| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Neuroradiology |
1 From the Department of Neurology, Division of Neuroradiology, University of Heidelberg Medical Center, Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany (B.K., C.S., K.S.); and Departments of Radiology (B.K., W.B.) and Neurology (F.G.), Armed Forces Hospital, Ulm, Germany. Received October 22, 2002; revision requested January 6, 2003; final revision received May 23; accepted June 16. Address correspondence to B.K. (e-mail: bodo_kress@med.uni-heidelberg.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: In a single-blinded study, MR images were obtained in 39 patients (32 men and seven women; age range, 18–75 years; average age, 37.9 years) with acute facial nerve palsy. MR images were obtained before the 6th day of illness, on the first day of standard inpatient treatment with high-dose steroids. Signal intensity (SI) was measured at ROIs in each of five segments (internal auditory canal [IAC]; geniculate ganglion; and labyrinth, tympanic, and mastoid segments) of the intratemporal portion of the facial nerve and quantitatively analyzed. The SI measurements in the five segments were summed and divided by 100 to provide a basis for establishing an MR imaging index. SI increases and MR imaging indexes were compared with available clinical findings and electrophysiologic data.
RESULTS: Data for all 39 patients could be analyzed. The MR imaging index was significantly higher in patients with poor outcomes than in patients with favorable outcomes (specificity, 97%; sensitivity, 75%; P < .01). The SI increases in the IAC were significantly different between patients who progressed to full recovery (mean increase, 45.7%) and patients who developed chronic facial paralysis (mean increase, 156.5%) (sensitivity, 100%; specificity, 97%; P < .001). The results of differentiating between patients with good and those with poor outcomes on the basis of SI measurements in the IAC were found to be in complete agreement with electrophysiologic data.
CONCLUSION: Quantitative analysis of ROI MR imaging data is a valid method of predicting the outcome of acute facial nerve palsy during the first days after onset of symptoms and thus at a time when it is not yet possible to obtain valuable prognostic information by using electrophysiologic methods.
© RSNA, 2003
Index terms: Nerves, diseases, 2123.299, 218.299 • Nerves, facial, 2123.299 • Nerves, MR, 2123.12143, 218.12141
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Against this background, many studies have been conducted over the past 14 years in an attempt to assess whether magnetic resonance (MR) imaging can provide prognostic information regarding the outcome of acute facial paralysis. A wide variety of study protocols have been described that are as diverse as the conclusions that were drawn from the study data. Some authors have deemed MR imaging to be of prognostic value (3–8), while other researchers have come to the conclusion that there is no indication for MR imaging in the acute phase of facial paralysis (9–15).
To our knowledge, no report has yet described a classification system with clearly defined MR imaging criteria that provide a basis for predicting outcome. The aforementioned studies, however, were identical in one respect: In all of the studies, signal intensity increases after contrast material administration were evaluated subjectively. Results of studies in which dynamic MR imaging measuring methods were used for evaluating other organs (eg, the breasts) have proved the high reliability of quantitative measurements. As a consequence, these measurements—which are partly computer controlled—are routinely performed in clinical examinations (16).
Initial findings suggest that region-of-interest (ROI) measurements can yield valuable prognostic information regarding the outcome of acute idiopathic facial paralysis (17). The objective of this study was to assess the prognostic value of quantitative analyses of ROI MR imaging data in patients with acute facial nerve palsy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MR Imaging
All examinations were performed by one operator (B.K.) by using a 1.5-T MR imaging unit (ACS INTERA Power Track 3000; Philips, Best, the Netherlands) and a commercially available type C4 8-cm-diameter high-spatial-resolution surface coil (Philips). A head coil helped immobilize the patient, and the surface coil was placed directly on the ear on the affected side of the face. Routine brain MR imaging was performed to rule out a brain stem infarct before MR images of the affected facial nerve were obtained with the following sequences: A T1-weighted fast-field-echo sequence (repetition time msec/echo time msec, 9.4/3.2; flip angle, 15°; field of view, 300 mm; matrix, 512 x 328; section thickness, 0.7 mm; imaging time, 2 minutes 51 seconds) was performed before and after intravenous administration of contrast material (0.1 mmol of gadopentetate dimeglumine [Magnevist; Schering, Berlin, Germany] per kilogram of body weight administered with a delay of 3 minutes), and a T2-weighted turbo spin-echo sequence (4,000/250; flip angle, 90°, field of view, 130 mm; matrix, 512 x 328; section thickness, 0.5 mm; imaging time, 3 minutes 40 seconds) was performed before the administration of contrast material only. The use of a surface coil made it impossible to obtain suitable images of the contralateral healthy side of the face. Once the image data were reconstructed, the image information was transmitted to a workstation (Easy Vision; Philips) and further processed.
ROI Measurements
The trial was single blinded—that is, the investigator (B.K.) who performed the quantitative measurements was unfamiliar with the clinical findings and data regarding each patient’s clinical course. Quantitative measurements on MR images obtained with the T1-weighted sequences were repeated five times at the ROIs by one investigator. ROIs were irregular—5 pixels in minimum size and 35 pixels in maximum size (average size, 12.7 pixels)—and five were placed to cover each of the five segments (internal auditory canal [IAC], labyrinth segment, geniculate ganglion, tympanic segment, mastoid segment) of the intratemporal portion of the facial nerve on both pre- and postcontrast MR images. The minimum size of the ROI was restricted to avoid randomly distributed results. Obvious vessels (ie, the anterior inferior cerebellar artery and the basilar artery) were able to be excluded from ROI measurements.
These results were used for calculating the mean signal intensity value and the coefficient of variation for each measurement site. The mean values on the contrast material–enhanced MR images were compared with those on the nonenhanced MR images by using the following equation (17,18):
|
The results served as a basis for calculating signal intensity increases (as a percentage) at each of the five segments of the intratemporal portion of the facial nerve for each patient. These values reflected the contrast enhancement in the canal of the examined segment of the neurovascular bundle.
The signal intensity increases measured on the contrast-enhanced MR images provided a basis for establishing an MR imaging index (Ind) in accordance with the following equation (17,18):
|
Clinical Data
The patients stayed in the hospital while they were treated with high-dose steroids (ie, 1 g of intravenous methylprednisolone [Urbason; Aventis, Frankfurt, Germany] per day; the dose was reduced from 1,000 to 20 mg over 18 days). After MR imaging, lumbar puncture was performed, and the cell and protein content of the cerebrospinal fluid were determined. Serum studies were performed (including a white blood cell count and testing for C-reactive protein and antibodies to herpes simplex, varicella, and Epstein-Barr viruses; Toxoplasma gondii; and Borrelia species). Before the patients were discharged from the hospital, their palsy was classified as "improved" if the Fisch score (13) had increased by at least 10 points. The mean follow-up interval for the patients with a poor outcome was 1.9 years (range, 1.6–2.9 years); none of the patients with a poor outcome experienced recovery from the palsy during this interval.
Electrophysiologic Examination
An electrophysiologic examination—that is, electroneuromyography with assessment of compound muscle action potentials (CMAPs)—was performed in 32 patients by a clinician (F.G.) 7 days after the onset of the palsy (1). The facial nerve was stimulated percutaneously by placing a cathode directly next to the mastoid process under the earlobe. The position of the surface electrode permitted the recording of the electric potentials evoked in the orbicular muscle of the eye or the nasal muscle; the recording electrode was placed on the lateral lower eyelid. This test is used to measure the amplitude of the CMAP and enables an assessment of the percentage difference in the amplitude of the CMAP between the unaffected and the affected sides of the face.
Statistical Analysis
The MR imaging index and the signal intensity increases measured on the contrast-enhanced MR images were compared with the clinical outcome of each of the 39 patients by using the t test, with a P value of less than .05 considered to indicate a statistically significant difference. The MR imaging index and the signal intensity increases were also compared with the electrophysiologic data by using the Spearman rank correlation coefficient.
The signal intensity increases measured in the five segments of the intratemporal portion of the facial nerve were correlated with the MR imaging index values by using the Spearman rank correlation coefficient. The signal intensity profile established for patients who progressed to full recovery was then compared with the signal intensity profile for patients with a poor outcome by using the t test, with a P value of less than .05 considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
For all patients, the mean value of the measured signal intensity increase was 64% (range, -49% to 302%), and the mean MR imaging index was 3.2 (range, 1.0–8.6). Signal intensity increases in the IAC on the contrast-enhanced MR images were higher in patients with unfavorable outcomes (mean value, 156.5%; range, 101%–302%) than in patients with favorable outcomes (mean value, 45.7%; range, 3%–95%) (P < .001). If the threshold value (which was chosen retrospectively) had been assumed to be a signal intensity increase of 90%, only one false-positive result but no false-negative result would have occurred in the group of patients with poor outcomes (sensitivity: 100%, specificity: 97%). If the threshold value had been assumed to be a signal intensity increase of 100%, patients with poor outcomes would have been clearly differentiated from patients with favorable outcomes, although the results for three patients with a poor outcome (in whom the signal intensity increase ranged from 101% to 112%) were relatively close to the value of 100% (Figs 2, 3).
|
|
|
|
For measurements performed in the distal segments, a lower degree of correlation between MR imaging indexes and signal intensity increases and lower specificity and sensitivity were observed. Correlation between the MR imaging index and the signal intensity increase was lower than 80% for the labyrinth portion, the geniculate ganglion, and the mastoid portion and was approximately 85% for the IAC (Figs 2, 5).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A comparison of MR imaging measurements with clinical findings and electrophysiologic data confirms our previously published results (17,18) regarding the quantitative analysis of signal intensity increases in the neurovascular bundle. The present data, which were obtained with a larger group of patients, suggest that MR imaging can provide reliable prognostic information at an early stage of disease. In comparison with our previous results (17,18), our present results indicate that the MR imaging index does not provide the level of sensitivity and specificity needed for patient care: If patient evaluation had been based on the MR imaging index, there would have been one false-positive and one false-negative result. This uncertainty was due to the poor correlation between signal intensity increases in the four distal nerve segments and the clinical findings.
Our preliminary results (18) regarding the signal intensity increase in the IAC, however, are now confirmed. There was a significant difference in the increase in signal intensity in the IAC (as observed between nonenhanced and contrast-enhanced T1-weighted MR imaging) between patients with a poor outcome and patients with a favorable outcome. No patient with a poor outcome had signal intensity increases that were lower than those for patients with a favorable outcome.
Accordingly, MR imaging of the facial nerve can be used as a routine clinical examination that easily and swiftly provides prognostic information at an early stage of illness after the onset of symptoms. Intraoperative findings confirm marked venous pooling in the region of the IAC, which, together with edema in the perineurium of the nerve, causes a compression of the facial nerve at the entrance to the fallopian canal (15). This pressure is considered to be an important pathophysiologic mechanism underlying progressive wallerian degeneration (15). One hypothesis is that the venous pooling may be the pathophysiologic explanation for the contrast enhancement observed in the IAC. If this hypothesis is correct, the enhancement does not represent a leakage of contrast material but rather represents intravascular enhancement.
By the time that electrophysiologic testing methods yield prognostic information, it is too late to proceed with surgical decompression in an attempt to improve the outcome of wallerian degeneration (2). With early quantitative analysis of MR imaging results, it would for the first time be possible to assess the need for surgical treatment at an early stage and thus prevent irreversible nerve damage. Whether this unproved technique improves the outcome for these patients at this stage should be tested with further investigation.
In the literature, MR imaging has been deemed to be a relatively unreliable follow-up method for acute facial nerve palsy (14). Can quantitative methods provide new information? What is the effect of different steroid regimens on signal intensity? And last but not least, can patients actually benefit from early surgical intervention, and is it possible to measure this benefit (19)?
There were some limitations to our study. The minimum size of the ROI was restricted to avoid randomly restricted results. A signal intensity increase of 302% observed in one patient affected the regression analysis results. But even when the results in this patient were excluded, the difference in signal intensity increases between patients with favorable and those with unfavorable outcomes was significant (P < .01). The small sample of four patients with an unfavorable outcome reduced the power of the statistical analysis.
In conclusion, it is technically feasible to quantitatively analyze signal intensity measurements in the neurovascular bundle of the facial nerve, and quantitative MR imaging measurements can provide reliable prognostic information regarding the clinical course of Bell palsy. A variety of algorithms can be used for this purpose. The measurement of signal intensity increases in the IAC was found to be an easy and clinically feasible method that enables clear differentiation between patients with favorable outcomes and patients with poor outcomes.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, B.K., F.G.; study concepts, B.K., F.G.; study design, B.K., F.G., W.B.; literature research, B.K., F.G., C.S.; clinical studies, F.G.; experimental studies, B.K.; data acquisition, B.K., F.G.; data analysis/interpretation, B.K., F.G., K.S.; statistical analysis, B.K., F.G., K.S.; manuscript preparation and editing, all authors; manuscript definition of intellectual content, B.K., F.G.; manuscript revision/review, B.K.; manuscript final version approval, B.K., K.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  I.-F. Talos, K. H. Zou, L. Ohno-Machado, J. G. Bhagwat, R. Kikinis, P. M. Black, and F. A. Jolesz Supratentorial Low-Grade Glioma Resectability: Statistical Predictive Analysis Based on Anatomic MR Features and Tumor Characteristics. Radiology, May 1, 2006; 239(2): 506 - 513. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
