Volume 116, Issue 6 p. 567-574
Original Article
Free Access

Cost-Effective Evaluation of Asymmetric Sensorineural Hearing Loss with Focused Magnetic Resonance Imaging

Ltc David A. Carrier MD

Ltc David A. Carrier MD

Department of Neuroradiology, Wilford Hall Medical Center, Lackland AFB, Texas

Search for more papers by this author
Moises A. Arriaga MD, FACS

Corresponding Author

Moises A. Arriaga MD, FACS

Department of Otology/Neurotology, Pittsburgh Ear Associates, Pittsburgh, Pennsylvania

Reprint requests: Moises A. Arriaga, MD, Pittsburgh Ear Associates, 420 East North Ave., Pittsburgh, PA 15212.Search for more papers by this author
First published: 01 February 2018
Citations: 3

Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, New Orleans, La., Sept. 17–20, 1995.

Abstract

The poor sensitivity of audiometric brain stem response for small vestibular schwannomas (acoustic neuromas) creates a dilemma for the physician evaluating a patient with signs and symptoms of retrocochlear disease. Magnetic resonance imaging is recognized as the gold standard for the evaluation of these problems, but if a complete examination of the internal auditory canals and head is done on every patient, the cost is high. Although less expensive, screening with audiometric brain stem response risks missing up to 33% of small tumors. Therefore we developed a focused magnetic resonance imaging sequence for evaluation of patients with asymmetric sensorineural hearing loss and/or nonpulsatile tinnitus. The protocol includes a T1-weighted sagittal localizer, pregadolinium and post-gadolinium T1-weighted 3-mm contiguous axial slices through the internal auditory canal and the region of the cerebellopontine angle, and T2-weighted axial images through the entire brain. Total scanning time is about 12 minutes, and the estimated cost is $300 to $500. We retrospectively reviewed the imaging records of 485 screening examinations done during an 18-month period. Twenty-four patients had diagnoses definitely or probably producing the hearing loss for an overall positive rate of 5%. By eliminating the need for follow-up audiometric or electrophysiologic studies, we believe a focused magnetic resonance imaging-based diagnostic scheme is actually more cost-effective on a cost-per-patient basis.

Audiometric brain stem response (ABR) is advocated by some authors as a screening test for selection of patients with asymmetric sensorineural hearing loss (ASNHL) or nonpulsatile tinnitus for further workup with magnetic resonance imaging (MRI).1-3 These recommendations are based on data obtained before the advent of enhanced MRI that placed the sensitivity of ABR between 92% and 100%.2, 4-9 We now know by using enhanced MRI as a gold standard that the sensitivity of ABR is very dependent on tumor size and that ABR frequently misses small schwannomas. The reported false-negative rate is between 17% and 33%.10-12 Moreover, because ABR is a functional test, the specificity for a particular anatomic lesion is near zero.

Hearing preservation surgery is now a realistic goal if schwannomas are diagnosed and treated before they are larger than 2 cm.13-16 Enhanced MRI can easily detect small intracanalicular schwannomas,16, 17 with a sensitivity of nearly 100%. It also has excellent specificity and is able to distinguish cochlear, internal auditory canal (IAC), cerebellopontine angle (CPA), and central lesions in the brain along the acoustic pathway. With these advantages, most would agree that enhanced MRI is the ideal test for the evaluation of ASNHL, with cost being the only downside.

Our purpose in this study was to develop a cost-competitive, focused MRI protocol that could replace ABR as the initial diagnostic test in the workup of patients with documented ASNHL and/or nonpulsatile tinnitus.

METHODS AND MATERIAL

All patients were imaged on a 1.5-tesla GE Signa scanner (General Electric, Milwaukee, Wis.). Initially a T1-weighted sagittal localizer was obtained. With the localizer as a guide, eight 3-mm-thick with zero gap T1-weighted axial slices were obtained through the region of the IACs before and after the administration of 0.05 mmol/kg gadolinium contrast agent. The head coil is used with the following parameters: TR 500; TE 15; BW 16 Hz; FOV 16 × 16 em; matrix 256 × 192; 2 NEX on the presequence; and 4 NEX on the postsequence. Lastly, a fast-spin echo T2-weighted axial sequence is obtained through the entire brain with parameters of TR 3600; TE 96/ef; BW 16 Hz; FOV 20 × 20 cm; slice thickness 5 mm with 2 mm gap; matrix 256 × 192; and 1 NEX. Total scanning time is 12 minutes, and total examination time including setup and contrast injection is approximately 20 to 25 minutes. The localizer sequence is not printed. The IAC sequences are printed on a 9 on 1 format with a magnification factor of 1.3, and the brain sequence is printed on a 20 on 1 format for a total of three sheets of film.

The imaging records of 485 patients who were imaged with this protocol between Jan. 1, 1994, and June 30, 1995, were retrospectively reviewed. Only patients with ASNHL or unilateral nonpulsatile tinnitus were included in the study. The examinations were categorized as negative, definitely accounts for ASNHL, or probably accounts for ASNHL.

RESULTS

Eleven of the 485 screening tests definitely accounted for the presenting signs and symptoms. These consisted of seven vestibular schwannomas and one case each of congenital common cavity, autoimmune labyrinthitis, choristoma, and labyrinthitis ossificans. Thirteen lesions that were believed to probably account for the findings on clinical examination consisted of six nonspecific foci of enhancement in the fundus of the IAC, three CPA arachnoid cysts, three ischemic foci in the auditory pathway in the brain stem, and one tortuous basilar artery impinging on the root exit zone of central nerve VIII (Table 1). The chance of obtaining a positive finding on the screening examination was 5%.

Table 1. Results of 485 screening-enhanced MRI scans
Results No. of scans
Negative 461
Explains clinical findings 11
Vestibular schwannoma 7
Autoimmune labyrinthitis 1
Congenital common cavity 1
Ganglioneuroma 1
Labyrinthitis ossificans 1
Probably explains clinical findings 13
Nonspecific fundal enhancement 6
CPA arachnoid cyst 3
Ischemic brain stem lesions 3
Macrovascular loop 1

DISCUSSION

Enhanced MRI is unquestionably the gold standard for the retrocochlear evaluation of ASNHL and/or nonpulsatile tinnitus.3, 10, 17-19 With respect to detecting surgical lesions (i.e., IAC and CPA tumors), MRI has a sensitivity of nearly 100% even for very small intra-canalicular schwannomas. We could find no report in the literature of a falsely negative enhanced MRI in such a case. The only barrier to its general application as an initial retrocochlear test in the workup of ASNHL is cost.

Hendrix et al.2 and Bu-Saba et al.1 proposed algorithms that use ABR as a screening test to select patients who should go on to evaluation with enhanced MRI. In the algorithm of Welling et al.,3 patients with high and intermediate clinical probability go directly to an enhanced MRI, whereas those with a low probability undergo ABR. In all of the above schemes, negative findings on ABR require follow-up evaluation with a repeat ABR, which at $200 to $300 per test rapidly erodes any cost advantage of the initial use of ABR. These recommendations were based on the assumption that ABR has a sensitivity of 92% to 100% for detection of schwannomas and, in Welling's paper, that MRI costs $1200. However, using enhanced MRI as the gold standard, more recent data have shown that ABR is a very poor screening test with respect to small schwannomas, with false-negative rates of 17% to 33%.10-12 Because a goal of evaluating ASNHL is to detect schwannomas smaller than 2 cm when hearing preservation surgery is feasible, we believe a focused enhanced MRI should be the initial retrocochlear test.

Our focused protocol can be done for $300 to $500 including interpretation, which is comparable to the cost of an ABR. By streamlining the examination to answer a specific question, we have reduced total scanner time to 20 to 25 minutes compared with 40 to 50 minutes for a more traditional examination. Even with this limited examination, the entire brain is covered by the T2-weighted axial sequence so that pathology remote from the temporal bone will not be missed. We use only half of the usual dose of gadolinium contrast agent, which has been shown to be more than adequate for the evaluation of schwannomas.20 Lastly we reduce cost by printing on only three sheets of film (Fig. 1). The old way of thinking is that every MRI takes about an hour and should cost approximately $1000. In the era of managed care and cost containment, we must begin thinking of MRI as a flexible tool that can be tailored to answer specific questions.

Details are in the caption following the image

Focused MRI protocol. A, Eight postgadolinium contiguous 3-mm axial T1-weighted images through the region of the IAC. The same sequence is also done before gadolinium. B, T2-weighted axial images of the entire brain.

Details are in the caption following the image

Schwannoma. Postgadolinium T1-weighted axial image through the IACs demonstrates an intensely enhancing mass nearly filling the left IAC.

Details are in the caption following the image

Choristoma. Postgadolinium T1-weighted axial image through the IACs demonstrates an enhancing mass in the fundus of the right IAC.

Details are in the caption following the image

Nonspecific fundal enhancement. Postgadolinium T1-weighted axial image through the IACs demonstrates a punctate focus of enhancement in the fundus of the right IAC (arrow). Because the probability of neoplasm is extremely low, we now reimage these patients in 3 to 6 months.23

Details are in the caption following the image

Autoimmune labyrinthitis. Post-gadolinium T1-weighted axial image through the IACs demonstrates diffuse enhancement of the right cochlea (arrow).

Details are in the caption following the image

Congenital common cavity. A, Pregadolinium T1-weighted image through the IACs demonstrates a globular appearance of the left vestibule (arrow), suggesting a congenital anomaly. B, Axial computed tomography scan confirmed the diagnosis (arrow).

Details are in the caption following the image

Labyrinthitis ossificans. A, Postgadolinium T1-weighted axial image through the IACs demonstrates absence of the left cochlea and vestibule (arrow). B, Subsequent T2-weighted axial image through the same area confirmed the finding (arrow). On further questioning the patient remembered a remote history of meningitis.

Details are in the caption following the image

Macrovascular loop. T2-weighted axial image demonstrates a tortuous left vertebral artery impinging on the root exit zone of nerve VIII (arrow).

Details are in the caption following the image

Ischemia. T2-weighted axial image demonstrates an ischemic focus in the midbrain extending into the right inferior colliculus (arrow).

Traditional cost comparisons of diagnostic algorithms for retrocochlear lesions assume a cost of $1200 for MRI when calculating cost per lesion. Such analysis favors an ABR-based scheme but usually ignores the cost of subsequent audiograms and ABRs in patients with equivocal or negative ABRs and continued symptoms. Our protocol avoids the need for subsequent studies because false-negative MRI findings for surgical lesions have not been reported. A focused MRI-based IAC algorithm is particularly cost-effective from the perspective of a cost per patient analysis. After identification of ASNHL, the total cost of evaluation is $400 per patient without “hidden costs” of subsequent audiograms, ABRs, and so forth. The cost-effectiveness of a focused MRI-based protocol is preferable to an ABR-based algorithm that requires additional testing.

Another advantage of MRI compared with ABR is a fairly high degree of specificity. By a lesion's location, morphology, signal characteristics, and enhancement pattern, a specific diagnosis can usually be offered when an anatomic lesion is detected. Vestibular schwannoma was the most common anatomic lesion in our series, with other entities as listed in Table 1 (Fig. 2-9). Many other lesions that can be detected by enhanced MRI have been described in the literature (Table 2).21-30       

Table 2. Differential diagnosis of lesions causing ASNHL detectable on enhanced MRI
Cochlear
Schwannoma
Infection—viral, bacterial, syphilitic
Autoimmune
Hemorrhage
Congenital
Fibrosis caused by old infection
IAC
Schwannoma—nerve VII or VIII
Hemangioma
Lipoma
Choristoma
Lymphoma
Metastasis
Sarcoid
Arteriovenous malformation
Nonspecific inflammatory changes
CPA
Schwannoma—nerve V or VIII
Meningioma
Epidermoid
Arachnoid cyst
Met
Vascular loop
Central
Tumor
Infarct
Multiple sclerosis
Vascular malformation
Trauma

In 176 patients with ASNHL, Armington et al.24 found 50 with lesions documented on enhanced MRI that definitely or possibly related to the hearing loss for a positive rate of 28%. In our study there was a much higher number of negative studies, with only 5% believed to be positive. This probably reflects our use of the test as a screening examination. However, negative examination results also have considerable value in terms of patient reassurance and reduction in cost of further workup.

In conclusion, we believe that MRI should be viewed as a flexible tool that can be tailored to answer specific questions. We have presented a protocol for use in patients with ASNHL and/or nonpulsatile tinnitus that should cost approximately $300 to $500. Enhanced MRI is nearly 100% sensitive in terms of excluding surgical disease and has high specificity when anatomic lesions are discovered. In contrast, ABR is insensitive for excluding small schwannomas, nonspecific, and at $200 to $300 is nearly as expensive as a focused enhanced MRI. We recommend that focused enhanced MRI replace ABR as the initial screening examination in patients with documented ASNHL and/or nonpulsatile tinnitus.

Seminar in Psychoacoustics

Arnold M. Small, Jr's 40 years in the field of hearing and psychoacoustics will be honored at the University of Iowa on Aug. 23, 1997. His students and colleagues will give presentations focusing on Professor Small's many contributions to the field. For further information, contact George Haskell at the VA Medical Center in Iowa City; phone, (319)339–7126; e-mail, [email protected].