The information provided below is developed by EBL librarians in response to questions from VUMC clinicians. The
syntheses provided are copyrighted by the EBL and should not be re-used without
permission. Please also review this disclaimer.
My patients with asymmetric hearing loss are inevitably referred for a brain MRI with contrast to look for acoustic neuroma or other rare pathology. In my experience, these MRI's always return normal. What is the evidence that this is a necessary or helpful thing to do? How much does it cost to diagnose a single acoustic neuroma and does it make a difference if this is diagnosed sooner rather than later?
Date created: 2010-12-07
Last Updated: 2010-12-07
Rate and pattern of acoustic neuroma growth
The general consensus in the current literature appears to be that the rate and the pattern of growth of acoustic neuroma (AN), also referred to as vestibular schwannoma (VS), vary widely and are not predictable (Fortnum et al, 2010; Nikolopoulos et al, 2010). A recent systematic review of the literature on AN growth reported that some studies have shown that nine to seventy-five percent of acoustic neuromas do not grow until a few years after diagnosis while as few as ten percent but up to twenty-two percent may regress in size (Nikolopoulos et al, 2010). The available body of research examined in this and other systematic reviews was noted to suffer from methodological weaknesses in the quality of reporting, however (Fortnum et al, 2010). The true incidence and natural history of acoustic neuroma remains questionable and no reliable predictors of growth or stability have yet been identified. Nikolopoulos and colleagues (2010) also reported that acoustic neuromas might stop or restart growth in an unpredictable manner; some tumors that have been stable for many years can exhibit tumor growth. They suggested that primary longitudinal studies are needed to better define the natural history and limit unnecessary interventions.
Effectiveness of MRI to detect acoustic neuroma
Based on systematic reviews and current and radiology practice guidelines (Turkish et al, 2008), magnetic resonance imaging (MRI) with gadolinium contrast appears to be the current gold standard diagnostic test to detect acoustic neuroma in adults with asymmetric hearing loss (AHL), though the incidence of AN in this population is generally low overall. Guidelines from the American College of Radiology (Turkish et al, 2008) note that routine use of gadolinium contrast depends on factors such as MRI coil size, field of view, field strength, and pulse sequences. Several protocols for selecting patients to screen have been developed and tested, however, none have proved highly accurate.
Some studies have compared the sensitivity and specificity of ultra high-resolution nonenhanced T2-weighted (T2W or T2*W) MRI sequencing with T1-weighted gadolinium contrast-enhanced (GdT1W) MRI. A recent systematic review from the United Kingdom (Fortnum et al, 2010) examined the accuracy of these two approaches and found that noncontrast, high-resolution, three-dimensional T2W or T2*W sequences appeared to have good accuracy (sensitivity 98% and 96% respectively) compared to GdT1W MRI. Assessment of specificity suffered, however, from wide heterogeneity in research design among studies that measured MRI specificity. Pooled analysis of selected studies comparing specificities of GdT1W MRI with T2W or T2*W MRI indicated the specificity of T2W versus GdT1W ranged from 90% to 100% while the specificity of T2*W MRI ranged from 86% to 99%. Their analyses also indicated that noncontrast, high-resolution, three-dimensional T2W or T2*W sequences achieved an accurate evaluation of the 7th and 8th cranial nerves within the cerebellopontine angle and internal auditory canal, cochlea, and labyrinth. The authors suggested GdT1W contrast sequences were unlikely to contribute information that would significantly change patient management when the above structures were clearly visualized with non-contrast enhanced MRI during screenings.
In contrast to these conclusions, Wilson et al (2009) suggest that ultra high-resolution noncontrast MRI may risk missing smaller lesions despite a similar sensitivity to that of gadolinium-enhanced MRI. They note that delayed diagnosis may also promote missed opportunities for hearing preservation surgery afforded by early detection and may increase the likelihood of facial nerve paralysis if tumor growth ensues before treatment. A few studies comparing the efficacy and safety of a "scan and wait" approach versus immediate surgery on patient morbidity and mortality were located during the search and a synthesis of their findings is available upon request.
Cost of diagnosing acoustic neuroma with MRI
In addition to reviewing the effectiveness of MRI for detecting AN in patients with asymmetric hearing loss, Fortnum and colleagues (2010) also examined the literature to determine its cost-effectiveness versus other diagnostic strategies. In pooled analyses comparing contrast-enhanced versus noncontrast MRI, they found noncontrast-enhanced MRI was more cost-effective than contrast-enhanced based on a practice model developed for the United Kingdom. Overall, they concluded that noncontrast T2W MRI should be used before GdT1W MRI in a "direct to contrast" MRI protocol to screen for AN in all patients with ASHL, noting that the efficacy of the noncontrast MRI also depends on radiologist experience and quality of the imaging chain. They also suggested that given its higher specificity, GdT1W MRI should remain the gold standard for screening results that are indeterminate and to characterize suspected pathology.
As an alternative to screening directly with MRI, Wilson and colleagues (2009) suggest conducting serum lab tests before MRI to explore potential treatable causes of ASNHL. Their small prospective study compared MRI versus an initial battery of serum lab tests for aiding the diagnosis of vestibular schwannoma in 247 adults with ASNHL from two hospitals in the United States. Results showed a positive detection rate (positive test results / number of tests performed) for MRI of 2.92% (4 of 137 tests) compared to 4.62% (16 of 345 tests) for the lab tests combined (see Table IV in summary linked below). The cost of identifying a patient with a positive result via MRI was $61,650 compared to $1,142 via the battery of selected lab tests. Use of lab tests for diagnosis resulted in a 60% higher positive detection rate compared to MRI while costing only 2% of the total expense (see Table IV). The authors concluded that MRI was 30 times more expensive than laboratory tests in their study. They also noted that while MRI was the single most ordered test at both hospitals, only one of the four pathologies positively identified in 137 MRI studies was diagnosed as VS (see Table II in summary linked below). They suggested that based on these findings, MRI tests should be ordered only after a thorough neurotologic history and exam have been sought and clinical suspicion is high, or when serial audiogram shows progression of the ASNHL. They recommended Lyme titers, syphilis/RPR, ANA, and ESR as the panel of serum lab tests performed.
Timing of acoustic neuroma detection
A few retrospective studies were located that assessed the effect of time delay between symptom presentation and diagnosis of AN on patient outcomes. Teppo and colleagues (2009) conducted a retrospective study using data gathered prospectively via records from primary care visits of 91 patients who initially presented with vertigo, tinnitus, or asymmetric hearing loss in one of two hospitals in Finland between 1992 and 2006. Analysis of 59 adult patients revealed that longer delays (mean 33 months, median 14, range 0 to 234) before diagnosis by MRI did not appear to significantly affect tumor size or post treatment morbidity in these patients. The authors did not specify the type of MRI sequencing used or whether or not contrast was administered. This study contains several design weaknesses, however. The small number of subjects analyzed may not have adequately powered analyses to detect statistically significant differences. Bias may also have been introduced given the retrospective nature of data collection as well as reliance upon patient recall on the timing of initial symptom development during primary care visits. Furthermore, only 41 (45%) of the 91 patients with tumors identified via MRI initially presented with unilateral hearing loss, making the sample size of patients presenting with this symptom even smaller for evaluating diagnostic delay in this population. Also, the study included only patients identified by positive MRI and therefore does not provide insight on the diagnostic delay and clinical outcome of any patients encountered who initially presented with asymmetric hearing loss but tested negative for a mass on MRI. Other studies located on diagnostic delay were also small, older and included analysis of other diagnostic modalities, such as auditory brainstem response and computed tomography. Larger prospective studies are needed to elucidate the effect that timing of AN diagnosis has, if any, on patient outcomes.
Click the link to the full summary below for additional data from these studies and to access the PubMed abstracts and full text of the studies cited.
[ Back To Top ]
Knowledge Management and Library Information Scientists attempt to provide accurate, inclusive, and informative reports based on the specific questions and relevant data provided to them. All information is treated with strict confidentiality, in keeping with Vanderbilt policies. Information Specialists attempt to report all points of view described in the literature, selecting what they consider the most representative examples of each. Summaries are prepared for the convenience of the clinician, and physicians should consult the full-text of the articles before taking action; however, the information provided is no substitute for clinical judgment. It is possible that any given search may omit some relevant information. More extensive bibliographic references for any topic are available upon request.