Neuroradiology (1992) 34:470-474
Neuro--
radiology 9 Springer-Verlag 1992
The demonstration of glomus tumours by subtraction MRI G. A. S. Lloyd, A. D. Cheesman, P. D. Phelps, and C. M. P. King Royal National Throat Nose and Ear Hospital, London, UK Received: 23 January 1992
Summary. Thirteen patients with 14 glomus tumours have been examined by subtraction gadolinium-enhanced magnetic resonance imaging (MRI), with T l - w e i g h t e d M R sequences before and after intravenous gadoliniumDTPA. To eliminate m o v e m e n t between subtraction pairs, the patient remains in the tunnel of the imager during administration of the contrast medium, and the venepuncture is m a d e into the dorsum of the foot. The effect of the subtraction process is to r e m o v e the N M R signal from the final image so that the photographic densities recorded are d e p e n d e n t on the vascularity of the tissue concerned, normal or abnormal. A particular advantage is the removal of fat signal: the low vascularity of adipose tissue ensures that it is recorded as of minimal density. T h e extent of skull base glomus tumours has been shown optimally by this technique. Subtraction can also help differentiate glomus t y m p a n i c u m f r o m glomus jugulare lesions, which m a y be of crucial importance when deciding the surgical approach. In addition to diagnosis, the technique is also important post-operatively, when imaging is needed to show residual or recurrent turnout and to monitor the effects of radiotherapy.
tigations, 2 post-radiotherapy follow-up studies, and 5 patients with recurrent tumours were examined. Tl-weighted images were obtained before and after injection of Gd-DTPA. For satisfactory subtraction it is essential that the post-injection images should reproduce exactly the cross-sections given by the initial series. In order to do this the patient's head must be secured so that there is no movement between the sequences, and the patient must not be moved from the tunnel of the imager or indeed in any way for administration of Gd. Venepuncture is therefore performed on the dorsum of the foot, prior to putting the patient into the imager, and the needle is irrigated with saline while the unenhanced images are obtained. The shortest acquisition times available are employed (TR 600, TE 10 ms at 1.5 T). Subtraction has proved more satisfactory using second-order photographic subtraction from the hard copy, rather than the subtraction programme in the computer software: minor errors of registration are more easily corrected by the photographic technique. The subtraction process removes the NMR signal completely from the final image; and since the enhanced images show the vascular phase of Gd enhancement (within 5 rain of injection), at which stage extravasation of contrast medium is minimal and itself partly dependent upon blood supply [2], the photographic densities are a record of the vascularity of the tissue concerned, normal or pathological: the end result is effectively a vasculogram.
Results Key words: Glomus tumours - Gadolinium - Subtraction - Magnetic resonance imaging The technique of subtraction gadolinium-enhanced magnetic resonance imaging ( G d M R I ) for demonstration of tumours of the skull base and paranasal sinuses has b e e n reported previously [1]. T h e purpose of this p a p e r is to describe its application to the imaging of glomus tumours of the skull base and middle ear (glomus jugulare, glomus t y m p a n i c u m and glomus vagale), emphasising its advantages over existing methods of investigation for initial diagnosis and patient management.
Materials and methods Thirteen patients with 14 glomus tumours (12 glomus jugulare, i glomus tympanicum and i glomus vagale) have been investigated using subtraction GdMRI. There were 6 pre-treatment initial inves-
The exact extent of the glomus tumour was better shown by the subtraction m e t h o d than by the unsubtracted G d M R I in all 13 patients examined. This was due to improved demonstration of the glomus tumour at the skull base; visualisation of any intracranial extension was not significantly improved. Subtraction was particularly important for demonstration of recurrent glomus jugulare tumours, which do not as a rule extend into the middle ear, and do not therefore present the typical features of a tum o u r examined ab initio (see below). Fat subtraction played an important role in improved tumour demonstration in five patients (Table 1). A second important feature of the subtraction studies was improved discrimination of t u m o u r from veins, which also produce a strong signal because of their G d content. Some of the veins of the cervical spinal plexus present a saccular appearance and m a y therefore be especially con-
471
Table 1. Advantages of subtraction MRI Patient
Age and sex
Clinical features
Subtraction results
1
24 F
Hearing loss; red mass behind drum
Coronal sections only; tumour delineation improved, mostly the result of identification of saccular veins in cervical plexus
2
49 M
Hearing loss, thumping tinnitus, red eardrum
Extent of tumour better shown on axial subtractions, which also revealed prominent veins in tumour mass
3
71 F
Injected tumour behind drum, facial palsy
Better demonstration of tumour extent in two planes; fat subtraction and identification of veins important
4
69 F
Previous radiotherapy, hearing loss, pulsatile tinnitus
Improved demonstration; tumour shown to involve jugular vein on subtraction
5
54 M
20-year history of glomus jugulare tumour treated by surgery and radiotherapy
Deep extension of recurrent tumour at skull base much better shown by subtraction
6
65 F
Left tinnitus, vascular mass in middle ear, pulsating eardrum
Glomus tympanicum. Subtraction showed exact relationship of tumour to jugular vein and floor of middle ear
7
44M
Pulsatile tinnitus, recurrent glomus tumour
Better evidence of recurrence on axial subtractions
8
33 F
Left carotid body tumour, recurrent glomus jugulare
Better shown on axial and coronal subtractions mostly due to fat subtraction
9
41 F
Previous embolisation and incomplete removal of glomus jugulare (left), glomus vagale (right)
Coronal subtractions only; boundaries of tumour optimally shown by subtraction
10
28 F
Hearing loss, red swelling in middle ear
Exact extent at skull base best demonstrated by subtraction, separating tumour from veins and fat signal
11
55 M
Bulge in external meatus floor, recurrent glomus jugulare tumour
Subtraction showed exact extent of tumour recurrence; fat used at previous surgery completely removed from image
12
69 F
Fullness in right ear, lower part of drum red
Exact extent of tumour at skull base better defined by subtraction and its relationship to spinal venous plexus shown
13
29 M
1984 glomus turnout left ear treated by radiotherapy; now tinnitus and red eardrum
Extent of tumour at skull base more accurately shown in two planes by subtraction
fusing a n d difficult to r e c o g n i s e o n u n s u b t r a c t e d images. Two f e a t u r e s h e l p e d to d i f f e r e n t i a t e g l o m u s t u m o u r f r o m n o r m a l veins: the t y p i c a l c o n f i g u r a t i o n o f a g l o m u s j u g u lare t u m o u r in c o r o n a l s e c t i o n (Fig. 1), a n d t h e p r e s e n c e o f l a r g e vessels in t h e t u m o u r mass, s h o w n as small r o u n d e d l a c u n a e on t h e s u b t r a c t i o n i m a g e (Fig.2). O v e r a l l , imp r o v e d r e c o g n i t i o n o f t h e j u g u l a r v e i n a n d cervical spinal p l e x u s m a d e an i m p o r t a n t c o n t r i b u t i o n to the i m a g i n g in 7 p a t i e n t s ( T a b l e 1).
Discussion I m a g i n g t e c h n i q u e s a r e r e q u i r e d at all stages of m a n a g e m e n t o f g l o m u s t u m o u r s : in p r e - t r e a t m e n t assessment, c o n c e r n e d with d i a g n o s i s o f t h e t y p e o f g l o m u s t u m o u r a n d its extent; d u r i n g t r e a t m e n t , w h e n a n g i o g r a p h y is r e q u i r e d if e m b o l i s a t i o n is u n d e r t a k e n p r i o r to resection; a n d p o s t - o p e r a t i v e l y , to s h o w r e c u r r e n t t u m o u r o r t h e effects o f r a d i o t h e r a p y . A t t h e p r e - t r e a t m e n t stage t h e
Fig.L a Gadolinium-enhanced coronal T1weighted image showing a left glomus jugulare tumour, b Same section after subtraction. The exact extent of the tumour is more clearly shown, including its extension down the jugular vein. Note the typical appearance of a glomus jugulare tumour, with a small lobule of tumour projecting into the middle ear cleft (arrow); also the strong signal from the gadolinium in the spinal venous plexus
472
Fig.2. Coronal subtraction sections showing a large glomus tumour. There are small rounded defects (arrows) within the tumour due to its arterial supply.Note also the associated dural enhancement
first requirement is to distinguish between a glomus tympanicum tumour localised to the middle ear cleft and a glomus jugulare tumour with varying degrees of skull base and intracranial involvement. This is important because the surgical approach is totally different: the glomus tympanicum tumour can usually be removed by a transmeatal approach, whereas most glomus jugulare tumours require some form of complex skull base resection. In the majority of patients differentiation between the two types of tumour can be made on CT: glomus jugulare lesions show a characteristic ragged erosion and enlargement of the jugular fossa, whereas glomus tympanicum tumours show a discrete mass in the middle ear arising from the promontory, without jugular fossa enlargement. However, there are borderline cases in which it is difficult to distinguish between a large glomus tympanicum tumour which has spread downwards to involve the jugular bulb, and a small glomus jugulare tumours which has spread upwards into the middle ear, without bone erosion demonstrable on CT [3]. The crucial question from the surgeon's point of view is whether the jugular bulb is significantly
involved by the tumour. Minor indentations on the bulb may be shown by vascular imaging, but subtraction G d M R I is now the best non-invasive technique for making this critical differentiation. Although no bony landmarks are visible on subtraction GdMRI, the vascular epithelial lining of the external auditory meatus and middle ear is clearly visible so that the relationship of the tumour mass to the middle ear cavity can be readily identified (Fig. 3). Demonstration of a glomus tympanicum lesion within the middle ear, discrete from the jugular bulb is also possible (Fig. 4). In coronal sections the main mass of a typical glomus jugulare tumour is seen below the line of the floor of the middle ear, with a small lobule projecting into the middle ear cleft (Figs. 1, 3). It has also been shown that G d M R I can enable differentiation of turnout from retained secretion resulting from eustachian tube obstruction [4]. The removal of fat signal is a most important advantage of the subtraction technique. In addition to being present naturally at the skull base, fat from the dermis is frequently used to pack the surgical cavity after removal of the tumour. The effect of Gd is to convert the M R I charac-
473
Fig.3. Coronal subtraction showing the typical configuration of a small glomus jugulare tumour (cf. Fig.la). The relationship of the tumour to the mucosa of the external auditory meatus is also demonstrated (arrows) Fig.4. Glomus tympanicum tumour the coronal subtraction study shows the turnout confined to the middle ear (arrow) and discrete from the jugular bulb
Fig.5. Coronal subtraction gadolinium-enhanced magnetic resonance imaging showing a recurrent glomus jugulare turnout which presented as a small bulge on the floor of the external auditory meatus
teristics of vascular tissue into those closely resembling normal fat, giving high signal on Tl-weighted images, and therefore making it difficult to differentiate fat from tumour on enhanced images. A partial ~olution to the problem is to use short tau inversion recovery (STIR) sequences [5, 6] which suppress the signal from fat. However, they also suppress the signal from Gd-enhanced tissues, resulting in diminished image contrast. Other methods of fat suppression include chemical shift imaging [7] and pre-saturation techniques [8], but subtraction provides the simplest method not requiring any additional imaging sequence. The low vascularity of adipose tissue ensures that it is recorded on the subtraction studies as an image of minimal density, which does not obscure the pathological enhancement. If surgery is to be planned optimally, accurate pretreatment assessment of the extent of the tumour is necessary, particularly in the region of the roof of the jugular fossa, the various segments of the carotid canal, the foramen lacerum and intracranially, and cases considered more appropriate for primary treatment by radiotherapy (vide infra) also require careful radiological assessment to ensure accurate planning of the radiation fields. Fisch et al. [9, 10] classified glomus jugulare tumours according to the extent of temporal bone erosion, and the degree of invasion of the carotid canal and posterior fossa, and their surgical approach is based upon this. All ttunours without, intracranial extension and those extending up to 2 cm intradurally can be resected via a single-stage infratemporal fossa approach; tumours extending more than 2 cm intradurally need a combined two-stage procedure by otologist and neurosurgeon. Our own surgical technique has been modified to allow a single stage otoneurological approach, and in selected cases the facial nerve can be preserved in the fallopian canal, removing the small risk of facial palsy that results from its transposition in the infratemporal fossa procedure.
There are, however, conflicting views concerning the best treatment for glomus jugulare tumours, especially the place of surgery and radiotherapy. This is primarily because of diagnostic inability to assess the effectiveness of any particular treatment in eradicating the tumour or the rate of growth of any recurrence. Radiotherapy has been claimed to be the ideal treatment modality because of its low morbidity, and although its supporters would agree that turnout eradication is not common, they claim that it is effective in stopping further growth. Unfortunately, diagnostic imaging has hitherto been inadequate for confirming this hypothesis. Most surgeons would agree that radiotherapy is the modality of choice for large symptomatic tumours in the elderly or infirm, while most radiotherapists accept that small tumours in the young are best treated surgically. Radical surgery, despite initial increased morbidity, can eradicate the tumour and any functional deficit resolves or is amenable to further surgery. However, the extremely complex anatomy of the skull base often makes it difficult to be sure of complete resection. In these circumstances there is need for follow-up imaging, to monitor the effects of radiotherapy or to show tumour recurrence (Fig. 5). CT and MRI have been the non-invasive tests most widely used for this purpose, but often producing uncertain evidence of turnour recurrence. The use of the subtraction method overcomes some of these problems and produces images similar to angiography in that they are dependent on tumour vascularity. Turnout presence and extent is thus shown optimally without the use of more invasive techniques.
References
1. Lloyd GAS, Barker PG (1991) Subtraction magnetic resonance for tumours of the skull base and sinuses: a new imaging technique. J Laryngol Oto1105: 628-631
474 2. Brix G, Semmler W, Port R, Schad R, Layer G, Lorentz WJ (1991) Pharmacokinetic parameters in CNS Gd-DTPA enhanced MR imaging. J Comput Assist Tomogr 15:621-628 3. Phelps PD, Stansbie JM (1988) Glomus jugulare or tympanicure? The role of CT and MR imaging with gadolinium DTPA. J Laryngol Oto1102: 766-776 4. Phelps PD (1990) Glomus tumours of the ear: an imaging regime. Clin Radio141:301-305 5. Bydder GM, Young IR (1985) MR imaging: clinical use of the inversion recovery sequence. J Comput Assist Tomogr 9:659-675 6. Dwyer AJ, Frank JA, Reinig JW, Hickey AM, Doppman JL (1988) Short T1 inversion recovery pulse sequences. Analysis and initial experience in cancer imaging. Radiology 168:827-836 7. Simon JH, Szumowski J (1989) Chemical shift imaging with paramagnetic contrast material enhancement for improved lesion depiction. Radiology 171:539-543
8. Barakos JA, Dillon WP, Chew WM (1991) Orbit, skull base and pharynx: contrast enhanced fat suppression MR imaging. Radiology 179:191-198 9. Fisch U (1982) Infratemporal fossa approach for glomus tumours of the temporal bone. Ann Otol Rhinol Laryngo191: 474479 10. Fisch U, Fagan R Valavanis A (1984) The infratemporal fossa approach for the lateral skull base. Otolaryngol Clin North Am 17: 513-552
Dr. G. A. S. Lloyd Royal National Throat Nose and Ear Hospital Grays Inn Road London WCIX 8DA, UK
Neuro--
radiology 9 Springer-Verlag 1992
The demonstration of glomus tumours by subtraction MRI G. A. S. Lloyd, A. D. Cheesman, P. D. Phelps, and C. M. P. King Royal National Throat Nose and Ear Hospital, London, UK Received: 23 January 1992
Summary. Thirteen patients with 14 glomus tumours have been examined by subtraction gadolinium-enhanced magnetic resonance imaging (MRI), with T l - w e i g h t e d M R sequences before and after intravenous gadoliniumDTPA. To eliminate m o v e m e n t between subtraction pairs, the patient remains in the tunnel of the imager during administration of the contrast medium, and the venepuncture is m a d e into the dorsum of the foot. The effect of the subtraction process is to r e m o v e the N M R signal from the final image so that the photographic densities recorded are d e p e n d e n t on the vascularity of the tissue concerned, normal or abnormal. A particular advantage is the removal of fat signal: the low vascularity of adipose tissue ensures that it is recorded as of minimal density. T h e extent of skull base glomus tumours has been shown optimally by this technique. Subtraction can also help differentiate glomus t y m p a n i c u m f r o m glomus jugulare lesions, which m a y be of crucial importance when deciding the surgical approach. In addition to diagnosis, the technique is also important post-operatively, when imaging is needed to show residual or recurrent turnout and to monitor the effects of radiotherapy.
tigations, 2 post-radiotherapy follow-up studies, and 5 patients with recurrent tumours were examined. Tl-weighted images were obtained before and after injection of Gd-DTPA. For satisfactory subtraction it is essential that the post-injection images should reproduce exactly the cross-sections given by the initial series. In order to do this the patient's head must be secured so that there is no movement between the sequences, and the patient must not be moved from the tunnel of the imager or indeed in any way for administration of Gd. Venepuncture is therefore performed on the dorsum of the foot, prior to putting the patient into the imager, and the needle is irrigated with saline while the unenhanced images are obtained. The shortest acquisition times available are employed (TR 600, TE 10 ms at 1.5 T). Subtraction has proved more satisfactory using second-order photographic subtraction from the hard copy, rather than the subtraction programme in the computer software: minor errors of registration are more easily corrected by the photographic technique. The subtraction process removes the NMR signal completely from the final image; and since the enhanced images show the vascular phase of Gd enhancement (within 5 rain of injection), at which stage extravasation of contrast medium is minimal and itself partly dependent upon blood supply [2], the photographic densities are a record of the vascularity of the tissue concerned, normal or pathological: the end result is effectively a vasculogram.
Results Key words: Glomus tumours - Gadolinium - Subtraction - Magnetic resonance imaging The technique of subtraction gadolinium-enhanced magnetic resonance imaging ( G d M R I ) for demonstration of tumours of the skull base and paranasal sinuses has b e e n reported previously [1]. T h e purpose of this p a p e r is to describe its application to the imaging of glomus tumours of the skull base and middle ear (glomus jugulare, glomus t y m p a n i c u m and glomus vagale), emphasising its advantages over existing methods of investigation for initial diagnosis and patient management.
Materials and methods Thirteen patients with 14 glomus tumours (12 glomus jugulare, i glomus tympanicum and i glomus vagale) have been investigated using subtraction GdMRI. There were 6 pre-treatment initial inves-
The exact extent of the glomus tumour was better shown by the subtraction m e t h o d than by the unsubtracted G d M R I in all 13 patients examined. This was due to improved demonstration of the glomus tumour at the skull base; visualisation of any intracranial extension was not significantly improved. Subtraction was particularly important for demonstration of recurrent glomus jugulare tumours, which do not as a rule extend into the middle ear, and do not therefore present the typical features of a tum o u r examined ab initio (see below). Fat subtraction played an important role in improved tumour demonstration in five patients (Table 1). A second important feature of the subtraction studies was improved discrimination of t u m o u r from veins, which also produce a strong signal because of their G d content. Some of the veins of the cervical spinal plexus present a saccular appearance and m a y therefore be especially con-
471
Table 1. Advantages of subtraction MRI Patient
Age and sex
Clinical features
Subtraction results
1
24 F
Hearing loss; red mass behind drum
Coronal sections only; tumour delineation improved, mostly the result of identification of saccular veins in cervical plexus
2
49 M
Hearing loss, thumping tinnitus, red eardrum
Extent of tumour better shown on axial subtractions, which also revealed prominent veins in tumour mass
3
71 F
Injected tumour behind drum, facial palsy
Better demonstration of tumour extent in two planes; fat subtraction and identification of veins important
4
69 F
Previous radiotherapy, hearing loss, pulsatile tinnitus
Improved demonstration; tumour shown to involve jugular vein on subtraction
5
54 M
20-year history of glomus jugulare tumour treated by surgery and radiotherapy
Deep extension of recurrent tumour at skull base much better shown by subtraction
6
65 F
Left tinnitus, vascular mass in middle ear, pulsating eardrum
Glomus tympanicum. Subtraction showed exact relationship of tumour to jugular vein and floor of middle ear
7
44M
Pulsatile tinnitus, recurrent glomus tumour
Better evidence of recurrence on axial subtractions
8
33 F
Left carotid body tumour, recurrent glomus jugulare
Better shown on axial and coronal subtractions mostly due to fat subtraction
9
41 F
Previous embolisation and incomplete removal of glomus jugulare (left), glomus vagale (right)
Coronal subtractions only; boundaries of tumour optimally shown by subtraction
10
28 F
Hearing loss, red swelling in middle ear
Exact extent at skull base best demonstrated by subtraction, separating tumour from veins and fat signal
11
55 M
Bulge in external meatus floor, recurrent glomus jugulare tumour
Subtraction showed exact extent of tumour recurrence; fat used at previous surgery completely removed from image
12
69 F
Fullness in right ear, lower part of drum red
Exact extent of tumour at skull base better defined by subtraction and its relationship to spinal venous plexus shown
13
29 M
1984 glomus turnout left ear treated by radiotherapy; now tinnitus and red eardrum
Extent of tumour at skull base more accurately shown in two planes by subtraction
fusing a n d difficult to r e c o g n i s e o n u n s u b t r a c t e d images. Two f e a t u r e s h e l p e d to d i f f e r e n t i a t e g l o m u s t u m o u r f r o m n o r m a l veins: the t y p i c a l c o n f i g u r a t i o n o f a g l o m u s j u g u lare t u m o u r in c o r o n a l s e c t i o n (Fig. 1), a n d t h e p r e s e n c e o f l a r g e vessels in t h e t u m o u r mass, s h o w n as small r o u n d e d l a c u n a e on t h e s u b t r a c t i o n i m a g e (Fig.2). O v e r a l l , imp r o v e d r e c o g n i t i o n o f t h e j u g u l a r v e i n a n d cervical spinal p l e x u s m a d e an i m p o r t a n t c o n t r i b u t i o n to the i m a g i n g in 7 p a t i e n t s ( T a b l e 1).
Discussion I m a g i n g t e c h n i q u e s a r e r e q u i r e d at all stages of m a n a g e m e n t o f g l o m u s t u m o u r s : in p r e - t r e a t m e n t assessment, c o n c e r n e d with d i a g n o s i s o f t h e t y p e o f g l o m u s t u m o u r a n d its extent; d u r i n g t r e a t m e n t , w h e n a n g i o g r a p h y is r e q u i r e d if e m b o l i s a t i o n is u n d e r t a k e n p r i o r to resection; a n d p o s t - o p e r a t i v e l y , to s h o w r e c u r r e n t t u m o u r o r t h e effects o f r a d i o t h e r a p y . A t t h e p r e - t r e a t m e n t stage t h e
Fig.L a Gadolinium-enhanced coronal T1weighted image showing a left glomus jugulare tumour, b Same section after subtraction. The exact extent of the tumour is more clearly shown, including its extension down the jugular vein. Note the typical appearance of a glomus jugulare tumour, with a small lobule of tumour projecting into the middle ear cleft (arrow); also the strong signal from the gadolinium in the spinal venous plexus
472
Fig.2. Coronal subtraction sections showing a large glomus tumour. There are small rounded defects (arrows) within the tumour due to its arterial supply.Note also the associated dural enhancement
first requirement is to distinguish between a glomus tympanicum tumour localised to the middle ear cleft and a glomus jugulare tumour with varying degrees of skull base and intracranial involvement. This is important because the surgical approach is totally different: the glomus tympanicum tumour can usually be removed by a transmeatal approach, whereas most glomus jugulare tumours require some form of complex skull base resection. In the majority of patients differentiation between the two types of tumour can be made on CT: glomus jugulare lesions show a characteristic ragged erosion and enlargement of the jugular fossa, whereas glomus tympanicum tumours show a discrete mass in the middle ear arising from the promontory, without jugular fossa enlargement. However, there are borderline cases in which it is difficult to distinguish between a large glomus tympanicum tumour which has spread downwards to involve the jugular bulb, and a small glomus jugulare tumours which has spread upwards into the middle ear, without bone erosion demonstrable on CT [3]. The crucial question from the surgeon's point of view is whether the jugular bulb is significantly
involved by the tumour. Minor indentations on the bulb may be shown by vascular imaging, but subtraction G d M R I is now the best non-invasive technique for making this critical differentiation. Although no bony landmarks are visible on subtraction GdMRI, the vascular epithelial lining of the external auditory meatus and middle ear is clearly visible so that the relationship of the tumour mass to the middle ear cavity can be readily identified (Fig. 3). Demonstration of a glomus tympanicum lesion within the middle ear, discrete from the jugular bulb is also possible (Fig. 4). In coronal sections the main mass of a typical glomus jugulare tumour is seen below the line of the floor of the middle ear, with a small lobule projecting into the middle ear cleft (Figs. 1, 3). It has also been shown that G d M R I can enable differentiation of turnout from retained secretion resulting from eustachian tube obstruction [4]. The removal of fat signal is a most important advantage of the subtraction technique. In addition to being present naturally at the skull base, fat from the dermis is frequently used to pack the surgical cavity after removal of the tumour. The effect of Gd is to convert the M R I charac-
473
Fig.3. Coronal subtraction showing the typical configuration of a small glomus jugulare tumour (cf. Fig.la). The relationship of the tumour to the mucosa of the external auditory meatus is also demonstrated (arrows) Fig.4. Glomus tympanicum tumour the coronal subtraction study shows the turnout confined to the middle ear (arrow) and discrete from the jugular bulb
Fig.5. Coronal subtraction gadolinium-enhanced magnetic resonance imaging showing a recurrent glomus jugulare turnout which presented as a small bulge on the floor of the external auditory meatus
teristics of vascular tissue into those closely resembling normal fat, giving high signal on Tl-weighted images, and therefore making it difficult to differentiate fat from tumour on enhanced images. A partial ~olution to the problem is to use short tau inversion recovery (STIR) sequences [5, 6] which suppress the signal from fat. However, they also suppress the signal from Gd-enhanced tissues, resulting in diminished image contrast. Other methods of fat suppression include chemical shift imaging [7] and pre-saturation techniques [8], but subtraction provides the simplest method not requiring any additional imaging sequence. The low vascularity of adipose tissue ensures that it is recorded on the subtraction studies as an image of minimal density, which does not obscure the pathological enhancement. If surgery is to be planned optimally, accurate pretreatment assessment of the extent of the tumour is necessary, particularly in the region of the roof of the jugular fossa, the various segments of the carotid canal, the foramen lacerum and intracranially, and cases considered more appropriate for primary treatment by radiotherapy (vide infra) also require careful radiological assessment to ensure accurate planning of the radiation fields. Fisch et al. [9, 10] classified glomus jugulare tumours according to the extent of temporal bone erosion, and the degree of invasion of the carotid canal and posterior fossa, and their surgical approach is based upon this. All ttunours without, intracranial extension and those extending up to 2 cm intradurally can be resected via a single-stage infratemporal fossa approach; tumours extending more than 2 cm intradurally need a combined two-stage procedure by otologist and neurosurgeon. Our own surgical technique has been modified to allow a single stage otoneurological approach, and in selected cases the facial nerve can be preserved in the fallopian canal, removing the small risk of facial palsy that results from its transposition in the infratemporal fossa procedure.
There are, however, conflicting views concerning the best treatment for glomus jugulare tumours, especially the place of surgery and radiotherapy. This is primarily because of diagnostic inability to assess the effectiveness of any particular treatment in eradicating the tumour or the rate of growth of any recurrence. Radiotherapy has been claimed to be the ideal treatment modality because of its low morbidity, and although its supporters would agree that turnout eradication is not common, they claim that it is effective in stopping further growth. Unfortunately, diagnostic imaging has hitherto been inadequate for confirming this hypothesis. Most surgeons would agree that radiotherapy is the modality of choice for large symptomatic tumours in the elderly or infirm, while most radiotherapists accept that small tumours in the young are best treated surgically. Radical surgery, despite initial increased morbidity, can eradicate the tumour and any functional deficit resolves or is amenable to further surgery. However, the extremely complex anatomy of the skull base often makes it difficult to be sure of complete resection. In these circumstances there is need for follow-up imaging, to monitor the effects of radiotherapy or to show tumour recurrence (Fig. 5). CT and MRI have been the non-invasive tests most widely used for this purpose, but often producing uncertain evidence of turnour recurrence. The use of the subtraction method overcomes some of these problems and produces images similar to angiography in that they are dependent on tumour vascularity. Turnout presence and extent is thus shown optimally without the use of more invasive techniques.
References
1. Lloyd GAS, Barker PG (1991) Subtraction magnetic resonance for tumours of the skull base and sinuses: a new imaging technique. J Laryngol Oto1105: 628-631
474 2. Brix G, Semmler W, Port R, Schad R, Layer G, Lorentz WJ (1991) Pharmacokinetic parameters in CNS Gd-DTPA enhanced MR imaging. J Comput Assist Tomogr 15:621-628 3. Phelps PD, Stansbie JM (1988) Glomus jugulare or tympanicure? The role of CT and MR imaging with gadolinium DTPA. J Laryngol Oto1102: 766-776 4. Phelps PD (1990) Glomus tumours of the ear: an imaging regime. Clin Radio141:301-305 5. Bydder GM, Young IR (1985) MR imaging: clinical use of the inversion recovery sequence. J Comput Assist Tomogr 9:659-675 6. Dwyer AJ, Frank JA, Reinig JW, Hickey AM, Doppman JL (1988) Short T1 inversion recovery pulse sequences. Analysis and initial experience in cancer imaging. Radiology 168:827-836 7. Simon JH, Szumowski J (1989) Chemical shift imaging with paramagnetic contrast material enhancement for improved lesion depiction. Radiology 171:539-543
8. Barakos JA, Dillon WP, Chew WM (1991) Orbit, skull base and pharynx: contrast enhanced fat suppression MR imaging. Radiology 179:191-198 9. Fisch U (1982) Infratemporal fossa approach for glomus tumours of the temporal bone. Ann Otol Rhinol Laryngo191: 474479 10. Fisch U, Fagan R Valavanis A (1984) The infratemporal fossa approach for the lateral skull base. Otolaryngol Clin North Am 17: 513-552
Dr. G. A. S. Lloyd Royal National Throat Nose and Ear Hospital Grays Inn Road London WCIX 8DA, UK
