Synchrotron Phase-Contrast Imaging and Cochlear Otosclerosis: A Case Report Open Access

Otosclerosis is a disorder that affects the capsular bone surrounding the human inner ear and results in inflammatory-like lesions at different stages of development [1, 2]. It is among the most common sources of acquired hearing loss and is histologically present in about 10% of Caucasians [3]. Otosclerosis causes conductive hearing loss due to bony fixation of the stapes, but it can also lead to sensorineural hearing loss, conceivably due to a disturbed inner ear fluid homeostasis or functional derangement of the fibrocyte network in the lateral wall of the cochlea [4].

The treatment of clinical otosclerosis is primarily surgical. Medical remedies have also been introduced with limited success, including the use of fluorides [5, 6] or systemic bisphosphonates [7]. In severe cases of sensorineural hearing loss, cochlear implantation is an important therapeutic option.

The etiology of clinical and histological otosclerosis remains unknown. Several causes have been presented, such as genetic factors, measles virus infection [8], hormonal changes [9], and autoimmunity [10]. Vascular abnormalities associated with otosclerosis and sensorineural hearing loss have also been shown [11‒14].

A proposed theory – other than those suggesting genetic, viral, autoimmune, and hormonal causes – is that the bone lacks the ability to remodel normally during life, which becomes a predisposing factor to developing otosclerosis [15, 16]. In addition to a poorly developed vascular channel system, the otic capsule contains a lacuno-canalicular network. Bone metabolism may be influenced by the action of the cytokine receptor osteoprotegerin (OPG) secreted from the inner ear, thereby reducing otic capsular remodeling [3, 9, 17]. A disturbance of this system has been associated with otospongiosis-initiating osteoclastogenesis [18].

High-resolution computed tomography can be used to clinically visualize the number and anatomical distribution of otosclerotic foci [19]. Recently, a cadaveric study was conducted in which an otosclerotic/otospongiotic cochlea was three-dimensionally imaged using micro-CT [20]. Three-dimensional (3D) reconstructions of otosclerosis and the inner ear spaces were also produced previously from histological serial sections [21].

In the present investigation, we aimed to further improve the imaging of otosclerotic lesions with the use of synchrotron radiation phase-contrast imaging (SR-PCI). A donated ear from a patient with otosclerosis was imaged and compared with a healthy control.

Two adult cadaveric human cochleae of unknown age were used in this study. One cochlea was from a patient with otosclerosis and conductive hearing loss due to stapes ankyloses, and the other was from a healthy control. The patient with otosclerosis had undergone partial stapedectomy with the insertion of a metal wire prosthesis. There were no data available on hearing, vertigo, or audiometric data. Both samples were obtained with permission from the body bequeathal program at Western University (London, ON, Canada) in accordance with the Anatomy Act of Ontario and Western’s Committee for Cadaveric Use in Research (approval # 06092020). The study adhered to the rules of the Declaration of Helsinki. Samples were fixed using formaldehyde (4%) and underwent SR-PCI at the Canadian Light Source Inc. (Saskatoon, SK, Canada) using the Biomedical Imaging and Therapy beamline (05ID-2). The SR-PCI technique used was previously described [22‒24]. In brief, the imaging field of view was set to 4,000 × 950 pixels, corresponding to 36.0 × 8.6 mm, and 3,000 projections over a 180° rotation were acquired per view. Reconstruction resulted in an effective isotropic pixel size of 9 μm. The acquisition time to capture all projections per view was ∼30 min.

Open-source medical imaging software, 3D Slicer (www.slicer.org, version 5.0.3) [25], was used to segment and create 3D representations. Otosclerotic/otospongiotic lesions were 3D rendered using the composite with shading technique. Scalar opacity and color mapping were adjusted to display volume properties with the removal of bones to enhance surfaces. Lesion erosions and vascular bony channels were manually segmented using the threshold painting technique, while surrounding cells were segmented using automatic threshold detection. Manual tracing was performed using threshold painting and 3D rendering to display the lytic and sclerotic changes of different foci engaging the otic capsule. Bony vascular connections between the otospongiotic foci, middle ear, and facial nerve were segmented and compared with those of the healthy control ear.

For 3D reconstructions of the cochlear anatomy, structures were traced and color-labeled manually (threshold painting) on each SR-PCI CT slice (approximately 1,400 slices per sample). To see possible regional differences, a tonotopic map of the basilar membrane (BM)/organ of Corti and the spiral ganglion was created by nonlinear least squares fitting a Greenwood-like function to the data [26‒28]. The BM tonotopic frequency at each dendrite location was calculated using Greenwood’s function [27], which relates each proportional length along the BM to its characteristic frequency. The dendrites were then traced to their respective locations in Rosenthal’s canal, which hosts the spiral ganglion.

Archival human cochleae obtained at surgery that had been directly fixed in buffered 3% glutaraldehyde were reevaluated for analyses of otic capsule blood vessels. These specimens were used for transmission electron microscopy and high-resolution light microscopy (HRLM). The specimens were decalcified in 10% Na-EDTA and stained with 1% osmium tetroxide. Semithin sections for HRLM were stained with toluidine blue and imaged [29].

Archival HRLM of the sections showed multiple blood vessels inside the human otic capsule. Vessels had a thin endothelium surrounded by loose connective tissue and a few free cells. The vessels frequently reached the inner layer of the bony capsule within 100 microns of the cochlear spiral ligament (Fig. 1). Their diameter ranged from 7 to 25 μm. The bony channels were surrounded by lacunar dendrite osteocytes, extending many purple-stained radiating canaliculi, seemingly connected to the vascular bone channels. This formed the so-called lacuno-canalicular network [18] (Fig. 1, insets). The canaliculi had a diameter of 0.7 microns, and the osteocytes residing in the bony lacunae had a diameter of 2.5–6 μm.

SR-PCI and 3D segmentation of a left human cochlea in a non-otosclerotic ear are shown in Figures 2 and 3. The inner ear soft tissues, including the Reissner´s membrane, and the blood vessels residing in the otic capsule bone channels are visualized. Many of these channels ended on the bony surface and mucosa of the middle ear. Figure 2 shows the vascular channel network in the otic capsule before and after semitransparency of the middle ear bony wall. Several bony channel openings to the middle ear mucosa were visualized. The projections of the different cochlear turns are shown with many vascular channels at the promontory wall near the round window (Fig. 2b–e). Segmented bone channels around the third turn are shown in Figure 2f.

The SR-PCI and 3D reconstructions also show that the vascular channels are connected to the perilabyrinthine air spaces (Fig. 3). A BM frequency map constructed according to Greenwood [27] determined the relative position of the channel network along the cochlea. A shallow bony pit could be identified in the bony capsule anterior to the oval window representing the fissula ante fenestram. This region contained few vascular channels. No connections were found between the otic capsule network and the inner ear.

Five otosclerotic/otospongiotic lesions were segmented and three dimensionally demonstrated in the otosclerotic ear (Fig. 4a). Lesions were located at the anterior and posterior parts of the oval window, round window, lateral semicircular canal, and anterior part of the internal auditory canal. The lesions showed a clear demarcation against the normal capsular bone. Four lesions were interconnected, while one lesion in the wall of the internal acoustic canal lay unconnected. The stapes footplate was partly removed, and a wire prosthesis had been inserted in the posterior part of the oval window (Fig. 4b). The steel wire produced few image artifacts. The window was surrounded by pathologic bone that extended to the lateral semicircular canal and ampulla nerve. The changes around the oval window and basal region of the cochlea were hyper-vascularized, suggesting active otospongiosis. A horizontal section of the temporal bone at the level of the oval window is shown in Figures 5 and 6. In this region, the bony changes extended from the endosteum at the perilymph/endolymph border to the middle ear. In the most basal region, channels are frequently connected to the ligament and the perilymphatic lining. The bony channels in the otosclerotic/otospongiotic lesions frequently opened up into the middle ear mucosa on the promontory wall (Fig. 7a–c). Automatic segmentation demonstrated a close relationship between otosclerotic bone channels and perilabyrinthine pneumatization (Fig. 7d). At the internal acoustic meatus, the lesion displayed osteophytes that protruded into the meatus. Several microfissures or microcracks were observed in the otic capsule. These were not found in the control ear.

The second lesion (lesion 2) engaged the basal part of the first turn of the cochlea, the round window region, and the acoustic crest. It reached the inferior cochlear vein (ICV) and the cochlear aqueduct (CA). The ICV drains all cochlear blood, and it was significantly dilated with a diameter of 0.25 mm near the floor of the scala tympani (Fig. 7-9). The posterior modiolar vein was also dilated. The ICV was angled perpendicular to the floor of the scala tympani where it was shielded by a thin bony ledge before it passed through the otic capsule. The mean diameter of the accessory canal housing the ICV was previously estimated to be 0.1 mm [30].

The stylomastoid artery, facial nerve, and venous plexus were manually segmented from the stylomastoid foramen to the internal acoustic meatus (Fig. 10-11). This showed that the facial nerve was surrounded by an extensive venous plexus that drained into veins at the skull base. While the venous plexus enclosed the nerve, the stylomastoid artery ran medially, joined by paired veins. At the lateral knee, several veins connected to the otosclerotic lesions located at the lateral semicircular canal (Fig. 10a, b, 11). Figure 10c and d shows the vascular connections to the otosclerotic lesions with and without the automatically segmented bone channels. The otosclerotic lesions were connected not only via blood vessels (Fig. 11) but also via bony channels linked to the perilabyrinthine pneumatization.

Here, we have presented the first SR-PCI 3D reconstruction of a temporal bone with otosclerotic/otospongiotic lesions. Image slices were used to three-dimensionally reconstruct and anatomically locate different otic capsule lesions with segmentation of vascular bone channels. SR-PCI and 3D rendering allowed unprecedented visualization of the pathologic changes in a cadaveric ear from a patient treated surgically for conductive hearing loss and stapes ankyloses caused by otosclerosis. The imaging allowed visualization and comparison of the distribution of vascular bone channels in the otic capsule compared with a control ear. The osteocyte lacunae and canaliculi could not be resolved. SR-PCI of an intact, unstained human cochlea was first performed in 2017 [23], and the technique has since been used to examine the cyto-architecture [31] and 3D tonotopic distributions of the human organ of Corti and spiral ganglion [24].

Despite diverse opinions, credible studies have indicated that extensive otosclerosis also leads to sensorineural hearing loss [32, 33]. Otosclerotic involvement of the endosteum may also directly damage neurosensory elements [34]. Some patients may suffer from progressive sensorineural hearing loss even without stapes fixation, which may be suspected in patients with a family history of clinical otosclerosis [4, 11]. The most damaging seem to be the lesions situated at the basal cochlea. An important role may be played by vascular shunts developed between otosclerotic foci and the “vena spiralis inferior” and ICV with signs of venous congestion [12]. Heavily dilated capillaries were described in the stria vascularis in the basal turn of the cochlea. In 24 temporal bones from patients with otosclerosis, the most extensive sensorineural degeneration was seen when the process involved the round window and the lower half of the basal turn [33]. Vascular shunts connected otosclerotic foci with the cochlear microvasculature, including arterial connections to the stria vessels. Nearly all bones showed sensorineural changes in the lower basal turn. Johnson et al. [35] investigated the 3D labyrinthine anatomy and vascular pathology in cochlear otosclerosis using microdissection. All specimens showed vascular shunts connecting otosclerotic foci to the cochlear vasculature, which was severely dilated in the scala tympani in the lower half of the basal turn with complete loss of the sensory organ in the first quadrant [35]. There was degeneration of nerve fibers in the osseous spiral lamina. The ICV was not investigated. Where the endosteum of the scala tympani was engaged, there was also a loss of cochlear venules and capillaries, suggesting changes in cochlear microcirculation leading to sensorineural degeneration.

The normal arterial blood supply to the embryonic otic capsule and ossification centers was thoroughly described by Bast and Anson [36]. Branches derive from the external carotid system, such as the ascending pharyngeal, posterior auricular, and inferior tympanic arteries. As there are normally no direct functional anastomoses between vessels inside the labyrinth and the otic capsule, newly developed shunts seem to have formed. In the present study of the non-otosclerotic ear, several vascular channels in the otic capsule were found to communicate to middle ear mucosa vessels and surrounding air cells. No direct communication pathways to the inner ear could be observed. Prominent vascular channels were located in the promontory. In the otosclerotic ear, however, vascular channels or shunts formed communication pathways between lesions and the inner ear blood system. This coincides with histological descriptions of vascular shunts present in cochlear otosclerosis [1, 11‒13]. In addition, rich arterial connections were found between lesions and vascular collaterals to the horizontal portion of the facial nerve and its outer knee. Such arterial connections were earlier described [37‒39]. Facial weakness, however, is seldom noted in patients with otosclerosis despite engagement of the bony wall of the facial canal [40].

A prominent feature of active otospongiotic lesions is the large number of dilated vascular channels together with inflammatory-like lytic changes with the accumulation of osteoblasts and active osteoclasts [2, 6, 41]. According to Nager [13], otosclerotic lesions at the oval window are supplied by vessels over the promontory mucoperiosteum and branches from the internal auditory, middle meningeal, superficial petrosal, and superior tympanic arteries [13]. The superficial petrosal and superior tympanic veins were said to drain the venous blood from the lesion at the oval window following the arteries. According to Jyung and Wacharasindhu [37], the increased angiogenesis, also named “angiotropism,” included vascular connections to inferior tympanic vessels [37]. Angiogenetic shunts could overload the inner ear, lead to dilation of stria vessels and degeneration of the organ of Corti, and cause sensory loss. Antiangiogenic agents were suggested to treat the disease. The authors found that vascular connections to major vessels could explain the predilection sites of the lesions. Otosclerotic lesions in the basal part of the cochlea together with oval window lesions are said to be present in up to 50% of cases [13, 42, 43]. According to Schuknecht and Barber [1], the round window region is affected in 30% of clinical otosclerosis and 17% in histological otosclerosis (absence of stapes fixation) [1]. The normal accumulation of capsular vessels at the basal cochlea in the present study could explain the classic location of Schwartze´s sign. It is believed to be due to an increased blood flow caused by larger, often pulsative, vessels through the promontory invasion of otosclerosis [44]. According to Ruedi [12], otosclerotic engagement of the endosteum causes venous shunts between the membranous labyrinth and active foci and is most prevalent in the basal turn of the cochlea. The venous congestion may cause damage to the spiral ganglion cells, organ of Corti, and stria vascularis. In 47% of his cases, histological signs of venous shunts were present.

Lesions at the basal turn may have a critical impact on the ICV and the CA at the round window [1]. In 44 ears with round window lesions, nine showed blockage of the CA and ICV. Collateral drainage routes were assumed to exist. Altmann et al. [45] described otosclerotic lesions that invaded the promontory and obliterated both the CA and the ICV, leading to venous congestion [45]. There were signs of disturbed cochlear microcirculation, and the organ of Corti was entirely degenerated with atrophy of the lamina nerve fibers in the basal turn. The inferior spiral vein was dilated with surrounding calcified fibrous tissue, and a similar situation was noted in the contralateral ears. There was apposition of bone at the inner opening of the channel of the ICV, and the dilation of the innermost portion of the ICV was believed to be caused by shunt formation. In the present study, the ICV and posterior modiolar vein were significantly dilated compared with those of the control ear, suggesting an increased ICV outflow and vascular congestion in the otosclerotic compared with the normal ear. There was no documented anatomic hindrance of the ICV at the otic capsule, but an obstruction further against the skull base could not be ruled out. The changes of the stria vascularis described by Altmann et al. [45] seem to be consistent with changes observed after experimental obstruction of the “inferior spiral vein” in guinea pigs [46]. Interference with the CA does not seem to have a major influence on inner ear physiology [47]. Yet, normal CA patency could be essential to maintain inner ear fluid homeostasis under pathological conditions. Altmann and coauthors [45] believed that the most valid explanation for sensorineural hearing loss was toxic substances released from the otosclerotic bone into the labyrinthine fluids or abnormal stria vascularis damaging the nerve elements.

Otosclerosis is still a disease of unknown etiology. Increased vascularity and bone changes could be due to a disrupted metabolism caused by abnormal bone remodeling. A better understanding of the molecular underpinnings and ways to inhibit the characteristic remodeling of the otic capsule could lead to new treatment strategies [3]. After early formation, the otic capsule shows few signs of remodeling that is different from that of ordinary bone [48]. The reason for this is unknown; however, it is believed to be due to inherent biological mechanisms, such as inhibitory stimulation from the chemokine OPG, a soluble factor of the tumor necrosis factor family that regulates bone mass and osteoclast activity [49]. It may be produced from the inner ear tissue by fibrocytes and root cells in the spiral ligament [9, 17]. The perilabyrinthine tubulo-canalicular network may disseminate chemical signaling, leading to a regulated osteoclast activity, thereby reducing capsular remodeling. The network consists of osteocyte canaliculi associated with blood vessel-containing bone channels. A reduced OPG signaling through the lacuno-canalicular network may lead to clustering of dead osteocytes, initiating bone remodeling and the otosclerosis process [18, 50].

Considering the extensive vascular communication pathways existing and remaining between the middle ear mucosa and the otic capsule, one may speculate if histological otosclerosis in patients with emerging sensorineural hearing loss could be treated by local middle ear pharmacotherapy (Fig. 12). Such intervention could aim to reduce bone angiogenesis, resume normal bone growth [51], and diminish osteoclastogenesis. The presence of otosclerotic plaques engaging critical regions, such as entry of arteries and venous and endolymphatic duct outflow corridors (endolymphatic hydrops), may be documented with current high-resolution computed tomography [19]. These foci could be targeted to halt the progression and sensory aggravation [45, 52, 53]. Externally induced chemical signals through OPG, bone enhancement, and antiangiogenetic therapy could help to restore bone biochemistry and microcirculation and offer a new viable therapeutic option for treating some patients.

We thank Karin Lodin for the skillful artwork and Lauren Siegel for language editing.

The study was approved by the Ethics Review Board (No. 99398, 22/9 1999, cont., 2003, no. C254/4; no. C45/7 2007, Dnr. 2013/190) at the Uppsala University Hospital (No. 99308). Written information was given to the patient, and informed consent was obtained. Samples used for synchrotron analysis were obtained with permission from the body bequeathal program at Western University (London, ON, Canada) in accordance with the Anatomy Act of Ontario and Western´s Committee for Cadaveric Use in Research (approval # 06092020). The study was conducted in accordance with the World Medical Association Declaration of Helsinki.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

This study was supported by the Swedish Research Council [2022-03339], Åke Wibergs Stiftelse, Tysta Skolan Foundation, and the Swedish Deafness Foundation (hrf). We also acknowledge the kind donations of private funds made by Arne Sundström, Sweden. The project was supported by MED-EL Elektromedizinische Geräte GmbH, Innsbruck, Austria, under the agreement and contract with Uppsala University. Part of the research described in this paper was performed at the Bio-Medical Imaging and Therapy (BMIT) facility at the Canadian Light Source Inc. (CLSI), which was funded by the Canada Foundation for Innovation, the Natural Sciences and Engineering Research Council of Canada, the National Research Council Canada, the Canadian Institutes of Health Research, the Government of Saskatchewan, Western Economic Diversification Canada, and the University of Saskatchewan. The sponsors did not involve in study design, execution and analysis, manuscript conception, planning, and writing, and decision to publish.

Sumit Agrawal and Hanif Ladak performed SR-PCI in Canada on the human cadavers. Hao Li and Dina Giese performed image processing and 3D visualization of scanned objects provided by Sumit Agrawal and Hanif Ladak. Helge Rask-Andersen was the head of the ear laboratory in Uppsala and planned the project. Dina Giese and Hao Li analyzed the images. Dina Giese, Sumit Agrawal, and Helge Rask-Andersen wrote the manuscript. All authors approved the submitted version.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.