Quantitative spectroscopic comparison of the optical properties of mouse cochlea microstructures using optical coherence tomography at 1.06 µm and 1.3 µm wavelengths

: Currently, the cochlear implantation procedure mainly relies on using a hand lens or surgical microscope, where the success rate and surgery time strongly depend on the surgeon’s experience. Therefore, a real-time image guidance tool may facilitate the implantation procedure. In this study, we performed a systematic and quantitative analysis on the optical characterization of ex vivo mouse cochlear samples using two swept-source optical coherence tomography (OCT) systems operating at the 1.06-µm and 1.3-µm wavelengths. The analysis results demonstrated that the 1.06-µm OCT imaging system performed better than the 1.3-µm OCT imaging system in terms of the image contrast between the cochlear conduits and the neighboring cochlear bony wall structure. However, the 1.3-µm OCT imaging system allowed for greater imaging depth of the cochlear samples because of decreased tissue scattering. In addition, we have investigated the feasibility of identifying the electrode of the cochlear implant within the ex vivo cochlear sample with the 1.06-µm OCT imaging. The study results demonstrated the potential of developing an image guidance tool for the cochlea implantation procedure as well as other otorhinolaryngology applications. this study demonstrated that 1.06-µm OCT can identify the presence and location of the metal electrode for a cochlear implant within the ex vivo mouse cochlear sample. The study results suggest that 1.06-µm OCT imaging has the potential to be used for developing a real-time imaging tool to guide the cochlear implantation procedure in the future.


Introduction
According to the World Health Organization (WHO), approximately 466 million people worldwide suffer from hearing impairments [1]. The management of impaired hearing or hearing loss is challenging because the inner hair cells, which are located on top of the basilar membrane, cannot regenerate once damaged because of exposure to excess noise, infection, or aging [2,3]. To overcome the aforementioned issues, various devices have been developed to enhance hearing ability [4], including hearing aids and hearing implants or cochlear implants. When hearing loss exacerbates, patients may require surgical intervention to place hearing implants into the cochlea. The cochlear implant comprises two parts, namely the outer device and the inner implant. The outer device serves as a microphone and transmits the sound signal to the receiver of the inner implant. Then, the electrodes of the inner implant pass the received signal to the auditory nerve. During cochlear implant surgery, the surgeon gradually drills into the spiral structure of the cochlea through the mastoid bone until the designated cochlea conduits is reached. Currently, the cochlear implant surgery mainly relies on using a hand lens or surgical microscope, where the success rate and surgery time strongly depend on the surgeon's experience [5,6]. Thus, a real-time image guidance tool might help improve the procedure for implanting the inner electrodes, for example, by easing the learning curve for the surgeons.
To guide the cochlear implant surgery, the desired technology should allow the differentiation of the fine anatomic features of the cochlea. Although MRI and CT have been used in previous studies to provide volumetric imaging of the cochlea, the imaging resolution of both techniques is limited to 0.5-1 mm, which does not allow for effective observation of the microscopic inner ear structure [7]. Moreover, the temporal resolution of MRI and CT is limited. On the other hand, ultrasound imaging allows for real-time imaging; however, the spatial resolution of ultrasound imaging is still relatively limited, approximately 30 µm with a state-of-the-art imaging setup [8]. Optical imaging modalities [9], such as optical coherence tomography (OCT), two-photon microscopy, and confocal microscopy, have been successfully used in various biomedical applications. Previous studies have demonstrated cochlea imaging with two-photon microscopy [10] or confocal microscopy [11]. Although the spatial resolution of both techniques is high, the imaging depth is limited to 300 µm. Several recent studies have also explored the feasibility of using OCT for otorhinolaryngology applications [12]; in these studies, the anatomic features of the guinea pia cochlea, such as the scala vestibuli (SV), scala media (SM), scala tympani (ST), modiolus, and spiral-like conduit inside the cochlea [13][14][15] can be clearly identified. Moreover, studies have investigated the potential of using OCT as the image guidance tool in the cochlear implant surgery [16,17].
In most studies that have employed OCT for cochlea imaging, a 1.3-µm light source was utilized to perform imaging with a sufficient imaging depth of 2-3 mm [18,19]. However, the axial imaging resolution of swept-source OCT (SS-OCT) at 1.3 µm is usually limited to 10 µm. Because the axial resolution of OCT scales with λ 0 2 /∆λ where λ 0 is the central wavelength of the light source, performing OCT imaging in a shorter wavelength regime allows for a higher axial imaging resolution, even for a similar spectral bandwidth ∆λ. For example, the axial resolution achieved at 1.06 µm is similar to that achieved at 1.3 µm even if the bandwidth of the 1.06-µm light source is 1.5 times narrower than that of the 1.3-µm light source, assuming a similar spectral envelope. However, light scattering decreases with wavelength. Thus, the imaging depth of an OCT system using a 1.06-µm light source is inferior to that of one using a 1.3-µm light source. Nevertheless, as an image guidance tool for the cochlear implantation procedure, we might not need to image through the entire cochlea structure. Thus, the imaging depth of the 1.06-µm OCT system might be sufficient. Moreover, 1.06-µm OCT exhibits lower signal attenuation because of water absorption compared with 1.3-µm OCT. Note that the cochlear conduits are filled with lymphatic fluid for in vivo condition.
Therefore, in this study, to facilitate the design of an image guidance tool with OCT for the cochlear implantation procedure, we developed two separate OCT systems with central wavelengths of 1.06 µm and 1.3 µm, respectively, and nearly identical detection sensitivities and lateral imaging resolutions. This would allow for a fair comparison of the imaging characteristics of mouse cochlear samples and the subsequent quantitative analysis of the OCT cochlear imaging. A customized cochlear sample holder was designed and manufactured using 3D printing technology to facilitate imaging procedures. In addition to performing OCT imaging of the freshly collected cochlear sample, we investigated how the decalcification of the cochlear sample affects its optical characteristics [20]. Two different metrics-image contrast and imaging depth-were used to quantify and compare the differences between the optical characteristics of the cochlear samples imaged at the 1.06-µm and 1.3-µm wavelengths and under different preservation conditions. Lastly, we have performed a separate OCT imaging session of an ex vivo mouse cochlea sample where we specifically introduced the metal electrode of a cochlear implant approved for human use to investigate the feasibility of using OCT to identify the metal electrode inside the cochlear sample.

OCT system description
We developed two separate SS-OCT systems (one operating at 1.06 µm and the other at 1.3 µm) to perform spectroscopic analysis of the OCT images of mouse cochleas ex vivo. Figure 1 shows a schematic of the 1.06-µm OCT system based on a Mach-Zehnder interferometer comprising four fiber-optic couplers, each with a power split ratio of 50:50. The swept source (AXP50125-3, Axsun Technologies) had a tuning range of 110 nm, an A-scan rate of 100 kHz, and an average output power of 18 mW. In the sample arm, light exiting the fiber terminal was collimated using an achromatic lens (AC-064-015-C, Thorlabs) and subsequently deflected toward the sample surface by a pair of closely spaced galvanometer scanning mirrors (GVS102, Thorlabs). The use of an achromatic lens (AC-254-045-C, Thorlabs) as the focusing lens allowed for a lateral resolution of approximately 9.9 µm (full-width at half-maximum). Except for the galvanometer scanner, the same optics used in the sample arm, including the collimating lens and the focusing lens, was used in the reference arm to minimize dispersion mismatch between the two arms. Lastly, the interference light signal was converted into an electrical signal using a dual-balanced photodiode detector (PDB-430C, Thorlabs). A high-speed digitizer (ATS9373, AlazarTech) with a 12-bit resolution was used for digitizing the detected interference signal using the sampling clock provided from the light source. Fig. 1. Schematic of the swept-source optical coherence tomography (SS-OCT) system employed in this study. Two separate OCT systems with central wavelengths of 1.06 µm and 1.3 µm were developed; both systems had the same A-scan rate of 100 kHz. The 1.06-µm and 1.3-µm OCT systems used two 50/50 couplers and two fiber-optic circulators, respectively (black dashed box), to deliver light output from the coupler connected to the light source to the reference arm or the sample arm of the OCT system. The sample holder was used during the scanning procedure to maintain the orientation of the cochlea. C: circulator; PC: polarization controller; Col: collimator; RM: reflection mirror; DBPD: dual-balanced photodetector; DAQ: data acquisition card; Galvo: galvanometer scanner; OFC: optical frequency clock. The 1.3-µm OCT system had a similar interferometer architecture to the 1.06-µm OCT system; however, the two couplers were replaced with two circulators to improve the light power efficiency. The circulators were not used in the 1.06-µm OCT system considering the mode dispersion at the 1.06 µm wavelength. The 1.3-µm light source (HSL-20, Santec Corp.) provided a tuning range of 100 nm, an A-scan rate of 100 kHz, and an average output power of 17 mW. In addition, single achromatic lenses were used as the collimating lens (AC-064-015-C, Thorlabs) and focusing lens (AC-254-035-C, Thorlabs), providing a lateral imaging resolution of approximately 11 µm (FHWM). We designed the two OCT systems such that they achieved a comparable lateral resolution and detection sensitivity to ensure a fair spectroscopic comparison of the OCT imaging characteristics of the mouse cochlea (Table 1). An in-house developed graphical user interface based on C++ was used to control beam scanning over the cochlear sample as well as data acquisition and real time imaging processing [21].

Animal specimen collection and imaging protocol
In this study, 4-to 5-week-old CBA/CaJ mice, which are commonly used in hearing research, were used. After the post-mortem collection procedure, fresh mouse cochlear samples were preserved in a phosphate-buffered saline (PBS) solution before the OCT imaging session. For the OCT imaging, we developed a customized sample holder, as depicted in Fig. 1, using a 3D printer (Form 2, Formlabs), which allowed us to image the mouse cochlear sample with the two SS-OCT systems in a co-registered manner. The end-user can rapidly switch between the two OCT systems while maintaining the same sample orientation under the same field-of-view (FOV). Upon completion of the first imaging session of the cochlear sample with both SS-OCT systems, the sample was submerged in 10% ethylenediaminetetraacetic acid (EDTA, J. T. Baker) pH 7.3, at 4°C on a rotating shaker. The decalcification of the cochlear sample with EDTA could enhance the OCT imaging depth and the visibility of the internal structure of the cochlea sample through tissue clearing [20]. Since we aimed to systematically investigate the changes of the optical characteristics of the cochlear samples after being treated with EDTA but not completely decalcifying the cochlear samples, the cochlear samples were only submerged in the EDTA solution for 7 days. After 7 days, the decalcified cochlear sample was imaged again following the same imaging protocol. The above animal imaging protocol was approved by the Institutional Animal Care and Use Committee (IACUC-19-274) of the National Defense Medical Center (Taipei, Taiwan).

Quantitative and spectroscopic analysis
In this study, we quantitatively analyzed the OCT images of the mouse cochlear sample acquired using the 1.06-µm and 1.3-µm OCT systems. Two different metrics were used to quantify and compare the differences in the optical characteristics of the cochlear samples: image contrast and imaging depth. First, the image contrast between the compartment indicating the cochlea conduit and regions corresponding to the neighboring bony cochlea wall was quantitatively computed with the contrast metric (CM) following the steps described subsequently. In brief, for each volumetric dataset, five cross-sectional OCT images with an imaging size of 2.5 mm × 2.5 mm were selected from the different locations of the cochlear sample. Then, four different regions (bone and compartments 1 to 3), were manually labeled and segmented in each selected cross-sectional OCT image. The mean OCT signal intensity values (logarithmic scale) corresponding to compartments 1 to 3 and bone were then computed. Lastly, the CM was derived using the following equation.
In order to evaluate how the axial imaging resolution affects the CM values, we also performed additional analysis on the change of the CM values if any, by lowering the axial resolution of 1.06-µm OCT imaging to 15.6 µm using a window function in post-processing.
For the imaging depth analysis, the OCT signal intensity profiles (logarithmic scale) were obtained with the 1.06-µm and 1.3-µm OCT imaging systems. The signal intensity profiles from the same imaging location and depth range (1 mm) were extracted for each cochlea sample, and a linear fitting model was used to compute the attenuation constant (α) of the OCT signal intensity profile, allowing for the characterization of the imaging depth under different settings; for example, between 1.06 µm and 1.3 µm OCT imaging as well as between PBS preservation and the EDTA-treated conditions. The CM and the attenuation constant (α) were computed using MATLAB 2019a (Mathworks). Prior to applying the linear fitting procedure, the OCT signal intensity profiles were smoothed by applying a moving average filter and then dividing by their maximum values.  Figure 2(c) shows the volumetric rendering of cochlear sample where cochlear conduits can be clearly identified as well. Figure 3 presents volumetric OCT images of the fresh mouse cochlear sample preserved in PBS solution prior to the imaging session; the images were acquired with the 1.06-µm and 1.3-µm OCT systems. Anatomic features similar to those of the mouse cochlear sample presented in Fig. 2(a) are observed in Figs. 3(a, c, d, f). However, because the sample was not treated with EDTA, the image depth is shallower than that of Fig. 2(a). In addition, a characteristic spiral contour following the connection of individual conduits inside the cochlea was observed in the en face OCT images, as indicated in Figs. 3(b, e) (blue dotted circles). Furthermore, we found that the image contrast between the compartments of the cochlea (red stars) and the neighboring bony cochlea wall was superior with the 1.06-µm OCT image data. By contrast, a greater imaging depth was observed with the 1.3-µm OCT imaging system, where features present at the deeper region within the cochlea sample could be identified (green triangles). Figure 4 presents volumetric OCT images of the same mouse cochlear sample used to obtain the images in Fig. 3; however, the images in Fig. 4 are those of the mouse cochlear sample treated with EDTA for 7 days. The image data of both systems (Figs. 4(a, c, d, f)) revealed a significant improvement in terms of penetration or imaging depth owing to decalcification from using EDTA; this allowed for the observation of cochlea compartments located deeper beneath  the tissue surface. In addition, the image contrast between the compartment and neighboring bony wall was still superior with the 1.06-µm OCT imaging system (red stars, Figs. 4(a, c, d, f)), whereas the imaging depth was still greater with the 1.3-µm OCT imaging system (green triangles, Figs. 4(a, c, d, f)). Lastly, just as in Figs. 3(b, e), a spiral contour following the connection of individual conduits inside the cochlea was observed in the en face OCT images (blue dotted circles, Figs. 4(b, e)).

Comparison of the image contrast with different wavelengths
As described in the previous section, the image contrast between the compartment and the neighboring bony wall of the mouse cochlea sample imaged with the 1.06-µm OCT system differed from that imaged with the 1.3-µm OCT system. Figure 5 presents a flowchart for computing the image contrast value, CM as described in section 2.3, and example cross-sectional OCT images of the mouse cochlea sample, where regions corresponding to the bone and compartments 1-3 were manually segmented and labeled. Figure 6 summarizes the measured CM values of the seven cochlear samples collected from four mice. Here, we compared the CM values obtained when imaging the cochlear samples using the 1.06-µm and 1.3-µm OCT systems. As depicted in Figs. 6 (a, b), in both cases (i.e., sample preserved in the PBS solution and sample treated with EDTA), the 1.06-µm OCT imaging system yielded higher CM values than did the 1.3-µm OCT imaging system, suggesting better differentiation between the cochlea compartment and neighboring bony wall.
Note that although the lateral resolution is comparable between 1.06-µm and 1.3-µm OCT, the axial resolution is different due the central wavelength and the optical bandwidth of the light sources. Therefore, we further compared the imaging contrast of the mouse cochlear sample where the axial resolution of the 1.06-µm OCT imaging was specifically lowered to 15.6 µm by applying a window function during the post processing. Thus, the axial resolution of the  1.06-µm OCT imaging is comparable to that of the 1.3-µm OCT imaging. As summarized in Fig. 7, whether the cochlear sample was under PBS preservation or being treated with EDTA, 1.06-µm OCT image of the cochlear samples still exhibits higher CM values. Figure 8 presents cross-sectional OCT images of the two mouse cochlea samples imaged in this study, where the difference between the imaging depths of the 1.06-µm and 1.3-µm OCT imaging systems can be clearly appreciated under different preservation conditions (PBS vs. EDTA). Thus, to quantitatively compare the difference between the imaging depths under different settings, we extracted the OCT signal intensity profiles from the same imaging location and depth range in the OCT images, as marked by the red dashed arrows in Fig. 8, for example, corresponding to the bony wall structure of the cochlea. Then, a linear fitting model was applied to compute the attenuation constant (α) of the axial OCT signal intensity profiles. The axial OCT signal intensity profiles were obtained after applying logarithmic compression and normalizing the data by the maximum intensity of the selected depth range. The extracted axial OCT signal intensity profiles and the overlaid fitted results for the images in Fig. 8 are presented in Fig. 9. As shown in Fig. 9, in both cases (i.e., sample preserved in the PBS solution and sample treated with EDTA), the attenuation constant was lower for the 1.3-µm OCT imaging system, which corresponded to a greater imaging depth compared with that of the 1.06-µm OCT imaging system. Moreover, the attenuation constant decreased with both imaging systems once the cochlear sample was treated with EDTA, which corresponded to greater imaging depth than that observed with the PBS-preserved cochlear sample, consistent with the findings of a previous study [20]. We quantitatively analyzed the imaging depth from images of the other five cochlear samples and summarized the analysis results of all cochlear samples in Fig. 10. In general, for all cochlea samples, the attenuation constant obtained with the 1.3-µm OCT imaging system was lower than that obtained with the 1.06-µm OCT imaging system. Moreover, the images of the EDTA-treated samples had a lower attenuation constant than did the PBS-preserved samples in both 1.06-µm and 1.3-µm OCT imaging.

OCT imaging of the cochlea sample with the electrode of a cochlear implant introduced
In order to investigate the feasibility of using 1.06-µm OCT imaging as the potential guidance tool for cochlear implantation procedure, we have performed 1.06-µm OCT imaging of a separate ex vivo mouse cochlear sample (under PBS preservation) where the electrode of a cochlear implant approved for human use (Cochlear Nucleus Profile, Cochlear) was introduced through the oval window. Figure 11 shows the volumetric OCT images and the photograph of the aforementioned cochlear sample. Note that the metal electrode of the cochlear implant exhibits strong OCT signal intensities and thus, it was clearly identified in the orthogonal views of the volumetric OCT images (pink-dashed lines, Fig. 11). As shown in Fig. 11 (b), the introduced electrode exhibits a U-shape contour (pink-dashed lines) following the spiral-like conduit anatomy from the bottom to the superior part of the cochlear conduits. Figure 11(c), the en face OCT image extracted from a deeper depth than that of Fig. 11(b), shows the electrode being introduced into the cochlear conduit. Due to the limiting size of the mouse cochlear conduit, particularly at the Here, based on the definition of the fitting model as described in Fig. 9, a larger value of the attenuation constant corresponds to a faster decay of OCT signal intensity over depth.
top of the cochlear sample compared to the size of the electrode for a human cochlear implant, we could only see the metal electrode at the bottom of the cochlea as shown in the cross-sectional OCT images (Figs. 11 (d, f)). Lastly, strong OCT signal of the electrode can be observed in the cross-sectional OCT images as well (Fig. 11 (e)).

Discussion
In this study, we developed two OCT systems operating at wavelengths of 1.06 µm and 1.3 µm. Both systems exhibited comparable lateral imaging resolutions and detection sensitivities. The developed imaging platform allowed for the identification of the anatomic features of the mouse cochlea sample (Fig. 2). In addition, we systematically investigated the optical characteristics of mouse cochlear samples by performing a spectroscopic and quantitative analysis on cochlear OCT images. As seen in Figs. 3 [23]. However, the aforementioned study only focused on imaging the guinea pig cochlea samples after they were treated with EDTA to enhance the image contrast, which may limit translating the study findings to in vivo settings. Moreover, only the analysis results from one single sample but not all samples (N = 10) were presented in the study by Kim et al. By contrast, in this study, we performed both 1.06-µm and 1.3-µm OCT imaging on seven cochlea samples collected from four mice under two different preservation conditions: freshly collected samples preserved with PBS and samples treated with EDTA for 7 days. Note that, in this study, we did not aim to completely decalcify the cochlear samples but temping to systematically investigate the changes of the optical characteristics of the cochlear sample before and after being treated with EDTA. Thus, the cochlear samples were only stored in the EDTA solution for 7 days, leading to the preservation of the bony wall of the cochlear sample still. Most importantly, we employed two different metrics-the CM and the attenuation constant-to systematically and quantitatively compare the spectroscopic characteristics of the mouse cochlear samples.
As shown in Figs. 6 and 7, 1.06-µm OCT imaging enabling a better differentiation between the bony wall and the conduit of the cochlear samples even at the setting where we specifically lower the axial resolution of 1.06-µm OCT imaging to be comparable to that of the 1.3-µm OCT imaging. By contrast, the 1.3-µm OCT imaging system provided better visibility of the anatomic features of the cochlea sample deep beneath the tissue surface owing to a lower attenuation constant. However, the 1.06-µm OCT imaging system did provide superior image contrast and sufficient imaging depth compared with the 1.3-µm OCT imaging system when a fresh mouse cochlea sample was imaged. Considering that the future goal of this study is to develop an image guidance tool with OCT for the cochlea implantation procedure in human patients, visualizing the anatomic features deep beneath the tissue surface might not be the most critical factor regarding the aforementioned application [16]. Therefore, the results obtained in this study suggest that 1.06-µm OCT may be more suitable than 1.3-µm OCT imaging for guiding the cochlea implantation procedure.
In order to further investigate the feasibility of using 1.06-µm OCT imaging as the guidance tool for the cochlear implantation procedure, we have performed OCT imaging of the ex vivo cochlear sample, particularly with the metal electrode of a commercially available cochlear implant being introduced into the cochlear conduit. Although the size of the metal electrode limiting it advancing into the top part of the cochlea, the presence and position of the metal electrode can be identified in the volumetric OCT images as shown in Fig. 11, suggesting the potential feasibility of 1.06-µm OCT imaging for cochlear implantation guidance.
In this study, we developed two SS-OCT systems to investigate the spectroscopic characteristics of a mouse cochlea sample ex vivo. Although spectral-domain OCT (SD-OCT) provides OCT imaging with even higher axial imaging resolution owing to the broader bandwidth of currently available light sources such as superluminescent diodes (SLEDs), the imaging speed of SD-OCT at 1.06 µm and 1.3 µm was limited when compared with that of SS-OCT [24]. A slower imaging speed may affect the feasibility of SD-OCT for the proposed application of surgical guidance using real-time imaging. Moreover, SS-OCT typically provides a better sensitivity roll-off than does SD-OCT [25], which is desired for surgical guidance with OCT.
Based on the experimental findings of this study, we will next design a guiding probe for cochlea implantation surgery based on 1.06-µm SS-OCT. Though the achromatic lens used in the sample arm of the 1.06-µm OCT system does provide ∼5-cm working space, suggesting the potential of using such a system during cochlear implantation surgery, a forward imaging probe might provide better ergonomics to perform OCT imaging during the surgery. In addition, the same tool could be used for other otorhinolaryngology applications, such as for imaging of the middle ear or tympanic membrane [26,27]. For instance, otitis media with effusion (OMEN) is a very common disease in children, where infection leads to the accumulation of tissue fluid in the middle ear space behind the tympanic membrane [28,29]. The lower water absorption at the 1.06-µm wavelength regime may allow for a more effective examination of OME compared with the 1.3-µm wavelength regime. In summary, given the results demonstrated in this study and the above discussion, we believe that 1.06-µm OCT imaging may be a suitable alternative to 1.3-µm OCT imaging, which has been widely used in otorhinolaryngology applications.

Conclusion
In this study, we developed two separate SS-OCT systems operating at wavelengths of 1.06 µm and 1.3 µm. Using this imaging platform, we performed a systematic and quantitative analysis on the OCT images of ex vivo mouse cochlear samples to obtain a better understanding of the optical characteristics of the cochlea. The 1.06-µm OCT imaging system provided superior performance in terms of the image contrast between the cochlear conduits and the neighboring cochlear bony wall structure compared with the 1.3-µm OCT imaging system. By contrast, the 1.3-µm OCT imaging system provided greater imaging depth of the cochlear samples. The above observations apply to the cochlea samples freshly collected (preserved in the PBS solution) as well as to the samples that were decalcified with EDTA. In addition, this study demonstrated that 1.06-µm OCT can identify the presence and location of the metal electrode for a cochlear implant within the ex vivo mouse cochlear sample. The study results suggest that 1.06-µm OCT imaging has the potential to be used for developing a real-time imaging tool to guide the cochlear implantation procedure in the future.