Diagnostic imaging

Diagnostic imaging

Otolaryngol Clin N Am 35 (2002) 239–253 Diagnostic imaging Dennis G. Pappas Jr, MDa,*, Joel K. Cure´, MDb a Pappas Ear Clinic, 2937 7th Avenue South...

95KB Sizes 0 Downloads 59 Views

Otolaryngol Clin N Am 35 (2002) 239–253

Diagnostic imaging Dennis G. Pappas Jr, MDa,*, Joel K. Cure´, MDb a

Pappas Ear Clinic, 2937 7th Avenue South, Birmingham, AL 35233, USA Neuroradiology Section, Department of Radiology, University of Alabama Medical Center, 619 19th Street South, Birmingham, AL 35249, USA


Although advancements in imaging technology have simplified the diagnosis of pathologic conditions of the temporal bone and cerebellopontine angle (CPA), clinicians face the growing challenge of learning a constantly evolving technology. Each year seems to bring something diagnostically new and exciting. No longer are traditional MR imaging and high-resolution CT (HRCT) the latest, best studies for every temporal bone disorder. As the list of available studies expands, diagnostic choices must continue to complement clinical suspicions while remaining responsible to both potential risk and cost. Clinicians must also acquire the skills to interpret these new studies, an ability that comes with practice and experience. Today’s otolaryngologist should continue to be the expert on temporal bone anatomy and pathology. Many clinicians have experienced the pleasure and frustration of participating in the training of a radiologist. More commonly, this is a give-and-take relationship in which each party benefits. This educational process is greatly enhanced by regularly scheduled film-review sessions, in which knowledge can be shared and the latest technology demonstrated. This article illustrates the newest imaging applications and strategies for the evaluation of neuro-otologic disorders. It is hoped that this article will temporarily fill a diagnostic niche until something more advanced comes along.

General imaging principles Before discussing newer technology, conventional MR imaging and HRCT basics are reviewed and frequently asked questions addressed. Traditional MR imaging usually consists of nonenhanced T1- and T2-weighted

* Corresponding author. E-mail: [email protected] (D.G. Pappas). 0030-6665/02/$ - see front matter Ó 2002, Elsevier Science (USA). All rights reserved. PII: S 0 0 3 0 - 6 6 6 5 ( 0 2 ) 0 0 0 1 1 - 7


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

imaging, followed by a gadolinium-enhanced T1-weighted study. Nonenhanced T1-weighted imaging, in axial and coronal planes, identifies bone marrow, fat, and subacute hemorrhage as high (bright) signal areas. Gadolinium-enhanced T1-weighted imaging in the same planes demonstrates enhancement when a neoplastic, vascular, or inflammatory process is present. Standard T2-weighted studies identify intra-axial disease, such as brainstem tumor, stroke, and multiple sclerosis. One also should be familiar with general MR imaging signal characteristics: fat is brighter on T1-weighted imaging, infection or inflammation intensifies from T1- to T2-weighted imaging, and cerebrospinal fluid (CSF) turns from black to white from T1- to T2-weighted imaging. Specific MR imaging scanning protocols continue to evolve and, therefore, are beyond the scope of this article. With regard to MR imaging safety, virtually all ossicular prosthetics and ventilation tubes are nonferromagnetic. The only exception is the McGee platinum stainless steel piston (Medtronic Xomed Surgical Products, Inc., Jacksonville, FL) produced in 1987. Contraindications to MR imaging include pacemakers and aneurysm clips that are ferromagnetic. The newer aneurysm clips are generally safe, but it is important to be absolutely sure of the clip identity before scanning. Patients who have been welders should undergo prescanning orbital radiography to rule out the presence of intraocular metallic fragments. In pregnancy, gadolinium has been shown to cross the placenta and even appear in the fetal bladder during MR imaging. Because the effects of gadolinium on the developing fetus are unknown, its use cannot be recommended during pregnancy [1]. MR imaging compatibility for cochlear implants varies according to the implant and manufacturer. The Nucleus 24 device (Cochlear Corp., Englewood, CO) contains a magnet that can be surgically removed for MR imaging. The Nucleus 24 (Cochlear Corp., Englewood, CO) can reportedly withstand a 1.5 Tesla imager [2]. MR imaging is contraindicated for the Nucleus 22 device (Cochlear Corp.). MR imaging cannot currently be recommended for the Clarion 1.2 device (Advanced Bionics, Sylmar, CA), although a recent study suggested 0.3 Tesla imager poses little to no risk [3]. Med EL (Durham, NC) cochlear implants are reportedly compatible with the MR imaging with a 1.0 Tesla imager [4]. In general, MR imaging examination of patients with a cochlear implant should be performed only if there is a strong medical indication. The manufacturer of the implant should be contacted before any MR imaging study is performed. Temporal bone HRCT is typically obtained in the axial and coronal planes, thin-section, 1-mm, contiguous slice thickness. To provide maximum detail, each temporal bone image can be reconstructed and magnified using software that emphasizes bony detail. Typically, the entire temporal bone is scanned from the tegmen tympani to the mastoid tip on axial imaging and from the carotid canal through the mastoid cell system on coronal imaging. Intravenous contrast is usually unnecessary because more information can be gleamed from gadolinium-enhanced MR imaging when needed.

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


Technology update High-resolution MR imaging MR imaging technology has seen significant advancement in recent years. With the development of HRMR imaging, detailed imaging of the membranous labyrinth and IAC can be achieved. Using various fast-spin or gradient echo techniques, these heavily T2-weighted studies provide high (bright) fluid signals so that even CSF pulsation artifacts are muted. CSF appears stark-white, providing a sharp contrast between CSF and adjacent soft tissue structures. The fluid signal of the membranous labyrinth also appears bright, allowing for similar high-resolution examination of labyrinthine anatomy and abnormalities. Axial, 0.7-mm, thin-section, three-dimensional Fourier transformation constructive interference in steady state (3DFT-CISS) imaging is one of the newest HRMR imaging applications. This thin-section study is obtained in a single plane. The images can then be reformatted into various planes, which allows for separate identification of each nerve within the IAC. In cases of small intracanalicular vestibular schwannomas, the nerve of origin can often be identified [5]. Thus, decisions regarding hearing-preservation surgery could be influenced. The signal intensity of intralabyrinthine fluid and that of intracanalicular CSF (between the vestibular schwannoma and the fundus) has also been identified as a predictive factor in hearing-preservation surgery with 3DFT-CISS MR imaging [6]. HRMR imaging screening of the IAC remains controversial. Equal sensitivity to traditional gadolinium-enhanced T1-weighted MR imaging has been claimed in the identification of vestibular schwannomas [7,8]. The advantages of HRMR imaging are shorter scanning time, the absence of gadolinium injection, and lower cost. HRMR imaging is also superior to traditional MR imaging in that it demonstrates labyrinthine dysplasias in the workup of sensorineural hearing loss [7]. Disadvantages include a limited screening field and the inability to demonstrate inflammatory lesions [9]. Many centers have incorporated HRMR imaging of the IAC into, rather than in place of, their standard MR imaging protocol. HRMR imaging has replaced HRCT as the preferred preoperative imaging study before cochlear implantation [10]. HRMR imaging can verify cochlear patency as well as confirm the presence and size of the cochlear nerve. This study is especially beneficial if there is a history of meningitis and subsequent cochlear scarring is suspected. HRCT adequately demonstrates bony cochlear obstruction but is less consistent with fibrous obstruction. Functional imaging Functional brain imaging allows for the measurement of cortical activity in response to specific stimuli. Three examples illustrate this recent imaging strategy: fMR imaging, SPECT, and positron emission tomography


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

(PET). The neuro-otologic value of functional studies has been demonstrated in preoperative and postoperative evaluation of cochlear implant candidates and recipients. fMR imaging has demonstrated benefit as a selection modality in implant cases (patient candidacy and side of implantation) [11]. Thus far, fMR imaging results appear to correlate with clinical and audiologic findings. Compared with SPECT and PET, fMR imaging has the advantage of better spatial resolution, being noninvasive, and the potential for recording images in real time. Without a need for contrast enhancement, fMR imaging takes advantage of the changing signal within blood and surrounding tissue of the auditory center. Specifically, the auditory pathway is stimulated with the patient in the scanner. This auditory activity leads to increased neuronal metabolism, subsequent vasodilatation, and increased blood flow. An increase in delivered oxygen then exceeds metabolic demands, leading to an increased oxyhemoglobin-to-deoxyhemoglobin ratio, which results in an increased signal in activated areas of the auditory center. PET and SPECT studies have also demonstrated significant differences in response between the auditory centers in cochlear implant candidates [12]. Post-implant PET and SPECT studies have made implant performance evaluation possible as well. These studies have allowed for more precise mapping of the auditory response, comparison to normal hearing patient pathways, and the evaluation of various speech-coding strategies [13,14]. Of course, implant performance cannot be measured with MR imaging due to ferromagnetic concerns. Imaging strategies Cerebellopontine angle and internal auditory canal Imaging of the CPA is typically obtained as a screen for neoplastic disease in the workup of sensorineural hearing loss. Although tumors of the CPA can be diverse in origin, vestibular schwannomas are overwhelmingly most common, representing more than 90% of all CPA lesions [15]. Because of this prevalence, most of this section is devoted to the imaging of vestibular schwannomas. Gadolinium-enhanced MR imaging readily demonstrates even the smallest of intracanalicular tumors, 1 mm in size. In fact, a negative study is generally considered reliable in excluding the presence of a vestibular schwannoma. As discussed earlier, recent advances in HRMR imaging are becoming integrated in this evaluation process. The physical imaging characteristics of vestibular schwannomas can vary. These tumors can be entirely intracanalicular, extracanalicular, or both depending on the tumor site of origin, at or near the glial–Schwann cell junction. Larger tumors typically demonstrate both an intracanalicular and extracanalicular component. These larger tumors are commonly centered

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


over the IAC. Vestibular schwannomas can be solid and uniformly enhancing or contain cystic components that may not enhance because cystic fluid is generally an avascular collection. More subtle and detailed findings that further characterize intracanalicular tumors are best imaged with HRMR imaging, with which even the nerve of origin can be discerned. Various methods of measuring the size of the imaged tumor exist. Intracanalicular tumors can be described as such, measured in millimeters or in descriptive terms that convey additional clinical significance. For instance, an intracanalicular lesion can also be described in terms of its relation to the fundus of the IAC. Clinically, this latter description would be more relevant in planning hearing preservation surgery. For larger tumors, inclusion of the intracanalicular component into the overall measurement of the tumor is probably irrelevant and misleading from a treatment standpoint. Extracanalicular dimensions carry greater clinical significance: medial to lateral dimension, maximum diameter, or the American Academy of Otolaryngology-Head Neck Surgery method [16]. The relevance of tumor size can also be conveyed in the description of its relationship to surrounding posterior fossa structures. Is there contact with, or compression of, the brainstem? Is the forth ventricle compressed as would be seen with larger lesions? Individual vestibular schwannoma growth rate can be established through serial MR imaging study. For conservatively managed tumors, initial follow-up studies can be obtained every 4 to 6 months for the first year and then at 1- to 2-year intervals if the tumor size is stable or very slow growing. Reported growth rates vary greatly, with significant growth defined as more than 1 mm per year [17]. Postoperative MR imaging can be obtained 1 year following surgery to rule out residual disease. Scar tissue may enhance but usually can be differentiated on MR imaging from residual tumor. When residual or recurrent lesions are demonstrated, the same serial scanning schedule can be applied. A fat-suppression MR imaging technique can be used for postoperative studies in cases in which the original surgical defect was obliterated with fat. Patients with neurofibromatosis type 2 demonstrate bilateral nerve VIII tumors. The importance of adequately imaging each nerve VIII bundle is stressed on every study. MR imaging can also reveal coexisting meningiomas and schwannomas of other cranial nerves in cases of neurofibromatosis type 2. In addition, MR imaging of the spine should be performed to screen for spinal lesions. Claustrophobic patients can be sedated or scanned with an ‘‘open’’ magnet. When MR imaging is contraindicated, contrast-enhanced HRCT is effective for the diagnosis of tumors greater than 2 cm in diameter. Smaller tumors and intracanalicular tumors can be difficult to identify with HRCT because the enhancing tumor signal can appear contiguous with the surrounding solid bone of the IAC. Thus, negative HRCT studies require serial follow-up scanning or gas-CT cisternography. Gas-CT cisternography is rarely performed because of its associated side effects, high false-positive rate, and the superior


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

sensitivity of gadolinium-enhanced MR imaging. Asymmetric expansion of the IAC is a secondary finding better appreciated on HRCT but is not diagnostically conclusive. Meningiomas generally give the same MR imaging signal characteristics as do vestibular schwannomas, but enhancement of an adjacent dural ‘‘tail’’ is usually present (72%) and an intracanalicular component is commonly absent [18]. Of note, a dural ‘‘tail’’ can also be seen with other tumors, including schwannomas, and, therefore, cannot be considered diagnostic of meningioma. Meningiomas of the CPA typically arise from the broad base of the posterior surface of the petrous bone, often eccentric to the axis of the IAC. Calcifications and associated hyperostosis are more commonly seen with meningiomas and are more readily visible with HRCT. Facial schwannomas make up a small percentage of CPA tumors but deserve mention from an imaging standpoint. These tumors are indiscernible from vestibular schwannomas with regard to MR imaging signal and enhancement characteristics when the tumor is confined to the IAC or CPA. HRMR imaging can distinguish the nerve of origin in smaller tumors. Clinically, a history of facial weakness or paralysis can be significant but not always present or conclusive. The labyrinthine segment of the facial nerve should be examined for asymmetry, enlargement, or enhancement on gadolinium-enhanced MR imaging. Distal nerve enhancement and enlargement can also be seen contiguous with the IAC tumor component or in the form of skip lesions. Here, HRCT can be useful in demonstrating adjacent bony erosion of the labyrinthine or mastoid segments. Schwannomas of cranial nerves V and IX to XII can also involve the CPA. Again, gadolinium-enhanced MR imaging and HRMR imaging are most sensitive. Clinically, any of these tumors can produce otologic symptoms when size is substantial. Thus, close examination of CPA tumor margins is important and usually demonstrates extension of the lesion into Meckel’s cave for a trigeminal neuroma or beneath the skull base toward the jugular foramen for lower cranial nerve tumors. Epidermoids of the CPA generally provide a low (dark) signal on T1weighted MR imaging and a high (bright) signal on T2-weighted imaging. Usually, the signal intensity of this lesion is slightly greater than that of the surrounding CSF. Rare so-called ‘‘white epidermoids’’ demonstrate the opposite signal characteristics on MR imaging. Epidermoids are nonenhancing on MR imaging and CT. Their shape is highly variable, and these lesions can extend into surrounding cisterns, resulting in a dumbbell shape. Arachnoid cysts demonstrate MR imaging characteristics similar to epidermoids but more identical to CSF. Again, there is no enhancement. In contrast to epidermoids, these cysts have a smoother surface and more homogenous consistency. Arachnoid cysts also can surround CPA tumors. Lipomas of the CPA or IAC are readily diagnosed by their characteristic MR imaging appearance. These lesions demonstrate a high (bright) signal

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


on T1-weighted imaging without gadolinium compared with vestibular schwannomas, which fail to demonstrate an intensified signal on the nonenhanced study. Thus, the importance of examining the nonenhanced T1weighted portion of all MR imaging studies is emphasized in cases in which vestibular schwannoma is suspected. T2-weighted imaging provides a lower (darker) signal for lipomas. Fat-suppression MR imaging can also confirm the diagnosis. Lipomas cannot be distinguished from vestibular schwannomas on HRMR imaging. Metastasis to the CPA or IAC is suspected when there is a history of malignancy, facial weakness in the presence of a small lesion, rapidly enlarging lesion seen with serial scanning, or multiple or bilateral disease. Because this is also a neoplastic process, these lesions can appear identical to an intracanalicular vestibular schwannoma. The presence of adjacent bony erosion is highly suspicious. Vascular lesions within the CPA include vascular anomalies, aneurysms, and neoplastic processes. Vascular loops and vertebrobasilar dolichoectasia can compress adjacent cranial nerves. T2-weighted MR imaging typically demonstrates a linear flow void in the area of the vertebrobasilar system. HRMR imaging provides detailed imaging of vascular structures in relation to adjacent cranial nerves and the brainstem. CPA aneurysms can be found incidentally or following a subarachnoid bleed. MR imaging appearance is dependent on aneurysm size and the presence or absence of thrombus. Angiography is diagnostic, but MR angiography has made significant strides in recent years. Hemangiomas of the IAC can be indistinguishable from vestibular schwannoma on contrast-enhanced and HRMR imaging. Clinical history can be suggestive when facial palsy is present with a small lesion. When the bony margins of the IAC are involved, HRCT can demonstrate erosion and possibly a ‘‘honeycomb’’ pattern of calcification. These tumors strongly enhance with contrast. Intra-axial neoplasms originating from the brainstem, cerebellum, or fourth ventricle can extend into the CPA and are readily identified on MR imaging. Although these lesions can present with otologic symptoms, cranial nerve deficits, or both, intra-axial tumors are rarely confused with vestibular schwannoma from an imaging standpoint. Facial nerve Choosing the appropriate imaging modality for evaluation of the facial nerve depends on both the suspected pathologic condition and the anatomic segment in question. When the entire course of the nerve is to be imaged in one setting, traditional MR imaging is the study of choice. The nerve should be imaged from the brainstem through the parotid gland in the screening of a neoplastic process. HRMR imaging is highly accurate within the CPA and IAC; however, inflammatory conditions, such as Bell’s palsy, are not identified with a high-resolution study alone. In addition, HRMR imaging is


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

inadequate for the evaluation of the infratemporal nerve segments. MR angiography can be added in the workup of hemifacial spasm. The role of HRCT is pivotal in the evaluation of the infratemporal facial nerve in most settings. HRCT delineates the relationship of the nerve to adjacent pathologic conditions of the temporal bone and surrounding anatomy, such as the labyrinth and stapedial footplate. Temporal bone trauma is also best evaluated by HRCT. When cervical injury is suspected, the study is limited to axial views. Fractures and ossicular dislocation are readily identified on HRCT. Temporal bone surgical trauma of the facial nerve can be difficult to image because these procedures typically alter or remove adjacent bony anatomy or landmarks. The entire infratemporal course of the facial nerve must be demonstrated with HRCT preoperatively for all atresia cases, in which anterior displacement of the descending nerve segment is common. Although HRCT is not obtained before every ear procedure, the course of the facial nerve should be verified on any scan obtained. Congenital and normal variations of the nerve are readily apparent on HRCT and most commonly involve the distal or descending mastoid segment [19]. The size and appearance of the geniculate region can vary, but it is generally symmetric with the opposite side [20]. When intratemporal neoplastic disease involving the facial nerve is suspected, both HRCT and gadolinium-enhanced MR imaging offer complementary information. A thickened, enhancing region of the horizontal or descending nerve segments on MR imaging is considered pathognomonic for facial nerve schwannoma [21]. Larger tumors can be more difficult to differentiate because surrounding landmarks and structures are destroyed by the disease process. Within the IAC or CPA, facial nerve schwannomas are indistinguishable on traditional MR imaging from vestibular schwannomas. HRMR imaging can differentiate smaller tumors, as mentioned earlier. HRCT can be obtained to examine the surrounding bony characteristics of the disease process. Facial nerve schwannomas are expansile in nature, demonstrating a smooth, distinct erosion of bone. In contrast, bony erosion accompanying hemangiomas is typically less distinct and irregular. Hemangiomas appear more heterogeneous on MR imaging compared with schwannomas. Inflammatory conditions that involve the facial nerve include Bell’s palsy, sarcoid, Lyme disease, and herpes-zoster oticus. Imaging of Bell’s palsy is not necessary in straightforward cases. When clinical presentation is atypical, gradually progressive, or recurrent, then gadolinium-enhanced MR imaging should be obtained to rule out a neoplastic process. Nerve enhancement, without focal enlargement, is often seen on traditional MR imaging in cases of Bell’s palsy. This enhancing pattern can be absent or present for months in duration and involve any or multiple segments but offers no prognostic or clinical significance in cases of Bell’s palsy [22]. Herpeszoster oticus can demonstrate similar MR imaging nerve enhancement. With zoster, nerve VIII and the labyrinth can also enhance. Lyme disease

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


can present with unilateral or bilateral facial nerve palsies. Of note, these enhancing characteristics cannot distinguish one inflammatory condition from the other. HRMR imaging can distinguish inflammatory conditions from a neoplastic process by demonstrating the presence or absence of diameter enlargement of the individual nerve within the IAC. Cerebrospinal fluid leaks Cerebrospinal fluid leaks of the temporal bone can result from a number of causes, including trauma, congenital defect, or spontaneous fistula. Imaging studies attempt to localize the leak site and, in some cases, establish the presence of the leak itself. CSF leaks that complicate a surgical procedure of the temporal bone do not generally require imaging because the site and mechanism of the leak are usually known. Nonsurgical trauma, blunt or penetrating, is generally imaged by HRCT to examine the extent of the associated injury rather than the leak itself. Occult or intermittent leaks require a high index of suspicion, relying solely on imaging for diagnosis in many cases. When these leaks present as CSF rhinorrhea, the entire skull base requires imaging. HRCT remains the most effective study in demonstrating bony dehiscence, erosion, or fracture that could be associated with a leak site. When intrathecal contrast is added to HRCT, the precise location of the leak can be confirmed if the leak is active during the study. Occult or intermittent CSF leaks can, therefore, be difficult to visualize. Suspicious HRCT findings, such as a labyrinthine deformity or a tegmen tympani defect, often direct successful surgical exploration and repair in such cases [23]. Traditional MR imaging has seen limited usefulness in the evaluation of CSF leaks other than quantitating herniated neural tissue. In recent years, MR cisternography has provided a noninvasive alternative to intrathecal contrast HRCT. This heavily T2-weighted fast-spin-echo study with fat suppression and video reversal has proven successful as a diagnostic technique. These images are striking, giving the CSF a stark-black appearance while surrounding structures and soft tissues are faded. Extravasation of CSF into the temporal bone is then apparent. In 1998, El Gammal et al [24] reported a sensitivity of 87% and accuracy of 78% with MR cisternography in 37 consecutive cases of CSF rhinorrhea. Shetty et al [25], also in 1998, demonstrated a higher accuracy rate of 98% in localizing the site of CSF fistula when a combination of MR cisternography and HRCT was performed. In this same study, HRCT alone demonstrated accuracy of 93% and MR cisternography alone demonstrated accuracy of 89%. Petrous apex Because the petrous apex is generally a clinically silent area, diagnostic imaging is often the only means of evaluating potential pathologic conditions


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

of the petrous apex. In many cases, petrous apex disease is discovered incidentally in the scanning workup of an unrelated problem. Together, HRCT and gadolinium-enhanced MR imaging are effective in characterizing the various disease processes involving the petrous apex. MR imaging typically establishes a diagnosis and identifies intracranial extension when present. HRCT is useful in characterizing bony destruction and delineating the relationship of the lesion to surrounding intratemporal structures, such as the carotid artery, labyrinth, and IAC. Review of any MR image or HRCT scan should include inspection and comparison of each petrous apex. In this manner, abnormalities of the petrous apex are typically discovered. Caution must be used to avoid labeling all imaging asymmetry as pathologic. The most common example is when one petrous apex is made of medullary bone, whereas the opposite side is made of pneumatized bone. The side made of medullary bone provides a bright signal on T1-weighted MR imaging because of fat content. Pathologic conditions can be excluded when the T2-weighted study reveals fading of the signal. A fat-suppression technique also rules out pathologic conditions in such cases. Cholesterol granuloma and epidermoid of the petrous apex can often be differentiated with T1-weighted MR imaging. Cholesterol granulomas demonstrate a bright T1-weighted signal, whereas epidermoids typically provide a low (darker) signal. The T2-weighted signal of both lesions is typically bright. Exceptions to this T2-weighted result exist, depending on the protein content of the lesion. The rim or borders of each lesion can enhance, whereas the body of each lesion does not enhance. HRCT typically demonstrates an expansile lesion in both cases. Mucoceles can muddy the differentiation process even further. Usually, mucoceles provide a low (dark) T1weighted signal like an epidermoid, but this signal can be bright when the protein content of the mucocele is high. Mucocele imaging characteristics are otherwise essentially the same as those of cholesterol granulomas and epidermoids. Petrous apicitis is probably the most consistently symptomatic petrous apex process. Although the diagnosis can be made clinically, imaging can play a role in identifying abscess formation within the petrous apex or an associated epidural or brain abscess. MR imaging signal intensifies from the T1- to T2-weighted imaging, and enhancement with gadolinium is intense. The surrounding dura may also enhance. HRCT demonstrates opacification and may or may not reveal bony septa destruction. Petrous carotid aneurysms can remain clinically silent until the aneurysm enlarges to compress surrounding structures or extends into the middle ear. MR imaging findings can be variable depending on turbulent blood flow or the amount of thrombus present. HRCT is reliable in demonstrating erosion or expansion of the bony walls of the petrous carotid canal. MR angiography confirms the diagnosis. When the aneurysm obstructs the eustachian tube, chronic otitis media and cholesteatoma formation can occur

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


secondarily. Thus, the importance of identifying the great vessels on every scan is emphasized. Neoplastic processes can primarily or secondarily involve the petrous apex. Imaging characteristics of these tumors are largely covered in the CPA-IAC section. Chordoma and chondrosarcoma imaging more typically involve the midline or clivus. HRCT clearly demonstrates such midline erosion and may also reveal the presence of calcifications. Characteristically, chordomas, and chondrosarcomas also provide a bright T2-weighted image on MR imaging. Vascular lesions and jugular foramen Discussion of temporal bone vascular lesions and the jugular foramen can be divided into the following topics: clinically evident disease, subjective pulsatile tinnitus, and asymmetric findings. Clinically evident disease is usually represented by a vascular-appearing middle ear mass or objective pulsatile tinnitus. Patients may also present with lower cranial nerve deficits in tumor cases. Both HRCT and gadolinium-enhanced MR imaging can be obtained to verify or rule out the presence of a pathologic condition. When a pathologic condition is demonstrated, these studies also define the extent of the disease process and its relationship to surrounding structures. Jugular bulb variants and an aberrant carotid artery are best visualized on HRCT, while diagnosis is more difficult with MR imaging. Gadolinium-enhanced MR imaging establishes the presence of neoplastic disease. Although all jugular foramen tumors enhance on MR imaging, larger paragangliomas or glomus tumors often demonstrate flow voids or a ‘‘salt-and-pepper’’ appearance within the lesion. Schwannomas and meningiomas lack such a finding and are generally smoother in shape. Schwannomas become more dumbbell-shaped as they enlarge, extending superiorly into the basal cistern or inferiorly into the carotid space. Meningiomas are more likely to extend into the CPA compared with paragangliomas. MR imaging also demonstrates tumor extension beyond the temporal bone: intracranial, infratemporal fossa, and neck involvement. Within the neck, multiple or contralateral lesions and nodal disease can be imaged. HRCT complements MR imaging findings by adding bony detail that further characterizes the disease process. Glomus jugularae demonstrate an irregular, ‘‘moth-eaten’’ pattern of bony erosion. Erosion of the jugular foramen and the bony septum separating the jugular foramen from the carotid canal (caroticojugular spine) can differentiate a glomus jugularae from other skull base tumors. Neuromas and meningiomas are more expansile with smooth, well-delineated bony margins. Although MR imaging provides detail of extratemporal extension, HRCT demonstrates disease extension within the temporal bone, supplementing the staging process. Of note, large tumors with a great deal of skull base destruction are more difficult to


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

characterize on either MR imaging or HRCT because of the lack of surrounding anatomical detail. In cases of extensive paraganglioma, postoperative MR imaging or HRCT (or both) should be obtained as baseline studies for future comparison. The evaluation of recurrent paraganglioma can be difficult, sometimes requiring angiography. Although imaging evaluation beyond MR imaging or HRCT is unnecessary for most glomus tympanicum tumors, angiography is obtained for most glomus jugularae. In addition to confirming the diagnosis, angiography demonstrates the relationship of the tumor to the adjacent great vessels, defines the feeding vasculature of the tumor, searches for multicentric lesions, and establishes collateral circulation to the brain. Angiography should be performed in a facility where embolization of the feeding vasculature can also be attempted. Thus, a second invasive procedure is avoided. MR angiography and MR venography are unnecessary when conventional angiography is to be obtained. Evaluation of subjective pulsatile tinnitus can require an extensive and costly workup without guarantee of securing a diagnosis. When clinical examination reveals a middle ear mass or cranial nerve deficit, the workup should follow the earlier-mentioned guidelines. When a pathologic condition is not clinically evident, contrast-enhanced MR imaging is the best initial screening modality for detecting a neoplastic process. HRCT may identify a great vessel anomaly when MR imaging is inconclusive. If both studies are unremarkable, angiography can be obtained when clinical suspicion is high. Arteriovenous malformation of the skull base, stenotic lesions of the carotid system, and aneurysm of the petrous carotid can be demonstrated with angiography. Both the carotid and vertebral systems should be studied. MR angiography and MR venography are becoming a more reliable, noninvasive alternative to angiography in the evaluation of pulsatile tinnitus. When examination demonstrates a clinically audible pulsation or objective pulsatile tinnitus, the workup should proceed until a pathologic condition or lack thereof is concluded. Diagnostic decisions in the evaluation of subjective pulsatile tinnitus must be weighed against potential morbidity associated with invasive studies, such as angiography. Asymmetry of the jugular bulb or foramen is commonly identified on MR imaging and HRCT, with the right side typically larger than the left. On HRCT, a normal corticated jugular foramen with preservation of the caroticojugular spine indicates a normal variant. Asymmetry can be mistaken for an abnormality on MR imaging when the jugular bulb is markedly different in size, high riding, or of slow flow. The issue is further confused because the bulb normally enhances with contrast. Careful examination and identification of the jugular bulb on all available MR imaging sequences usually identifies or rules out true pathologic conditions. Formal workup or follow-up study may be necessary in suspicious cases. MR venography would also settle the issue in such cases.

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


Labyrinth Twenty percent of patients with congenital sensorineural hearing loss demonstrate radiographic anomalies of the inner ear. For many years, HRCT has been the study of choice for evaluation of congenital inner ear deformities, such as classic Mondini’s deformity. HRCT has the added advantage of simultaneously imaging the remainder of the temporal bone. HRMR imaging has expanded imaging capability to include the membranous labyrinth. Traditional T1- and T2-weighted MR imaging is inadequate in this capacity. HRMR imaging often visualizes dilatation of the endolymphatic duct and sac, as seen with large vestibular aqueduct syndrome. From a more practical standpoint, HRMR imaging may predict a ‘‘gusher’’ when the CSF signal of the IAC is identical to that of the perilymphatic fluid signal and there is an absence of a bony partition between these fluid compartments [26]. The role of HRMR imaging in the preoperative assessment of cochlear patency for cochlear implant was discussed earlier. Traditional MR imaging demonstrates inflammatory conditions, such as labyrinthitis, in which faint enhancement of the labyrinth can be diffusely seen. Brighter, more localized enhancement is seen with intralabyrinthine schwannoma. Nonenhanced T1-weighted findings can include intralabyrinthine hemorrhage or lipoma.

Complicated chronic ear disease Uncomplicated chronic ear disease, as an imaging topic, is more appropriately discussed elsewhere. Complicated chronic ear disease is best imaged with HRCT. Notable exceptions include cases in which intracranial complication is suspected. Here, traditional MR imaging is best for the evaluation of epidural abscess, brain abscess, sigmoid sinus thrombosis, intracranial extension of cholesteatoma, and brain herniation. MR imaging signal characteristics for cholesteatoma were discussed earlier as dermoids.

External auditory canal Imaging of the external auditory canal is performed when an inflammatory process does not respond to appropriate medical treatment or when neoplasm is suspected. HRCT is the initial imaging study of choice in looking for signs of bony erosion. Malignant external otitis and squamous cell carcinoma cannot be differentiated without biopsy. When neoplastic disease is diagnosed, both HRCT and MR imaging again assume roles in intratemporal and extratemporal evaluation, respectively. Follow-up evaluation continues to be a core issue in the management of malignant external otitis. To date, advancements in nuclear medicine studies have not delivered a ‘‘gold standard’’ for malignant external otitis resolution.


D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253

Atresia of the external auditory canal is a congenital bony condition that requires HRCT evaluation. In preparation of surgical reconstruction, the entire intratemporal course of the facial nerve should be mapped.

References [1] Elster AD, Burdette JH. Safety and biological effects. In: Questions and answers in magnetic resonance imaging. St. Louis: Mosby; 2001. p. 287–307. [2] Cochlear Corporation. Package insert. Englewood, CO. [3] Weber BP, Goldring JE, Santogrossi T, et al. Magnetic resonance imaging compatibility testing of the clarion 1.2 cochlear implant. Am J Otol 1998;19:584–90. [4] Youssefzadeh S, Baumgartner W, Dorffner R, et al. MR compatibility of Med EL cochlear implants: clinical testing at 1.0 T. J Comput Assist Tomogr 1998;22:346–50. [5] Ludman CN, O’Donoghue GM. Magnetic resonance imaging demonstrating the origin of a vestibular schwannoma to be the superior division of the vestibular nerve. Am J Otol 2000;21:135–6. [6] Somers T, Casselman J, de Ceulaer G, et al. Prognostic value of magnetic resonance imaging findings in hearing preservation surgery for vestibular schwannoma. Otol Neurotol 2001;22:87–94. [7] Daniels RL, Swallow C, Shelton C, et al. Causes of unilateral sensorineural hearing loss screened by high-resolution fast spin echo magnetic resonance imaging: review of 1,070 consecutive cases. Am J Otol 2000;21:173–80. [8] Linker SP, Ruckenstein MJ, Acker J, et al. An accurate cost-effective approach for diagnosing retrocochlear lesions utilizing the T2-weighted, fast spin-echo magnetic resonance imaging scan. Laryngoscope 1997;107:1525–9. [9] Cure JK, Cromwell LD, Case JL, et al. Auditory dysfunction caused by multiple sclerosis: detection with MR imaging. AJNR Am J Neuroradiol 1990;11:817–20. [10] Ellul S, Shelton C, Davidson HC, et al. Preoperative cochlear implant imaging: is magnetic resonance imaging enough? Am J Otol 2000;21:528–33. [11] Van HT, Fraysse B, Berry I, et al. Functional magnetic resonance imaging may avoid misdiagnosis of cochleovestibular nerve aplasia in congenital deafness. Am J Otol 2000;21: 663–70. [12] Roland PS, Tobey EA, Devous MD. Preoperative functional assessment of auditory cortex in adult cochlear implant users. Laryngoscope 2001;111:77–83. [13] Fujiki N, Naito Y, Hirano S, et al. Influence of speech-coding strategy on cortical activity in cochlear implant users: a positron emission tomography study. Acta Otolaryngol 1998; 118:797–802. [14] Miyamoto RT, Wong D, Pisoni DB, et al. Positron emission tomography in cochlear implant and auditory brain stem implant recipients. Am J Otol 1999;20:596–601. [15] Brackmann DE, Bartels LJ. Rare tumors of the cerebellopontine angle. Otolaryngol Head Neck Surg 1980;88:555–9. [16] Committee on Hearing and Equilibrium. Guidelines for the evaluation of hearing preservation in acoustic neuroma (vestibular schwannoma). Otolaryngol Head Neck Surg 1995; 113:179–80. [17] Nedzelski JM, Schessel DA, Pfleiderer A, et al. Conservative management of acoustic neuromas. Otolaryngol Clin North Am 1992;25:691–705. [18] Aoki S, Sasaki Y, Machida T, et al. Contrast-enhanced MR images in patients with meningioma: importance of enhancement of the dura adjacent to the tumor. AJNR Am J Neuroradiol 1990;11:935–8. [19] Nager GT, Proctor B. Anatomic variations and anomalies involving the facial canal. Otolaryngol Clin North Am 1991;24:531–54.

D.G. Pappas, J.K. Cure´ / Otolaryngol Clin N Am 35 (2002) 239–253


[20] Swartz JD. High-resolution computed tomography of the middle ear and mastoid: I. Normal radioanatomy including normal variation. Radiology 1983;148:449–54. [21] Parnes LS, Lee DH, Peerless SJ. Magnetic resonance imaging of facial nerve neuromas. Laryngoscope 1991;101:31–5. [22] Murphy TP. MRI of the facial nerve during paralysis. Otolaryngol Head Neck Surg 1991; 104:47–51. [23] Pappas DG, Hoffman RA, Holliday RA, et al. Evaluation and management of spontaneous temporal bone cerebrospinal fluid leaks. Skull Base Surg 1995;5:1–7. [24] El Gammal T, Sobol W, Wadlington VR, et al. Cerebrospinal fluid fistula: detection with MR cisternography. AJNR Am J Neuroradiol 1998;19:627–31. [25] Shetty PG, Shroff MM, Dushyant VS, et al. Evaluation of high-resolution CT and MR cisternography in the diagnosis of cerebrospinal fluid fistula. AJNR Am J Neuroradiol 1998;19:633–9. [26] Casselman JW, Kuhweide R, Ampe W, et al. Inner ear malformations in patients with sensorineural hearing loss: detection with gradient-echo (3DFT-CISS) MR imaging. Neuroradiology 1996;38:278–86.