- Email: [email protected]
The effective use of magnetic resonance imaging in the diagnosis of acoustic neuromas
ClinicalRadiology(1993) 48, 25-28
The Effective Use of Magnetic Resonance Imaging in the Diagnosis of Acoustic Neuromas S. A. R E N O W D E N and P. A N S L O W
Radcliffe Infirmary, Oxford The diagnostic approach to acoustic neuromas is complex but these tumours are infrequent in patients even with a suggestive clinical picture. Easy access to an efficient, reliable and cost effective investigation is desirable. Magnetic resonance imaging is now the imaging modality of choice and this paper shows that it may be used effectively as the sole investigation. T2-weighted (T2W) fast spin echo axial images taking 1 min 37 s to acquire were compared with Tl-weighted (T1W) gadolinium-enhanced axial images, taking 5 min 11 s to acquire, in 157 patients. The T2W images were satisfactory alone in 43% of patients and allowed confident diagnosis of seven of the nine acoustic neuromas. Partial volume artefact and CSF flow artefact resulted in equivocal examinations in the remaining patients. Using both sequences in every patient, imaging time was 7 min 41 s and at least 40 patients could be examined in I day, thus ensuring efficient patient throughput. We suggest that T2W fast spin echo axial images be routinely obtained and that only when they are equivocal should TIW gadolinium-enhanced axial sequences be employed. Renowden, S.A. & Anslow, P. (1993). Clinical Radiology 48, 25-28. The Effective Use of Magnetic Resonance Imaging in the Diagnosis o f Acoustic Neuromas
Accepted for Publication 15 February 1993
Until recently our protocol for the investigation of sensorineural hearing loss was as shown in Table 1. Conventional tomography was not performed. Contrastenhanced computed t o m o g r a p h y (CT) was performed on patients in whom two of the three tests shown were abnormal. A similar protocol is probably followed in m a n y departments but it is far from ideal. A b o u t half the patients cannot be tested by auditory brain stem responses because their heating loss is too great [1] and caloric tests are not always satisfactory. Both require a technician's time and expertise. Contrast-enhanced CT is 80-88% sensitive [2,3] reliably demonstrating tumours greater than 15 m m in diameter. Beam attenuation or petrous bone artefacts cause false negative or equivocal results and acoustic neuromas which are completely intracanalicular are difficult to detect. The cost of such a protocol has been calculated at about s per patient (Table 2) and some o f these patients will proceed to M R I thus increasing the cost further. M R I is currently the imaging modality of choice [4-6] and should ideally be used as the sole investigation for sensorineural hearing loss but to achieve this M R I would need to be both cost effective and efficient. The purpose o f this study was to evaluate these two factors. METHOD We examined 157 patients (68 men, 89 women) with an age range of 15-78 years using a 1.5 Tesla system. A dedicated head coil Was used. All patients were referred from the E N T department with clinical features requiting exclusion of an acoustic neuroma. 9A sagittal localizer was first acquired in 53 s. Then 3 m m axial interleaved images (matrix size 256 x 256; Correspondence to: Dr P. Anslow, Consultant Neuroradiologist, Radcliffe Infirmary, Woodstock Road, Oxford.
2 N E X ) were obtained using a fast spin echo (FSE) sequence with a T R of 3000 ms and an effective TE of 100 ms (echo train length, 16, interecho spacing, 20 ms). After this, 4 m m interleaved axial images (matrix size 256 x 256; 2 N E X ) were obtained, using a T R of 400 ms and a TE of 11 ms, immediately following intravenous gadoliniumDTPA.
RESULTS
Imaging Times and Throughput The imaging time was 1 min 37 s for the T2W FSE sequence, and for the gadolinium-enhanced T I W images, 5 min 11 s; total imaging time for each patient was therefore 7 min 41 s.
FSE T2W Versus Gadolinium-Enhanced T1W Images Axial FSE T2W images alone were satisfactory for exclusion of an acoustic neuroma in 60 patients (Fig. 1). Nine patients were shown to have an acoustic neuroma and FSE T2W images allowed confident diagnosis in seven. The neuromas varied from 8 m m in diameter (Fig. 2) to larger lesions which extended into the cerebellopontine angle (Fig. 3). Pathological confirmation was obtained in all seven cases. In one patient, a 3 m m intracanalicular lesion was seen on the FSE T2W image (Fig. 4a) and gadoliniumenhanced T l W images were required to support the
Table 1- Traditional protocol for the investigation of sensorineural hearingloss Pure tone audiogram Evoked response audiometry Hallpike calorics
26
CLINICAL RADIOLOGY
Table 2 - Approximate cost per patient of the investigation of sensorineural hearing loss Hallpike calorics Evoked response audiometry Contrast-enhanced CT
s 12 s s
Total
s
Fig. 3 - FSE T2W axial image demonstrating a larger right-sided acoustic neuroma with cystic change and extension into the cerebellopontine angle.
radiological suspicion of a small acoustic neuroma (Fig. 4b) (the eighth in the study). Pathological confirmation is not available and the patient is currently under review. In 66 patients, partial volume artefact from the petrous bone (Fig. 5) made it impossible absolutely to exclude an acoustic neuroma on the FSE T2W images and in one of Fig. 1 - FSE axial T2W image through the left IAM demonstrating the normal course of the seventh and eighth nerves from the internal auditory canal to the medulla.
,t
(
,.
,
9
.r
I
/
(a)
(b)
Fig. 2 FSE axial T2W image through the IAM demonstrating an 8 mm right-sided acoustic neuroma.
Fig. 4 - (a) FSE T2W axial image through the IAMs suggests the presence o f a 3 mm intracanalicular fight-sided acoustic .neuroma (arrow). (b) The corresponding TI W gadolinium-enhanced (5 ml) image to (a), in the same patient demonstrates a small enhancing intracanalicular lesion on the right compatible with an acoustic neuroma.
27
MRI IN ACOUSTIC NEUROMAS
Fig, 5 FSE T2W axial image through the IAMs demonstrates how partial volume artefact (arrow) from the petrous bones can render it difficult to completely exclude an acoustic neuroma.
these, T1W gadolinium-enhanced images revealed an intracanalicular neuroma (the ninth in the study) (Fig. 6a, b). Pathological verification is available. In 24 patients, CSF flow artefact in or around the internal auditory meatus on the FSE T2W images (Fig. 7) made it impossible to exclude a small acoustic neuroma but in none of these was a lesion demonstrated on the T 1W gadolinium-enhanced images. In most patients with sensorineural hearing loss in whom M R ! did not reveal an acoustic neuroma, a vascular or viral aetiology was presumed. In 5-6%, a diagnosis of Meniere's was made. These patients have been followed for 12 months.
(a)
DISCUSSION Acoustic neuromas occur predominantly in middleaged and older patients, women outnumbering men by 2: 1. They usually originate from the peripheral portion of the vestibular nerve and are circumscribed, well-encapsulated lesions which may be lobulated and are frequently elongated. They are always located at the porus acousticus and m a y grow into the cerebellopontine angle or the internal auditory canal. Expansion of the bony canal, particularly its medial end,As a c o m m o n feature of most acoustic neuromas. They grow slowly and in the past most patients had a cerebellopontine angle mass at the time of diagnosis. In their early stages, acoustic neuromas m a y cause sensorineural deafness which may be of sudden, gradual or fluctuating onset, with or without tinnitus. Vestibular symptoms appear later. Larger lesions may compress other cranial nerves, the cerebellum or brain stem and may cause hydrocephalus secondary to compression of the fourth ventricle or to high CSF protein. Moffat e t al. [7] demonstrated the benefits of early detection in terms of surgical morbidity and mortality. Traditionally, the diagnostic approach to acoustic neuromas was complex, involving tomography of the internal auditory meatuses, audiovestibular and electrophysiological tests, but the detection rate in intracanalicular tumours is poor [8]. "MRI is now the preferred imaging modality. The features are well described [9,10], the tumours being hypointense to isointense with brain on T1W images and are usually isointense or hyperintense on T2W sequences,
(b) Fig. 6 - (a) FSE T2W axial image through the IAM demonstrates how the presence of an intracanalicular acoustic neuroma (left) may exhibit a similar appearance to that of the partial volume artefact. (b) A T I W gadolinium-enhanced (5 ml) axial T I W image comparable to (a), in the same patient confirms the presence of a left-sided intracanalicular acouslic neuroma.
on which they may be isointense with CSF. They exhibit marked contrast enhancement with gadolinium. Smaller lesions are homogeneous but larger ones may contain loci of calcification, vascular channels, haemorrhage, cysts or necrosis, resulting in a more heterogeneous appearance. T2W images are best for studying the size of the internal auditory canal, the contour of which is well delineated by the bright CSF. The normal internal auditory canal contains the facial and acoustic nerves, CSF and in 30-50% of cases, the anterior inferior
28
CLINICAL RADIOLOGY
Fig. 7 - FSE T 2 W axial image demonstrates how CSF flow artefact (arrow) in a n d around the I A M m a y make it difficult to exclude the presence o f an acoustic neuroma.
canal to the upper medulla. The FSE T2W images were adequate in this respect in 67 patients (43 %) and showed a neuroma In seven patients. The internal auditory canal varies from 2-12 m m in diameter, with an average width o f 5 m m and one o f the main problems in interpretation of the FSE T2W images was partial volume artefact from the petrous bone, resulting in equivocal results in 42%. CSF flow artefact in the basal cisterns and cervical subarachnoid space interfered with interpretation in 15% of patients. The gadolinium-enhanced T l W images showed only one intracanalicular acoustic neuroma not confidently diagnosed o n the T2W FSE images due to partial volume artefact. It was necessary in one case to administer gadolinium to Support our suspicion o f a 3 m m intracanalicular neuroma. We believe that we have shown that M R I should be used as the sole investigation for sensorineurai hearing loss, replacing audiovestibular and electrophysiological tests and contrast-enhanced CT. Fast spin echo interleaved T2W axial images should be obtained routinely (see Table 3). Using a 256 • 256 matrix, these images alone will suffice in many cases; they ensure a large throughput of patients, do not require gadolinium and enable alternative diagnoses to be made. I f the FSE T2W images are not diagnostic, then gadolinium-enhanced axial T l W images should be obtained. Using a 256 • 512 matrix, with its superior resolution, we have reason to believe that the number of diagnostic FSE T2W images is dramatically increased. As a result of short imaging times and the use of lowdose gadolinium, the cost of an acoustic neuroma protocol M R study in our unit is s or s if gadolinium is required (cf. Table 2). M R I is therefore both efficient and cost effective as the sole investigation for sensorineural hearing loss. Acknowledgement. We wish to thank I G E Medical Systems, Ltd, Slough, for their gift of fast spin echo. REFERENCES
Fig. 8 - FSE T 2 W axial image through the I A M s suggests that the cause of this patient's right-sided sensorineural hearing loss is due to a demyelinating plaque in the restiform body rather than an acoustic neuroma. The right I A M is normal. Table 3 - Suggested protocol for the investigation of sensorineural hearing loss
3 m m interleaved FSE T 2 W axial scans through the I A M s Review 4 m m T I W gadolinium-enhanced scans if the FSE T2W images are not diagnostic
cerebellar artery loops into the canal where it may be identified by its signal void. T2W images are also important to exclude other lesions which may be responsible for the patient's symptoms (Fig. 8). FSE T2W axial images were regarded as diagnostic in this study when the facial and acoustic nerves were visualized in their entire length from the internal auditory
1 Pfleiderer AG, Evans KL, Grace A R H , Lloyd GAS. A screening protocol for the detection o f acoustic neuromas: a clinical evaluation. Clinics in Otolaryngology and Allied Sciences 1988; 13:145 151. 2 Valvanis A, Dabir K, Hamdi R, Oguz M, Wellauer J. Current state of the radiological diagnosis of acoustic neuroma+ Neuroradiology 1982;23:7-13. 3 Davies K R , Parker SW, New PJF. Computed tomography of acoustic neuroma. Radiology1977; 124:81 86. 4 Mafee M F , K u m a r A, Yannias DA, Valvassori GE, Appelbaum EL. Computed tomography of the middle ear in the evaluation of cholesteatomas and other soft tissue masses: comparison with pluridirectional tomography. Radiology 1983;148:465-472. 5 Press G A , Hesselink JR. M R imaging of the cerebellopontine angle and internal auditory canal lesions at 1.5 T. American Journal of Neuroradiology 1988;9:241 245. 6 Valvassori GE, Morales FG, Palacios E, Dobben GE. M R o f the normal and abnormal internal auditory canal. American Journal of Neuroradiology 1988;9:115-119. 7 Moffat DA, Hardy DG, Baguley DM. Strategy and benefits of acoustic neuroma screening. Laryngology and Otology 1988; 103:5 I 59. 8 R a m s d e n RT. Acoustic neuromas. In: Kerr AG, ed. Scott-Brown's Otolaryngology, Vol. 3, 5th ed. London: Butterworth's, 1981:500-553. 9 Ishii K, Takahashi K, M a t s u m o t o T, lshibashi T, Tazawa S, Sakamoto K. M R I of acoustic neuromas. Neuroradiology 1991;33(Suppl.):307 308. 10 Stack JP, Ramsden RT, A n t o u n NM, Lye RH, lsherwood I, Jenkins JPR. Magnetic resonance imaging of acoustic neuromas: the role of gadolinium-DTPA. British Journal of Radialagy 1988;61:800 805.
The Effective Use of Magnetic Resonance Imaging in the Diagnosis of Acoustic Neuromas S. A. R E N O W D E N and P. A N S L O W
Radcliffe Infirmary, Oxford The diagnostic approach to acoustic neuromas is complex but these tumours are infrequent in patients even with a suggestive clinical picture. Easy access to an efficient, reliable and cost effective investigation is desirable. Magnetic resonance imaging is now the imaging modality of choice and this paper shows that it may be used effectively as the sole investigation. T2-weighted (T2W) fast spin echo axial images taking 1 min 37 s to acquire were compared with Tl-weighted (T1W) gadolinium-enhanced axial images, taking 5 min 11 s to acquire, in 157 patients. The T2W images were satisfactory alone in 43% of patients and allowed confident diagnosis of seven of the nine acoustic neuromas. Partial volume artefact and CSF flow artefact resulted in equivocal examinations in the remaining patients. Using both sequences in every patient, imaging time was 7 min 41 s and at least 40 patients could be examined in I day, thus ensuring efficient patient throughput. We suggest that T2W fast spin echo axial images be routinely obtained and that only when they are equivocal should TIW gadolinium-enhanced axial sequences be employed. Renowden, S.A. & Anslow, P. (1993). Clinical Radiology 48, 25-28. The Effective Use of Magnetic Resonance Imaging in the Diagnosis o f Acoustic Neuromas
Accepted for Publication 15 February 1993
Until recently our protocol for the investigation of sensorineural hearing loss was as shown in Table 1. Conventional tomography was not performed. Contrastenhanced computed t o m o g r a p h y (CT) was performed on patients in whom two of the three tests shown were abnormal. A similar protocol is probably followed in m a n y departments but it is far from ideal. A b o u t half the patients cannot be tested by auditory brain stem responses because their heating loss is too great [1] and caloric tests are not always satisfactory. Both require a technician's time and expertise. Contrast-enhanced CT is 80-88% sensitive [2,3] reliably demonstrating tumours greater than 15 m m in diameter. Beam attenuation or petrous bone artefacts cause false negative or equivocal results and acoustic neuromas which are completely intracanalicular are difficult to detect. The cost of such a protocol has been calculated at about s per patient (Table 2) and some o f these patients will proceed to M R I thus increasing the cost further. M R I is currently the imaging modality of choice [4-6] and should ideally be used as the sole investigation for sensorineural hearing loss but to achieve this M R I would need to be both cost effective and efficient. The purpose o f this study was to evaluate these two factors. METHOD We examined 157 patients (68 men, 89 women) with an age range of 15-78 years using a 1.5 Tesla system. A dedicated head coil Was used. All patients were referred from the E N T department with clinical features requiting exclusion of an acoustic neuroma. 9A sagittal localizer was first acquired in 53 s. Then 3 m m axial interleaved images (matrix size 256 x 256; Correspondence to: Dr P. Anslow, Consultant Neuroradiologist, Radcliffe Infirmary, Woodstock Road, Oxford.
2 N E X ) were obtained using a fast spin echo (FSE) sequence with a T R of 3000 ms and an effective TE of 100 ms (echo train length, 16, interecho spacing, 20 ms). After this, 4 m m interleaved axial images (matrix size 256 x 256; 2 N E X ) were obtained, using a T R of 400 ms and a TE of 11 ms, immediately following intravenous gadoliniumDTPA.
RESULTS
Imaging Times and Throughput The imaging time was 1 min 37 s for the T2W FSE sequence, and for the gadolinium-enhanced T I W images, 5 min 11 s; total imaging time for each patient was therefore 7 min 41 s.
FSE T2W Versus Gadolinium-Enhanced T1W Images Axial FSE T2W images alone were satisfactory for exclusion of an acoustic neuroma in 60 patients (Fig. 1). Nine patients were shown to have an acoustic neuroma and FSE T2W images allowed confident diagnosis in seven. The neuromas varied from 8 m m in diameter (Fig. 2) to larger lesions which extended into the cerebellopontine angle (Fig. 3). Pathological confirmation was obtained in all seven cases. In one patient, a 3 m m intracanalicular lesion was seen on the FSE T2W image (Fig. 4a) and gadoliniumenhanced T l W images were required to support the
Table 1- Traditional protocol for the investigation of sensorineural hearingloss Pure tone audiogram Evoked response audiometry Hallpike calorics
26
CLINICAL RADIOLOGY
Table 2 - Approximate cost per patient of the investigation of sensorineural hearing loss Hallpike calorics Evoked response audiometry Contrast-enhanced CT
s 12 s s
Total
s
Fig. 3 - FSE T2W axial image demonstrating a larger right-sided acoustic neuroma with cystic change and extension into the cerebellopontine angle.
radiological suspicion of a small acoustic neuroma (Fig. 4b) (the eighth in the study). Pathological confirmation is not available and the patient is currently under review. In 66 patients, partial volume artefact from the petrous bone (Fig. 5) made it impossible absolutely to exclude an acoustic neuroma on the FSE T2W images and in one of Fig. 1 - FSE axial T2W image through the left IAM demonstrating the normal course of the seventh and eighth nerves from the internal auditory canal to the medulla.
,t
(
,.
,
9
.r
I
/
(a)
(b)
Fig. 2 FSE axial T2W image through the IAM demonstrating an 8 mm right-sided acoustic neuroma.
Fig. 4 - (a) FSE T2W axial image through the IAMs suggests the presence o f a 3 mm intracanalicular fight-sided acoustic .neuroma (arrow). (b) The corresponding TI W gadolinium-enhanced (5 ml) image to (a), in the same patient demonstrates a small enhancing intracanalicular lesion on the right compatible with an acoustic neuroma.
27
MRI IN ACOUSTIC NEUROMAS
Fig, 5 FSE T2W axial image through the IAMs demonstrates how partial volume artefact (arrow) from the petrous bones can render it difficult to completely exclude an acoustic neuroma.
these, T1W gadolinium-enhanced images revealed an intracanalicular neuroma (the ninth in the study) (Fig. 6a, b). Pathological verification is available. In 24 patients, CSF flow artefact in or around the internal auditory meatus on the FSE T2W images (Fig. 7) made it impossible to exclude a small acoustic neuroma but in none of these was a lesion demonstrated on the T 1W gadolinium-enhanced images. In most patients with sensorineural hearing loss in whom M R ! did not reveal an acoustic neuroma, a vascular or viral aetiology was presumed. In 5-6%, a diagnosis of Meniere's was made. These patients have been followed for 12 months.
(a)
DISCUSSION Acoustic neuromas occur predominantly in middleaged and older patients, women outnumbering men by 2: 1. They usually originate from the peripheral portion of the vestibular nerve and are circumscribed, well-encapsulated lesions which may be lobulated and are frequently elongated. They are always located at the porus acousticus and m a y grow into the cerebellopontine angle or the internal auditory canal. Expansion of the bony canal, particularly its medial end,As a c o m m o n feature of most acoustic neuromas. They grow slowly and in the past most patients had a cerebellopontine angle mass at the time of diagnosis. In their early stages, acoustic neuromas m a y cause sensorineural deafness which may be of sudden, gradual or fluctuating onset, with or without tinnitus. Vestibular symptoms appear later. Larger lesions may compress other cranial nerves, the cerebellum or brain stem and may cause hydrocephalus secondary to compression of the fourth ventricle or to high CSF protein. Moffat e t al. [7] demonstrated the benefits of early detection in terms of surgical morbidity and mortality. Traditionally, the diagnostic approach to acoustic neuromas was complex, involving tomography of the internal auditory meatuses, audiovestibular and electrophysiological tests, but the detection rate in intracanalicular tumours is poor [8]. "MRI is now the preferred imaging modality. The features are well described [9,10], the tumours being hypointense to isointense with brain on T1W images and are usually isointense or hyperintense on T2W sequences,
(b) Fig. 6 - (a) FSE T2W axial image through the IAM demonstrates how the presence of an intracanalicular acoustic neuroma (left) may exhibit a similar appearance to that of the partial volume artefact. (b) A T I W gadolinium-enhanced (5 ml) axial T I W image comparable to (a), in the same patient confirms the presence of a left-sided intracanalicular acouslic neuroma.
on which they may be isointense with CSF. They exhibit marked contrast enhancement with gadolinium. Smaller lesions are homogeneous but larger ones may contain loci of calcification, vascular channels, haemorrhage, cysts or necrosis, resulting in a more heterogeneous appearance. T2W images are best for studying the size of the internal auditory canal, the contour of which is well delineated by the bright CSF. The normal internal auditory canal contains the facial and acoustic nerves, CSF and in 30-50% of cases, the anterior inferior
28
CLINICAL RADIOLOGY
Fig. 7 - FSE T 2 W axial image demonstrates how CSF flow artefact (arrow) in a n d around the I A M m a y make it difficult to exclude the presence o f an acoustic neuroma.
canal to the upper medulla. The FSE T2W images were adequate in this respect in 67 patients (43 %) and showed a neuroma In seven patients. The internal auditory canal varies from 2-12 m m in diameter, with an average width o f 5 m m and one o f the main problems in interpretation of the FSE T2W images was partial volume artefact from the petrous bone, resulting in equivocal results in 42%. CSF flow artefact in the basal cisterns and cervical subarachnoid space interfered with interpretation in 15% of patients. The gadolinium-enhanced T l W images showed only one intracanalicular acoustic neuroma not confidently diagnosed o n the T2W FSE images due to partial volume artefact. It was necessary in one case to administer gadolinium to Support our suspicion o f a 3 m m intracanalicular neuroma. We believe that we have shown that M R I should be used as the sole investigation for sensorineurai hearing loss, replacing audiovestibular and electrophysiological tests and contrast-enhanced CT. Fast spin echo interleaved T2W axial images should be obtained routinely (see Table 3). Using a 256 • 256 matrix, these images alone will suffice in many cases; they ensure a large throughput of patients, do not require gadolinium and enable alternative diagnoses to be made. I f the FSE T2W images are not diagnostic, then gadolinium-enhanced axial T l W images should be obtained. Using a 256 • 512 matrix, with its superior resolution, we have reason to believe that the number of diagnostic FSE T2W images is dramatically increased. As a result of short imaging times and the use of lowdose gadolinium, the cost of an acoustic neuroma protocol M R study in our unit is s or s if gadolinium is required (cf. Table 2). M R I is therefore both efficient and cost effective as the sole investigation for sensorineural hearing loss. Acknowledgement. We wish to thank I G E Medical Systems, Ltd, Slough, for their gift of fast spin echo. REFERENCES
Fig. 8 - FSE T 2 W axial image through the I A M s suggests that the cause of this patient's right-sided sensorineural hearing loss is due to a demyelinating plaque in the restiform body rather than an acoustic neuroma. The right I A M is normal. Table 3 - Suggested protocol for the investigation of sensorineural hearing loss
3 m m interleaved FSE T 2 W axial scans through the I A M s Review 4 m m T I W gadolinium-enhanced scans if the FSE T2W images are not diagnostic
cerebellar artery loops into the canal where it may be identified by its signal void. T2W images are also important to exclude other lesions which may be responsible for the patient's symptoms (Fig. 8). FSE T2W axial images were regarded as diagnostic in this study when the facial and acoustic nerves were visualized in their entire length from the internal auditory
1 Pfleiderer AG, Evans KL, Grace A R H , Lloyd GAS. A screening protocol for the detection o f acoustic neuromas: a clinical evaluation. Clinics in Otolaryngology and Allied Sciences 1988; 13:145 151. 2 Valvanis A, Dabir K, Hamdi R, Oguz M, Wellauer J. Current state of the radiological diagnosis of acoustic neuroma+ Neuroradiology 1982;23:7-13. 3 Davies K R , Parker SW, New PJF. Computed tomography of acoustic neuroma. Radiology1977; 124:81 86. 4 Mafee M F , K u m a r A, Yannias DA, Valvassori GE, Appelbaum EL. Computed tomography of the middle ear in the evaluation of cholesteatomas and other soft tissue masses: comparison with pluridirectional tomography. Radiology 1983;148:465-472. 5 Press G A , Hesselink JR. M R imaging of the cerebellopontine angle and internal auditory canal lesions at 1.5 T. American Journal of Neuroradiology 1988;9:241 245. 6 Valvassori GE, Morales FG, Palacios E, Dobben GE. M R o f the normal and abnormal internal auditory canal. American Journal of Neuroradiology 1988;9:115-119. 7 Moffat DA, Hardy DG, Baguley DM. Strategy and benefits of acoustic neuroma screening. Laryngology and Otology 1988; 103:5 I 59. 8 R a m s d e n RT. Acoustic neuromas. In: Kerr AG, ed. Scott-Brown's Otolaryngology, Vol. 3, 5th ed. London: Butterworth's, 1981:500-553. 9 Ishii K, Takahashi K, M a t s u m o t o T, lshibashi T, Tazawa S, Sakamoto K. M R I of acoustic neuromas. Neuroradiology 1991;33(Suppl.):307 308. 10 Stack JP, Ramsden RT, A n t o u n NM, Lye RH, lsherwood I, Jenkins JPR. Magnetic resonance imaging of acoustic neuromas: the role of gadolinium-DTPA. British Journal of Radialagy 1988;61:800 805.