- Email: [email protected]
Magnetic resonance findings in 92 acoustic neuromas
European Journal of Radiology, 13 ( 199 1) 96-l 02 0
96
EURRAD
1991 Elsevier
Science Publishers
B.V. All rights reserved. 0720-048X/91/$03.50
00 186
Magnetic resonance findings in 92 acoustic neuromas Bertrand Duvoisin”4, Julio Fern andes’, Dominique Doyen’, Jean-Marc Sterkers2 and Serge Bobin Departments
of ‘Neuroradiology
30to-Rhino-Laryngology,
and CIERM,
20to-Rhino-Laryngology,
Hospital Cochin Park
University Hospital of Bicetre, Le Kremlin Bicetre, France,
and 4Diagnostic Radiology. University Hospital CHUV,
(Received 29 November
Key words: Magnetic resonance
Alban Denys’,
1990; accepted
imaging, brain; Magnetic resonance
Lausanne,
Switzerland
after revision 3 April 1991)
imaging, acoustic neuroma;
Brain, MRI; Brain, neoplasm
Abstract A retrospective analysis of the MR findings in 92 cases of acoustic neuromas is presented. The method of examination included in all cases intravenous injection of Gadolinium (Gd-DTPA or DOTA) with realization of sections in the axial and coronal planes. In 21 cases native MR studies were performed in the axial plane, before Gadolinium injection, with TlWI (n = 21), and T2WI (n = 6) images. Tumors were strictly intracanalar in 19 cases (20.7%), only localized in the cerebellopontine angle (CPA) in 5 cases (5.4%), and in 68 cases (73.9%) the tumors had intra- and extracanalar components. In this last group of lesions, 63.2% completely filled the internal auditory canal (IAC), and 36.8% occupied the internal portion of the IAC. In most cases (85.3 %) the mean diameter of the CPA component was less than or equal to 2.5 cm. Lesions were more frequently homogeneous (58.8%) after Gd iv.-enhancement. Heterogeneity was noted mainly in large lesions (>2.5 cm: 100%). In all cases but one, the tumors were round, or oval-shaped, well-delineated, and did not present significant contact with the petrous bone. In 80.9% of CPA lesions, the center of the tumor was posteriorly excentered in relation to the internal auditory canal, whereas it was centered in 16.2 y0 of cases. Widening of the IAC was observed in 5 1.1 y0 of cases. In 45.2 y0 of tumors of the CPA, smoothing of the posterior edge of the porus was visible.
Introduction Acoustic neuromas are common extra-axial tumors, that usually arise from the vestibular division of the eighth cranial nerve, and characteristically occur in the cerebellopontine angle (CPA) and internal auditory canal (IAC). The clinical work-up of these lesions is complex; despite the variety of audiological and electrophysiological tests used, the clinical detection rate of small acoustic neuromas (mainly intrameatal) is fairly low [ 11. Computed tomography (CT), with intravenous contrast medium, and gas CT cisternography are valuable diagnostic tools to detect large or mediumsized acoustic neuromas [2-81. However CT studies bear morbidity due to radiation exposure and iodine injection, and beam-hardening artefacts often limit the
Address for reprints: Bertrand Duvoisin, M.D., Department of Diagnostic Radiology, University Hospital CHUV, 1011 Lausanne, Switzerland.
quality of the study; furthermore gas CT cisternography only depicts the internal limit of the tumor, and does not give any information about tumoral extent to the fundus of the IAC. Over recent years magnetic resonance (MR) imaging has proved to be the most sensitive tool for diagnosis of acoustic neuromas; it is more effective than CT with iodine injection or gas CT cisternography [ 9- 131. However small intracanalar neuromas, and post-operative tumoral recurrences can be difficult to assess on plain MR [ 14,161. Intravenous injection of Gadolinium (Gd-DTPA or DOTA) provides an optimal demonstration of acoustic neuromas, which usually enhance significantly, and exhibit a typical hyperintense signal. This feature allows exact assessment of tumoral extent into the lateral portion of the IAC, and a good demonstration of post-operative recurrences [ 16- 191. Having examined over the last years a fairly large number of patients with acoustic neuromas (92 tumors) in our MR center, we decided to review these studies retrospectively to determine the most typical MR features of acoustic neuromas.
Patients and Methods Consecutive patients with a diagnosis of acoustic neuroma, who underwent MR studies between December 1986 and May 1990, were assessed retrospectively. Eighty-four patients (37 men and 47 women) aged 15-81 years (mean, 53.7 years) were included in this study. The clinical symptoms included unilateral or bilateral deafness of variable duration, tinnitus and/or vertigo. Audiometric tests (audiograms and acoustic reflex testing), as well as labyrinthine tests (vestibular response, electronystagmography) were positive in the majority of patients. Auditory brainstem responses (ABRs) were obtained in 68 patients, and in all cases indicated a retrocochlear abnormality, suggesting an acoustic neuroma. In eight cases of neurofibromatosis, the lesions were bilateral Sixteen patients previously underwent operative excision of an acoustic neuroma; of these, one patient was operated bilaterally; clinical recurrence of the symptoms urged MR study. A CT study was obtained in all cases, four of these before MR imaging. In 70 cases, CT depicted abnormalities, including a mass in the CPA (n = 13), widening of the porus and/or IAC (n = 34) intracanalar or intrameatal contrast enhancement (n = 32), abnormal gas cisternography (n = 4). In 10 c ases CT did not disclose any abnormality. Fifty-nine patients were operated and in all of them the histological study depicted acoustic neuroma. Twenty-five patients were treated conservatively; these
patients were considered as having acoustic neuromas on the basis of clinical criteria, audiological and electrophysiological tests, and typical MR findings. Of these patients, follow-up MR studies were performed in 22 cases, showing stability of the lesions in 14 cases, and an increasing size in 8 cases (slight: y1= 5, moderate: n=3,marked:n= 1). All MR examinations were performed on a 1.5 T imaging unit (Signa; GE Medical Systems, Milwaukee, U.S.A.), with a head coil, having a 24-cm field of view, 256 x 256 matrices corresponding to a pixel size of 0.95 x 0.95 mm, and consecutive 3-mm-thick slices (in some rare cases with 5-mm-thick slices). Spin-echo (SE) MR images were obtained routinely in the axial and coronal planes. In all cases Gd-DTPA (or -DOTA) was injected intravenously (dose: 0.1 mmol/kg); among these, post-contrast sequences were selectively obtained in 63 patients. In the other 21 patients, native MR studies were also obtained before Gd injection, with T- 1 weighted (n = 21), and T-2 weighted (n = 5) images. All acoustic neuromas were separately evaluated by two of us to assess the following parameters: (1) localization (intra- and/or extracanalar); (2) size; (3) homogeneous vs heterogeneous; (4) borders and shape of the extracanalar portion of the tumor; (5) extension of the tumor to the fundus of the IAC; (6) position of the center of the extracanalar component with respect to the IAC (Fig. 1 shows the spectrum ofintra- and extracanalar types of lesions); and (7) widening of the IAC.
a
Fig. 1. Drawing of the different sites of intra- and extracanalar acoustic neuromas. Intracanalar lesions include types a, b and c, according to the tumoral extent into the internal auditory canal (IAC). Note that types a and b reach the ftmdus. Extracanalar tumors include types 1 to 6, according to the position ofthe lesion in respect to the IAC. Types 2 and 4 are posteriorly excentered; the former does not extend anteriorly to the porus. whereas type 4 does. Note that not all types present a significant contact with the petrous bone, except type 5. Difference between types 1 and 3 is only in size, 3 being larger than 1. Tumors with intra- and extracanalar components are labelled in the text: a 1 to a 4, and c 1 to c5.
98
a
b
Fig. 2. Strictly intracanalar right acoustic neuroma, type a. Gadolinium i.v.-enhanced axial (a) and coronal (b) MR sections, depicting extension of the tumor to the entire IAC (arrows), reaching the fundus, without widening of the IAC. Note hypersignal of both petrous bone apex, due to fatty bone marrow (black star).
Results In 19 cases (20.7%) the acoustic neuroma was strictly intracanalar. The tumor occupied the entire IAC (type a, Fig. 2) in 14 cases (73.7%) and the external portion of the IAC (type b, Fig. 3) in 5 cases (26.3%). In all cases the fundus of the IAC was reached. A slight to moderate widening of the IAC was noted in 5 cases (26.3 %). In one case a bilateral neuroma was observed, occupying the whole IAC; the patient suffered Von Recklinghausen disease, and had been operated eight years previously on the left side for an acoustic neuroma mainly localized in the CPA cistern. Both neuromas had approximately the same size on MR; as no apparent growth was evident since a CT performed two years previously, a conservative treatment was conducted. In 6 cases, including the patient with bilateral lesions, TlWI axial MR images were obtained before Gd-DTPA injection. On these studies the lesions were homogeneous, isointense to the pons in 5 cases, hypointense in one case and heterogeneous with a central hypointense zone in the last case. In this case T2WI sections showed that the tumor was isointense. In all 19 cases the tumor exhibited a hyperintense signal on Gd i.v.-enhanced TlWI sections, and was homogeneous in 18 cases. In five cases (5.4%) the tumor was selectively extracanalar. The size of the lesions ranged between 4 mm and 45 mm in mean diameter. The localization included one case each of types 2-6. In a single case the posterior edge of the IAC opening was smoothed. Three cases
had been previously operated for an acoustic neuroma, recurrence of which urged the MR study. After GdDTPA injection, the lesions showed significant enhancement and all but one were homogeneous. The last tumor exhibited a heterogeneous aspect, with a large central necrotic zone. In all other 68 cases (73.9%) the tumor had extra-
Fig. 3. Strictly intracanalar right acoustic neuroma, type b. Gadolinium iv.-enhanced axial MR section, showing a lesion occupying 2/3 of the external part of the right IAC (small arrows). Note a controlateral larger lesion, type a (large arrow), which enlarges the left IAC; the patient suffered Von Recklinghausen’s disease.
a
C
b
Fig. 4. Various types of lesions in the cerebellopontine angle, all Gadolinium i.v.-enhanced axial sections. (a) Left tumor, type 2, (arrow) which almost entirely fills the IAC; an oval-shaped, hyperintense lesion, with a broad base on the right petrous apex, is typical of a meningioma (small arrows). The patient suffered Von Recklinghausen’s disease. (b) Right tumor, type 4, which causes a moderate impression on the pons (arrows), and extends into the internal portion of the IAC. (c) Left recurrent acoustic neuroma, type 5 (arrows), which has a fairly large surface of contact with the petrous bone; note hyperintense signal of the left petrous bone due to fat packing, after a previous translabyrinthine resection of an acoustic neuroma.
and intracanalar components. According to the classilication depicted in Fig. 1, the most frequent tumors included types a 2 (n = 22) (Fig. 4a), a 4 (n = 16) (Fig. 4b), c 2 (n = 1 l), and c 4 (n = 6) representing respectively 32.4%, 23.5%, 16.2% and 8.8% of this group of lesions. Other types occurred less frequently. Fifty-eight lesions (85.3 %) had a mean diameter inferior or equal to 2.5 cm, seven (10.3%) measured between 2.5 cm and 4 cm, whereas three tumors (4.4%) were larger than 4 cm. In all cases but one, the tumors exhibited a round, or ovoid shape, with clearly defined borders, and did not present a significant contact with the petrous bone. A single neuroma presented a large surface of contact with the posterior petrous bone (Fig. 4~). In 15 cases, including two patients with bilateral lesions, native MR studies were obtained. Fourteen neuromas were homogeneous, and exhibited, on TlWI sequences, an isosignal to the pons in 7 cases, and a hyposignal in 7 cases. In live cases T2WI images were also obtained; the tumors depicted a hyposignal respective to the pons in 3 cases, and were not clearly recognized in the other 2 cases. The other 3 neuromas were heterogeneous on Tl WI and /or T2WI sequences. One lesion showed a peripheral hyperintense signal in both series, which was considered as post-hemorrhagic. After Gd i.v.enhancement, 40 lesions (58.8%) were homogeneous and exhibited a diffuse hyperintense signal. In 28 cases (41.2%) the tumors were heterogeneous, with internal hypointense zones (Fig. 5). In most of the lesions less than 2.5 cm this heterogeneity was absent or discrete,
whereas in all large lesions (>2.5 cm) it was fairly marked. Due to its absence of signal the cortical surface of the porus was easily recognized on MR images; in 32 cases (47.1%) a clear smoothing of the posterior edge of the IAC was noted, associated to neuromas type 2 and 4. A widening of the IAC, recognizable due to signilicant asymmetry relative to the controlateral side, was observed in 41 tumors (60.3 %) (Fig. 6). In eight patients a large CPA mass caused a compression on the pons, with displacement of the fourth ventricle in three cases, and obstructive hydrocephalus in two cases. Among these 68 acoustic neuromas, 14 were observed in patients with neurolibromatosis, and bilateral lesions. Nine lesions constituted post-operative recurin patients with rences, including four neuromas neurolibromatosis. Discussion Until the introduction of MR, i.v.-enhanced CT and gas CT cisternography were considered the definitive tests to rule out acoustic neuromas [2-81. However false-negative or equivocal results occur on i.v.-enhanced CT studies, due to beam attenuation and/or petrous bone artefacts. The application of gas CT cistemography has improved the sensitivity of detection or exclusion of small tumors, but this procedure is not innocuous, and occasional false-negative and falsepositive results occur [7]. Furthermore, this type of investigation cannot demonstrate the lateral intra-
100
canalar extent of the tumor, particularly with respect to the fundus of the IAC. Over recent years MR has proven to be the most sensitive diagnostic tool in cases of acoustic neuromas, being more effective than CT with iodine injection or gas CT cistemography [9-133. However, small intracanalar neuromas, and post-operative tumoral recurrences are difficult to assess on plain MR [ 14,151. After iv injection of Gd-DTPA (-DOTA) acoustic neuromas significantly enhance [ 19-221. This feature allows detection of tumoral recurrences, and clear depiction of the lateral extent of intracanalar lesions, which is man-
datory to determine the prognosis and the type of surgical procedure[ 17-191. During a 42-month period we examined in our center 92 acoustic neuromas in 84 patients, including eight patients with neurofibromatosis. In 59 cases surgical confirmation of the MR diagnosis was obtained, whereas in 25 patients a conservative treatment was conducted. In these latter patients the lesions were considered as acoustic neuromas, following clinical criteria, audiological and electrophysiological tests, and a typical MR aspect. Control MR studies were obtained in 22 cases, showing stability of the lesions in 14 cases, and an increasing size in 8 cases. The distribution of the site of neuromas showed 19 strictly intracanalar tumors (20.7%), five pure extracanalar masses (5.4%), and 68 lesions with extra- and intracanalar components (73.9%). Our primary goal in the present study was to evaluate the lesions which presented components both in the IAC and CPA in order to establish the typical MR findings observed in such lesions. Sixty-eight (73.9%)
Fig. 5. Heterogeneous acoustic neuromas. (a) Left tumor, which depicts a large central hypointensity, due to cystic necrosis, and a regular peripheral enhancing rim (arrow). (b and c), (same patient) large right acoustic neuroma, exhibiting significant heteregeneous features, due to necrosis and hemorrhage; on TlWI unenhanced section (b), a posterior hyperintense biloculate zone (arrowheads) represents hemorrhage within cystic spaces. Other hypointense cystic lesions are recognized. After Gadolinium i.v.-enhancement, the solid portions of the tumor are well-delineated of the necroticohemorrhagic components.
a
b
a b Fig. 6. Widening of the right porus and IAC, consecutive to a large type 3 tumor; (a) CT and (b) MR axial sections. A strong correlation observed between CT and MR: both methods well define the cortical surfaces (white on CT, and black on MR). of the neuromas reviewed had such a feature. Among these tumors, 43 (63.2%) completely filled the IAC, till the fundus, whereas 25 (36.8 %) only occupied the internal part of the IAC. In the majority of cases (85.3%), the CPA component had a mean diameter less than or equal to 2.5 cm. All CPA lesions presented a spherical or ovoid shape, with well-defined borders. In all cases the contact between the petrous bone and the round extracanalar component of the neuroma was small, or absent. These particulars, in conjunction with statistically significant higher frequency of acoustic neuromas allows to differentiate neuromas from meningiomas, which are classically crescentic, elongated, dural-based, and eccentric to the porus [ 8, 11, 231. Fourty (58.8%) tumors were homogeneous after Gd-iv enhancement, whereas 28 (4 1.2%) exhibited a variable inhomogeneity. This feature was rarely observed (25 %) in small (< 1 cm) lesions, whereas its frequency and extent increased parallel to lesion size with, respectively, 53.6% of lesions inferior to 2.5 cm, and 100% in lesions larger than 2.5 cm. The heterogeneity of the lesion is caused by necrosis, or cystic transformation. In the literature cystic acoustic neuromas are rarely reported [24]. The topographic relationship between acoustic neuromas and the IAC classically describes that the tumor is centered on the IAC [ 8, 111. In our group of tumors with extra- and intracanalar components, most lesions (80.9%) were posteriorly eccentered to the IAC, whereas only 11 (16.2%) tumors were centered on the porus. This can be easily explained by the fact that acoustic neuromas primarily involve the vestibular
is
nerve, which is posteriorly placed. Thus most CPA tumors have their equator posteriorly eccentered respective to the IAC. This feature seems a very reliable sign of an acoustic neuroma, in cases of CPA tumors. In two post-operative cases the lesions exhibited an atypical position, being situated along the posterior surface of the petrous bone, at a distance from the IAC. The IAC , which exhibits absence of signal, is clearly recognized, due to its high contrast respective to the enhanced acoustic neuroma. Thus erosion of the bone surfaces, or widening of the IAC can be nicely depicted [ 181. In 33 (35,9x) lesions, an enlargement of the IAC due to smoothing of its posterior edge was noted, particularly in forms type 2 and 4. A widening of the IAC was recognizable in 47 (5 1.1%) tumors, associated with a strictly intracanalar lesion in 5 cases (5.4%). The widening of the IAC is a little less frequent in our collective than in a collective reported by Wu et al. in 1986, which published a rate of 7 1.8 y0 of enlargement of the IAC, in a collective of 78 acoustic neuromas examined with CT [ 81. Bilateral neuromas occur in neurolibromatosis type-2 [26-281. In our group, neurolibromatosis was noted in 8 patients (9.5 %), who all presented a bilateral neuroma. Most tumors were treated conservatively; two neuromas were operated, to alleviate significant compression on the pons. Usually the treatment is conservative in these cases, except when a risk of bilateral facial paralysis is present, or in cases of compression of the pons [26].
102
References 1 Ramsden RT. Acoustic tumors. In: Kerr AG, ed. Scott-Brown’s Otolaryngology. London: Butterworth’s, 1987; 500-533 2 Thomsen J, Tos M. Acoustic neuroma: clinical aspects, audiovestibular assessment, diagnostic delay, and growth rate. Am J Otol 1990; 11: 12-19 3 Portmann M, Dauman R, Duriez F, Portmann D, Dhillon RS. Modern diagnostic strategy for acoustic neuroma. Arch Otorhinolaryngol 1989; 246: 286-291 4 Windle-Taylor PC, Dhillon RS, Kenyon GS, Morrisson AW. Acoustic neuroma suspicion index: an aid to investigation and diagnosis. Laryngoscope 1984; 94: 1464-1467 5 Moffat DA, Hardy DC. Early diagnosis and surgical treatment of acoustic neuroma: is it cost effective. J R Sot Med 1989; 82: 329-332 6 Sarrat P, Zanaret M, Triglia JM, Koskas A, Giraud G, Cannoni M, Pech P, Chevrot L. Le diagnostic radiologique du neurinome de l’acoustique. A propos de 130 patients. J Radio1 1985; 66: 31-37 7 Barrs DM, Luxford WM, Becker TS, Brackmann DE. Computed tomography with gas cisternography for detection of small acoustic tumors. Arch Otolaryngol 1984; 110: 535-537 8 Wu EH, Tang YS, Zhang YT. CT in diagnosis of acoustic neuromas. AJNR 1986; 7: 645-650 9 Randell CP, Collins AG, Young JR, Haywood R, Thomas DJ, McDonnel MJ, Orr JS, Bydder GM, Steiner RE. Nuclear magnetic resonance imaging of posterior fossa tumors. AJR 1983; 141: 489-496 10 Armington WG, Harnsberger HR, Smoker WRK, Osborn AG. Normal and diseased acoustic pathway: evaluation with MR imaging. Radiology 1988; 167: 509-515 11 Gentry LR, Jacoby CG, Turski PA, Houston LW, Srother CM, Sackett JF. Cerebellopontine angle-petromastoid lesions; comparative study of diagnosis with MR imaging and CT. Radiology 1987; 162: 513-520 12 Mickhael MA, Ciric IS, Wolff AP. MR diagnosis of acoustic neuromas. J Comput Assist Tomogr 1987; 11: 232-235 13 Mahwhinney MB, Buckley JH, Worthington JS. Magnetic resonance imaging of the cerebellopontine angle. Br J Radio1 1986; 59: 961-969 14 Delillip GJ, Buchheit WA. Magnetic resonance imaging of acoustic neuromas. Neurosurgery 1985; 16: 763-765
15 Valvassori GE, Morales FG, Palacios E, Dobben GE. MR imaging of the normal and abnormal internal auditory canal. AJNR 1988; 9: 115-119 16 Wilms G, Decrop E, Plets C, Demaerel P, Marchal G, Van Hecke P, Baert AL. Magnetic resonance imaging in acoustic neuroma. Comparison with CT. J Belge Radio1 1989; 72: 151-158 17 Daniels DL, Millen SJ, Meyer GA, Pojunas KW, Kilgore DP, Shalfer KA, Williams AL, Haughton VM. MR detection of tumor in the internal auditory canal. AJNR 1987; 8: 249-252 18 Press GA, Hesselink JR. MR imaging of cerebellopontine angle and internal auditory canal lesions at 1.5 T. AJNR 1988; 9: 241-251 19 Curati WL, Graif M, Kingsley DPE, Niendorf HP, Young IR. Acoustic neuromas: Gd-DTPA enhancement in MR imaging. Radiology 1986; 158: 447-451 20 Stack JP, Ramsden RT, Antoun NM, Lye RH, Isherwood I, Jenkins JP. Magnetic resonance imaging of acoustic neuromas: the role of Gadolinium-DTPA. Br J Radio1 1988; 61: 800-805 21 Breger RK, Papke RA, Pojunas KW, Haughton VN, Williams AL, Daniels DL. Benign extra-axial tumors: contrast enhancement with Gd-DTPA. Radiology 1987; 163: 427-429 22 Watabe T, Azdma T. Tl and T2 measurements of meningiomas and neuromas before and after Gd-DTPA. AJNR 1989; 10: 463-470 23 Mikhael MA, Ciric IS, Wolff AP. Differentiation of cerebellopontine angle neuromas and meningiomas with MR imaging. J Comput Assist Tomogr 1985; 9: 852-856 24 Drapkin AJ, Rose WS. Multicystic acoustic neuroma: case report and differential diagnosis. Acta Radio1 1989; 30: 7-9 25 Thomsen J, Klinken L, Tos M. Calcified acoustic neurinoma. J Laryngol Otol 1984; 98: 727-732 26 Poignonec S, Lamas G, Sichez JP, Bokowy C, Soudant J, Van Effenterre R. Neurinome du VIII: Indications chirurgicales au tours des neurolibromatoses. Ann Oto-Laryngol (Paris) 1990; 107: 231-236 27 Aoki S, Barkovich AJ, Nishimura K, Kjos BO, Machida T, Logen P, Edwards M, Norman D. Neurofibromatosis types 1 and 2: cranial MR findings. Radiology 1989; 172: 527-534 28 Costantino PD, Friedman CD, Pelzer HJ. Neurofibromatosis type II of the head and neck. Arch Otolaryngol Head Neck Surg 1989; 115: 380-383.
96
EURRAD
1991 Elsevier
Science Publishers
B.V. All rights reserved. 0720-048X/91/$03.50
00 186
Magnetic resonance findings in 92 acoustic neuromas Bertrand Duvoisin”4, Julio Fern andes’, Dominique Doyen’, Jean-Marc Sterkers2 and Serge Bobin Departments
of ‘Neuroradiology
30to-Rhino-Laryngology,
and CIERM,
20to-Rhino-Laryngology,
Hospital Cochin Park
University Hospital of Bicetre, Le Kremlin Bicetre, France,
and 4Diagnostic Radiology. University Hospital CHUV,
(Received 29 November
Key words: Magnetic resonance
Alban Denys’,
1990; accepted
imaging, brain; Magnetic resonance
Lausanne,
Switzerland
after revision 3 April 1991)
imaging, acoustic neuroma;
Brain, MRI; Brain, neoplasm
Abstract A retrospective analysis of the MR findings in 92 cases of acoustic neuromas is presented. The method of examination included in all cases intravenous injection of Gadolinium (Gd-DTPA or DOTA) with realization of sections in the axial and coronal planes. In 21 cases native MR studies were performed in the axial plane, before Gadolinium injection, with TlWI (n = 21), and T2WI (n = 6) images. Tumors were strictly intracanalar in 19 cases (20.7%), only localized in the cerebellopontine angle (CPA) in 5 cases (5.4%), and in 68 cases (73.9%) the tumors had intra- and extracanalar components. In this last group of lesions, 63.2% completely filled the internal auditory canal (IAC), and 36.8% occupied the internal portion of the IAC. In most cases (85.3 %) the mean diameter of the CPA component was less than or equal to 2.5 cm. Lesions were more frequently homogeneous (58.8%) after Gd iv.-enhancement. Heterogeneity was noted mainly in large lesions (>2.5 cm: 100%). In all cases but one, the tumors were round, or oval-shaped, well-delineated, and did not present significant contact with the petrous bone. In 80.9% of CPA lesions, the center of the tumor was posteriorly excentered in relation to the internal auditory canal, whereas it was centered in 16.2 y0 of cases. Widening of the IAC was observed in 5 1.1 y0 of cases. In 45.2 y0 of tumors of the CPA, smoothing of the posterior edge of the porus was visible.
Introduction Acoustic neuromas are common extra-axial tumors, that usually arise from the vestibular division of the eighth cranial nerve, and characteristically occur in the cerebellopontine angle (CPA) and internal auditory canal (IAC). The clinical work-up of these lesions is complex; despite the variety of audiological and electrophysiological tests used, the clinical detection rate of small acoustic neuromas (mainly intrameatal) is fairly low [ 11. Computed tomography (CT), with intravenous contrast medium, and gas CT cisternography are valuable diagnostic tools to detect large or mediumsized acoustic neuromas [2-81. However CT studies bear morbidity due to radiation exposure and iodine injection, and beam-hardening artefacts often limit the
Address for reprints: Bertrand Duvoisin, M.D., Department of Diagnostic Radiology, University Hospital CHUV, 1011 Lausanne, Switzerland.
quality of the study; furthermore gas CT cisternography only depicts the internal limit of the tumor, and does not give any information about tumoral extent to the fundus of the IAC. Over recent years magnetic resonance (MR) imaging has proved to be the most sensitive tool for diagnosis of acoustic neuromas; it is more effective than CT with iodine injection or gas CT cisternography [ 9- 131. However small intracanalar neuromas, and post-operative tumoral recurrences can be difficult to assess on plain MR [ 14,161. Intravenous injection of Gadolinium (Gd-DTPA or DOTA) provides an optimal demonstration of acoustic neuromas, which usually enhance significantly, and exhibit a typical hyperintense signal. This feature allows exact assessment of tumoral extent into the lateral portion of the IAC, and a good demonstration of post-operative recurrences [ 16- 191. Having examined over the last years a fairly large number of patients with acoustic neuromas (92 tumors) in our MR center, we decided to review these studies retrospectively to determine the most typical MR features of acoustic neuromas.
Patients and Methods Consecutive patients with a diagnosis of acoustic neuroma, who underwent MR studies between December 1986 and May 1990, were assessed retrospectively. Eighty-four patients (37 men and 47 women) aged 15-81 years (mean, 53.7 years) were included in this study. The clinical symptoms included unilateral or bilateral deafness of variable duration, tinnitus and/or vertigo. Audiometric tests (audiograms and acoustic reflex testing), as well as labyrinthine tests (vestibular response, electronystagmography) were positive in the majority of patients. Auditory brainstem responses (ABRs) were obtained in 68 patients, and in all cases indicated a retrocochlear abnormality, suggesting an acoustic neuroma. In eight cases of neurofibromatosis, the lesions were bilateral Sixteen patients previously underwent operative excision of an acoustic neuroma; of these, one patient was operated bilaterally; clinical recurrence of the symptoms urged MR study. A CT study was obtained in all cases, four of these before MR imaging. In 70 cases, CT depicted abnormalities, including a mass in the CPA (n = 13), widening of the porus and/or IAC (n = 34) intracanalar or intrameatal contrast enhancement (n = 32), abnormal gas cisternography (n = 4). In 10 c ases CT did not disclose any abnormality. Fifty-nine patients were operated and in all of them the histological study depicted acoustic neuroma. Twenty-five patients were treated conservatively; these
patients were considered as having acoustic neuromas on the basis of clinical criteria, audiological and electrophysiological tests, and typical MR findings. Of these patients, follow-up MR studies were performed in 22 cases, showing stability of the lesions in 14 cases, and an increasing size in 8 cases (slight: y1= 5, moderate: n=3,marked:n= 1). All MR examinations were performed on a 1.5 T imaging unit (Signa; GE Medical Systems, Milwaukee, U.S.A.), with a head coil, having a 24-cm field of view, 256 x 256 matrices corresponding to a pixel size of 0.95 x 0.95 mm, and consecutive 3-mm-thick slices (in some rare cases with 5-mm-thick slices). Spin-echo (SE) MR images were obtained routinely in the axial and coronal planes. In all cases Gd-DTPA (or -DOTA) was injected intravenously (dose: 0.1 mmol/kg); among these, post-contrast sequences were selectively obtained in 63 patients. In the other 21 patients, native MR studies were also obtained before Gd injection, with T- 1 weighted (n = 21), and T-2 weighted (n = 5) images. All acoustic neuromas were separately evaluated by two of us to assess the following parameters: (1) localization (intra- and/or extracanalar); (2) size; (3) homogeneous vs heterogeneous; (4) borders and shape of the extracanalar portion of the tumor; (5) extension of the tumor to the fundus of the IAC; (6) position of the center of the extracanalar component with respect to the IAC (Fig. 1 shows the spectrum ofintra- and extracanalar types of lesions); and (7) widening of the IAC.
a
Fig. 1. Drawing of the different sites of intra- and extracanalar acoustic neuromas. Intracanalar lesions include types a, b and c, according to the tumoral extent into the internal auditory canal (IAC). Note that types a and b reach the ftmdus. Extracanalar tumors include types 1 to 6, according to the position ofthe lesion in respect to the IAC. Types 2 and 4 are posteriorly excentered; the former does not extend anteriorly to the porus. whereas type 4 does. Note that not all types present a significant contact with the petrous bone, except type 5. Difference between types 1 and 3 is only in size, 3 being larger than 1. Tumors with intra- and extracanalar components are labelled in the text: a 1 to a 4, and c 1 to c5.
98
a
b
Fig. 2. Strictly intracanalar right acoustic neuroma, type a. Gadolinium i.v.-enhanced axial (a) and coronal (b) MR sections, depicting extension of the tumor to the entire IAC (arrows), reaching the fundus, without widening of the IAC. Note hypersignal of both petrous bone apex, due to fatty bone marrow (black star).
Results In 19 cases (20.7%) the acoustic neuroma was strictly intracanalar. The tumor occupied the entire IAC (type a, Fig. 2) in 14 cases (73.7%) and the external portion of the IAC (type b, Fig. 3) in 5 cases (26.3%). In all cases the fundus of the IAC was reached. A slight to moderate widening of the IAC was noted in 5 cases (26.3 %). In one case a bilateral neuroma was observed, occupying the whole IAC; the patient suffered Von Recklinghausen disease, and had been operated eight years previously on the left side for an acoustic neuroma mainly localized in the CPA cistern. Both neuromas had approximately the same size on MR; as no apparent growth was evident since a CT performed two years previously, a conservative treatment was conducted. In 6 cases, including the patient with bilateral lesions, TlWI axial MR images were obtained before Gd-DTPA injection. On these studies the lesions were homogeneous, isointense to the pons in 5 cases, hypointense in one case and heterogeneous with a central hypointense zone in the last case. In this case T2WI sections showed that the tumor was isointense. In all 19 cases the tumor exhibited a hyperintense signal on Gd i.v.-enhanced TlWI sections, and was homogeneous in 18 cases. In five cases (5.4%) the tumor was selectively extracanalar. The size of the lesions ranged between 4 mm and 45 mm in mean diameter. The localization included one case each of types 2-6. In a single case the posterior edge of the IAC opening was smoothed. Three cases
had been previously operated for an acoustic neuroma, recurrence of which urged the MR study. After GdDTPA injection, the lesions showed significant enhancement and all but one were homogeneous. The last tumor exhibited a heterogeneous aspect, with a large central necrotic zone. In all other 68 cases (73.9%) the tumor had extra-
Fig. 3. Strictly intracanalar right acoustic neuroma, type b. Gadolinium iv.-enhanced axial MR section, showing a lesion occupying 2/3 of the external part of the right IAC (small arrows). Note a controlateral larger lesion, type a (large arrow), which enlarges the left IAC; the patient suffered Von Recklinghausen’s disease.
a
C
b
Fig. 4. Various types of lesions in the cerebellopontine angle, all Gadolinium i.v.-enhanced axial sections. (a) Left tumor, type 2, (arrow) which almost entirely fills the IAC; an oval-shaped, hyperintense lesion, with a broad base on the right petrous apex, is typical of a meningioma (small arrows). The patient suffered Von Recklinghausen’s disease. (b) Right tumor, type 4, which causes a moderate impression on the pons (arrows), and extends into the internal portion of the IAC. (c) Left recurrent acoustic neuroma, type 5 (arrows), which has a fairly large surface of contact with the petrous bone; note hyperintense signal of the left petrous bone due to fat packing, after a previous translabyrinthine resection of an acoustic neuroma.
and intracanalar components. According to the classilication depicted in Fig. 1, the most frequent tumors included types a 2 (n = 22) (Fig. 4a), a 4 (n = 16) (Fig. 4b), c 2 (n = 1 l), and c 4 (n = 6) representing respectively 32.4%, 23.5%, 16.2% and 8.8% of this group of lesions. Other types occurred less frequently. Fifty-eight lesions (85.3 %) had a mean diameter inferior or equal to 2.5 cm, seven (10.3%) measured between 2.5 cm and 4 cm, whereas three tumors (4.4%) were larger than 4 cm. In all cases but one, the tumors exhibited a round, or ovoid shape, with clearly defined borders, and did not present a significant contact with the petrous bone. A single neuroma presented a large surface of contact with the posterior petrous bone (Fig. 4~). In 15 cases, including two patients with bilateral lesions, native MR studies were obtained. Fourteen neuromas were homogeneous, and exhibited, on TlWI sequences, an isosignal to the pons in 7 cases, and a hyposignal in 7 cases. In live cases T2WI images were also obtained; the tumors depicted a hyposignal respective to the pons in 3 cases, and were not clearly recognized in the other 2 cases. The other 3 neuromas were heterogeneous on Tl WI and /or T2WI sequences. One lesion showed a peripheral hyperintense signal in both series, which was considered as post-hemorrhagic. After Gd i.v.enhancement, 40 lesions (58.8%) were homogeneous and exhibited a diffuse hyperintense signal. In 28 cases (41.2%) the tumors were heterogeneous, with internal hypointense zones (Fig. 5). In most of the lesions less than 2.5 cm this heterogeneity was absent or discrete,
whereas in all large lesions (>2.5 cm) it was fairly marked. Due to its absence of signal the cortical surface of the porus was easily recognized on MR images; in 32 cases (47.1%) a clear smoothing of the posterior edge of the IAC was noted, associated to neuromas type 2 and 4. A widening of the IAC, recognizable due to signilicant asymmetry relative to the controlateral side, was observed in 41 tumors (60.3 %) (Fig. 6). In eight patients a large CPA mass caused a compression on the pons, with displacement of the fourth ventricle in three cases, and obstructive hydrocephalus in two cases. Among these 68 acoustic neuromas, 14 were observed in patients with neurolibromatosis, and bilateral lesions. Nine lesions constituted post-operative recurin patients with rences, including four neuromas neurolibromatosis. Discussion Until the introduction of MR, i.v.-enhanced CT and gas CT cisternography were considered the definitive tests to rule out acoustic neuromas [2-81. However false-negative or equivocal results occur on i.v.-enhanced CT studies, due to beam attenuation and/or petrous bone artefacts. The application of gas CT cistemography has improved the sensitivity of detection or exclusion of small tumors, but this procedure is not innocuous, and occasional false-negative and falsepositive results occur [7]. Furthermore, this type of investigation cannot demonstrate the lateral intra-
100
canalar extent of the tumor, particularly with respect to the fundus of the IAC. Over recent years MR has proven to be the most sensitive diagnostic tool in cases of acoustic neuromas, being more effective than CT with iodine injection or gas CT cistemography [9-133. However, small intracanalar neuromas, and post-operative tumoral recurrences are difficult to assess on plain MR [ 14,151. After iv injection of Gd-DTPA (-DOTA) acoustic neuromas significantly enhance [ 19-221. This feature allows detection of tumoral recurrences, and clear depiction of the lateral extent of intracanalar lesions, which is man-
datory to determine the prognosis and the type of surgical procedure[ 17-191. During a 42-month period we examined in our center 92 acoustic neuromas in 84 patients, including eight patients with neurofibromatosis. In 59 cases surgical confirmation of the MR diagnosis was obtained, whereas in 25 patients a conservative treatment was conducted. In these latter patients the lesions were considered as acoustic neuromas, following clinical criteria, audiological and electrophysiological tests, and a typical MR aspect. Control MR studies were obtained in 22 cases, showing stability of the lesions in 14 cases, and an increasing size in 8 cases. The distribution of the site of neuromas showed 19 strictly intracanalar tumors (20.7%), five pure extracanalar masses (5.4%), and 68 lesions with extra- and intracanalar components (73.9%). Our primary goal in the present study was to evaluate the lesions which presented components both in the IAC and CPA in order to establish the typical MR findings observed in such lesions. Sixty-eight (73.9%)
Fig. 5. Heterogeneous acoustic neuromas. (a) Left tumor, which depicts a large central hypointensity, due to cystic necrosis, and a regular peripheral enhancing rim (arrow). (b and c), (same patient) large right acoustic neuroma, exhibiting significant heteregeneous features, due to necrosis and hemorrhage; on TlWI unenhanced section (b), a posterior hyperintense biloculate zone (arrowheads) represents hemorrhage within cystic spaces. Other hypointense cystic lesions are recognized. After Gadolinium i.v.-enhancement, the solid portions of the tumor are well-delineated of the necroticohemorrhagic components.
a
b
a b Fig. 6. Widening of the right porus and IAC, consecutive to a large type 3 tumor; (a) CT and (b) MR axial sections. A strong correlation observed between CT and MR: both methods well define the cortical surfaces (white on CT, and black on MR). of the neuromas reviewed had such a feature. Among these tumors, 43 (63.2%) completely filled the IAC, till the fundus, whereas 25 (36.8 %) only occupied the internal part of the IAC. In the majority of cases (85.3%), the CPA component had a mean diameter less than or equal to 2.5 cm. All CPA lesions presented a spherical or ovoid shape, with well-defined borders. In all cases the contact between the petrous bone and the round extracanalar component of the neuroma was small, or absent. These particulars, in conjunction with statistically significant higher frequency of acoustic neuromas allows to differentiate neuromas from meningiomas, which are classically crescentic, elongated, dural-based, and eccentric to the porus [ 8, 11, 231. Fourty (58.8%) tumors were homogeneous after Gd-iv enhancement, whereas 28 (4 1.2%) exhibited a variable inhomogeneity. This feature was rarely observed (25 %) in small (< 1 cm) lesions, whereas its frequency and extent increased parallel to lesion size with, respectively, 53.6% of lesions inferior to 2.5 cm, and 100% in lesions larger than 2.5 cm. The heterogeneity of the lesion is caused by necrosis, or cystic transformation. In the literature cystic acoustic neuromas are rarely reported [24]. The topographic relationship between acoustic neuromas and the IAC classically describes that the tumor is centered on the IAC [ 8, 111. In our group of tumors with extra- and intracanalar components, most lesions (80.9%) were posteriorly eccentered to the IAC, whereas only 11 (16.2%) tumors were centered on the porus. This can be easily explained by the fact that acoustic neuromas primarily involve the vestibular
is
nerve, which is posteriorly placed. Thus most CPA tumors have their equator posteriorly eccentered respective to the IAC. This feature seems a very reliable sign of an acoustic neuroma, in cases of CPA tumors. In two post-operative cases the lesions exhibited an atypical position, being situated along the posterior surface of the petrous bone, at a distance from the IAC. The IAC , which exhibits absence of signal, is clearly recognized, due to its high contrast respective to the enhanced acoustic neuroma. Thus erosion of the bone surfaces, or widening of the IAC can be nicely depicted [ 181. In 33 (35,9x) lesions, an enlargement of the IAC due to smoothing of its posterior edge was noted, particularly in forms type 2 and 4. A widening of the IAC was recognizable in 47 (5 1.1%) tumors, associated with a strictly intracanalar lesion in 5 cases (5.4%). The widening of the IAC is a little less frequent in our collective than in a collective reported by Wu et al. in 1986, which published a rate of 7 1.8 y0 of enlargement of the IAC, in a collective of 78 acoustic neuromas examined with CT [ 81. Bilateral neuromas occur in neurolibromatosis type-2 [26-281. In our group, neurolibromatosis was noted in 8 patients (9.5 %), who all presented a bilateral neuroma. Most tumors were treated conservatively; two neuromas were operated, to alleviate significant compression on the pons. Usually the treatment is conservative in these cases, except when a risk of bilateral facial paralysis is present, or in cases of compression of the pons [26].
102
References 1 Ramsden RT. Acoustic tumors. In: Kerr AG, ed. Scott-Brown’s Otolaryngology. London: Butterworth’s, 1987; 500-533 2 Thomsen J, Tos M. Acoustic neuroma: clinical aspects, audiovestibular assessment, diagnostic delay, and growth rate. Am J Otol 1990; 11: 12-19 3 Portmann M, Dauman R, Duriez F, Portmann D, Dhillon RS. Modern diagnostic strategy for acoustic neuroma. Arch Otorhinolaryngol 1989; 246: 286-291 4 Windle-Taylor PC, Dhillon RS, Kenyon GS, Morrisson AW. Acoustic neuroma suspicion index: an aid to investigation and diagnosis. Laryngoscope 1984; 94: 1464-1467 5 Moffat DA, Hardy DC. Early diagnosis and surgical treatment of acoustic neuroma: is it cost effective. J R Sot Med 1989; 82: 329-332 6 Sarrat P, Zanaret M, Triglia JM, Koskas A, Giraud G, Cannoni M, Pech P, Chevrot L. Le diagnostic radiologique du neurinome de l’acoustique. A propos de 130 patients. J Radio1 1985; 66: 31-37 7 Barrs DM, Luxford WM, Becker TS, Brackmann DE. Computed tomography with gas cisternography for detection of small acoustic tumors. Arch Otolaryngol 1984; 110: 535-537 8 Wu EH, Tang YS, Zhang YT. CT in diagnosis of acoustic neuromas. AJNR 1986; 7: 645-650 9 Randell CP, Collins AG, Young JR, Haywood R, Thomas DJ, McDonnel MJ, Orr JS, Bydder GM, Steiner RE. Nuclear magnetic resonance imaging of posterior fossa tumors. AJR 1983; 141: 489-496 10 Armington WG, Harnsberger HR, Smoker WRK, Osborn AG. Normal and diseased acoustic pathway: evaluation with MR imaging. Radiology 1988; 167: 509-515 11 Gentry LR, Jacoby CG, Turski PA, Houston LW, Srother CM, Sackett JF. Cerebellopontine angle-petromastoid lesions; comparative study of diagnosis with MR imaging and CT. Radiology 1987; 162: 513-520 12 Mickhael MA, Ciric IS, Wolff AP. MR diagnosis of acoustic neuromas. J Comput Assist Tomogr 1987; 11: 232-235 13 Mahwhinney MB, Buckley JH, Worthington JS. Magnetic resonance imaging of the cerebellopontine angle. Br J Radio1 1986; 59: 961-969 14 Delillip GJ, Buchheit WA. Magnetic resonance imaging of acoustic neuromas. Neurosurgery 1985; 16: 763-765
15 Valvassori GE, Morales FG, Palacios E, Dobben GE. MR imaging of the normal and abnormal internal auditory canal. AJNR 1988; 9: 115-119 16 Wilms G, Decrop E, Plets C, Demaerel P, Marchal G, Van Hecke P, Baert AL. Magnetic resonance imaging in acoustic neuroma. Comparison with CT. J Belge Radio1 1989; 72: 151-158 17 Daniels DL, Millen SJ, Meyer GA, Pojunas KW, Kilgore DP, Shalfer KA, Williams AL, Haughton VM. MR detection of tumor in the internal auditory canal. AJNR 1987; 8: 249-252 18 Press GA, Hesselink JR. MR imaging of cerebellopontine angle and internal auditory canal lesions at 1.5 T. AJNR 1988; 9: 241-251 19 Curati WL, Graif M, Kingsley DPE, Niendorf HP, Young IR. Acoustic neuromas: Gd-DTPA enhancement in MR imaging. Radiology 1986; 158: 447-451 20 Stack JP, Ramsden RT, Antoun NM, Lye RH, Isherwood I, Jenkins JP. Magnetic resonance imaging of acoustic neuromas: the role of Gadolinium-DTPA. Br J Radio1 1988; 61: 800-805 21 Breger RK, Papke RA, Pojunas KW, Haughton VN, Williams AL, Daniels DL. Benign extra-axial tumors: contrast enhancement with Gd-DTPA. Radiology 1987; 163: 427-429 22 Watabe T, Azdma T. Tl and T2 measurements of meningiomas and neuromas before and after Gd-DTPA. AJNR 1989; 10: 463-470 23 Mikhael MA, Ciric IS, Wolff AP. Differentiation of cerebellopontine angle neuromas and meningiomas with MR imaging. J Comput Assist Tomogr 1985; 9: 852-856 24 Drapkin AJ, Rose WS. Multicystic acoustic neuroma: case report and differential diagnosis. Acta Radio1 1989; 30: 7-9 25 Thomsen J, Klinken L, Tos M. Calcified acoustic neurinoma. J Laryngol Otol 1984; 98: 727-732 26 Poignonec S, Lamas G, Sichez JP, Bokowy C, Soudant J, Van Effenterre R. Neurinome du VIII: Indications chirurgicales au tours des neurolibromatoses. Ann Oto-Laryngol (Paris) 1990; 107: 231-236 27 Aoki S, Barkovich AJ, Nishimura K, Kjos BO, Machida T, Logen P, Edwards M, Norman D. Neurofibromatosis types 1 and 2: cranial MR findings. Radiology 1989; 172: 527-534 28 Costantino PD, Friedman CD, Pelzer HJ. Neurofibromatosis type II of the head and neck. Arch Otolaryngol Head Neck Surg 1989; 115: 380-383.