1996, The British Journal of Radiology, 69, 114-121 Ganglioneuroma: computed tomography and magnetic resonance features 1 T ICHIKAWA, MD, ""K OHTOMO, MD, ""T ARAKI, MD, 2 H FUJIMOTO, MD, 3 K NEMOTO, MD, 1 A NANBU, MD, 4 M ONOUE, MD and 5 K AOKI, MD Departments of Radiology, 1 Yamanashi Medical College, 1110 Shimokato, Tamaho-cho, Nakakoma-Gun, Yamanashi 409-38, 2 Numazu City Hospital, 550 Harunoki, Higashishiiji, Numazu, Shizuoka 410, 3 Haibara Synthetic Hospital, 2887-1 Hosoe, Haibara, Shizuoka 421-04, 4 Yamanashi Prefectural Central Hospital, 1-1-1 Hujimi, Koufu, Yamanashi 400, and 5 Shizuoka Prefectural Children Hospital, 860 Shituyama, Shizuoka 420, Japan Abstract 12 patients who had histological proven ganglioneuromas were investigated by computed tomography (CT) and magnetic resonance (MR) imaging. CT scans (n = ll), conventional spin-echo MR images (n = 10) and dynamic MR images (n = 5) were acquired. All lesions showed a well defined, oval shape. Five lesions (42%) snowed calcification which was punctate in four and coarse in one on CT. CT attenuation was predominantly low in three of 10 (30%) and intermediate in the remaining seven (70%). In all lesions MR signals were mainly of low intensity on T t weighted images (7iWI) and of high intensity on T 2 weighted images (T 2 WI). Dynamic MR studies in five cases showed a lack of early enhancement but gradual increasing enhancement. One case had a ganglioneuroblastoma component which showed soft-tissue density and coarse calcifications on CT scans, MR images with intermediate intensity on T{WI and T 2 WI and early enhancement and little washout on dynamic MR images. In conclusion, ganglioneuroma typically shows punctate calcification and low attenuation on CT and marked hyperintensity on T 2 WI with gradual increasing enhancement on dynamic MR images. If a ganglioneuroma has atypical CT and MR features, coexistence of a malignant component should be considered. Neurogenic tumours originating from autonomic gan- glia include ganglioneuromas, ganglioneuroblastomas and neuroblastomas. Ganglioneuromas are benign, whereas neuroblastomas are high grade malignant tumours which are common in children [1-7]. Ganglioneuroblastomas are considered to have low grade malignancy, between that of ganglioneuromas and neuroblastomas [2]. Approximately 25% of ganglioneu- romas are not truly benign but contain poorly differen- tiated components including ganglioneuroblastoma and neuroblastoma, or more aggressive neurogenic tumours, such as phaeochromocytomas [8]. If an aggressive com- ponent is found in a surgical specimen, additional chemo- therapy is instituted [9]. Diagnostic difficulties may be compounded by reliance on the histological evaluation of biopsy specimens which may represent only a small portion of the lesion. It is therefore important to differ- entiate ganglioneuromas from malignant components by radiological means, if possible. Computed tomography (CT) has been found to be useful for evaluating the size, extent and composition of such tumours [1]. Magnetic resonance (MR) imaging has the advantage of multiplanar imaging and improved soft tissue contrast [10,11]. However, there have been only a Received 22 May 1995 and in revised form 21 September 1995, accepted 27 September 1995. few case reports describing CT and MR imaging in gang- lioneuromas [1,7,10-22]. In this study, we evaluate the CT and MR imaging features of ganglioneuromas. Materials and methods We retrospectively reviewed the CT and MR images in 12 patients with ganglioneuroma who presented between 1989 and 1993. Patient age and distribution of tumours are shown in Table I. The mean age of all patients was 31 years. In six patients the masses were discovered incidentally. In the others, initial symptoms were hypertension (n = 2), abdominal pain (n = 2) and a palpable mass (n = 2). Histopathological diagnoses were confirmed by complete surgical resection in all patients. Unenhanced (n = l0) and enhanced CT examinations (n = ll) were reviewed. MR images were available for comparison in 10 patients. MR images were obtained with a superconducting 1.5 T MR unit (Signa; GE Medical Systems, Milwaukee, USA or MRT-200A; Toshiba Medical, Tokyo) in eight cases; 0.5 T (MRT-50A; Toshiba Medical, Tokyo) in one; and 0.15 T (Mark J; Siemens-Asahi, Tokyo) in one. Transaxial images were obtained in all patients with additional coronal and sag- ittal images in some patients. A 7-10 mm slice thickness with a 0-3 mm interslice gap, and an image matrix of 256 x 128-192 were used. MR pulse sequences included 7j weighted (repetition time ms (TR)/echo time ms (TE), 114 The British Journal of Radiology, February 1996
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Ganglioneuroma: computed tomography and magnetic resonance features1996, The British Journal of Radiology, 69, 114-121
Ganglioneuroma: computed tomography and magnetic resonance features 1T ICHIKAWA, MD, ""K OHTOMO, MD, ""T ARAKI, MD, 2H FUJIMOTO, MD, 3K NEMOTO, MD, 1A NANBU, MD, 4M ONOUE, MD and 5K AOKI, MD
Departments of Radiology, 1Yamanashi Medical College, 1110 Shimokato, Tamaho-cho, Nakakoma-Gun, Yamanashi 409-38, 2Numazu City Hospital, 550 Harunoki, Higashishiiji, Numazu, Shizuoka 410, 3Haibara Synthetic Hospital, 2887-1 Hosoe, Haibara, Shizuoka 421-04, 4Yamanashi Prefectural Central Hospital, 1-1-1 Hujimi, Koufu, Yamanashi 400, and 5Shizuoka Prefectural Children Hospital, 860 Shituyama, Shizuoka 420, Japan
Abstract 12 patients who had histological proven ganglioneuromas were investigated by computed tomography (CT) and magnetic resonance (MR) imaging. CT scans (n = l l ) , conventional spin-echo MR images (n = 10) and dynamic MR images (n = 5) were acquired. All lesions showed a well defined, oval shape. Five lesions (42%) snowed calcification which was punctate in four and coarse in one on CT. CT attenuation was predominantly low in three of 10 (30%) and intermediate in the remaining seven (70%). In all lesions MR signals were mainly of low intensity on Tt weighted images (7iWI) and of high intensity on T2 weighted images (T2WI). Dynamic MR studies in five cases showed a lack of early enhancement but gradual increasing enhancement. One case had a ganglioneuroblastoma component which showed soft-tissue density and coarse calcifications on CT scans, MR images with intermediate intensity on T{WI and T2WI and early enhancement and little washout on dynamic MR images. In conclusion, ganglioneuroma typically shows punctate calcification and low attenuation on CT and marked hyperintensity on T2WI with gradual increasing enhancement on dynamic MR images. If a ganglioneuroma has atypical CT and MR features, coexistence of a malignant component should be considered.
Neurogenic tumours originating from autonomic gan- glia include ganglioneuromas, ganglioneuroblastomas and neuroblastomas. Ganglioneuromas are benign, whereas neuroblastomas are high grade malignant tumours which are common in children [1-7] . Ganglioneuroblastomas are considered to have low grade malignancy, between that of ganglioneuromas and neuroblastomas [2] . Approximately 25% of ganglioneu- romas are not truly benign but contain poorly differen- tiated components including ganglioneuroblastoma and neuroblastoma, or more aggressive neurogenic tumours, such as phaeochromocytomas [8] . If an aggressive com- ponent is found in a surgical specimen, additional chemo- therapy is instituted [9] . Diagnostic difficulties may be compounded by reliance on the histological evaluation of biopsy specimens which may represent only a small portion of the lesion. It is therefore important to differ- entiate ganglioneuromas from malignant components by radiological means, if possible.
Computed tomography (CT) has been found to be useful for evaluating the size, extent and composition of such tumours [1] . Magnetic resonance (MR) imaging has the advantage of multiplanar imaging and improved soft tissue contrast [10,11]. However, there have been only a
Received 22 May 1995 and in revised form 21 September 1995, accepted 27 September 1995.
few case reports describing CT and MR imaging in gang- lioneuromas [1,7,10-22]. In this study, we evaluate the CT and MR imaging features of ganglioneuromas.
Materials and methods We retrospectively reviewed the CT and MR images
in 12 patients with ganglioneuroma who presented between 1989 and 1993. Patient age and distribution of tumours are shown in Table I. The mean age of all patients was 31 years. In six patients the masses were discovered incidentally. In the others, initial symptoms were hypertension (n = 2), abdominal pain (n = 2) and a palpable mass (n = 2). Histopathological diagnoses were confirmed by complete surgical resection in all patients.
Unenhanced (n = l0) and enhanced CT examinations (n = l l ) were reviewed. MR images were available for comparison in 10 patients. MR images were obtained with a superconducting 1.5 T MR unit (Signa; GE Medical Systems, Milwaukee, USA or MRT-200A; Toshiba Medical, Tokyo) in eight cases; 0.5 T (MRT-50A; Toshiba Medical, Tokyo) in one; and 0.15 T (Mark J; Siemens-Asahi, Tokyo) in one. Transaxial images were obtained in all patients with additional coronal and sag- ittal images in some patients. A 7-10 mm slice thickness with a 0-3 mm interslice gap, and an image matrix of 256 x 128-192 were used. MR pulse sequences included 7j weighted (repetition time ms (TR)/echo time ms (TE),
114 The British Journal of Radiology, February 1996
>3
TIchikawa, K Ohtomo, T Araki et al
numbers of excitations (NEX): 500/14-20, 2-4 NEX) spin-echo (SE) MR images (T^WI); T2 weighted (2000/80-100, 1-4 NEX) SE images (T2WI), and T2* weighted (300/22/flip angle (FA) 20°, 8 NEX) gradient- echo (GRE) MR images (T2*WI). Post-contrast MR images were obtained after Gd-DTPA in all cases. Dynamic contrast-enhanced MR images were performed in five patients (one at 0.5 T, and four at 1.5 T), and conventional contrast-enhanced 71WI (500/20, 2 NEX) were obtained in the other five.
Dynamic MR examinations were performed with a 7i weighted GRE sequence (TR/TE/FA: 100/14/60°) at 0.5 T in one patient and with a Tx weighted SE sequence (TR/TE: 100/12-15) at 1.5 T in four patients. Two contiguous slices were obtained simultaneously during one breath-hold. The following dynamic MR protocol was performed in these four patients: (i) a pre-contrast series; (ii) intravenous (IV) bolus of 0.1 mmol/kg of gado- linium diethylenetriaminepentaacetic acid (Gd-DTPA) (Magnevist, Schering, Berlin); (iii) images every minute during the first 3 min, followed by (iv) images every 2 min up to 21 min after IV administration of Gd-DTPA.
All CT and MR images were reviewed independently by two radiologists (TI, AN). When there was inter- observer disagreement a consensus view was reached. We evaluated the following imaging features of the tumours: size, margin, shape, attenuation (unenhanced CT), presence of calcification on CT, MR signal intensity and pattern of contrast enhancement.
Predominant CT attenuation was classified as follows: low (<20 HU), intermediate (20-40 HU) and soft tissue (>40 HU). MR signal intensity was also classified on the basis of the predominant signal intensity of the tumours, as follows: low (equal to or less than that of muscles on both TjWI and T^WI); intermediate (greater than that of muscles but less than that of fat on both 7;WI and T2WI); high (equal to or higher than that of fat on TjWI, but definitely lower than that of water on 7^WI); or markedly high (equal to that of water on 7^WI). On post-contrast
CT images and conventional post-contrast T[WI, the degree (slight, moderate, marked) and distribution (homo- geneous, inhomogeneous) of contrast enhancement were analysed on visual appearances. On dynamic MR images, the pattern of contrast enhancement (early and/or gradually increasing enhancement, washout of contrast material) was noted.
Results The CT and MR features, as well as surgical and
pathological findings in the 12 patients are summarised in Table I. 11 patients had true ganglioneuromas, while one patient with a retroperitoneal ganglioneuroma had a ganglioneuroblastoma component on histology. All resected specimens were very soft. The diameter of the masses ranged from 3.8 to 18 cm (mean largest diameter 10 cm). The lesion margin was well defined in all cases. 10 of the 12 ganglioneuromas had an oval shape (Figures 1, 2 and 3), while two in the posterior mediastinum showed a crescent shape.
CT attenuation on unenhanced scans was predomi- nantly low in three of the 10 patients (30%) (Figure 1) and intermediate in the remaining seven (70%) (Figure 2). Calcification was present in five patients (three were retro- peritoneal and two were adrenal in location). Calcification was punctate in four (Figure 2) and coarse in the other case (Figure 3). On enhanced CT, little enhancement was seen in eight of 11 patients (73%) (Figures 2 and 4) and moderate inhomogeneous enhancement was seen in the remaining three (27%) (Figure 1).
TiWI showed homogeneous low intensity in all patients. T2WI or T2*WI showed inhomogeneous but predominantly markedly high signal intensity in all patients (Figures 2-4). In addition, enhanced TjWI showed inhomogeneous moderate enhancement in all five patients (Figure 4). Five patients who had dynamic MR studies showed a lack of early enhancement but gradually increasing enhancement (Figures 2 and 3).
One of the retroperitoneal ganglioneuromas revealed a
Figure 1. Case 2. 24-year-old man with a retroperitoneal ganglioneuroma. (a) Unenhanced CT demonstrates well defined, oval shaped mass, CT attenuation is low (19 HU) and no calcification is seen, (b) Enhanced CT shows inhomogeneous moderate enhancement.
116 The British Journal of Radiology, February 1996
CT and MR features of ganglioneuroma
Figure 2. Case 3. 32-year-old man with an adrenal ganglioneur- oma. (a) Enhanced CT shows little enhancement, irregular fatty infiltration of the liver is also seen, (b) T2WI shows inhomo- geneous but markedly high signal intensity, (c-e) Dynamic MR images: (c) before, (d) 2min, (e) 15 min, show lack of early enhancement and gradually increasing enhancement.
different component in the centre of the lesion (Figure 4). This showed CT attenuation greater than that of muscle (soft-tissue density), coarse calcifications and inter- mediate signal intensity on TjWI and T^WI with early enhancement and little washout on dynamic MR images
lioneuroblastoma (potentially malignant component) represented as neuroblast-like cell proliferation.
Discussion Ganglioneuroma is a rare benign neurogenic tumour
Histologically, this component corresponded to a gang- arising from sympathetic ganglia. 60% of all patients
Vol. 69, No. 818 117
TIchikawa, K Ohtomo, T Araki et al
Figure 3. Case 9. 42-year-old woman with a retroperitoneal ganglioneuroma which included a component of a ganglio- neuroblastoma in the centre of the mass, (a) Unenhanced CT demonstrates well defined, oval shaped mass. Peripheral por- tions of the mass show low density, while the central portion of the mass shows soft-tissue density with coarse calcification, (b) On enhanced CT, peripheral portions of the mass show little enhancement, while the central portion shows moderate enhancement, (c) T2WI shows inhomogeneous signal intensity. Peripheral portions of the mass show markedly high signal intensity, while central portion shows intermediate or high signal intensity. (d,e) Dynamic MR images: (d) before, (e) 1 min. On the early phase (e), the central portion of the mass shows marked enhancement, while the peripheral portion of the mass shows no enhancement.
with these tumours are younger than 20 years, most being less than 10 years of age [1-7] . 10 of 12 cases (83%) in our series, however, occurred in patients older than 20 years. Although female patients are slightly more common in other series [3,4,11], eight of the patients in
our series (67%) were male. Ganglioneuromas most frequently occur in the posterior mediastinum (43%), followed by the retroperitoneum (32%) and the neck (8%) [7] . In our series, eight cases (67%) arose in the retroperitoneum. Approximately 25% of
118 The British Journal of Radiology, February 1996
CTand MR features of ganglioneuroma
Figure 4. Case 5. 38-year-old woman with a ganglioneuroma of the neck (left carotid space), (a) TjWI shows low-signal intensity less than muscle, (b) T2WI shows higher signal intensity than that of cerebrospinal fluid, (c) Enhanced TiWI shows inhomogeneous moderate enhancement.
Vol.69, No. 818 119
ganglioneuromas are not truly benign but contain poorly differentiated components including ganglioneuroblas- tomas and neuroblastomas, or more aggressive neuro- genic tumours such as phaeochromocytomas [8] . In the present study, one case showed elements of gang- lioneuroblastoma.
There have been approximately 15 reports in the English literature describing the CT and MR features of ganglioneuromas [1,7,10-22]. Armstrong et al reported that CT is useful for detecting calcification within tumours, which occurs in 20% of ganglioneuromas [1] . Jasinski et al also reported that CT is more sensitive than ultrasound for detecting calcification within ganglioneuromas. The tumour margin on CT is usually well-defined [1,7] as was also found in our series. An oval or crescent shape, with little mass-effect relative to the size of the tumour, may be important imaging fea- tures. All the resected tumours in our series were very soft. The shape of the tumours may therefore reflect tumour extension between tissue planes.
All cases in the present study, except for a component of ganglioneuroblastoma in one tumour, showed charac- teristic CT attenuation with low attenuation on unen- hanced CT, and slight or inhomogeneously moderate enhancement on contrast-enhanced CT. Iwata et al reported a case of retroperitoneal ganglioneuroma and reviewed nine other Japanese cases in regard to CT attenuation and enhancement pattern of ganglio- neuromas [23]. Their results are similar to ours; eight of their 10 (80%) cases showed low attenuation relative to muscle, while the remainder (20%) showed soft-tissue attenuation. Lesions showed little (80% of cases) or no (20% of cases) enhancement.
In our study all cases, except for the lesion containing an element of ganglioneuroblastoma, revealed character- istic MR signal intensity patterns: predominantly low intensity on T^WI and inhomogeneous but marked high intensity on 7^WI. Sakai et al have reported two medias- tinal ganglioneuromas [11], both of which showed marked high signal intensity on 7^WI. Serra et al emphasized that the most important MR characteristic of ganglioneuromas is an inhomogeneous signal intensity on T2WI [20]. Histologically, ganglioneuromas show abundant myxoid matrices and a relatively small amount of ganglion cells [11,24]. The low CT attenuation and markedly high signal intensity on T2WI of our cases may be due to these features.
The time-related biodistribution of Gd-DTPA is deter- mined by the relative vascularity of the different tissues and by the capillary permeability. Early enhancement occurs with increased vascularity and capillary per- meability. The enlarged extracellular space in some tumours, such as fibrous tissue or myxoid tissue, results in greater accumulation of the contrast medium in the late phase [25]. Therefore, the lack of early enhancement and gradually increasing enhancement on dynamic MR images may correlate with abundant myxoid matrices.
Since ganglioneuroma is the benign form of the spec- trum that includes neuroblastoma, it is important to distinguish these two different forms on CT and MR
images. In addition, some hormonally active ganglio- neuromas show elevated excretion of catecholamines or vanilylmandelic acid (VMA), and cause episodic hypertension, sweating, or flushing that mimic phaeo- chromocytoma [8,15,20-22].
In previous reports, CT features of neuroblastoma or phaeochromocytoma include intermediate or soft-tissue density (20-50 HU, mean value 25 HU) on unenhanced CT scans, and inhomogeneous enhancement on enhanced CT scans [26-28]. Some cases of ganglio- neuromas cannot therefore be distinguished from neuro- blastomas or phaeochromocytomas by clinical and CT features. In the present study, however, calcifications within ganglioneuromas were discrete and punctate in three cases, while coarse calcifications in one case were present in a component of ganglioneuroblastoma. It has been suggested that the pattern of calcification in neuro- blastomas on CT is more commonly amorphous and coarse rather than discrete and punctate in form [7]. The pattern of calcifications on CT may therefore be helpful in distinguishing ganglioneuromas from malig- nant forms, including ganglioneuroblastomas and neuroblastomas.
Neuroblastoma generally shows intermediate or high signal intensity, similar to that of kidney, but phaeo- chromocytoma shows markedly high signal intensity on T2WI in previous reports [28,29]. In addition, in our series, a component of ganglioneuroblastoma showed a lower signal intensity than that of ganglioneuromas. We suggest that these malignant forms may have a higher cellular content and less myxoid matrices, resulting in lower signal intensity on 7^WI. Markedly high signal intensity of ganglioneuromas may differentiate ganglio- neuromas from neuroblastomas or ganglioneuroblas- tomas, although phaeochromocytomas could not be ruled out.
The dynamic MR features of phaeochromocytomas have been reported in a total of 10 cases [30-32]. In all cases, the phaeochromocytomas show marked early enhancement due to their hypervascularity within 1 min after administration of Gd-DTPA. Kornreich et al reported that neuroblastomas (« = 3) and ganglioneuro- blastomas (n = 2) show marked enhancement on contrast- enhanced T^WI, perhaps within 5 min after Gd-DTPA [33]. Although there have been no previous reports of dynamic MR study of neuroblastomas, we have performed dynamic MR studies in three children with posterior mediastinal neuroblastomas. These patients showed marked early enhancement similar to that of phaeochromocytomas (unpublished data). We therefore emphasize that lack of early enhancement and gradually increasing enhancement on dynamic MR images may distinguish ganglioneuromas from neuroblastomas, gan- glioneuroblastomas and phaeochromocytomas, all of which usually show marked early enhancement. However, larger series are needed to confirm these findings.
We conclude that the characteristic CT and MR fea- tures of ganglioneuroma are (i) a well defined, oval or crescent shape mass; (ii) low or intermediate CT attenu- ation; (iii) discrete punctate calcification; (iv) inhomo-
120 The British Journal of Radiology, February 1996
CT and MR features of ganglioneuroma
geneous, but markedly high intensity on T2WI; and (v) lack of early enhancement with increasing enhance- ment on dynamic MR images. In addition, if suspected ganglioneuromas have components which show atypical CT and MR findings, malignant or aggressive elements due to neuroblastoma, ganglioneuroblastoma, or phaeo- chromocytoma should be considered.
References 1. ARMSTRONG, E A, HARWOOD-NASH, D C F and
RITZ, C R, CT of neuroblastomas and ganglioneuromas in children, AJR, 139, 571-576 (1982).
2. KILTON, L J, ASCHENBRENER, C and BURNS, C P, Ganglioneuroblastoma in adults, Cancer, 37, 974-983 (1976).
3. BUNN, N D Jr and KING, A B, Cervical ganglioneuroma. A case report and review of the literature, Guthrie Clin. Bull., 330,5-14(1961).
4. CARPENTER, W B and KERNOHAN, J W, Retro- peritoneal ganglioneuromas and neurofibromas. A clinico- pathological study, Cancer, 16, 788-797 (1963).
5. HAMILTON, J P and KOOP, C E, Ganglioneuromas in children, Surg. Gynecol. Obstet., 121, 803-812 (1965).
6. TALERMAN, A and GRATAMA, S, Primary ganglio- neuroblastoma of the anterior mediastinum in a 61-year-old woman, Histopathol, 7, 967-975 (1983).
7. JASINSKI, R W, SAMUELS, B I and SILVER, T M, Sonographic features of retroperitoneal ganglioneuroma, J. Ultrasound Med., 3, 413-415 (1984).
8. STOUT, A P and ARTHUR, P S, Ganglioneuroma of the sympathetic nervous system, Surg. Gynecol. Obstet., 84, 101-110(1947).
9. HAYES, F A, GREEN, A A and RAO, B N, Clinical mani- festations of ganglioneuroma, Cancer, 63,1211-1214(1989).
10. BREIDAHL, W H and KHANGURE, M S, MRI of lumbar and sacral plexus nerve sheath tumors, Australas. Radioi, 35,140-144(1991).
11. SAKAI, F, SONE, S, KIYONO, K ET AL, Intrathoracic neurogenic tumors: MR-pathologic correlation, AJR, 159, 279-283(1992).
12. SAMPSON, M A, MITCHELL, R and MORELY, T R, Case report: Painless thoracic scoliosis due to dumbbell ganglioneuroma, Clin. Radioi, 44, 359-360 (1991).
13. IKEZOE, J, SONE, S, HIGASHIHARA, T ET AL, CT of intrathoracic neurogenic tumors, Eur. J. Radioi, 6, 266-269 (1986).
14. KHAFAGI, F A, GROSS, M D, SHAPIRO, B ET AL, Clinical significance of the large adrenal mass, Br. J. Surg., 78,828-833(1991).
15. DAVIES, R P, SLAVOTINEK, J P and DORNEY, S F, VIP secreting tumors in infancy. A review of radiological appearances, Pediatr. Radioi, 20, 504-508 (1990).
i 16. CHEUNG, Y K, FUNG, C F, CHAN, F L and LEONG, [ L L, MRI features of spinal ganglioneuroma, Clin. Imaging, \ 15,109-112(1991).
17. GRAHAM, R T, HERSCHORN, S and SRIGLEY, J, Calcified ganglioneuroma of the bladder, Urol. Radioi, 9, 177-180(1987).
18. MARKS, F, YOUNG, B K, RAGHAVENDRA, B N ET AL, Diagnosis of adrenal ganglioneuroma in pregnancy with magnetic resonance imaging and ultrasonography, J. Reproduct. Med., 34, 59-60 (1993).
19. SIRINELLI, D, SILBERMAN, B, BAUDON, J J ET AL, Beckwith-Wiedemann syndrome and neural crest tumors, Pediatr. Radioi., 19, 242-245 (1989).
20. SERRA, A D, RAFAL, R B and MARKISZ, J A, MRI characteristics of two cases of adrenal ganglioneuromas, Clin. Imaging, 16, 37-39 (1992).
21. ICHIKAWA, T, KOYAMA, A, FUJIMOTO, H ET AL, Retroperitoneal ganglioneuroma extending across the mid- line, Clin. Imaging, 17, 19-21 (1993).
22. ANTONIE, T E, Ganglioneuroma presenting as an abdominal mass: report of a case, Aust. NZ J. Surg., 42, 71-74(1972).
23. IWATA, R, MORIYAMA, N, TAKAYASU, K ET AL, A case of ganglioneuroma in the retroperitoneum, Jpn J. Clin. Radioi, 37, 515-518 (1992).
24. DUNNICK, N R and CASTELLINO, R A, Arteriographic…
Ganglioneuroma: computed tomography and magnetic resonance features 1T ICHIKAWA, MD, ""K OHTOMO, MD, ""T ARAKI, MD, 2H FUJIMOTO, MD, 3K NEMOTO, MD, 1A NANBU, MD, 4M ONOUE, MD and 5K AOKI, MD
Departments of Radiology, 1Yamanashi Medical College, 1110 Shimokato, Tamaho-cho, Nakakoma-Gun, Yamanashi 409-38, 2Numazu City Hospital, 550 Harunoki, Higashishiiji, Numazu, Shizuoka 410, 3Haibara Synthetic Hospital, 2887-1 Hosoe, Haibara, Shizuoka 421-04, 4Yamanashi Prefectural Central Hospital, 1-1-1 Hujimi, Koufu, Yamanashi 400, and 5Shizuoka Prefectural Children Hospital, 860 Shituyama, Shizuoka 420, Japan
Abstract 12 patients who had histological proven ganglioneuromas were investigated by computed tomography (CT) and magnetic resonance (MR) imaging. CT scans (n = l l ) , conventional spin-echo MR images (n = 10) and dynamic MR images (n = 5) were acquired. All lesions showed a well defined, oval shape. Five lesions (42%) snowed calcification which was punctate in four and coarse in one on CT. CT attenuation was predominantly low in three of 10 (30%) and intermediate in the remaining seven (70%). In all lesions MR signals were mainly of low intensity on Tt weighted images (7iWI) and of high intensity on T2 weighted images (T2WI). Dynamic MR studies in five cases showed a lack of early enhancement but gradual increasing enhancement. One case had a ganglioneuroblastoma component which showed soft-tissue density and coarse calcifications on CT scans, MR images with intermediate intensity on T{WI and T2WI and early enhancement and little washout on dynamic MR images. In conclusion, ganglioneuroma typically shows punctate calcification and low attenuation on CT and marked hyperintensity on T2WI with gradual increasing enhancement on dynamic MR images. If a ganglioneuroma has atypical CT and MR features, coexistence of a malignant component should be considered.
Neurogenic tumours originating from autonomic gan- glia include ganglioneuromas, ganglioneuroblastomas and neuroblastomas. Ganglioneuromas are benign, whereas neuroblastomas are high grade malignant tumours which are common in children [1-7] . Ganglioneuroblastomas are considered to have low grade malignancy, between that of ganglioneuromas and neuroblastomas [2] . Approximately 25% of ganglioneu- romas are not truly benign but contain poorly differen- tiated components including ganglioneuroblastoma and neuroblastoma, or more aggressive neurogenic tumours, such as phaeochromocytomas [8] . If an aggressive com- ponent is found in a surgical specimen, additional chemo- therapy is instituted [9] . Diagnostic difficulties may be compounded by reliance on the histological evaluation of biopsy specimens which may represent only a small portion of the lesion. It is therefore important to differ- entiate ganglioneuromas from malignant components by radiological means, if possible.
Computed tomography (CT) has been found to be useful for evaluating the size, extent and composition of such tumours [1] . Magnetic resonance (MR) imaging has the advantage of multiplanar imaging and improved soft tissue contrast [10,11]. However, there have been only a
Received 22 May 1995 and in revised form 21 September 1995, accepted 27 September 1995.
few case reports describing CT and MR imaging in gang- lioneuromas [1,7,10-22]. In this study, we evaluate the CT and MR imaging features of ganglioneuromas.
Materials and methods We retrospectively reviewed the CT and MR images
in 12 patients with ganglioneuroma who presented between 1989 and 1993. Patient age and distribution of tumours are shown in Table I. The mean age of all patients was 31 years. In six patients the masses were discovered incidentally. In the others, initial symptoms were hypertension (n = 2), abdominal pain (n = 2) and a palpable mass (n = 2). Histopathological diagnoses were confirmed by complete surgical resection in all patients.
Unenhanced (n = l0) and enhanced CT examinations (n = l l ) were reviewed. MR images were available for comparison in 10 patients. MR images were obtained with a superconducting 1.5 T MR unit (Signa; GE Medical Systems, Milwaukee, USA or MRT-200A; Toshiba Medical, Tokyo) in eight cases; 0.5 T (MRT-50A; Toshiba Medical, Tokyo) in one; and 0.15 T (Mark J; Siemens-Asahi, Tokyo) in one. Transaxial images were obtained in all patients with additional coronal and sag- ittal images in some patients. A 7-10 mm slice thickness with a 0-3 mm interslice gap, and an image matrix of 256 x 128-192 were used. MR pulse sequences included 7j weighted (repetition time ms (TR)/echo time ms (TE),
114 The British Journal of Radiology, February 1996
>3
TIchikawa, K Ohtomo, T Araki et al
numbers of excitations (NEX): 500/14-20, 2-4 NEX) spin-echo (SE) MR images (T^WI); T2 weighted (2000/80-100, 1-4 NEX) SE images (T2WI), and T2* weighted (300/22/flip angle (FA) 20°, 8 NEX) gradient- echo (GRE) MR images (T2*WI). Post-contrast MR images were obtained after Gd-DTPA in all cases. Dynamic contrast-enhanced MR images were performed in five patients (one at 0.5 T, and four at 1.5 T), and conventional contrast-enhanced 71WI (500/20, 2 NEX) were obtained in the other five.
Dynamic MR examinations were performed with a 7i weighted GRE sequence (TR/TE/FA: 100/14/60°) at 0.5 T in one patient and with a Tx weighted SE sequence (TR/TE: 100/12-15) at 1.5 T in four patients. Two contiguous slices were obtained simultaneously during one breath-hold. The following dynamic MR protocol was performed in these four patients: (i) a pre-contrast series; (ii) intravenous (IV) bolus of 0.1 mmol/kg of gado- linium diethylenetriaminepentaacetic acid (Gd-DTPA) (Magnevist, Schering, Berlin); (iii) images every minute during the first 3 min, followed by (iv) images every 2 min up to 21 min after IV administration of Gd-DTPA.
All CT and MR images were reviewed independently by two radiologists (TI, AN). When there was inter- observer disagreement a consensus view was reached. We evaluated the following imaging features of the tumours: size, margin, shape, attenuation (unenhanced CT), presence of calcification on CT, MR signal intensity and pattern of contrast enhancement.
Predominant CT attenuation was classified as follows: low (<20 HU), intermediate (20-40 HU) and soft tissue (>40 HU). MR signal intensity was also classified on the basis of the predominant signal intensity of the tumours, as follows: low (equal to or less than that of muscles on both TjWI and T^WI); intermediate (greater than that of muscles but less than that of fat on both 7;WI and T2WI); high (equal to or higher than that of fat on TjWI, but definitely lower than that of water on 7^WI); or markedly high (equal to that of water on 7^WI). On post-contrast
CT images and conventional post-contrast T[WI, the degree (slight, moderate, marked) and distribution (homo- geneous, inhomogeneous) of contrast enhancement were analysed on visual appearances. On dynamic MR images, the pattern of contrast enhancement (early and/or gradually increasing enhancement, washout of contrast material) was noted.
Results The CT and MR features, as well as surgical and
pathological findings in the 12 patients are summarised in Table I. 11 patients had true ganglioneuromas, while one patient with a retroperitoneal ganglioneuroma had a ganglioneuroblastoma component on histology. All resected specimens were very soft. The diameter of the masses ranged from 3.8 to 18 cm (mean largest diameter 10 cm). The lesion margin was well defined in all cases. 10 of the 12 ganglioneuromas had an oval shape (Figures 1, 2 and 3), while two in the posterior mediastinum showed a crescent shape.
CT attenuation on unenhanced scans was predomi- nantly low in three of the 10 patients (30%) (Figure 1) and intermediate in the remaining seven (70%) (Figure 2). Calcification was present in five patients (three were retro- peritoneal and two were adrenal in location). Calcification was punctate in four (Figure 2) and coarse in the other case (Figure 3). On enhanced CT, little enhancement was seen in eight of 11 patients (73%) (Figures 2 and 4) and moderate inhomogeneous enhancement was seen in the remaining three (27%) (Figure 1).
TiWI showed homogeneous low intensity in all patients. T2WI or T2*WI showed inhomogeneous but predominantly markedly high signal intensity in all patients (Figures 2-4). In addition, enhanced TjWI showed inhomogeneous moderate enhancement in all five patients (Figure 4). Five patients who had dynamic MR studies showed a lack of early enhancement but gradually increasing enhancement (Figures 2 and 3).
One of the retroperitoneal ganglioneuromas revealed a
Figure 1. Case 2. 24-year-old man with a retroperitoneal ganglioneuroma. (a) Unenhanced CT demonstrates well defined, oval shaped mass, CT attenuation is low (19 HU) and no calcification is seen, (b) Enhanced CT shows inhomogeneous moderate enhancement.
116 The British Journal of Radiology, February 1996
CT and MR features of ganglioneuroma
Figure 2. Case 3. 32-year-old man with an adrenal ganglioneur- oma. (a) Enhanced CT shows little enhancement, irregular fatty infiltration of the liver is also seen, (b) T2WI shows inhomo- geneous but markedly high signal intensity, (c-e) Dynamic MR images: (c) before, (d) 2min, (e) 15 min, show lack of early enhancement and gradually increasing enhancement.
different component in the centre of the lesion (Figure 4). This showed CT attenuation greater than that of muscle (soft-tissue density), coarse calcifications and inter- mediate signal intensity on TjWI and T^WI with early enhancement and little washout on dynamic MR images
lioneuroblastoma (potentially malignant component) represented as neuroblast-like cell proliferation.
Discussion Ganglioneuroma is a rare benign neurogenic tumour
Histologically, this component corresponded to a gang- arising from sympathetic ganglia. 60% of all patients
Vol. 69, No. 818 117
TIchikawa, K Ohtomo, T Araki et al
Figure 3. Case 9. 42-year-old woman with a retroperitoneal ganglioneuroma which included a component of a ganglio- neuroblastoma in the centre of the mass, (a) Unenhanced CT demonstrates well defined, oval shaped mass. Peripheral por- tions of the mass show low density, while the central portion of the mass shows soft-tissue density with coarse calcification, (b) On enhanced CT, peripheral portions of the mass show little enhancement, while the central portion shows moderate enhancement, (c) T2WI shows inhomogeneous signal intensity. Peripheral portions of the mass show markedly high signal intensity, while central portion shows intermediate or high signal intensity. (d,e) Dynamic MR images: (d) before, (e) 1 min. On the early phase (e), the central portion of the mass shows marked enhancement, while the peripheral portion of the mass shows no enhancement.
with these tumours are younger than 20 years, most being less than 10 years of age [1-7] . 10 of 12 cases (83%) in our series, however, occurred in patients older than 20 years. Although female patients are slightly more common in other series [3,4,11], eight of the patients in
our series (67%) were male. Ganglioneuromas most frequently occur in the posterior mediastinum (43%), followed by the retroperitoneum (32%) and the neck (8%) [7] . In our series, eight cases (67%) arose in the retroperitoneum. Approximately 25% of
118 The British Journal of Radiology, February 1996
CTand MR features of ganglioneuroma
Figure 4. Case 5. 38-year-old woman with a ganglioneuroma of the neck (left carotid space), (a) TjWI shows low-signal intensity less than muscle, (b) T2WI shows higher signal intensity than that of cerebrospinal fluid, (c) Enhanced TiWI shows inhomogeneous moderate enhancement.
Vol.69, No. 818 119
ganglioneuromas are not truly benign but contain poorly differentiated components including ganglioneuroblas- tomas and neuroblastomas, or more aggressive neuro- genic tumours such as phaeochromocytomas [8] . In the present study, one case showed elements of gang- lioneuroblastoma.
There have been approximately 15 reports in the English literature describing the CT and MR features of ganglioneuromas [1,7,10-22]. Armstrong et al reported that CT is useful for detecting calcification within tumours, which occurs in 20% of ganglioneuromas [1] . Jasinski et al also reported that CT is more sensitive than ultrasound for detecting calcification within ganglioneuromas. The tumour margin on CT is usually well-defined [1,7] as was also found in our series. An oval or crescent shape, with little mass-effect relative to the size of the tumour, may be important imaging fea- tures. All the resected tumours in our series were very soft. The shape of the tumours may therefore reflect tumour extension between tissue planes.
All cases in the present study, except for a component of ganglioneuroblastoma in one tumour, showed charac- teristic CT attenuation with low attenuation on unen- hanced CT, and slight or inhomogeneously moderate enhancement on contrast-enhanced CT. Iwata et al reported a case of retroperitoneal ganglioneuroma and reviewed nine other Japanese cases in regard to CT attenuation and enhancement pattern of ganglio- neuromas [23]. Their results are similar to ours; eight of their 10 (80%) cases showed low attenuation relative to muscle, while the remainder (20%) showed soft-tissue attenuation. Lesions showed little (80% of cases) or no (20% of cases) enhancement.
In our study all cases, except for the lesion containing an element of ganglioneuroblastoma, revealed character- istic MR signal intensity patterns: predominantly low intensity on T^WI and inhomogeneous but marked high intensity on 7^WI. Sakai et al have reported two medias- tinal ganglioneuromas [11], both of which showed marked high signal intensity on 7^WI. Serra et al emphasized that the most important MR characteristic of ganglioneuromas is an inhomogeneous signal intensity on T2WI [20]. Histologically, ganglioneuromas show abundant myxoid matrices and a relatively small amount of ganglion cells [11,24]. The low CT attenuation and markedly high signal intensity on T2WI of our cases may be due to these features.
The time-related biodistribution of Gd-DTPA is deter- mined by the relative vascularity of the different tissues and by the capillary permeability. Early enhancement occurs with increased vascularity and capillary per- meability. The enlarged extracellular space in some tumours, such as fibrous tissue or myxoid tissue, results in greater accumulation of the contrast medium in the late phase [25]. Therefore, the lack of early enhancement and gradually increasing enhancement on dynamic MR images may correlate with abundant myxoid matrices.
Since ganglioneuroma is the benign form of the spec- trum that includes neuroblastoma, it is important to distinguish these two different forms on CT and MR
images. In addition, some hormonally active ganglio- neuromas show elevated excretion of catecholamines or vanilylmandelic acid (VMA), and cause episodic hypertension, sweating, or flushing that mimic phaeo- chromocytoma [8,15,20-22].
In previous reports, CT features of neuroblastoma or phaeochromocytoma include intermediate or soft-tissue density (20-50 HU, mean value 25 HU) on unenhanced CT scans, and inhomogeneous enhancement on enhanced CT scans [26-28]. Some cases of ganglio- neuromas cannot therefore be distinguished from neuro- blastomas or phaeochromocytomas by clinical and CT features. In the present study, however, calcifications within ganglioneuromas were discrete and punctate in three cases, while coarse calcifications in one case were present in a component of ganglioneuroblastoma. It has been suggested that the pattern of calcification in neuro- blastomas on CT is more commonly amorphous and coarse rather than discrete and punctate in form [7]. The pattern of calcifications on CT may therefore be helpful in distinguishing ganglioneuromas from malig- nant forms, including ganglioneuroblastomas and neuroblastomas.
Neuroblastoma generally shows intermediate or high signal intensity, similar to that of kidney, but phaeo- chromocytoma shows markedly high signal intensity on T2WI in previous reports [28,29]. In addition, in our series, a component of ganglioneuroblastoma showed a lower signal intensity than that of ganglioneuromas. We suggest that these malignant forms may have a higher cellular content and less myxoid matrices, resulting in lower signal intensity on 7^WI. Markedly high signal intensity of ganglioneuromas may differentiate ganglio- neuromas from neuroblastomas or ganglioneuroblas- tomas, although phaeochromocytomas could not be ruled out.
The dynamic MR features of phaeochromocytomas have been reported in a total of 10 cases [30-32]. In all cases, the phaeochromocytomas show marked early enhancement due to their hypervascularity within 1 min after administration of Gd-DTPA. Kornreich et al reported that neuroblastomas (« = 3) and ganglioneuro- blastomas (n = 2) show marked enhancement on contrast- enhanced T^WI, perhaps within 5 min after Gd-DTPA [33]. Although there have been no previous reports of dynamic MR study of neuroblastomas, we have performed dynamic MR studies in three children with posterior mediastinal neuroblastomas. These patients showed marked early enhancement similar to that of phaeochromocytomas (unpublished data). We therefore emphasize that lack of early enhancement and gradually increasing enhancement on dynamic MR images may distinguish ganglioneuromas from neuroblastomas, gan- glioneuroblastomas and phaeochromocytomas, all of which usually show marked early enhancement. However, larger series are needed to confirm these findings.
We conclude that the characteristic CT and MR fea- tures of ganglioneuroma are (i) a well defined, oval or crescent shape mass; (ii) low or intermediate CT attenu- ation; (iii) discrete punctate calcification; (iv) inhomo-
120 The British Journal of Radiology, February 1996
CT and MR features of ganglioneuroma
geneous, but markedly high intensity on T2WI; and (v) lack of early enhancement with increasing enhance- ment on dynamic MR images. In addition, if suspected ganglioneuromas have components which show atypical CT and MR findings, malignant or aggressive elements due to neuroblastoma, ganglioneuroblastoma, or phaeo- chromocytoma should be considered.
References 1. ARMSTRONG, E A, HARWOOD-NASH, D C F and
RITZ, C R, CT of neuroblastomas and ganglioneuromas in children, AJR, 139, 571-576 (1982).
2. KILTON, L J, ASCHENBRENER, C and BURNS, C P, Ganglioneuroblastoma in adults, Cancer, 37, 974-983 (1976).
3. BUNN, N D Jr and KING, A B, Cervical ganglioneuroma. A case report and review of the literature, Guthrie Clin. Bull., 330,5-14(1961).
4. CARPENTER, W B and KERNOHAN, J W, Retro- peritoneal ganglioneuromas and neurofibromas. A clinico- pathological study, Cancer, 16, 788-797 (1963).
5. HAMILTON, J P and KOOP, C E, Ganglioneuromas in children, Surg. Gynecol. Obstet., 121, 803-812 (1965).
6. TALERMAN, A and GRATAMA, S, Primary ganglio- neuroblastoma of the anterior mediastinum in a 61-year-old woman, Histopathol, 7, 967-975 (1983).
7. JASINSKI, R W, SAMUELS, B I and SILVER, T M, Sonographic features of retroperitoneal ganglioneuroma, J. Ultrasound Med., 3, 413-415 (1984).
8. STOUT, A P and ARTHUR, P S, Ganglioneuroma of the sympathetic nervous system, Surg. Gynecol. Obstet., 84, 101-110(1947).
9. HAYES, F A, GREEN, A A and RAO, B N, Clinical mani- festations of ganglioneuroma, Cancer, 63,1211-1214(1989).
10. BREIDAHL, W H and KHANGURE, M S, MRI of lumbar and sacral plexus nerve sheath tumors, Australas. Radioi, 35,140-144(1991).
11. SAKAI, F, SONE, S, KIYONO, K ET AL, Intrathoracic neurogenic tumors: MR-pathologic correlation, AJR, 159, 279-283(1992).
12. SAMPSON, M A, MITCHELL, R and MORELY, T R, Case report: Painless thoracic scoliosis due to dumbbell ganglioneuroma, Clin. Radioi, 44, 359-360 (1991).
13. IKEZOE, J, SONE, S, HIGASHIHARA, T ET AL, CT of intrathoracic neurogenic tumors, Eur. J. Radioi, 6, 266-269 (1986).
14. KHAFAGI, F A, GROSS, M D, SHAPIRO, B ET AL, Clinical significance of the large adrenal mass, Br. J. Surg., 78,828-833(1991).
15. DAVIES, R P, SLAVOTINEK, J P and DORNEY, S F, VIP secreting tumors in infancy. A review of radiological appearances, Pediatr. Radioi, 20, 504-508 (1990).
i 16. CHEUNG, Y K, FUNG, C F, CHAN, F L and LEONG, [ L L, MRI features of spinal ganglioneuroma, Clin. Imaging, \ 15,109-112(1991).
17. GRAHAM, R T, HERSCHORN, S and SRIGLEY, J, Calcified ganglioneuroma of the bladder, Urol. Radioi, 9, 177-180(1987).
18. MARKS, F, YOUNG, B K, RAGHAVENDRA, B N ET AL, Diagnosis of adrenal ganglioneuroma in pregnancy with magnetic resonance imaging and ultrasonography, J. Reproduct. Med., 34, 59-60 (1993).
19. SIRINELLI, D, SILBERMAN, B, BAUDON, J J ET AL, Beckwith-Wiedemann syndrome and neural crest tumors, Pediatr. Radioi., 19, 242-245 (1989).
20. SERRA, A D, RAFAL, R B and MARKISZ, J A, MRI characteristics of two cases of adrenal ganglioneuromas, Clin. Imaging, 16, 37-39 (1992).
21. ICHIKAWA, T, KOYAMA, A, FUJIMOTO, H ET AL, Retroperitoneal ganglioneuroma extending across the mid- line, Clin. Imaging, 17, 19-21 (1993).
22. ANTONIE, T E, Ganglioneuroma presenting as an abdominal mass: report of a case, Aust. NZ J. Surg., 42, 71-74(1972).
23. IWATA, R, MORIYAMA, N, TAKAYASU, K ET AL, A case of ganglioneuroma in the retroperitoneum, Jpn J. Clin. Radioi, 37, 515-518 (1992).
24. DUNNICK, N R and CASTELLINO, R A, Arteriographic…
Related Documents
