Development of a virtual reality clinically oriented temporal bone anatomy module with randomised control study of three-dimensional display technology

Abstract

Objective

To investigate the effectiveness of a virtual reality (VR), three-dimensional (3D) clinically orientated temporal bone anatomy module, including an assessment of different display technologies.

Methods

A clinically orientated, procedural and interactive anatomy module was generated from a micro-CT of a cadaveric temporal bone. The module was given in three different display technologies; 2D, 3D with monoscopic vision, and 3D with stereoscopic vision. A randomised control trial assessed the knowledge acquisition and attitudes of 47 medical students though a pretutorial and post-tutorial questionnaire. The questionnaire included questions identifying anatomic structures as well as understanding structural relations and clinical relevance. Furthermore, a five-point Likert scale assessed the students’ attitudes to the module and alternative learning outcomes, such as interest in otology and preparedness for clinical rotations.

Results

As a whole cohort, the total test score improved significantly, with a large effect size (p≤0.005, Cohen’s d=1.41). The 23 students who returned the retention questionnaire had a significant improvement in total test score compared with their pretutorial score, with a large effect size (p≤0.005, Cohen’s d=0.83). Display technology did not influence the majority of learning outcomes, with the exception of 3D technologies, showing a significantly improvement in understanding of clinical relevance and structural relations (p=0.034). Students preferred 3D technologies for ease of use, perceived effectiveness and willingness to use again.

Conclusions

The developed VR temporal bone anatomy tutor was an effective self-directed education tool. 3D technology remains valuable in facilitating spatial learning and superior user satisfaction.

Introduction

Anatomy training in medical schools has suffered the pressures of reduced training hours^1–4 leading to a decline in the time allocated to acquiring a breadth of anatomical knowledge.^5–7 In order to effectively use the available time, alternatives to traditional teaching techniques such as three-dimensional (3D) visual technologies are increasingly being used. These technologies have been shown to benefit anatomy education, particularly in the reinforcement of spatial knowledge, and have reported high user satisfaction.^8–10 However, in a review of anatomy education in Australian and New Zealand medical schools, Craig et al remarked on the under-representation of 3D technologies.⁴ Additionally, in medical education, clinically oriented anatomy education has been shown to improve knowledge retention,¹¹ motivation¹² and students’ interest in pursuing a surgical career.¹³ Unfortunately, in the classroom environment, clinical integration of anatomy education can be resource intense^{13 14} and logistically challenging.¹⁵

Knowledge of the 3D complex anatomy of the temporal bone is essential in understanding the pathophysiology and management of otological disorders and forms the basis of therapeutic surgical procedures.^{16 17} In addition, a working knowledge of ear anatomy is also important to general practitioners and emergency physicians, where ear disease is often a presenting complaint.^{18 19} Although there have been several virtual reality (VR) simulators developed to train otology surgery,²⁰ there has been little work on VR simulation to facilitate learning of temporal bone anatomy. For example, Nicholson et al developed a 3D inner ear anatomy module from MRI images that could be manipulated through a web browser.²¹ In a randomised control study, they found that students learning in this 3D environment performed better on summative anatomical assessment than students who learnt from 2D images and text. Fang et al evaluated the use of an interactive 3D VR temporal bone simulator as an anatomy teaching aid.²² The seven medical students who participated reported an increased understanding of temporal bone structures, and a heightened interest in otology.

This limited evidence base suggests that 3D modules may assist training in inner ear and temporal bone anatomy. Fang et al’s study highlights the potential for VR training to foster an interest in otology and to benefit procedural style learning. This is echoed in Sugand et al, whose review of anatomy teaching emphasised that procedural and clinical anatomy are core components of an optimised approach to anatomy education.²³ In a review of 3D technologies for anatomy education, Hackett and Proctor noted that despite the beneficial results of 3D anatomy education technologies, there is a lack of literature regarding the retention of knowledge and transfer to clinical skills.²⁴ In addition, the comparison of 3D modalities (in particular monoscopic 3D vs stereoscopic 3D) for content delivery is unexplored. The aim of this study was to bridge these gaps by investigating whether a VR 3D clinically oriented, procedural and interactive temporal bone anatomy module would lead to improved knowledge, retention of knowledge and increased clinical understanding. In addition, this study compared the module’s effectiveness when delivered in stereoscopic 3D and monoscopic 3D displays.

Methods

Setting

This study used the University of Melbourne VR temporal bone surgery simulator. The experiments were conducted in a quiet simulator centre (figure 1). The simulator displayed a virtual model of a 3D cadaveric human temporal bone reconstructed from manual segmentation of a micro-CT scan. A haptic device (Sensable Phantom Omni) provided tactile interaction and feedback, and was represented in the simulation as a surgical drill. Stereoscopic 3D visualisation (providing depth perception) was rendered with NVIDIA 3D vision glasses. A MIDI device (musical instrument digital interface) controlled the zoom and drill burr size. Face validity of this simulator had previously been established.^25–27

Figure 1.

The University of Melbourne virtual reality temporal bone simulator setup.

The surgery simulator was adapted to create an anatomy module, for the purpose of anatomy education, by segmenting anatomically relevant structures and colour-highlighting each structure within the temporal bone (figure 2, video demonstration provided in online suppplemental appendix video). The structures segmented were: mastoid, tegmen tympani, dura, sigmoid sinus, spine of Henle, aditus ad antrum, facial nerve, chorda tympani, malleus, incus, stapes, cochlea, round window, oval window, stylomastoid foramen, digastric fossa, facial recess, sinus tympani, promontory, tympanic bone and scutum. In addition, a prerecorded audio was embedded to the simulator explaining the clinically oriented anatomy relevant to each structure. The content of this module was developed by a doctor in specialist training with anatomy education experience, overseen by a consultant otologist. When drilling a highlighted structure during simulation, the name of the structure was displayed on the screen and the audio accompaniment was triggered (online supplemental appendix I). The student was expected to drill away each anatomical point of interest to reveal deeper structures. In addition, the entire temporal bone could be made transparent which permitted visualisation of the deep and membranous structures (eg, ossicles, cochlea and semicircular canals).

Figure 2.

Screenshots of the anatomy simulator showing the highlighted virtual bone (right) and transparentbone (left). The drill shows varying burr sizes. A cursor that is used to manipulate the bone angle. Coloured highlighted bone displays anatomic structure name when it is drilled. In the transparent bone mode, anatomic structure name is displayed when the burr contacts the structure. With the display of the anatomic structure name, the corresponding audio explanation is triggered.

The development of the VR module was guided by the components of ‘an optimal anatomy curriculum’, as proposed by Sugand et al for a modern and compact learning environment.²³ The system components as discussed below were introduced to ensure that this was an independent and self-guided learning environment.

Dissection/prosection

Sugand et al identified that dissection remains a valuable tool in aiding an understanding of anatomy in 3D. Dissection also fosters reflective education through curiosity and self-exploration.^{28 29} To this end, the anatomy module developed relied on student dissection of the VR temporal bone in order to expose clinically relevant anatomic structures. The automatically triggered audio was used in lieu of an anatomy demonstrator.

Interactive multimedia

The use of interactive multimedia programmes for revision of anatomy has been shown to positively affect examination scores.³⁰ It has been suggested that visual aids should be considered compulsory for anatomy teaching, and that virtual dissection should be offered to all students.²³ The VR temporal bone simulator presents clinically oriented anatomy education in an interactive model that students can access via either a worksheet outlining a step-by-step dissection or through self-guided exploration. The simulator could be rotated in 3D virtual space and a ‘fresh’ temporal bone could be generated after virtual dissection, with or without the highlighting of the anatomy.

Procedural and clinical anatomy

Integrating the clinical relevance of anatomy into education has resulted in effective and engaging anatomy courses that are well received by medical students.¹⁴ It is, therefore, understandable that there has been a push toward clinically relevant anatomy education in modern medical curricula.¹⁷ This VR training module integrated clinically relevant information into the audio presented, and defined specific anatomic structures as surgically relevant. A worksheet was given which allowed students to follow a surgical-style approach to the middle ear (online supplemental appendix II). For example, the participant was directed to drill away the mastoid bone and proceed to explore the relation of the facial nerve within the mastoid bone prior to exposing the middle ear.

Study design

This randomised control trial follows an initial pilot study.³¹ In the pilot study, participants were randomised to the aforementioned stereoscopic VR teaching module (S3D) or to a 2D, non-interactive version which presented the same content (PPT). The 2D version was developed by generating screenshots of the simulator at each step of the worksheet. These were displayed on a Microsoft Powerpoint presentation (V.16.18), with relevant audio accompaniment. In the current study, to further evaluate the effects of differing display technologies, participants were block randomised using Qualtrics software (Provo, UT) into one of three groups. Two of the groups were the same as that of the pilot study: S3D and PPT. The third group was allocated to the monoscopic 3D display (M3D), where the participants did not use the NVIDIA 3D glasses and thus had no depth perception. In this mode, the temporal bone appeared in 3D on a computer screen and the haptic device was still used for interaction.

An initial pretutorial questionnaire was completed by all participants (figure 3). The questionnaire comprised demographic questions, followed by questions related to the participants’ self-assessed ability in temporal bone anatomy and clinical otolaryngology based on a five-point Likert scale. Lastly, the questionnaire included 13 quiz questions regarding temporal bone anatomy (including spatial relationships and clinically orientated anatomy) (online supplemental appendix III). The participants then proceeded to the training module. All participants were given orientation to the lie of the temporal bone (eg, anterior/posterior borders). The S3D and M3D groups were orientated to the use of the haptic device and virtual environment with a 5 min explanation of the components of the simulator followed by 5 min’ practice time on a simulated temporal bone. All participants were given a transcript of the embedded audio and allowed to make notes. After proceeding through the worksheet/PPT presentation, the module was reset, and the participants were allowed to redo it in part or in full. A time limit of 2 hours was set, and no additional feedback, prompting or debriefing was given during or after the module. Directly after the training module was completed, participants completed a post-tutorial questionnaire. This included five-point Likert scale type questions similar to the initial survey, as well as 10 temporal bone anatomy quiz questions (including spatial relationships and clinically orientated anatomy) (online supplemental appendix III). As there are no validated temporal bone anatomical questionnaires, the initial and postmodule anatomy questions were derived from analysis of those included the pilot study. The 50 pilot questions were analysed for discrimination and difficulty indices. Question items with Difficulty Indices of 0.4–0.6 (acceptable)^{32 33} and Discrimination Indices of 0.4–0.6 (good-excellent discrimination)^{32 33} were included. The pretutorial and post-tutorial temporal bone anatomy quiz questions were similar in theme but different in content. All participants were emailed a retention survey at 6 weeks after training which included a repeat of all 23 anatomy questions previously presented. All quiz questionnaires were distributed using Qualitrics software, with the participant’s name blinded to the reviewer while analysing results. The participants were recruited and followed up between October 2018 and April 2019.

Figure 3.

Overview of the methods. M3D, monoscopic 3D group; PPT, presented the same content; S3D, stereoscopic 3D group.

Statistical analysis

Stata/SE V.13.0 was used for all analyses.³⁴ Baseline demographic characteristics were analysed using one-way analysis of variance (age, medical school year, mastoid experience, VR experience, and gaming experience) and χ² test (gender and handedness) (table 1). The baseline five-point Likert scale questionnaire was analysed using χ² tests after the answers were condensed (agree and strongly agree allocated to a single category of response) (table 2). The Likert type questions in the post-tutorial questionnaire were similarly condensed and analysed using χ² tests. The group as a whole was compared with their pretutorial Likert questionnaire answers using paired t-tests, with effect sizes analysed using Cohen’s d. A Cohen’s d result of 0.2 was considered a small effect size, 0.5 medium and 0.8 large.³⁵ The total scores in the post-tutorial and retention tests were compared with the pretutorial scores using paired t tests and effect sizes were analysed using Cohen’s d (table 3). The post-tutorial questionnaire results were analysed by group, comparing pretutorial scores to an adjusted post-tutorial test score with analysis of covariance (ANCOVA) regression. Similarly, the adjusted results of the retention test questionnaire were analysed by group, comparing to pretutorial scores with ANCOVA regression (table 4). Sub analyses were performed with pairwise comparisons.

Table 1.

Baseline participant characteristics

Baseline characteristics		PPT	M3D	S3D	P value
Age	Mean (SD)	22.69 (2.36)	23.94 (4.77)	22.4 (0.99)	0.36
Gender	Female n (%)	9 (56.25)	7 (43.75)	6 (40)	0.63
MD year	Mean (SD)	1.63 (0.72)	1.63 (.5)	1.53 (0.64)	0.9
Mastoid experience (no)	Mean (SD)	0 (0)	0 (0)	0 (0)
VR experience (no)	Mean (SD)	0 (0)	0.06 (0.25)	0 (0)	0.39
Gaming experience (subjective)	Mean (SD)	1.75 (0.86)	1.5 (0.63)	1.67 (0.61)	0.6
Handed	Right n (%)	15 (93.75)	14 (87.5)	15 (100)	0.36

Age, medical student year (MD year), previous mastoid experience (previousobservation of mastoid surgery), previous VR experience (previous use of a VR temporal bone simulator), gamingexperience (subjective experience; none, casual, skilled or extreme), handedness.

Mean and SDs given with one-way ANOVA p value for age, MD year, mastoid experience, VR experience and gaming experience. Number of respondents and percentage given with χ² pvalues for gender and handedness.

ANOVA, analysis of variance; M3D, monoscopic 3D group; PPT, 2D group; S3D, stereoscopic 3D group; VR, virtual reality.

Table 2.

Baseline education experience

Baseline educational experience	Agree or strongly agree N (%)			P value
	PPT	M3D	S3D
I feel confident in my knowledge of the temporal bone and inner ear anatomy	1 (6.25)	1 (6.25)	4 (26.67)	0.15
I understand the temporal bone and inner ear as a 3D structure	5 (31.25)	8 (50)	7 (46.67)	0.52
I prefer to learn anatomy alone	3 (31.25)	10 (62.5)	8 (53.33)	0.192
I prefer to learn anatomy in a group	7 (43.75)	2 (12.5)	8 (53.33)	0.045
I find textbook revision an effective way to learn temporal bone and inner ear anatomy	8 (50)	7 (43.75)	5 (33.33)	0.64
I find didactic lectures an effective way to learn temporal bone and inner ear anatomy	8 (5)	4 (25)	6 (40)	0.34
I feel prepared for an ENT rotation	2 (12.5)	0 (0)	1 (6.67)	0.35
I understand the possible complications of ear surgery	3 (31.25)	3 (31.25)	1 (6.67)	0.55
I feel prepared to perform an ear examination in ED or GP	3 (31.25)	2 (12.5)	2 (13.33)	0.87
I am interested in a career in ENT	6 (37.5)	6 (37.5)	4 (26.67)	0.77

P values of χ² test.

Number and percentage of participants who responded agree or strongly agree to the given statements.

P value significance <0.05 (bold)

3D, three dimensional; ENT, ear, nose and throat; GP, general practitioner; M3D, monoscopic 3D group; PPT, presented the same content; S3D, stereoscopic 3D group.

Table 3.

Summary of post-tutorial questionnaire results and comparison to pretutorial results

Post-tutorial questionnaire	Agree or strongly agree N (%)			P value	Whole group comparison to pre-tutorial (paired t-test)				M diff	95% CI	Cohen’s d
	PPT	M3D	S3D		M pre (SD)	M post (SD)	t	P value
Confident knowledge PTB anat	8 (50)	7 (43.75)	11 (73.33)	0.221	0.13 (0.34)	0.55 (0.5)	5.39	<0.005	0.43	0.27 to 0.58	0.99
Understand 3D PTB	8 (50)	11 (68.75)	14 (93.33)	0.031	0.43 (0.50)	0.70 (0.46)	3.51	0.001	0.28	0.12 to 0.44	0.57
Prefer module alone	4 (25)	8 (50)	7 (46.67)	0.296
Prefer module group	8 (50)	8 (50)	10 (66.67)	0.563
Textbooks effective	8 (50)	5 (31.25)	8 (53.33)	0.406	0.43 (0.50)	0.45 (0.50)	0.33	0.743	0.02	−0.11 to 0.15	0.04
Lectures effective	8 (50)	4 (25)	7 (46.67)	0.296	0.38 (0.49)	0.40 (0.50)	0.33	0.743	0.02	−0.11 to 0.15	0.04
Prepared ENT rotation	4 (25)	4 (25)	5 (33.33)	0.838	0.06 (0.25)	0.28 (0.45)	2.87	0.006	0.21	0.06 to 0.36	0.58
Understand complicaitons	14 (87.5)	10 (62.5)	11 (73.33)	0.267	0.15 (0.36)	0.74 (0.44)	8.23	<0.005	0.60	0.45 to 0.74	1.48
Prepared for ENT examination	5 (31.25)	5 (31.25)	6 (40)	0.840	0.15 (0.36)	0.34 (0.48)	2.87	0.006	0.19	0.06 to 0.32	0.45
Career interest ENT	7 (43.75)	12 (75)	4 (26.67)	0.024	0.34 (0.48)	0.49 (0.50	2.45	0.017	0.15	0.03 to 2.7	0.30
Module easy	4 (25)	7 (43.75)	13 (86.67)	0.024
Module effective	4 (25)	11 (68.75)	12 (0.8)	0.004
Would use again	4 (25)	14 (87.5)	13 (86.67)	<0.005
Total score					33.44 (17.91)	58.09 (17.12)	0.09	<0.005	24.65	19.19 to 30.11	1.41
Anatomy score					45.50 (28.88)	60.00 (20.11)	3.65	<0.005	14.50	6.50 to 22.50	0.58
Clinical and relations score					23.10 (4.76)	56.17 (20.28)	9.58	<0.005	33.07	26.12 to 40.02	1.86
Comparison to retention n=23					M pre (SD)	M ret (SD)
Total score					35.28 (19.04)	50.09 (6.48)	4.96	<0.005	14.81	8.62 to 21.00	0.83
Anatomy score					51.81 (30.25)	57.71 (21.73)	1.08	0.292	5.90	−5.443 to 17.23	0.22
Clinical and relations score					21.12 (8.89)	43.12 (20.20)	5.58	<0.005	22.00	13.83 to 30.17	1.24

Number and percentage of participants who responded agree orstrongly agree to the given statements. Whole group means and standard deviations for both post tutorial questionnaire andretention test score, mean difference and confidence interval, with whole group comparison to pretutorial with paired t test. Effect size Cohen’s d test. Cohen’s d effect size small =0.2, medium=0.5, large 0.8 (48).

P value significance <0.05 (bold)

3D, three dimensional; ENT, ear, nose and throat; M3D, monoscopic 3D group; PPT, presented the same content; PTB, petrous temporal bone; S3D, stereoscopic 3D group.

Table 4.

Summary of questionnaire results by group

	Mean pretest (%)	95% CI	P value	Adj post-test (%)	95% CI	P value	Adj ret test (%)	95% CI	P value
Total			0.238			0.103			0.937
PPT	27.31	18.50 to 36.11		61.87	47.69 to 63.25		51.08	41.41 to 60.75
M3D	35.82	27.03 to 44.61		53.79	46.45 to 61.73		48.9	39.89 to 57.90
S3D	37.44	27.96 to 46.92		64.63	57.21 to 73.07		50.58	39.45 to 61.71
Anat			0.210			0.235			0.209
PPT	35.21	20.42 to 50.00		61.87	52.49 to 71.25		67.37	53.82 to 80.92
M3D	49.48	35.91 to 63.04		53.79	44.59 to 63.00		52.38	39.67 to 65.10
S3D	52.22	37.09 to 67.36		64.63	55.07 to 74.18		52.81	37.09 to 68.52
Clin/relat			0.697			0.034			0.303
PPT	20.54	12.67 to 28.40		47.76	38.00 to 57.51		34.87	21.10 to 48.64
M3D	24.11	17.30 to 30.91		54.90	45.19 to 64.61		45.95	33.06 to 58.64
S3D	24.76	16.60 to 32.93		66.50	56.46 to 76.54		50.01	34.38 to 65.64

Mean pretutorial (pretest) scores with 95% CIs. ANCOVA regression analysis with adjusted post-tutorial (post-test) and adjusted retention test (ret test) means with 95% CIs. Comparison betweengroups with χ² test, p values given.

ANCOVA, analysis of covariance; M3D, monoscopic 3D group; PPT, presented the same content; S3D, stereoscopic 3D group.

Results

Participant demographics

A total of 47 medical students from the University of Melbourne postgraduate medical programme completed the training module, as well as the pretutorial and post-tutorial questionnaire. There were 16 participants in the PPT group, 16 participants in the M3D group and 15 participants in the S3D group. The retention questionnaire was completed by 23 participants (48.9%). There were 22 female participants and 25 male participants, of which 22 were in their first year and 25 in their second year. The mean age of participants was 23.0 years. No participants had previously observed mastoid surgery and one student had previously used a VR simulator. The mean subjective gaming experience was 1.64, which falls between casual and skilled on the four-point scale presented (none, casual, skilled, extreme). There were three left-handed participants. There was no significant difference between groups in any of these baseline characteristics (table 1). Students’ attitudes to their style of learning, anatomy knowledge and preparedness for clinical practice in ear, nose and throat (ENT) were explored (table 2). Prior to participating in the teaching module, the students exhibited low confidence in their knowledge of temporal bone anatomy and were not confident in their preparedness for a clinical rotation in ENT, with no significant difference between groups. The M3D group were significantly less comfortable learning anatomy in a group setting (p=0.045).

Subjective experience of the training module

Compared with their pretutorial experience, students in all groups felt significantly more confident with temporal bone and inner ear anatomy after completing their respective module. Scores reported by those in the S3D group suggested they felt a greater subjective understanding of the anatomy as a 3D structure (p=0.031) (table 3). As a whole group, students felt more comfortable with their preparedness to use the anatomical knowledge in a clinical context, including clinical examination and understanding how pathology arises. Both the M3D and S3D groups felt that the training module was effective and supported future exposure to this style of learning (figure 4). The S3D learning platform showed a significant benefit in ease of use.

Figure 4.

Results of the Likert scale for subjective experience. Percentage of respondents who agreed or strongly agreed with the statements ‘I found this module easy to use’, ‘I would use this way of learning again’, ‘I found this an effective way to learn temporal bone and inner ear anatomy’. M3D, monoscopic 3D group; PPT, presented the same content; S3D, stereoscopic 3D group.

Quantitative questionnaire results

As a whole group, the participants showed significant improvements in test scores with a medium to large effect size (p≤0.005) (table 3). The 23 participants who returned the retention questionnaire exhibited a significant improvement in total scores (retention score M=50.09, SD=6.45, pretutorial score M=35.28, SD=19.04) and clinical and spatial relations scores (retention score M=43.12, SD 20.20, pretutorial score M=21.12, SD=8.89) with large effect sizes (total score Cohen’s d=0.83, clinical and spatial relations score Cohen’s d=1.24). There was no significant improvement in retention of participants’ identification of anatomic structures scores across the whole group (retention score M 57.71, SD=21.73, pretutorial score M 51.81, SD=30.25) (table 3).

When analysed by group (table 4, figure 5), there were no significant differences between groups with respect to post-tutorial or retention questionnaire ‘total’ and ‘anatomy’ scores. Pairwise comparisons between groups revealed a significant difference between the adjusted post-tutorial questionnaire ‘total’ scores of the M3D and S3D groups (p=0.048; S3D 64.63, 95% CI 57.21 to 73.07 compared with M3D 53.79%, 95% CI 46.45 to 61.73). There was also a significant difference in the results of the post-tutorial questionnaire scores for the clinical and spatial relation domain between the PPT and S3D (p=0.010, S3D 66.50%, 95% CI 56.46 to 76.54, PPT 47.76%, 95% CI 38.00 to 57.51). Neither of these differences were seen in the retention subgroup analyses. For total questions and anatomy domains, the pretutorial scores had a significant effect on the post-tutorial scores and retention scores (total post-tutorial p=0.005, total retention p=0.003, anatomy post-tutorial p=0.030, anatomy retention p=0.005). There was no significant effect of the pretutorial results on the post-tutorial or retention scores for the clinical and spatial relations domains.

Figure 5.

Mean pretest results, with CIs. Adjusted means post tutorial, linear regression analysis accounting for pretest scores, with CIs. Adjusted means retention score of the 23 respondents, linear regression analysis accounting for the pretest scores of these 23 respondents. M3D, monoscopic 3D group; PPT, presented the same content; S3D, stereoscopic 3D group.

Discussion

This study presents an effective training module in clinically oriented temporal bone anatomy, which facilitates significant acquisition of knowledge. These results are consistent with those identified in the pilot study.³¹ In addition, as a group there was adequate retention of overall knowledge (as evidenced by ‘total’ score outcomes). This was mainly due to the retention of clinically oriented information and the spatial relationships between anatomical structures.

The display technology applied did not bear significant effect in the majority of knowledge outcomes, except that 3D technologies exhibited significant benefit in facilitating learning of clinically oriented and structural relational outcomes. Although the training of spatial relationships has not been assessed in previous temporal bone anatomy modules, Estevez et al demonstrated the benefit of 3D physical modelling in teaching spatial relationships in similarly complex neuroanatomy.³⁶ While the importance of clinically oriented anatomy modules has been established,³⁷ this study suggests a particular benefit to the 3D display technology in imparting clinically oriented anatomy knowledge. Hariri et al previously explored the benefit of VR simulation in teaching clinical anatomy of the shoulder joint, and although results were not inferior to the comparison textbook group, the participants in the simulation group exhibited greater motivation for learning in the VR environment.³⁸ Further to this, Jurgaitis et al reported the superiority of 3D hepatic visualisations in students’ ability to understand liver segmental anatomy, as well as the clinical tasks of identifying liver resection margins and predict possible intraoperative difficulties related to hepatic anatomy.³⁹ Similar to the learning of clinically oriented temporal bone anatomy, these tasks required the understanding of complex 3D anatomy and spatial relationships. This is likely explained by the benefit of 3D display technologies to train anatomy with reduced cognitive load required by participants, as has been previously demonstrated.^40–42 In addition, both the 3D displays were subjectively more effective than the 2D display and the participants were significantly more likely to report they would use these modules again. In a meta-analysis of 2226 participants in 2015, Yammine and Violato suggested the superiority of user satisfaction and perceived effectiveness of 3D visualisation technologies may reflect the novelty of the method.⁸ However, in this study, a significantly higher percentage of students also found the S3D module easier to use than the other displays. This correlates to the findings of Huang et al that immersion features of stereoscopic 3D VR anatomy courses positively impact perceived usefulness and ease of use.⁴³ The unique benefits of stereoscopic 3D displays have been explored by McIntire et al in a comprehensive literature review, concluding that S3D is of particular benefit in complex tasks for novice learners and in tasks requiring spatial manipulations of objects, spatial understanding, memory and recall.⁴⁴

Turney suggests that in modern medical curricula, the innate fascination with anatomy should be encouraged and that anatomy courses should promote student interest through clinical relevance and technology.¹⁷ To ensure medical practitioners have sufficient anatomy knowledge to practise safely, and to aid in retention of this knowledge, Turney suggested that anatomy education be delivered in a vertical approach, where the student initially learns core anatomy, revisiting the topics in a more clinically relevant, specialised approach later on in medical school and junior doctor years. The clinically oriented VR temporal bone anatomy module studied here lends itself to this style of vertical learning, and would be an appropriate module to bridge the gap between core anatomy knowledge of basic sciences and specialty-specific anatomy training.¹⁴ In addition, while Morgan et al noted the high labour associated with development and implementation of specialty-specific anatomy courses, this study exhibits a self-directed module that facilitated improvement in anatomy knowledge despite the absence of anatomy demonstrators. With regards to the benefits of differing display modalities, it was observed that there was a significant benefit of S3D over M3D in immediate overall knowledge gain, and S3D over PPT in immediate knowledge gain and understanding clinical and spatial relationships. This benefit did not extend to retention of the naming of individual anatomical structures which may reflect the under-representation of participant responses or could be an indication that this aspect of anatomy was not as important to the learners. The latter interpretation is supports the conclusions drawn from a review of medical curricula, where it was argued that greater importance should be placed on the participant’s subjective experience, in which both the M3D and S3D modules were evidently superior.¹⁷

It was observed that this clinically oriented module also improved students’ subjective confidence in their knowledge of anatomy as well as their interest in and preparedness for clinical work relating to otolaryngology. These alternative learning outcomes have previously been identified as benefits of anatomy education through cadaveric dissection.⁴⁵ Given that the face validity of our simulator has previously been demonstrated,⁴⁶ it is expected that these additional learning outcomes associated with cadaveric dissection courses would also be seen in VR simulation training. This is particularly important for otolaryngology education, as it has been identified that medical school graduates have poor comfort levels in otolaryngology despite this discipline forming an essential component of primary healthcare.⁴⁷

The findings of this study suggest that 3D technologies should continue to be integrated in anatomy education and can be successfully used to present clinically relevant content. Furthermore, this content can be delivered in a self-directed platform, facilitating reallocation of valuable anatomy demonstrator resources to more advanced and complex anatomic discussions.

There were several limitations identified in this study. First, given there is no standardised Australian anatomy curriculum,⁴ the information given in this module relied on expert opinion as to what was sufficient and appropriate. In addition, due to the absence of previously validated assessment tools for temporal bone anatomy education, the assessment questions were those evaluated as appropriately difficult and effective discriminators in the pilot study. To allow assessment of understanding, different questions were asked in the pretutorial and post-tutorial questionnaires. The themes of these questions were similar, although in particular for the clinical questions, the content was varied, particularly for the clinical material. The retention test was completed by under half of the participants, introducing a bias of respondents, thus limiting the assessment of retention of knowledge.

Conclusion

The VR temporal bone anatomy tutor was developed along the guidelines for an optimal anatomy curriculum. Notwithstanding the above limitations, it is an effective self-guided clinically oriented anatomy tutor in both 3D and 2D visual displays for imparting factual knowledge. The 3D display technologies had greater user satisfaction and were more effective in training clinically oriented anatomy and spatial relationships. Stereoscopic 3D exhibited particular benefit in ease of use. With the flexibility to incorporate varying levels of detail and a standardised curriculum, this module lends itself to a range of applications from undergraduate medical education to specialty specific anatomy education.

What is already known on this subject.

• Three-dimensional (3D) technologies and clinically oriented content are important adjuncts to time-limited anatomy teaching.

• A limited evidence base suggests that 3D technologies are promising in training inner ear and temporal bone anatomy.

What this study adds.

• A clinically orientated, procedural, and interactive virtual reality temporal bone anatomy module was developed and found to be efficient in facilitating knowledge acquisition of medical students.

• Of the three display technologies compared (two-dimensional (2D), monoscopic 3D, and stereoscopic 3D), there was no benefit in the majority of learning outcomes.

• 3D display exhibited a marginal benefit for clinically oriented and structural relation outcomes, as well as ease of use, perceived effectiveness, and students’ preference to use again.

Data availability statement

Data are available on reasonable request. Data are available on request.

Ethics approval

Ethical approval was given by the Royal Victorian Eye and Ear Hospital Human Research Ethics Committee (17–1313H).