Eur Arch Otorhinolaryngol
DOI 10.1007/s00405-017-4572-1
OTOLOGY
Modifications to a 3D-printed temporal bone model
for augmented stapes fixation surgery teaching
Yann Nguyen1,2 • Elisabeth Mamelle1,2 • Daniele De Seta1,2 • Olivier Sterkers1,2
Daniele Bernardeschi1,2 • Renato Torres1
Received: 30 January 2017 / Accepted: 17 April 2017
Ó Springer-Verlag Berlin Heidelberg 2017
Abstract Functional outcomes and complications in oto-
sclerosis surgery are governed by the surgeon’s experience.
Thus, teaching the procedure to residents to guide them
through the learning process as quickly as possible is
challenging. Artificial 3D-printed temporal bones are
replacing cadaver specimens in many institutions to learn
mastoidectomy, but these are not suitable for middle ear
surgery training. The goal of this work was to adapt such an
artificial temporal bone to aid the teaching of otosclerosis
surgery and to evaluate this tool. We have modified a
commercially available 3D-printed temporal bone by
replacing the incus and stapes of the model with in-house
3D-printed ossicles. The incus could be attached to a 6-axis
force sensor. The stapes footplate was fenestrated and
attached to a 1-axis force sensor. Six junior surgeons
(residents) and seven senior surgeons (fellows or consul-
tants) were enrolled to perform piston prosthesis placement
and crimping as performed during otosclerosis surgery.
The time required to perform the tasks and the forces
applied to the incus and stapes were collected and ana-
lyzed. No statistically significant differences were observed
between the junior and senior groups for time taken to
Electronic supplementary material The online version of this
article (doi:10.1007/s00405-017-4572-1) contains supplementary
material, which is available to authorized users.
& Yann Nguyen
yann.nguyen@inserm.fr
1
Sorbonne Universités, Université Pierre et Marie Curie Paris
6, Inserm, UMR-S 1159 ‘‘Minimally Invasive Robot-based
Hearing Rehabilitation’’, Paris, France
2 restore the conductive properties of the middle ear. The
AP-HP, Groupe Hospitalier Pitié Salpêtrière, Otolaryngology
Department, Unit Otology, Auditory Implants and Skull Base
Surgery, Paris, France
•
perform the tasks and the forces applied to the incus during
crimping and placement of the prosthesis. However, sig-
nificantly lower forces were applied to the stapes by the
senior surgeons in comparison with the junior surgeons
during prosthesis placement (junior vs senior group,
328 ± 202.9 vs 80 ± 99.6 mN, p = 0.008) and during
prosthesis crimping (junior vs senior group, 565 ± 233 vs
66 ± 48.6 mN, p = 0.02). We have described a new
teaching tool for otosclerosis surgery based on the modi-
fication of a 3D-printed temporal bone to implement force
sensors on the incus and stapes. This tool could be used as
a training tool to help the residents to self-evaluate their
progress with recording of objective measurements.
Keywords Training
Educational
Stapedotomy
Middle
ear
Piston
Prosthesis
Forces
Ossicular chain

Otosclerosis
Introduction
Otosclerosis is an inflammatory bone disease leading to
bone dystrophy and involving the stapes footplate and the
bony labyrinth. It will result in stapes fixation and con-
ductive hearing loss at an early stage and inner ear lesions
at a later stage giving mixed hearing loss.
Hearing rehabilitation can be performed either by
hearing aids or surgical replacement of the stapes function.
Surgical treatment consists of resection of the stapes
superstructure, the fenestration or removal of the stapes
footplate, and the placement and crimping of an ossicular
prosthesis between the incus and the fenestrated stapes to
procedure is very challenging and is performed through a
narrow field exposure with a speculum placed in the outer
123

ear canal and involves millimetric and light fragile struc-
tures represented by the ossicular chain. When performed
by experienced surgeons, the technique yields excellent
results, with hearing improvement and a postoperative air-
bone gap of less than 10 dB in 90% of cases.
The worst complication is sensorineural hearing loss
which may be partial or total and irreversible. Functional
outcomes and occurrence of complications vary with the
surgeon’s experience [1, 2]. To improve the success rate
and diminish the incidence of complications among less
experienced surgeons, technical modifications have pro-
gressively been adopted in the last three decades. Stape-
dotomy instead of stapedectomy was proposed to lower
sensorineural hearing loss [3]. Laser was used alone [4] or
in combination with a microdrill [5] to assist with footplate
fenestration and lower the risk of footplate fracture. Other
authors have proposed the use of robotic assistance to
increase safety during surgery [6]. Nevertheless, no tool or
technique will be as valuable as the training, skills, and
experience of the surgeon to guarantee the success of the
surgery.
Most otologist have been trained through an
apprenticeship in the operating room with the tradi-
tional adage ‘‘see one, do one, teach one’’, known as the
Halstedian method [7]. This clinical training can be
completed with a temporal bone dissection course or
free practice in a temporal bone laboratory [8]. How-
ever, access to cadaver specimens varies from one
faculty to another and even more from one country to
another. Changes in legislation on body donation [9],
medical restrictions due to risk of prion diseases [10],
and high costs have further reduced the access of res-
idents to temporal bone in many training programs.
Because of these restrictions, innovative teaching
methods such as artificial simulators using, for example,
3D-printed temporal bones [11], animal models such as
sheep [12] or pig, or virtual simulators [13] are
employed to reduce the use of temporal bone specimens
and shorten the learning curve. Three-dimensional
printed temporal bones offer a realistic anatomical
representation that can reproduce real-case anatomy
based on DICOM images acquired with a CT scan.
Artificial bones are very beneficial in learning posterior
cavity drilling techniques but are less valuable in
teaching middle ear cleft surgery as the ossicular chain
is fixed and sometimes poorly reproduced.
In the present work, we describe how we have mod-
ified a commercially available 3D-printed artificial
temporal bone to provide objective measures for the
learner to raise his interest in extended training for
otosclerosis prostheses placement and crimping on such
an artificial simulator.
123
Eur Arch Otorhinolaryngol
Materials and methods
Population
Thirteen surgeons were enrolled in the study. All of them
were right-handed. The population was divided into two
groups depending on their otological experience. A junior
group was represented by six residents (two women and
four men) and a senior group was represented by six fel-
lows and a consultant from our ENT department (three
women and four men).
Modification of the artificial temporal bone
A commercially available 3D-printed temporal bone
(Schmidt model, left ear, Phacon, Leipzig, Germany) [14]
was modified to provide measurements of force applied to
the incus and stapes. The malleus and stapes were removed
from the commercial device. An access point was drilled
superiorly to expose the epitympanum and anteriorly to
expose the oval window from an inner ear perspective.
These two access points allowed placement of a modified
incus (Fig. 1a) and stapes (Fig. 1b) that were printed with a
stereolithographic technique. These two ossicle models
were obtained from an ossicular chain segmentation
reported in a previous work [15]. Ablation of the super-
structures and a 500 lm fenestration were performed with
a microdrill to obtain a perforated footplate (Fig. 1b).
A Teflon tubing (length 3 mm, outer diameter 1 mm, inner
diameter 0.8 mm) was glued to the inner ear side of the
footplate (Fig. 1d). This tubing could then be placed on the
contact ball of the 1-axis force sensor to ensure the con-
nection between the footplate model and a 1-axis force
sensor (range 0–1 N, resolution: 10 mN, Millinewton force
sensor, EPFL, Lausanne, Switzerland, http://lpm.epfl.ch/,
Fig. 2) and the incus was 3D printed connected to a rod to
be mounted on a 6-axis force sensor (ATI Nano 17, cali-
bration type SI-12-0.12, resolution: 3 mN, Apex, NC,
USA, Figs. 1c, 3). Sensor data were recorded in real time
via the same analog-to-digital interface card (NI 6210,
National Instruments, Austin, TX, USA) and in-house
software. Only components Dx, Dy, and Dz, provided by
the 6-axis force sensor, were averaged to obtain the norm
of the force applied to the incus.
Experimental setup
Participants were asked to place the prosthesis into the
fenestrated footplate and around the long process of the
incus under microscopic view (Kaps, Asslar, Germany) in a
transcanal approach through a 5 mm diameter ear specu-
lum. An image of the simulated surgical exposure is
Eur Arch Otorhinolaryngol
Fig. 1 Modified incus and stapes. The incus (a) and stapes (b) were
modified so as to be attached to force sensors. The incus was printed
jointly with a rod to ease attachment to a 6-axis force sensor (c). A
stapes footplate was printed and then drilled with a 500 lm
fenestration. A Teflon tubing (length 3 mm, outer diameter 1 mm,
Fig. 2 Internal view of the modified artificial temporal bone. A
commercially available 3D-printed temporal bone (Schmidt model,
left ear, Phacon, Leipzig, Germany) was modified to include
measurement of force applied to the incus and stapes. An access
point was drilled to the vestibule to expose the oval window niche
from the inner ear side. The stapes was removed from the commercial
temporal bone and replaced by an in-house 3D-printed stapes. This
stapes model was mounted on a 1-axis force sensor (Millinewton
force sensor, EPFL, Lausanne, Switzerland). This sensor allowed
measurement of force applied to the stapes
represented in Fig. 4. The prosthesis used was a titanium
K-Piston with a 4.5 mm length and 0.4 mm shaft diameter
(Kurtz, Dusslingen, Germany). A Hartmann alligator micro
forceps (Hartmann ear forceps, ref 221150, Storz,
inner diameter 0.8 mm) was cut and glued to the inner ear side of the
footplate (d). This tubing could then be placed on the contact ball of
the 1-axis force sensor to ensure the connection between the footplate
model and the 1-axis force sensor. The stapes could then be attached
to a 1-axis force sensor
Fig. 3 External view of the modified artificial temporal bone. A
commercially available 3D-printed temporal bone (Schmidt model,
left ear, Phacon, Leipzig, Germany) was modified to include
measurement of force applied to the incus and stapes. An access
point was drilled from the tegmen to the epitympanum and the incus
was removed from the commercial artificial temporal bone. It was
replaced by an in-house incus 3D-printed jointly with a rod that could
be attached to a 6-axis force sensor (ATI Nano 17, Apex, NC, USA).
This sensor allowed measurement of force applied to the incus
Tuttlingen, Germany) was used to place the piston into the
stapedotomy and around the incus, and a McGee wire
crimping micro forceps (McGee wire crimper straight, ref
227400, Storz) was used to crimp the piston. The prosthesis
placement and crimping were assessed by an external
123

Fig. 4 Middle ear cleft exposure through the external auditory canal.
The use of an artificial temporal bone (Schmidt model, left ear,
Phacon, Leipzig, Germany) allowed the surgical and anatomical
environment to be reproduced with high accuracy. The incus in this
model was replaced by an in-house 3D-printed incus to attach it to a
force sensor. This image shows the implementation of our incus
model into the commercial artificial temporal bone
evaluator. After each trial, the prosthesis was removed,
visually inspected, and replaced by a new one if damaged.
Analysis
Numbers of trials required to complete the procedure were
collected. The procedure was considered completed when
the prosthesis could be placed and crimped in a single step.
Only the first completed procedure by the participant
including prosthesis placement and crimping was analyzed.
Time taken to complete these two steps was also collected.
Considering the force measurements, we investigated the
shape of the curve corresponding to the force versus time
during the two steps of the procedure. The peak force
applied to the incus and footplate during placement of the
prostheses on the long process of the incus was collected.
We believed that it was reasonable to assume that this metric
would represent potential damage to the ossicular chain or
the cochlea if an excessive force was applied. Results were
expressed as mean ± standard deviation. Data were ana-
lyzed and graphics were generated with R version 3.1.3 (R
Core Team, 2013, R, Vienna, Austria). Data are presented as
mean ± standard deviation. We used the Mann–Whitney
test to evaluate statistical significance for duration of the
procedure and force applied to the stapes and incus. p values
\0.05 indicated a statistically significant difference.
Results
In the senior group, 3 ± 2.8 trials were required to achieve
the first completed trial. In the novice group, 9 ± 9.7 trials
were necessary.
123
Eur Arch Otorhinolaryngol
Duration of the procedure
No statistically significant differences were observed
between the two groups with regard to duration of the
procedure for prosthesis placement (junior vs senior group,
26 ± 13 vs 15 ± 4.5 s, p = 0.13) or for prosthesis
crimping (junior vs senior group, 28 ± 10.5 vs 20 ± 7.4 s,
p = 0.31).
Force applied to the incus
No statistically significant differences were observed
between the two groups with regard to forces applied to the
incus during prosthesis placement (junior vs senior group,
720 ± 203.4 vs 338 ± 233.1 mN, p = 0.07) or during
prosthesis crimping (junior vs senior group, 433 ± 334.1
vs 182 ± 169.3 mN, p = 0.11).
Force applied to the stapes
Statistically significant differences were observed between
the two groups with regard to forces applied to the stapes
by the two groups. Lower forces were applied by the senior
surgeons in comparison with the junior surgeons during
prosthesis placement (junior vs senior group, 328 ± 202.9
vs 80 ± 99.6 mN, p = 0.008) and during prosthesis
crimping (junior vs senior group, 565 ± 233 vs
66 ± 48.6 mN, p = 0.02).
The results of duration and force measurements for
prosthesis placement and crimping are, respectively,
reported in Figs. 5 and 6.
Discussion
This study describes a training model based on modified
3D temporal bone with measurement of force applied to the
incus and stapes during prosthesis placement and crimping
in simulated otosclerosis surgery. We have shown that,
during these steps in the surgical procedure, surgeons with
otological experience would apply a gentler force to the
stapes in comparison to surgeons with short or no previous
otological experience.
Advantages and limits of this simulator
3D-printed temporal bones generally offer a high fidelity
for visual and anatomical representation. Thus, the
dimensions, approach, and exposure of the surgical field
could be easily reproduced in this simulator using
commercially available temporal bone. With the possi-
bility to customize printed temporal bone, it would be
easy to change the surgical scenario from a left to a right
Eur Arch Otorhinolaryngol
Fig. 5 Collected metrics for evaluation of prosthesis placement in an
otosclerosis surgery model. Six junior surgeons (residents) and seven
senior surgeons (fellows or consultants) were enrolled to perform a
piston prosthesis placement in our modified temporal bone model.
The duration of the task and forces collected via two force sensors and
applied to the incus and stapes were investigated. No significant
ear or simulate difficult cases for advanced surgeons
such as a narrow oval niche with facial nerve overhang
or necrosis of the long process of the incus. Furthermore,
use of a physical simulator allows the learner to use real
tools (microscope, surgical tools, and prosthesis) to
become accustomed to the tools available in the oper-
ating room. Another advantage of such a simulator is its
versatility to compare the user friendliness of different
techniques (e.g., different types of prosthesis, robot-
based versus manual technique, etc.), although compli-
ance of the ossicular chain does not reflect human
physiology and no prediction on hearing outcomes can
be estimated from a comparison of techniques or pros-
theses with our simulator.
Nevertheless, some drawbacks limit the value of this
simulator. Its realism for ossicular chain palpation has not
been objectively compared to a real ossicular chain. Fur-
thermore, absolute force measurement values cannot be
considered to compare them with the previous reports on
ossicular chain manipulation. Indeed, the incus in our setup
is mounted on a long rod for attachment to the 6-axis force
difference was observed between the groups for duration of the task
or force applied to the incus. The senior surgeons applied significantly
lower forces to the stapes in comparison with the junior surgeons
during prosthesis placement (junior vs senior group, 328 ± 202.9 vs
80 ± 99.6 mN, p = 0.008)
sensor and this creates a leverage effect that does not
directly reflect the effort applied to the incus. The exact
force applied to the incus could not be calculated from the
moment of the force as the angle of the applied force may
vary constantly during prosthesis placement and crimping.
The previous studies have already reported that expert sur-
geons would apply less force during otosclerosis surgery
simulation compared to junior surgeons [16, 17]. Moreover,
some other critical steps in the otosclerosis surgery such as
scutum lowering or footplate fenestration were not simu-
lated. Difficult intraoperative conditions such as bleeding
were not reproduced. In addition, intraoperative complica-
tions such as incus fracture, incudo-malleolar joint rupture,
and floating footplate, could not be reproduced in this sim-
ulator. Finally, the cost of the force sensors (around
30–50 eu for the 1-axis force sensor and 10,000 eu for the
6-axis force sensor) and the need for a computer may
hamper widespread academic use of this simulator, even
though it can be reused without deterioration through time.
This cost could be reduced with the use of a 3-axis force
sensor (around 1000 eu) instead of a 6-axis force sensor.
123

Fig. 6 Collected metrics for evaluation of prosthesis crimping in an
otosclerosis surgery model. Six junior surgeons (residents) and seven
senior surgeons (fellows or consultants) were enrolled to perform a
piston prosthesis crimping in our modified temporal bone model. The
duration of the task and forces collected via two force sensors and
applied to the incus and stapes were investigated. No significant
Other models of otosclerosis training surgery
The estimated learning curve for otosclerosis surgery is
between 60 and 80 cases [18] and complications can occur
even after the first successful cases [19]. For these reasons,
simulators have been designed to train residents. Some
authors have created simple artificial surgery boxes to per-
form training outside the operating room [20–22]. A more
complex model reproducing not only otosclerosis but also
chronic otitis scenarios was proposed by Mills and Lee [23].
These simulators are low cost and easy to transport but have
a limited representation of the anatomy of the middle ear
ossicles. Thus, Murrant and Gatland proposed to glue the
oval window to reproduce pathological conditions observed
in otosclerosis in cadaver models [24]. Another totally dif-
ferent approach is represented by a virtual simulator. This
has become very popular for temporal bone drilling teaching
[13, 25] but less so for otosclerosis surgery training [26].
With such simulations, residents could train without
restrictions even with personal computers outside medical
faculties. The main drawback of virtual simulators is the
poor haptic feedback that is limited by the prohibitive cost of
123
Eur Arch Otorhinolaryngol
difference was observed between the groups for duration of the task
or force applied to the incus. The senior surgeons applied significantly
lower forces to the stapes in comparison with the junior surgeons
during prosthesis crimping (junior vs senior group, 565 ± 233 vs
66 ± 48.6 mN, p = 0.02)
efficient haptic devices with multiple degrees of freedom
and force feedback. Thus, a 3-axis force feedback haptic
interface (three translations) cost around 1500–2000 eu and
a 6-axis force feedback interface (three translations and
three translations) cost around 30,000 eu.
Conclusion
We report a new teaching tool for otosclerosis surgery
based on modification of 3D-printed temporal bone to
implement force sensors on the incus and stapes and
measure the forces applied to these ossicles. We have
observed that senior surgeons apply a lower peak force on
the stapes during prosthesis manipulation. The best use of
this simulator would be to use it as a training tool to help
residents self-evaluate their progress with recording of
objective measurements. On one hand, this training would
raise their confidence but also allow them to improve their
hand positioning, accuracy, and steadiness as reported in
other models of middle ear surgery training [27, 28]. The
next steps will be to modify the model to create additional
Eur Arch Otorhinolaryngol

surgical scenarios such incus necrosis, or a narrow oval stapedectomy training. Eur Arch Otorhinolaryngol
window. 272(9):2167–2174
13. Sorensen MS, Mosegaard J, Trier P (2009) The visible ear sim-
Compliance with ethical standards ulator: a public PC application for GPU-accelerated haptic 3D
simulation of ear surgery based on the visible ear data. Otol
Neurotol 30(4):484–487
Human and animal rights statement This article does not contain
14. Roosli C, Sim JH, Mockel H, Mokosch M, Probst R (2013) An
any studies with human participants or animals performed by any of
artificial temporal bone as a training tool for cochlear implanta-
the authors. This research received no specific grant from any funding
tion. Otol Neurotol 34(6):1048–1051
agency in the public, commercial, or not-for-profit sectors.
15. Kazmitcheff G, Miroir M, Nguyen Y, Ferrary E, Sterkers O,
Cotin S, Duriez C, Grayeli AB (2014) Validation method of a
Conflict of interest The authors indicate no potential conflict of
middle ear mechanical model to develop a surgical simulator.
interest. The authors would like to thank Armand Czaplinski for his
Audiol Neurootol 19(2):73–84
assistance with the drawing of the 3D models, the modified incus, and
16. Bergin M, Sheedy M, Ross P, Wylie G, Bird P (2014) Measuring
stapes used in this study.
the forces of middle ear surgery; evaluating a novel force-de-
tection instrument. Otol Neurotol 35(2):e77–e83
17. Rothbaum DL, Roy J, Hager GD, Taylor RH, Whitcomb LL,
Francis HW, Niparko JK (2003) Task performance in stapedo-
References tomy: comparison between surgeons of different experience
levels. Otolaryngol Head Neck Surg 128(1):71–77
18. Yung MW, Oates J, Vowler SL (2006) The learning curve in
1. Caldart AU, Terruel I, Enge DJ Jr, Kurogi AS, Buschle M,
stapes surgery and its implication to training. Laryngoscope
Mocellin M (2007) Stapes surgery in residency: the UFPR clin-
116(1):67–71
ical hospital experience. Braz J Otorhinolaryngol 73(5):647–653
19. Sergi B, Paludetti G (2016) Can the learning curve in stapes
2. Lial PI, Soares VY, Viana LM, Sampaio AL, Oliveira CA,
surgery predict future functional outcome? Acta Otorhinolaryn-
Bahmad FJ (2013) Stapedotomy in a residency training program.
gol Ital 36(2):135–138
Int Tinnitus J 18(2):163–167
20. Mathews SB, Hetzler DG, Hilsinger RL Jr (1997) Incus and stapes
3. Kursten R, Schneider B, Zrunek M (1994) Long-term results after
footplate simulator. Laryngoscope 107(12 Pt 1):1614–1616
stapedectomy versus stapedotomy. Am J Otol 15(6):804–806
21. Monfared A, Mitteramskogler G, Gruber S, Salisbury JK Jr,
4. Perkins RC (1980) Laser stepedotomy for otosclerosis. Laryn-
Stampfl J, Blevins NH (2012) High-fidelity, inexpensive surgical
goscope 90(2):228–240
middle ear simulator. Otol Neurotol 33(9):1573–1577
5. Nguyen Y, Bozorg Grayeli A, Belazzougui R, Rodriguez M,
22. Owa AO, Gbejuade HO, Giddings C (2003) A middle-ear sim-
Bouccara D, Smail M, Sterkers O (2008) Diode laser in oto-
ulator for practicing prosthesis placement for otosclerosis surgery
sclerosis surgery: first clinical results. Otol Neurotol
using ward-based materials. J Laryngol Otol 117(6):490–492
29(4):441–446
23. Mills R, Lee P (2003) Surgical skills training in middle-ear sur-
6. Nguyen Y, Miroir M, Kazmitcheff G, Ferrary E, Sterkers O,
gery. J Laryngol Otol 117(3):159–163
Grayeli AB (2011) From conception to application of a tele-
24. Murrant NJ, Gatland DJ (1989) Temporal bone laboratory
operated assistance robot for middle ear surgery. Surg Innov
training for stapedectomy. J Laryngol Otol 103(9):833–834
19(3):241–251
25. Linke R, Leichtle A, Sheikh F, Schmidt C, Frenzel H, Graefe H,
7. Kerr B, O’Leary JP (1999) The training of the surgeon: Dr.
Wollenberg B, Meyer JE (2013) Assessment of skills using a
Halsted’s greatest legacy. Am Surg 65(11):1101–1102
virtual reality temporal bone surgery simulator. Acta Otorhino-
8. Levenson MJ (1999) Methods of teaching stapedectomy.
laryngol Ital 33(4):273–281
Laryngoscope 109(11):1731–1739
26. Kazmitcheff G, Nguyen Y, Miroir M, Pean F, Ferrary E, Cotin S,
9. George AP, De R (2010) Review of temporal bone dissection
Sterkers O, Duriez C (2014) Middle-ear microsurgery simulation
teaching: how it was, is and will be. J Laryngol Otol
to improve new robotic procedures. Biomed Res Int 2014:891742
124(2):119–125
27. Murbe D, Huttenbrink KB, Zahnert T, Vogel U, Tassabehji M,
10. Scott A, De R, Sadek SA, Garrido MC, Courteney-Harris RG
Kuhlisch E, Hofmann G (2001) Tremor in otosurgery: influence
(2001) Temporal bone dissection: a possible route for prion
of physical strain on hand steadiness. Otol Neurotol
transmission? J Laryngol Otol 115(5):374–375
22(5):672–677
11. Bakhos D, Velut S, Robier A, Al zahrani M, Lescanne E (2010)
28. Ovari A, Nemenyi D, Just T, Schuldt T, Buhr A, Mlynski R,
Three-dimensional modeling of the temporal bone for surgical
Csokay A, Pau HW, Valalik I (2016) Positioning Accuracy in
training. Otol Neurotol 31(2):328–334
Otosurgery Measured with Optical Tracking. PLoS One
12. Cordero A, Benitez S, Reyes P, Vaca M, Polo R, Perez C, Alonso
11(3):e0152623
A, Cobeta I (2015) Ovine ear model for fully endoscopic

123