Temporal Bone Imaging in Osteogenesis Imperfecta Patients With Hearing Loss

Transcription

1 The Laryngoscope VC 2013 The American Laryngological, Rhinological and Otological Society, Inc. Temporal Bone Imaging in Osteogenesis Imperfecta Patients With Hearing Loss Freya K. R. Swinnen, MSc, PhD; Jan W. Casselman, MD, PhD; Els M. R.De Leenheer, MD, PhD; Cor W. R. J. Cremers, MD, PhD; Ingeborg J. M. Dhooge, MD, PhD Objectives/Hypothesis: Osteogenesis imperfecta (OI) is an autosomal-dominant connective-tissue disorder, predominantly characterized by bone fragility. Conductive hearing loss develops in half of the OI patients and often progresses to mixed loss. Findings of computed tomography (CT) and magnetic resonance (MR) imaging of the temporal bone in the largest series of OI patients to date are presented and correlated with the audiograms. Study Design: Retrospective case series. Methods: CT images and audiograms of 17 hearing-impaired OI patients, aged 9 to 67 years, were analyzed retrospectively. In four patients, MR imaging was performed as well. Imaging abnormalities were correlated with type and severity of hearing loss deduced from the audiograms. Results: CT revealed fenestral hypodense foci in the fissula ante fenestram (25 of 33 ears), oval window (23 of 33 ears), and round window (20 of 33 ears). Retrofenestral hypodensities were observed, affecting the cochlear turns (16 of 33 ears), facial nerve canal (10 of 33 ears), or semicircular canals (6 of 33 ears), or appearing like the fourth turn of the cochlea (11 of 33 ears). The site of hypodensities corresponded to the type of hearing loss in 72.2% of the OI ears. The air-bone gap and bone-conduction thresholds showed significant positive associations with the number of affected fenestral (P <.05) and retrofenestral structures (P <.01), respectively. Gadolinium-enhanced MR images demonstrated active lesions in three patients with mixed hearing loss or deafness. Conclusions: The site of hypodensities on temporal bone CT images in OI corresponds to presence and type of hearing loss determined by audiometry. The more severe the hearing loss, the more affected temporal bone structures in OI. Key Words: Osteogenesis imperfecta, computed tomography, magnetic resonance imaging, temporal bone. Level of Evidence: 4. Laryngoscope, 123: , 2013 INTRODUCTION Osteogenesis imperfecta (OI) is a connective-tissue disorder, in most cases caused by an autosomal-dominant mutation in the COL1A1 or COL1A2 gene. These genes code for elements of type I collagen, an important constituent of bone, blood vessels, skin, joints, and tendons. The cardinal feature in OI is bone fragility. Four major types of dominantly inherited OI, varying in disease severity, are distinguished on the basis of clinical and genetic characteristics. 1 Type I is the most common and mildest phenotype, presenting with blue sclerae, multiple fractures in childhood, and hearing From the Department of Otorhinolaryngology, Ghent University Hospital (F.K.R.S., E.M.R.D.L., I.J.M.D.), Ghent, Belgium, Department of Medical Imaging, Sint-Jan Hospital (J.W.C.), Bruges, Belgium, FC Donders Institute for Neurosciences, Radboud University Nijmegen Medical Centre, Department of Otorhinolaryngology (C.W.R.J.C.), Nijmegen, The Netherlands. Editor s Note: This Manuscript was accepted for publication December 6, Freya K. R. Swinnen, MSc, holds a PhD fellowship of the Research Foundation Flanders (Fonds Wetenschappelijk Onderzoek Vlaanderen), Belgium. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Freya Swinnen, Department of Otorhinolaryngology (1P1), Ghent University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium. freya.swinnen@ugent.be DOI: /lary loss development in a considerable number of patients. Type II is perinatally lethal because of extreme bone fragility affecting the respiratory tract and insufficient ossification of the skull. Short stature, progressive bone deformity, scoliosis, and multiple fractures characterize the patients with severe type III. Finally, the moderate-to-severe type IV is associated with normal sclerae, short stature, and mild-to-moderate osseous fragility. More uncommon, autosomal-recessive forms of OI caused by mutations in other genes have recently been recognized. Consequently, several attempts have been made to expand the initial Sillence classification with types V to XI. 2 Irrespective of the underlying OI genotype or phenotype, half of the OI patients develop conductive hearing loss between the second and fourth decade of life, associated with bony changes affecting the oval window region and causing stapes footplate fixation. 3 5 In addition, thinning, atrophy, and fractures of the stapes suprastructures and long process of the incus have been described in several series reporting on the intraoperative findings of stapes surgery in OI The conductive hearing loss often evolves to a mixed hearing loss because of a gradually arising concomitant sensorineural hearing loss component. 4,12,13 Postmortem histopathologic analyses of temporal bones have been performed in OI infants with a severe,

2 perinatally lethal OI type in whom hearing had not been evaluated. They demonstrated deficient ossification of the cortical layers of the otic capsule, semicircular canals, ossicles, and bony walls of the middle ear In scarce reports on histopathologic temporal bone findings in adult OI patients with late-onset hearing loss, the deficient ossification of the cortical layers in the inner and middle ear was accompanied by vascular and otosclerosis (OS)-like changes encroaching on the stapes footplate and causing stapes ankylosis In the present study, we report on the computed tomography (CT) and magnetic resonance (MR) temporal bone findings in the largest series of OI patients to date. In addition, it is hypothesized that the occurrence, type, and severity of hearing loss in OI are associated with the occurrence and location of OS-like lesions visualized on CT images of the temporal bones. MATERIALS AND METHODS The CT scans of the temporal bones and the audiograms of 17 hearing-impaired OI patients who participated in a large study in which the correlation between audiologic phenotype and underlying OI genotype was determined were acquired at the same time and were reviewed retrospectively. Informed consent was obtained from all patients. In each patient, the diagnosis of OI was genetically confirmed. Fifteen patients demonstrated OI type I, whereas OI types III and IV were diagnosed in one patient each. MR images of the temporal bones were also available in four patients. From one patient we only obtained CT images of the right ear. The patients were between 9 and 67 years old at the time of radiologic and audiometric evaluation. Their audiograms were analyzed for determination of type and severity of hearing loss. Pure-tone averages (PTAs) for air-conduction (AC) and bone-conduction (BC) thresholds, expressed in decibel hearing level (db HL), from frequencies 0.5, 1.0, 2.0, and 4.0 khz were calculated, as well as the corresponding average air-bone gap (ABG), expressed in decibels. Following these calculations, hearing was categorized as follows: 1) normal: PTA (AC) < 15 db HL; 2) conductive hearing loss: PTA (BC) < 15 db HL and corresponding average ABG > 15 db; 3) mixed hearing loss: PTA (BC) > 15 db HL and corresponding average ABG > 15 db; 4) pure sensorineural hearing loss: PTA (AC) > 15 db HL and corresponding average ABG < 15 db; and 5) deafness: PTA (AC) > 120 db HL. The severity of the hearing loss was established as mild (15 db HL PTA [AC] < 40 db HL), moderate (40 db HL PTA [AC] < 70 db HL), severe (70 db HL PTA [AC] < 95 db HL), or profound (95 db HL PTA [AC]). Asymmetric hearing loss was defined as more than 10 db difference between the ears for at least two frequencies with a difference in PTA between the ears of more than 20 db. The CT images originated from eight different radiology centers, using different CT systems. Axial CT images with a bone window setting were acquired in all patients. Slice thickness ranged from 0.1 to 1.2 mm. Coronal 0.1- to 1.5-mm-thick reformatted images were made in all but one patient. The MR images of the brain and inner ear structures in four patients were performed in three different radiology centers. Axial 4.0- to 5.0-mm-thick two-dimensional turbo-spin-echo (TSE) T2-weighted images of the cerebrum were acquired, as well as axial 0.4- to 1.0-mm-thick three-dimensional TSE or gradientecho (GE) T2-weighted images through the inner ear and internal auditory canal. In addition, axial and coronal 0.6- to 3.0- mm-thick images (submillimeter 3-dimensional T1-weighted TSE or GE images or thicker two-dimensional spin-echo or TSE images) before and after gadolinium contrast administration were acquired in these four patients. All CT and MR images were evaluated by the same radiologist (J.C.) experienced in head and neck pathology, who was blinded to the hearing levels. Different significant anatomic temporal bone regions were explored in detail on the CT images. Fenestral structures, including the fissula ante fenestram, oval window, and round window, were separately evaluated on the presence of hypodense foci. The occurrence and location of hypodense lesions at retrofenestral temporal bone structures were determined as well. These structures included the cochlear turns, semicircular canals, and facial nerve canal. In addition, retrofenestral temporal bone structures were evaluated on the presence of a hypodense focus anteroinferior to the fundus of the internal auditory canal (HAIIAC), an important early indicator of retrofenestral involvement, and on the appearance of a hypodense band around the cochlea, also called the fourth turn of the cochlea or the double ring sign Finally, abnormalities of the middle ear ossicles and mastoid were traced. All audiometric data and radiologic observations were entered into SPSS, version 19.0 (IBM Statistics, Chicago, IL) to evaluate whether occurrence, type, and degree of hearing loss were associated with the occurrence and location of hypodense foci on CT images of the temporal bones. RESULTS Audiometric Evaluation An overview of the demographic, audiologic, and radiologic data in our 17 OI patients is provided in Table I. At the time of CT evaluation, all OI patients demonstrated hearing loss, which manifested bilaterally in all but two of them. In 10 ears, stapes surgery had been performed before CT evaluation. Short- and longterm results of surgery in these ears have been described previously. 11,23 In the 33 evaluated OI ears, normal hearing (6.1%), conductive hearing loss (12.1%), mixed hearing loss (60.6%), sensorineural hearing loss (6.1%), and deafness (15.1%) were observed. In the five deaf ears, retrieval of audiometric data from the past revealed that mixed hearing loss had preceded the deafness. The hearing loss severity varied from mild to profound. Two patients without history of stapes surgery demonstrated asymmetric hearing loss. Radiologic Evaluation Fenestral bony changes were radiologically confirmed in 26 of 33 ears (78.9%). These fenestral changes are illustrated in Figure 1 (patient 1) and included the presence of hypodense bone in the fissula ante fenestram in 25 of 33 ears (75.8%), a thickened, hypodense stapes footplate at the level of the oval window in 23 of 33 ears (69.7%), and a narrowed or obliterated round window in 20 of 33 ears (60.6%). Involvement of the round window always coincided with oval window narrowing or obliteration. Retrofenestral hypodense foci were identified in 21 of 33 ears (63.6%), all of which showed fenestral involvement as well. An HAIIAC (Fig. 2; patient 17) was seen in 16 of 33 ears (48.5%), and the double ring sign was seen in 11 of 33 ears (33.3%). Otospongiotic irregularities 1989

3 TABLE I. Demographic, Audiologic, and Radiologic Data From 17 Osteogenesis Imperfecta Patients. Audiologic Data CT Data Patient No. Age (yr) Ear AC BC Hearing Loss Type Fenestral Hypodensities Retrofenestral Hypodensities 1 9 R 29 6 Conductive FAF, OW, RW L Mixed FAF, OW, RW HAIIAC 2 21 R Mixed FAF, OW, RW CoT, FNC, HAIIAC, 4th turn L Mixed FAF, OW, RW CoT, FNC, HAIIAC, 3 28 R Mixed FAF L Sensorineural 4 28 R 33 9 Conductive L 19 3 Conductive 5 28 R Mixed FAF, OW, RW CoT, HAIIAC, 4th turn L Mixed FAF CoT, HAIIAC, 4th turn 6 37 R Mixed FAF, OW, RW CoT, FNC, HAIIAC, 4th turn, SCC L Sensorineural (after SS) FAF, OW, RW CoT, FNC, HAIIAC, SCC 7 37 R Mixed FAF, OW, RW CoT, FNC, 4th turn L Mixed FAF, OW, RW CoT, FNC, 4th turn 8 44 R Mixed (after SS) FAF, OW, RW CoT, HAIIAC L Mixed (after SS) FAF, OW, RW CoT, FNC, HAIIAC 9 47 R Mixed FAF L Mixed FAF R Mixed (after SS) L Mixed R Conductive FAF, OW L 11 8 Normal (after SS) R Mixed (after SS) FAF, OW, RW CoT, FNC, HAIIAC L Mixed FAF, OW, RW HAIIAC R Normal OW L Mixed FAF, OW HAIIAC R 120 Deafness FAF, OW, RW CoT, fundus, SCC, 4th turn L 120 Deafness (after SS) FAF, OW, RW CoT, fundus, 4th turn R 120 Deafness (after SS) FAF, OW, RW CoT, SCC, 4th turn R Mixed (after SS) OW, RW SSC, HAIIAC L Mixed FAF, OW, RW HAIIAC R 120 Deafness (after SS) FAF, OW, RW CoT, fundus, FNC, SCC, HAIIAC, 4th turn L 120 Deafness FAF, OW, RW CoT, FNC, HAIIAC, 4th turn AC 5 average air conduction threshold in db HL calculated from thresholds at 0.5, 1.0, 2.0, and 4.0 khz; BC 5 average bone conduction threshold in db HL calculated from thresholds at 0.5, 1.0, 2.0, and 4.0 khz; CoT 5 cochlear turns; CT 5 computed tomography; FAF 5 fissula ante fenestram; FNC 5 facial nerve canal; HAIIAC 5 hypodensity anteroinferior to the internal auditory canal; L 5 left; OW 5 oval window; R 5 right; RW 5 round window; SCC 5 semicircular canals; SS 5 stapes surgery. in the basal, middle, and apical cochlear turns were perceived in 11 ears (33.3%), all of which showed at least 15 db sensorineural loss. Spongy, hypodense bone occupied the intracochlear scalae in five deaf ears (15.2%). Axial and coronal temporal bone CT images from a bilaterally deaf OI patient are presented in Figure 3. The posterior or lateral semicircular canals were affected in six ears (18.2%). Hypodensities were seen around the facial nerve canal in 10 ears (30.3%). Nonenhanced and gadolinium-enhanced MR images of the temporal bones were normal in both ears from patient 11, with normal hearing in the left ear after stapes surgery and a moderate conductive hearing loss in the right ear. CT images only showed fenestral changes in the right ear and an accurately positioned piston in the left ear. In patient 13, who presented with unilateral severe mixed hearing loss, gadolinium-enhanced MR images demonstrated unilateral enhancement around the left cochlea, compatible with the retrofenestral hypodensities seen on CT. Finally, in four ears originating from two patients with bilateral deafness (patients 14 and 17), bilateral enhancement was obtained on gadolinium-enhanced MR images, compatible with active otospongiosis (Fig. 4; patient 17). 1990

4 was positioned unusually close to the incudostapedial joint, which probably was a sequel of chronic otitis media. Furthermore, we discovered an enlarged vestibular aqueduct bilaterally, which was more prominent on the right side, and a superior canal dehiscence on the left side (Fig. 5). Both pathologies might have been the underlying cause for a pseudoconductive hearing loss in this patient, who was the only type III OI patient in this series. In three patients with mixed hearing loss, relatively thick slices were used for temporal-bone CT imaging, which may explain the discrepancies between the audiograms suggesting mixed hearing loss and the underrepresentation of hypodense foci in the temporal bones. Three ears (patients 3 and 9) showed fenestral changes only. In both ears of patient 10, with mixed hearing loss, no fenestral or retrofenestral hypodensities were observed. Both fenestral and retrofenestral changes were observed on CT images from the left ear of patient 6, for Fig. 1. Fenestral focus in an osteogenesis imperfecta patient with moderate, mixed hearing loss in the right ear (patient 1). Axial images through the right temporal bone show an occlusive, thickened stapes footplate (arrowhead in A), a focus of hypodense bone in the fissula ante fenestram (arrow in A), and obliterative hypodense changes at the level of the round window (arrowhead in B). Correlation Between Hearing Loss and Hypodensities on CT The occurrence of fenestral or retrofenestral hypodensities corresponded to the expectations based on the type of the hearing loss in 24 of 33 ears (72.7%), in that fenestral involvement was associated with a conductive hearing loss component and a combination of fenestral and retrofenestral involvement with both a conductive and a sensorineural hearing loss component (Table I). In nine of 33 ears (27.3%), the findings of CT evaluation of the temporal bones were not in agreement with the audiograms. CT images in one ear associated with a normal audiogram revealed a thickened stapes footplate. This ear was from patient 13, who presented with severe, mixed hearing loss in the contralateral ear, associated with a thickened stapes footplate and a hypodensity affecting the fissula ante fenestram. In two ears presenting with mild conductive hearing loss (patient 4), no otospongiotic alterations were detected. In the right ear, the manubrium of the malleus Fig. 2. Retrofenestral hypodensity anteroinferior to the fundus of the internal auditory canal. Axial computed tomography scans through the right (A) and left (B) temporal bone of patient 13 with normal hearing on the right and unilateral severe, mixed hearing loss on the left side show a hypodense focus located anteroinferior to the fundus of the internal auditory canal on the left side (arrow in B) but not on the right side. 1991

5 Fig. 3. Fenestral and retrofenestral changes in a bilaterally deaf osteogenesis imperfecta patient (patient 17). Axial (A, B) and coronal reformatted (C, D) computed tomography images of the right (A, C) and left (B, D) ear demonstrate fenestral changes (single white arrow in A) and pericochlear ringlike hypodensities (white arrowheads in A D). Otospongiotic alterations of the right horizontal semicircular canal are observed in image A (gray arrow). whom the audiogram demonstrated a pure sensorineural hearing loss. The observation of a stapes prosthesis disclosed a history of stapes surgery for the reduction of a conductive hearing loss component in this ear. Table II summarizes the involvement of the different middle ear structures as a function of the average ABG. The ABG showed a significant positive association with the number of affected fenestral structures (r s ; P <.05; Spearman rank correlation coefficient). No direct association was observed between the magnitude of the ABG and a specific structure being affected, as hypodensities affecting the fissula ante fenestram as well as the oval and round window were seen in ears with both small and large ABGs. The radiographic involvement of the different retrofenestral structures in relation to the magnitude of the average BC threshold is presented in Table III. Each retrofenestral structure may already be affected by hypodense bone in ears with only slightly elevated BC thresholds. Nevertheless, a significant positive correlation was observed between the magnitude of the average BC threshold and the number of affected retrofenestral structures (r s ; P <.01; Spearman rank correlation coefficient). In the majority of the OI patients, bilateral symmetric involvement was observed on the CTs, which was in agreement with the observation of bilateral symmetric hearing loss on the audiograms. However, two nonoperated patients included in the present study demonstrated asymmetric hearing loss. In patient 3, with moderate mixed hearing loss in the right ear and mild sensorineural hearing loss in the opposite ear, CT images revealed a hypodense lesion in the fissula ante fenestram on the side of the mixed hearing 1992 loss. No other abnormalities could be detected on either side. The second, nonoperated patient with asymmetric hearing loss (patient 13) showed severe, mixed hearing loss in the left ear and normal thresholds in the right ear. CT images demonstrated a thickened footplate bilaterally as well as fenestral and pericochlear hypodensities in the left ear. MR images showed a high signal intensity assuming active pericochlear otospongiosis on the left side. Fig. 4. Magnetic resonance image of the temporal bone in a bilaterally deaf osteogenesis imperfecta patient (patient 17). Coronal submillimeter gadolinium-enhanced three-dimensional turbo-spinecho T1-weighted image through the inner ear shows enhancement around the cochlea on the left side and around the fundus of the internal auditory canal on the right side (arrowheads), pointing toward active otospongiotic lesions.

6 Fig. 5. Enlarged vestibular aqueduct and dehiscence of the superior semicircular canal in an osteogenesis imperfecta type III patient (patient 4) presenting with bilateral conductive hearing loss. Axial computed tomography (CT) images of the right (A) and left (B) temporal bone show an enlarged vestibular aqueduct on both sides (white arrows), more prominent on the right side. (C) Axial CT image of the left temporal bone shows a dehiscent superior semicircular canal (white arrowhead). DISCUSSION Hearing loss is common in the adult OI population. Usually the hearing loss develops as a conductive hearing loss in the second to fourth decade of life and evolves to a mixed hearing loss. It is often associated with a fixation of the stapes footplate due to a process of abnormal bone remodeling in the temporal bone. As this process progresses with time, the pericochlear bone and other temporal bone structures may be affected as well. Temporal bone characteristics in hearing-impaired OI patients show similarities with those found in patients with classic OS, as has been reported before on the basis of temporal bone imaging in a number of OI cases Otosclerosis refers to a disease unique to the otic capsule and surrounding bone, characterized by abnormal remodeling of temporal bone structures, which normally undergo very little remodeling. In the active, otospongiotic stage, dense pericochlear and perilabyrinthine bone adjacent to the oval and round window is replaced by highly vascular and cellular woven bone. 28 The latter appear as hypodense foci on CT images. The active stage may be followed by intermediate, and inactive, otosclerotic phases. Inactive lesions may escape detection by CT, because their density might approximate that of the surrounding bone. 29 In OS, the fissula ante fenestram is suggested to be the site of predilection for the hypodense lesions causing conductive hearing loss, subsequently proliferating and affecting the oval window region, resulting in a thickened and fixed stapes footplate. 30 When the retrofenestral otic capsule structures become involved, irreversible sensorineural hearing loss may arise owing to otospongiotic or otosclerotic bone interfering with the normal function of the cochlea. Consistent with the finding that the majority of hearing-impaired OI patients have mixed hearing loss, it has been suggested that the labyrinthine bone is more frequently involved in the otospongiotic process in OI patients compared to OS patients and that its involvement is more severe and more symmetric in OI ,31,32 In the present study, the type of hearing loss corresponded to the location of the hypodensities on temporal bone CT in 72.7% of the OI ears. Fenestral hypodensities, referring to a thickened footplate, round-window narrowing or obliteration, and alterations in the fissula ante fenestram, were diagnosed in 78.9% of all evaluated OI ears and in 82.8% of the ears with an ABG of at least 15 db or deafness. Retrofenestral involvement was identified on CT images in addition to fenestral changes in 80.0% of the ears with mixed hearing loss or deafness. Tabor et al. 25 reported facial nerve paralysis in an OI patient due to narrowing of the facial nerve canal by dysplastic bone. Although otospongiotic involvement of the facial nerve canal is detected by CT in a number of patients in the present study, facial nerve paralysis was not observed in these patients and is estimated to be rare in OI. In OI patients with deafness, hypodense foci are observed adjacent to all turns of the otic capsule, sometimes in association with calcifications that are obliterating or narrowing the basal cochlear turn. Gadolinium-enhanced MR images showed pericochlear enhancement in two deaf patients, indicating the presence of active otospongiotic lesions. Complete obstruction of the cochlear lumen or total loss of definition of the 1993

7 TABLE II. Fenestral Temporal Bone Involvement With Regard to the Magnitude of the Average Air-Bone Gap in 28 Ears From Osteogenesis Imperfecta Patients. ABG < 15 db 15 db ABG 25 db 25 db < ABG 35 db 35 db < ABG 45 db 45 db < ABG Total Fissula ante fenestram 2/4 6/9 4/7 4/4 4/4 22/28 Oval window 2/4 3/9 6/7 4/4 3/4 18/28 Round window 1/4 3/9 5/7 3/4 3/4 15/28 Air-bone gap determined by subtracting the average bone conduction thresholds from the average air conduction thresholds at frequencies 0.5, 1.0, 2.0, and 4.0 khz. Five ears with undeterminable air-bone gaps were excluded. ABG 5 air-bone gap. cochlear architecture has previously been reported in a few deaf OI patients 33,34 but was not observed in the present study. Therefore, our deaf patients were estimated to be good candidates for cochlear implantation. Besides extreme otospongiotic involvement of the cochlea, high vascularity and brittleness of the bone structures inherent to the disease render cochlear implantation in OI patients a challenging procedure for the otologic surgeon. Nevertheless, a number of successful implantations in OI patients with profound hearing loss or deafness have been reported. 33,35,36 As the conductive and sensorineural hearing loss components progress, more fenestral and retrofenestral temporal bone structures are affected by hypodense lesions, respectively. However, direct relationships between the affected fenestral or retrofenestral structures and the ABG or average BC threshold cannot be established, as each of these structures may already be involved when the ABG or BC thresholds, respectively, are relatively close to normal. Discrepancies between the hearing loss type determined by audiometry and the observation or location of hypodense foci were found in 27.3% of the evaluated OI ears. In some of them, the hearing loss might have been related to temporal bone pathologies other than otospongiosis-like changes, such as an enlarged vestibular aqueduct or a dehiscent semicircular canal. In patient 4, who presented with bilateral conductive hearing loss, CT images revealed no hypodense foci, but they did show a number of other peculiarities, such as a deviating position of the ossicles, which might be a sequel of chronic otitis media, an enlarged vestibular aqueduct bilaterally, and a superior canal dehiscence. Because tympanometry and acoustic stapedial reflex measurements in this patient showed reduced admittance values and absent reflexes bilaterally, it remains uncertain whether the hearing loss results from ossicular displacement only or ossicular displacement in combination with a third mobile window. Because of the inability to detect hypodense foci with submillimeter CT imaging, the hearing loss was not suspected to be related to the typical ossicular fixation in OI. However, precise follow-up was recommended in this patient. In a few other patients, CT images did not reveal any abnormality but were obtained with poor spatial resolution. It is advised to apply high-resolution, multidetector or cone-beam CT to obtain multiple slices that are adequate for detection of small alterations in the temporal bone or stapes footplate and to avoid any false-negative findings. As there is a partial overlap between the radiologic characteristics of the temporal bone in OI and OS, temporal bone imaging is not of decisive importance for the diagnosis of OI, which should rely rather on the identification of a mutation in COL1A1 or COL1A2 and on clinical features. The surplus value of temporal bone imaging in hearing-impaired OI patients lies in confirming the diagnosis of OI, identifying the underlying cause for hearing loss, determining the extent of the temporal bone pathology, and analyzing the cause for unsatisfactory hearing gain after surgery. Before deciding on stapes surgery, we may perform temporal bone imaging to assist in the detection of potential round-window obliteration, which may substantially impede the intended hearing gain. Furthermore, when stapes surgery or cochlear implantation is considered, CT and MR may be helpful to anticipate possible surgical complications or to rule out the possibility that hearing loss is caused by pathologies other than otosclerotic bone TABLE III. Retrofenestral Temporal Bone Involvement With Regard to the Magnitude of the Average Bone-Conduction Thresholds in 33 Ears From Osteogenesis Imperfecta Patients. BC < 15 db HL 15 db HL BC 25 db HL 35 db HL < BC 45 db HL 45 db HL < BC Total Cochlear turns 0/6 6/13 4/7 6/7 16/33 Facial nerve canal 0/6 4/13 3/7 3/7 10/33 Semicircular canals 0/6 2/13 1/7 3/7 6/33 HAIIAC 0/6 9/13 5/7 2/7 16/33 Fourth turn 0/6 4/13 1/7 6/7 11/33 Average bone-conduction threshold at frequencies of 0.5, 1.0, 2.0, and 4.0 khz. BC 5 bone conduction; HAIIAC 5 hypodensity anteroinferior to the fundus of the internal auditory canal. 1994

8 formation, such as temporal bone fracture, semicircular canal dehiscence, enlarged vestibular aqueduct, and so on. Although a semicircular canal dehiscence was observed in two patients and an enlarged vestibular aqueduct was seen in one patient in the present study, it remains unclear whether there is a higher incidence of these pathologies in OI. Such an association should be investigated in a larger sample of OI patients using high-quality CT with submillimeter spatial resolution. CONCLUSION Temporal bone CT imaging in hearing-impaired OI patients reveals pathologic lesions of reduced bone density, whereas gadolinium-enhanced MR images demonstrate foci of high metabolic activity. The extent of the hypodense foci on temporal bone CT corresponds to the type of the hearing loss, as conductive hearing loss is associated with fenestral changes, and mixed hearing loss is associated with both fenestral and retrofenestral hypodensities. The larger the ABG and the average BC threshold, the more fenestral and retrofenestral structures involved, respectively. The use of a submillimeter spatial resolution when performing CT imaging of the temporal bone is recommended to increase the ability to detect small hypodense foci or other abnormalities that substantially contribute to the origin of the hearing loss and to the decision whether surgery should be performed. ACKNOWLEDGMENTS The authors express their gratitude to all the participating patients as well as to their otolaryngologists and radiologists providing audiometric and radiologic data. BIBLIOGRAPHY 1. Sillence DO, Senn A, Danks DM. Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 1979;16: Forlino A, Cabral WA, Barnes AM, Marini JC. New perspectives on osteogenesis imperfecta. Nat Rev Endocrinol 2011;7: Hartikka H, Kuurila K, Korkko J, et al. Lack of correlation between the type of COL1A1 or COL1A2 mutation and hearing loss in osteogenesis imperfecta patients. Hum Mutat 2004;24: Pedersen U. Hearing loss in patients with osteogenesis imperfecta. A clinical and audiological study of 201 patients. Scand Audiol 1984;13: Garretsen TJ, Cremers CW. Clinical and genetic aspects in autosomal dominant inherited osteogenesis imperfecta type I. Ann N Y Acad Sci 1991;630: Pedersen U, Elbrond O. Surgical findings and results of stapedectomy in patients with osteogenesis imperfecta. J Laryngol Otol 1979;93: Vincent R, Gratacap B, Oates J, Sperling NM. Stapedotomy in osteogenesis imperfecta: a prospective study of 23 consecutive cases. Otol Neurotol 2005;26: Shea JJ, Postma DS. Findings and long-term surgical results in the hearing loss of osteogenesis imperfecta. Arch Otolaryngol 1982;108: Garretsen TJ, Cremers CW. Ear surgery in osteogenesis imperfecta. Clinical findings and short-term and long-term results. Arch Otolaryngol Head Neck Surg 1990;116: van der Rijt AJ, Cremers CW. Stapes surgery in osteogenesis imperfecta: results of a new series. Otol Neurotol 2003;24: Swinnen FK, De Leenheer EM, Coucke PJ, Cremers CW, Dhooge IJ. Audiometric, surgical, and genetic findings in 15 ears of patients with osteogenesis imperfecta. Laryngoscope 2009;119: Kuurila K, Kaitila I, Johansson R, Grenman R. Hearing loss in Finnish adults with osteogenesis imperfecta: a nationwide survey. Ann Otol Rhinol Laryngol 2002;111: Pillion JP, Shapiro J. Audiological findings in osteogenesis imperfecta. J Am Acad Audiol 2008;19: Altman F. The temporal bone in osteogenesis imperfecta congenita. A contribution to the question of the etiology of otosclerosis. Arch Otolaryngol 1962;75: Bretlau P, Jorgensen MB. Otosclerosis and osteogenesis imperfecta. Arch Otolaryngol 1969;90: Berger G, Hawke M, Johnson A, Proops D. Histopathology of the temporal bone in osteogenesis imperfecta congenita: a report of 5 cases. Laryngoscope 1985;95: Zajtchuk JT, Lindsay JR. Osteogenesis imperfecta congenita and tarda: a temporal bone report. Ann Otol Rhinol Laryngol 1975;84: Sando I, Myers D, Harada T, Hinojosa R, Myers EN. Osteogenesis imperfecta tarda and otosclerosis. A temporal bone histopathology report. Ann Otol Rhinol Laryngol 1981;90: Nager GT. Osteogenesis imperfecta of the temporal bone and its relation to otosclerosis. Ann Otol Rhinol Laryngol 1988;97: Kawase S, Naganawa S, Sone M, Ikeda M, Ishigaki T. Relationship between CT densitometry with a slice thickness of 0.5 mm and audiometry in otosclerosis. Eur Radiol 2006;6: Valvassori GE, Dobben GD. CT densitometry of the cochlear capsule. AJNR Am J Neuroradiol 1985;6: Tringali S, Bertholon P, Pouget JF, et al. Cochlear otosclerosis and pseudo fourth turn of the cochlea. Ann Otolaryngol Chir Cervicofac 2004;121: Swinnen FK, De Leenheer EM, Coucke PJ, Cremers CW, Dhooge IJ. Stapes surgery in osteogenesis imperfecta: retrospective analysis of 34 operated ears. Audiol Neurootol 2012;3: Alkadhi H, Rissmann D, Kollias SS. Osteogenesis imperfecta of the temporal bone: CT and MR imaging in Van der Hoeve-de Kleyn syndrome. AJNR Am J Neuroradiol 2004;25: Tabor EK, Curtin HD, Hirsch BE, May M. Osteogenesis imperfecta tarda: appearance of the temporal bones at CT. Radiology 1990;175: Ross UH, Laszig R, Bornemann H, Ulrich C. Osteogenesis imperfecta: clinical symptoms and update findings in computed tomography and tympano-cochlear scintigraphy. Acta Otolaryngol 1993;113: Heimert TL, Lin DD, Yousem DM. Case 48: osteogenesis imperfecta of the temporal bone. Radiology 2002;224: Mafee MF, Valvassori GE, Deitch RL, et al. Use of CT in the evaluation of cochlear otosclerosis. Radiology 1985;156: Guneri EA, Ada E, Ceryan K, Guneri A. High-resolution computed tomographic evaluation of the cochlear capsule in otosclerosis: relationship between densitometry and sensorineural hearing loss. Ann Otol Rhinol Laryngol 1996;105: Mafee MF, Henrikson GC, Deitch RL, et al. Use of CT in stapedial otosclerosis. Radiology 1985;156: d Archambeau O, Parizel PM, Koekelkoren E, Van de HP, De Schepper AM. CT diagnosis and differential diagnosis of otodystrophic lesions of the temporal bone. Eur J Radiol 1990;11: Jardin C, Ghenassia M, Vignaud J. Tomographic and CT features of the petrous bone in Lobstein s disease. One case and a review of the literature. J Neuroradiol 1985;12: Rotteveel LJ, Beynon AJ, Mens LH, Snik AF, Mulder JJ, Mylanus EA. Cochlear implantation in 3 patients with osteogenesis imperfecta: imaging, surgery and programming issues. Audiol Neurootol 2008;13: Szilvassy J, Jori J, Czigner J, Toth F, Szilvassy Z, Kiss JG. Cochlear implantation in osteogenesis imperfecta. Acta Otorhinolaryngol Belg 1998;52: Streubel SO, Lustig LR. Cochlear implantation in patients with osteogenesis imperfecta. Otolaryngol Head Neck Surg 2005;132: Huang TS, Yen PT, Liu SY. Cochlear implantation in a patient with osteogenesis imperfecta and otospongiosis. Am J Otolaryngol 1998;19: