Who profits from three-dimensional optics in endoscopic surgery? Analysis of manual tasks under two-dimensional/three-dimensional optic vision using a pelvic trainer model.

BACKGROUND: In endoscopic operations, direct binocular view, tissue sensation and depth perception get lost. It is still unclear whether the novel three-dimensional (3D) high-definition (HD) cameras are able to compensate the limited senses and how this affects the skill set of users with different endoscopic experience. This study aimed first to evaluate if the 3D technology improves depth perception, precision and space orientation as compared to conventional two-dimensional (2D) HD technology. The second aim was to determine the 3D influence on participants with different endoscopic experience.
METHODS: A total of 24 participants of different experience levels performed three different tasks on a pelvic trainer using the same thoracoscopic unit in 2D and 3D modes. Results were statistically analysed using Student's t-test and Pearson's product-moment correlation.
RESULTS: Across all the participants, we found that 3D optic vision significantly reduced the needed time to perform a defined difficult task in comparison to 2D. This difference was less pronounced in participants with higher experience level. Participants with eyeglasses performed slower in both 2D and 3D in comparison to participants with normal vision. Only participants with normal vision could significantly improve their completion times with 3D optic vision.
CONCLUSIONS: By testing the novel generation of 3D HD cameras, we could demonstrate that the 3D optic of these systems improves depth perception and space orientation for novices and experienced users and especially inexperienced users benefit from 3D optic.

INTRODUCTION

In contrast to open surgery, in endoscopic operations, the surgeon has to give up the direct binocular view, tissue sensation and depth perception.[1] Even if minimal invasive surgery has reduced complications and made smaller incisions possible, two-dimensional (2D) vision lacks visual depth, which may cause image distortion and may also lead to a decreased ability to estimate size.[2]

The development of new three-dimensional (3D) high-definition (HD) cameras compensates the above-mentioned disadvantages, and a reduction of errors, an increasing precision and shorter completion time of tasks were proofed in several studies analysing laparoscopic skills.[2345] All of the above-mentioned studies[45] evaluated 3D influence only within a defined group of users (students or experts). Therefore, the variation of how the third dimension affects the skills of users with different endoscopic experience was not assessable.

We hypothesised that the novel HD generation of 3D cameras improves the depth perception, precision and space orientation compared to the 2D technology even for systems offering the same HD image quality. Furthermore, we supposed that surgically inexperienced persons do profit significantly from the third dimension and are therefore able to perform tasks of different precisions in the same time and quality, as experienced surgeons.

SUBJECTS AND METHODS

Participants

The present prospective study was conducted at the Department of Orthopedics and Trauma Surgery at the University Hospital Bonn, Germany. A total of 24 participants of different occupations and functions performed three different tasks on a pelvic trainer using the thoracoscopic unit (Einstein vision 3D Full HD, Aesculap, Tuttlingen, Germany). To exclude a training effect, half of the participants began in 2D and the other half in 3D mode. After completion of the three tasks, the viewing mode was changed and the same three tasks were performed again. The same camera and monitor were used for both 2D and 3D modes and allows perception in full HD quality.

The three analysed tasks have not been validated before to prove the exclusion of all factors except 2D/3D on hand–eye coordination. Even if all tasks can be performed without any surgical experience, a certain bias cannot be excluded.

Epidemiologic details of participants are given in Table 1.

Table 1

Epidemiologic details of the participants

Age (years)	Sex	Function	Education (years)	Laterality	Experience	Vision
34.5 (range: 19-57)	Male (n=14)	5 students	6.55±5.46	Right (n=21)	<10 (n=11)	Normal (n=15)
	Female (n=10)	12 residents		Left (n=3)	10-100 (n=8)	Glasses (n=9)
		5 consultants			>100 (n=5)
		2 Chairmen

Tasks

The three tasks were performed with rising levels of difficulty [Figure 1]. Task 1: Three rubber bands had to be pulled off of a rack outfitted with six different screws [Figure 1a]. Task 2: All rubber bands had to be rearranged again in the same order [Figure 1b]. Task 3: Six drawing pins had to be placed head down on the screws without falling off. The task was completed if all five drawing pins were placed on the screws correctly [Figure 1c]. All tasks were performed in 2D and 3D vision and respective times were taken from start to finish.

Figure 1

Experimental setup of tested tasks with rising levels of difficulty (tasks 1–3). (a) Task 1: Three rubber bands had to be pulled off of a rack outfitted with six different screws. (b) Task 2: All rubber bands had to be rearranged again in the same order. (c) Task 3: Six drawing pins had to be placed head down on the screws without falling off. The task was completed if all five drawing pins were placed on the screws correctly

Statistics

All data were processed with GraphPad Prism 6 (GraphPad Inc., La Jolla, CA, USA). Comparisons between means were made using paired Student's t-tests with a significance level set at P ≤ 0.05 for task times. Data indicate means ± standard error. Graphics were produced using the same software package.

RESULTS

Analysis of completion times for each task irrespective of the participants’ surgical experience [Figure 2a–c] yielded a significant time improvement for all three tasks with the 3D optic (task 1: ***P = 0.0001; task 2: **P = 0.0029; task 3: *P = 0.0107).

Figure 2

All three tasks [as depicted in Figure 1] were performed in two-dimensional and three-dimensional vision and respective times were taken from start to finish. (a) Mean completion time of all participants for the two-dimensional and three-dimensional modes of task 1 (Rubber band pull of), (b) task 2 (Rubber band hoist) and (c) task 3 (Drawing pins)

With regard to the preexisting surgical experience, three groups of participants were separately analysed: Beginners (<10 endoscopies), intermediate (10–100 endoscopies) and experts (>100 endoscopies). All three groups improved their own completion times for all tasks using the 3D optic [Figure 3a–c].

Figure 3

Completion times were analysed according to the participants’ surgical experience. Three groups were generated: Beginners (<10 endoscopies), medium experienced (10–100 endoscopies) and experts (>100 endoscopies). All tasks were performed in two-dimensional and three-dimensional vision. (a) Mean completion time for the two-dimensional and three-dimensional mode of task 1 (Rubber band pull-off), (b) task 2 (Rubber band hoist) and (c) task 3 (Drawing pins)

The individual time improvement for each group and task are given in Table 2.

Table 2

Analysis of completion times for tasks 1-3 in two-dimensional and three-dimensional vision

	Mean time (s)	SE	95% CI of mean	3D/2D ratio
Task 1
2D – beginner	36.1	4.1	26.5-45.8	0.7
3D – beginner	25.4	2.1	20.3-30.4
2D – intermediate	30.7	3.1	23.1-38.3	0.8
3D – intermediate	24.6	1.7	20.3-28.8
2D – expert	30.5	4.0	20.2-40.8	0.67
3D – expert	20.5	2.6	13.9-27.1
Task 2
2D – beginner	76.3	12.1	47.7-104.8	0.5
3D – beginner	38.0	4.3	27.9-48.1
2D – intermediate	45.7	3.0	38.5-53.0	0.81
3D – intermediate	37.1	3.1	29.6-44.7
2D – expert	39.8	4.3	28.7-51.0	0.82
3D – expert	32.5	3.6	23.3-41.7
Task 3
2D – beginner	149.5	21.1	99.6-199.4	0.92
3D – beginner	138.1	25.4	78.2-198.1
2D – intermediate	159.6	20.0	110.7-208.4	0.66
3D – intermediate	105.6	6.4	89.8-121.3
2D – expert	118.8	29.3	43.6-194.0	0.74
3D – expert	88.2	21.4	33.2-143.1

2D: Two-dimensional, 3D: Three-dimensional, SE: Standard error, CI: Confidence interval

The mean completion times as well as standard error and confidence interval of mean are determined for each level of experience. The greatest time improvement with the lowest 3D/2D ratio could be noted for beginners performing task 2 (0.5).

For beginners, a ssignificant time improvement with use of the 3D optic was measured for task 1 (3D/2D ratio = 0.7, **P = 0.0032) and task 2 (3D/2D ratio = 0.5, *P = 0.0158). The time improvement for task 3 was not statistically significant (3D/2D ratio = 0.92, P = 0.63). The greatest improvement was measured for task 2 (3D/2D ratio = 0.5).

Intermediates improved their completion times significantly in all the three tasks using the 3D optic (task 1: 3D/2D ratio = 0.8, *P = 0.0325; task 2: 3D/2D ratio = 0.81, *P = 0.0375; task 3: 3D/2D ratio = 0.66, *P = 0.0176). The greatest improvement was measured for task 3 (3D/2D ratio = 0.66).

Laparoscopic experts improved their own completion times significantly for each of the three tasks using the 3D optic (task 1: 3D/2D ratio = 0.67, *P = 0.0385; task 2: 3D/2D ratio = 0.82, **P = 0.0041; task 3: 3D/2D ratio = 0.74, *P = 0.0499).

Experts performed faster than beginners and intermediate experienced participants. The shortest completion time for each task was determined in the expert group both with the 2D optic (task 1: 30.5 s; task 2: 39.8 s; task 3: 118.8 s) and with the 3D optic (task 1: 20.5 s; task 2: 32.5 s; task 3: 88.2 s).

The order of completion times (experts

Eyeglasses

For task 1 (rubber band pull-off), no correlations for participants with or without glasses could be noted. For task 2 (rubber band hoist), a positive correlation could be noted between 2D vision and eyeglasses (0.5319; P = 0.0280) as well as 3D vision and eyeglasses (0.7189; P = 0.0011). Participants with corrective lenses performed slower than those without for the 2D and 3D vision part of the tasks. Also for task 3 (drawing pins), a positive correlation could be noted between 2D vision and eyeglasses (0.5563; P = 0.0204) as well as 3D vision and eyeglasses (0.6239; P = 0.0074). Participants with eyeglasses performed slower than those without using either 2D or 3D vision.

Contingency analysis of normal vision versus glasses in relation to 2D and 3D vision revealed a significant improvement from 2D to 3D for participants with normal vision (P = 0.0023) as compared to those with glasses (n.s. P =0.0854) [Table 3].

Table 3

Contingency analysis of normal vision versus glasses in relation to two-dimensional and three-dimensional vision

	2D vision	3D vision	3D/2D ratio
Normal	191.4 (Sec)	135.5 (Sec)*	0.7*
Glasses	224.0 (Sec)	173.4 (Sec)	0.77 (NS)

NS: Not significant, Sec: Seconds, *: Significant, 2D: Two-dimensional, 3D: Three-dimensional

DISCUSSION

The advance of 3D cameras has been evaluated in a few studies, and a greater completion rate, shorter completion time and a shorter learning curve could be noted.[45] Obviously, possible advantages of the 3D system are dependent on the examined group. Surgeons with multiple years of endoscopic experience are used to the limited view and compensate for lack of depth perception in a better way than students or inexperienced users. Chiu et al. compared novices and experienced participants to show effects of groups with different experience.[6] Certainly, only one basic task was examined to demonstrate the influence of 3D vision. The authors summarised that the time difference to the experienced group could be significantly narrowed with the use of a 3D system. Surprisingly, completion times between 2D and 3D were not significantly reduced.

In our opinion, a possible advantage of the 3D technology cannot be evaluated by comparing one basic task between different experienced groups. Tasks with increasing difficulty require a higher precision and may emphasise the real effect of the 3D technology for groups of different experience levels. Another methodological requirement for comparability is that 2D and 3D cameras and monitors with the same perception should be used. Otherwise, the higher precision may not be explained by the binocular view alone.

Task setting

The most important result of this study is significantly reduced completion times with the use of 3D optics for all the three tasks. The greatest time improvement considering completion times of all participants was noted for task 2 (medium difficult).

Completion times for basic tasks (task 1) were also significantly shorter but not as remarkable. This result is comparable to the findings of Chiu et al.[6] They described slightly faster completion times for the tested basic task (peg transfer) with the use of 3D optics as compared to conventional 2D optics. In contrast to our study, the reported difference in their study was not statistically significant, neither for novices nor for experienced users.

A closer look to the completion times of different experienced participants yielded interesting results. Besides the above-mentioned positive effect for all groups, especially beginners profited from the 3D optic during medium difficult tasks (task 2). Using the 3D optic, beginners improved their own completion times significantly. While experts completed task 2 significantly faster than beginners in 2D vision, with the use of 3D vision, no significant time difference could be noted anymore. With the 3D optic, beginners reached completion times of experts and obviously profited more from the third dimension while performing task 2 [Figure 3b]. We therefore agree with Chiu et al. that the 3D system has a positive influence on performance for novice trainees and may considerably shorten the initial steep learning curve.[6] Limitations of the study by Chiu et al. are that only one basic task was tested and different effects of the 3D optics, like the point accuracy for tasks of higher difficulty, could not be assessed.[6]

Tanagho et al. investigated the influence of 3D optics with different tasks of the laparoscopic surgery (Fundamentals of Laparoscopic Surgery) skill set under individuals with varying laparoscopic experience.[7] According to our own data, all participants (irrespective of the surgical experience) improved significantly for all the three tasks with the use of 3D optics. Tanagho et al. also reported a significant time reduction under 3D vision, but looking at times of the different experienced groups, surprisingly 7 out of 33 still needed multiple attempts to complete one of the tasks.[7] It can be assumed that the tasks were not particularly suited to examine the practical effects of the third dimension. Furthermore, preexisting surgical experience may have played an important role (e.g., suturing and knotting) to complete the difficult exercises. To avoid preexisting surgical experience as an additional bias, in the present study, we examined tasks of different difficulty, which are feasible without any surgical pre-experience. Interestingly, our results are comparable with the results of Tanagho et al.[7] For the most difficult task (task 3), only intermediate and experts improved their completion times significantly [Figure 3c].

In a randomised controlled prospective trial by Alaraimi et al., the median completion time for the 3D group was not significantly faster, but the corresponding median numbers of repetitions and errors were significantly lower which may advocate for a better safety and efficacy.[4] Surprisingly, the 3D group needed more time to perform the basic peg transfer task, which was interpreted as possible adapting problems to the 3D system. In the present study, we cannot confirm any significant adapting problems, and only three participants complained about adapting problems for a few seconds. We also did not notice any significant nausea or dizziness. Interestingly, almost 50% of participants of Alaraimi et al. reported a significant tiredness in the 2D group.[4] We and Feng et al. cannot confirm these results with our data.[8] The authors concluded that 3D vision improves surgery by reducing errors and facilitating patient safety. Improved accuracy in laparoscopic surgery is, however, more pronounced for advanced tasks. Recent studies comparing 2D and 3D vision as well as our own results support this observation.[91011]

The different effects of the 3D vision tasks on participants with and without eyeglasses have not been reported yet. Interestingly, participants without eyeglasses were significantly faster in challenging tasks (tasks 2 and 3) than users with glasses. This result was determined with the 2D and 3D vision. In fact, even participants with glasses improved their own completion times and performed faster in 3D vision than in 2D, but this result was not statistically significant. Participants with normal eyesight profited significantly more from the third dimension and enlarged their time advance. A possible explanation is a disturbing effect of eyeglasses becoming relevant in challenging tasks.

With these results, the hypothesis of a better depth perception, precision and space orientation under 3D optic can be confirmed. In addition, the results support the hypothesis that inexperienced users do profit significantly from the visible third dimension and are therefore able to complete medium difficult tasks in times comparable to those of experts. On the other hand, intermediate and experts do benefit significantly from the 3D optic also for tasks of high difficulty.

Limitations of the present study are the non-clinical setting and the small number of participants. As our presented model system is an ex vivo test, there are clear limitations of how these results can be transferred into an in vivo situation (e.g., endoscopic operations). Experts in laparoscopic surgery are used to the 2D view and, by moving the scope proximal and distal, they subliminally recognise changing surfaces (different shadows and accumulation of fluids) which help the brain to create a 3D impression and judge about depth perception. This compensating mechanism does not work in an ex vivo setting, and therefore, experts can struggle to perform tasks in this artificial setup. To analyse the true effect of optics on operations, timings of the same type of surgeries by an expert surgeon using 2D and 3D optical systems should be compared.

CONCLUSION

By testing the novel generation of 3D HD cameras, we could demonstrate that the 3D optic of these systems improves depth perception and space orientation for novices and experienced users and especially inexperienced users benefit from 3D optic.

Financial support and sponsorship

None.

Conflicts of interest

There are no conflicts of interest.