, Australia. Informed consent was obtained from participants. In total, 22 interns and medical students were recruited to the study. Participants were included if they had no previous laparoscopic experience or simulated training. Two participants were excluded because of previous laparoscopic experience.
Experimental design
The remaining 20 participants were randomly assigned to either the 2D or 3D group via a computer-generated randomization schedule. Sex, handedness, use of eyeglasses, and hours spent playing computer games per week were recorded. Practice was not allowed before the evaluation. A crossover study was not used as participant performance could be confounded owing to experience gained by moving between imaging systems.
In this study we used the four-step framework of Walker and Peyton 10 for teaching procedural skills. An instructional DVD was shown outlining the tasks required of the participant. The task was demonstrated to the participant and deconstructed into discrete steps. Comprehension was checked by the observer, and the tasks were performed under supervision. Feedback was given at the completion of each individual repetition of the task. The laparoscopic tasks are described in Table 1. Each task was completed inside a laparoscopic trainer. Participants repeated each individual tasks four times. The pegboard transfer (PT) was performed first, followed by continuous suturing (CS), and the intracorporeal knot-tying (IK) was completed last. Time to complete the task in seconds and the number of errors made were recorded. Participants were monitored during completion of all tasks, and there were no reports of nausea or headaches using the 2D or 3D system.
Imaging systems
The 3D group used an Olympus 10-mm EndoEYE Flex 3D high-definition laparoscope with the Exera III imaging platform and 3D visualization unit (Olympus Corp., Tokyo, Japan). The participants used passive 3D glasses. The 2D group used a Fundamentals of Laparoscopic Surgery standard color camera with 4.3-mm focus lens.
Statistical analysis
All analyses were completed using SAS version 9.3 software (SAS Institute Inc, Cary, NC). Time to completion of each task was treated as a continuous measure and analyzed using a linear mixed-effects model. Log transformations were applied to the data in order to meet the distributional assumptions of linear models. The data were transformed back to the original scale prior to reporting. Error data were treated as count measures and analyzed using generalized linear mixed-effects models with a Poisson distribution and a log link function. Subject was treated as a random factor in all models. The estimates provided by all models represent the ratio of the mean (time or errors) in the 2D group relative to the corresponding mean in the 3D group. Group differences were presented as unadjusted (univariate) effects, and effects were adjusted for potential confounders (sex, glasses, and attempt). All tests were two-tailed, and statistical significance was assessed at the 5% alpha level.
Results
All 20 participants recruited into the study completed all laparoscopic tasks. In total, 240 tasks were performed during the study. The sample consisted of 11 males and 9 females. Nineteen subjects were right-handed, and no subject reported prior experience. When compared with the 3D group, the 2D group had a higher proportion of males and a lower proportion of subjects who wore glasses. However, group differences were not statistically significant in either case (sex, P=.37; glasses, P=.65). Data concerning computer game use were collected; however, we observed that all users were male. As a result, this measure was not included in any further analyses.
The 2D laparoscope was associated with significantly longer mean times to completion on the CS and IK tasks (Table 2). Following adjustment for potential confounders (sex, glasses, and attempt), the mean time to completion of the CS task was 1.58 times longer among subjects using the 2D scope compared with those using the 3D scope (95% confidence interval, 1.28–1.96; P<.0001); mean time to completion of the IK task was 1.74 times higher in the 2D group (95% confidence interval, 1.48–2.05; P<.0001). There were no differences between the 2D and 3D groups with respect to PT time in the unadjusted model; however, mean time to completion was lower among subjects in the 2D group compared with those in the 3D group (mean=0.73; 95% confidence interval, 0.54–0.99; P=.04).
Adjusted for sex, glasses, and attempt.
2D, two-dimensional; 3D, three-dimensional; CS, continuous suturing; IK, intracorporeal knot-tying; PT, pegboard transfer.
The mean number of errors on the IK task was significantly higher among subjects using the 2D scope compared with those using the 3D scope (Table 3). This amounted to a 2.62-fold increase in the mean number of errors in the 2D group following adjustment for confounders (95% confidence interval, 1.90–3.62; P<.0001). There were no significant differences between groups with respect to mean errors during performance of the PT and CS tasks (P=.35 and P=.26, respectively).
Adjusted for sex, glasses, and attempt.
2D, two-dimensional; 3D, three-dimensional; CS, continuous suturing; IK, intracorporeal knot-tying; PT, pegboard transfer.
Discussion
Laparoscopic surgical training traditionally uses 2D imaging systems for learning surgical skills. There are limited numbers of studies examining whether 3D systems are superior to 2D systems in obtaining surgical skills, and they have contradictory results. The literature contains varying levels of laparoscopic experience, differing types of laparoscopic tasks tested, and attempted randomization and statistical analysis to compensate for the learning effect. The 2D and 3D technology has evolved over time with improvements in laparoscopic systems from the original first-generation equipment described in the literature. There has also been development of high-definition screens and passive eyewear replacing active eyewear.
Literature supporting the benefits of 3D has demonstrated that 2D endoscopic vision impairs performance over direct visualization and that 3D reduces this effect by 41%–53% in novices as well as experienced surgeons. 3 Storz et al. 4 reported fewer errors and decreased completion time in four out of five laparoscopic tasks in 3D versus 2D; the results were obtained in students as well as experienced surgeons. Peitgen et al. 5 demonstrated completion time using 3D decreased 24.4% and the number of mistakes decreased 52.5% compared with 2D regardless of previous laparoscopic experience. Tanagho et al. 6 reported greater speed in peg transfer, pattern cutting, and suturing/knot-tying independent of the participant's level of technical expertise.
Conversely, there is literature suggesting that it is laparoscopic experience rather than imaging systems that influence performance in performing laparoscopic tasks. 7 Jones et al. 8 demonstrated no significant advantage of 3D over 2D for the novice or expert surgeon in performing five different laparoscopic tasks. Volantopoulos et al. 9 showed inexperienced subjects required longer time to complete three of six tasks in 2D compared with those using 3 D. These included ring board exchange, duct cannulation, and instrument spatial navigation. The number of errors was also higher for rope passing, instrument spatial navigation, and suturing. For experienced participants there was no difference in performance.
In our study, after adjustment for potential confounders, time to complete CS and IK was significantly longer among participants who used the 2D laparoscope compared with those who used the 3D laparoscope (CS, P<.0001; IK, P<.0001). The same effect was not demonstrated in time to perform PT (PT, P=.04). The 2D laparoscope was associated with a significant increase in the number of errors on the IK task (P<.0001) but not on the PT or CS tasks (PT, P=.35; CS, P=.26). These results could be explained by the 3D laparoscope maintaining depth perception for the novice operator. Performing IK is challenging and involves several more complex maneuvers compared with PT and CS. Using number of errors as a marker for performance, the completion of the more complex IK task was improved by the 3D laparoscope.
Our study details the steps taken to demonstrate and instruct the tasks to the participants. This has not been detailed in previous literature on the topic. This was based on Cognitive Load Theory, which addresses the limitations of working memory when learning new tasks.11–13 This limited working memory can process no more than five to nine information elements and actively process no more than two to four simultaneously. A surgical learning curve is similar to the human performance curve with respect to motor skills, with rapid improvement during early experience. 14 Learning laparoscopic skills in the simulated or low-fidelity environment can address the limitations of working memory by managing the intrinsic nature of the learning task before moving to the operating room or high-fidelity environment. 15 Another advantage of this study is that potential confounding of results by a learning effect was avoided by not performing a crossover study and adjusting for attempt in the statistical analysis.
In conclusion, the 3D system assists novice operators perform more complex laparoscopic tasks in a decreased amount of time and with fewer errors. This technology could be applied to training novices for laparoscopic skills in a decreased amount of time and preparing them for the operating room.
Footnotes
Acknowledgments
We thank Ms. Julie Gray from Olympus Australia for allowing us to use the Olympus EndoEYE Flex 3D high-definition laparoscope with the Exera III imaging platform and 3D visualization unit used in this study.
Disclosure Statement
No competing financial interests exist.