Page 185 - IJB-9-4

International Journal of Bioprinting 3D-printed middle ear prostheses

photography, the distance of the prosthesis from the Connect VFX-F-110, Polytec GmbH, Waldbronn, DE).
camera was sequentially changed by approximately Afterward the velocity was converted to displacement
0.1 mm by attaching the camera to a macro-focusing rail. with Polytec Vibrometer software (Polytec GmbH). The
Typically, around 30–40 photographs were taken from each LDV sensor head (VibroFlex Compact VFX-I-130, Polytec
prosthesis. The series of photographs, i.e., a stack of frames, GmbH) was attached to the operation microscope M320
was then processed by calculating the final sharp image (Leica Microsystems, Heerbrugg, Switzerland) using a
with a high depth of field using CombineZP software (Alan micromanipulator (A-HLV-MM40, Polytec GmbH), which
Hadley, United Kingdom). made it easier to aim the laser beam at the measurement
point marked at the medial side of the stapes footplate.
2.6. Preparation of temporal bones The measuring range was adjusted to the smallest possible
Four anonymous cadaveric temporal bones were used. (10 mm/s) and the samples were averaged 50 times to
The temporal bones were removed from the cadaver heads increase the SNR of the measurements.
as whole. No evidence of macroscopic or microscopic
pathology was detected. The fresh-frozen temporal bones Cadaver temporal bones were always moistened with
were thawed only once. Standard mastoidectomy was saline before measurements. An approximately 1 × 1 mm
performed with a high-speed otosurgical drill (Stryker, plastic reflective glitter was placed on the middle of the
5400-50 Core, Kalamazoo, MI). Labyrinthectomy was stapes footplate to improve the reflection of the laser beam
then performed to gain access to the medial side of the (Figure 2B and C). The first measurement was performed
stapes footplate (Figure 2A and B). A wide posterior with intact ossicular chain, followed by a measurement
tympanotomy was performed to allow inspection and where incus was removed. Then, the measurement
manipulation of the whole ossicular chain and medial was performed with a reconstructed ME together with
aspect of the tympanic membrane (Figure 2D and E). 3D-printed PORP or a commercial titanium PORP (Clip
Partial Flexibal Prosthesis, Heinz Kurz GmbH).
2.7. Middle ear transfer function
The mechanics of the ossicular chain and the acoustical The lengths of the prostheses were determined by
performance of the 3D-printed prostheses were determined testing different lengths for each specimen, and the most
by single-axis LDV. The measurement setup was based on suitable prostheses (1.5, 1.75, 2, 2.25, 2.5, 2.75, and 3 mm)
Stoppe et al. The setup is shown in Figure 2A. as deemed by the otosurgeon and the received METF
[18]
results were selected for the comparisons.
Based on pilot measurements on the transduction
of sound from the tympanic membrane to the stapes 2.8. Statistics
footplate in an intact ME, it was determined that the METF A two-way analysis of variance (ANOVA) was performed
stabilized after approximately 2 h from the beginning of with ME transduction mode (intact ME, titanium PORP,
the thawing process. Similarly, the lack of moisture in the 3D-printed PORP and incus removed) and frequency
ME due to evaporation started to affect the METF after (averaged on octave bands centered around 125, 250, 500,
3.5 h from a thawed temporal bone. Therefore, 2 h after 1,000, 2,000, and 4,000 Hz) as dependent variables and
thawing, a speculum and an associated adapter were placed METF as an independent variable. Multiple comparisons
in the ear canal. An earphone (ER1, Etymotic Research of the means were adjusted with the Bonferroni correction.
Inc., Elk Grove Village, IL) and a probe microphone (ER- The analysis was conducted using IBM SPSS version 25
7C, Etymotic Research Inc., Elk Grove Village, IL) were software (IBM, Armonk, NY).
connected to the adapter. Probe tube was inserted to the
end of the speculum as close as possible to the tympanic 3. Results
membrane. The adapter had a sealing window that allowed The first-generation 3D-printed PORPs (Figure 3A)
visibility to the tympanic membrane. The ear canal was were visually analyzed using an operating microscope.
sealed with silicone wax. Due to sealing, the signal-to- The printed prostheses were evaluated by their general
noise ratio (SNR) was at a minimum of about 25 dB. A overview, cup structure, and straightness of the shaft. With
multisine excitation signal was used in the frequency range these viewpoints in mind, all prostheses were classified as
from 100 Hz to 5 kHz. The sound signal was generated with successfully or unsuccessfully printed. The problems faced
MATLAB (MathWorks, Natick, MA) and reproduced by with the first-generation PORPs were deviations in the cup
the earphone through a USB sound card and an amplifier. structure (Figure 3B) and broken shafts (Figure 3C). No
The LDV measurements were performed at an overall level other issues were detected. As a result, 43 of the 80 printed
of approximately 105 dB SPL in the ear canal. PORPs (54%) were classified as successfully printed.
Velocity of the medial side of the stapes footplate was From the successfully printed first-generation
measured with a laser Doppler vibrometer (VibroFlex PORPs, 14 consecutive prostheses were selected by visual

Volume 9 Issue 4 (2023) 177 https://doi.org/10.18063/ijb.727