MAGNETIC RESONANCE IMAGING OF A MEDULLARY DERMOID CYST WITH SECONDARY HYDROCEPHALUS IN A DOG MIKEP. TARGETT, MA, VETMB,PHD,MRCVS, ELIZABETH MCINNES BVSc, PHD,MRC(PATH),MRCVS, RUTHDENNIS, MA, VETMB,DVR, MRCVS

Magnetic resonance images were acquired of the brain of a 7-year-old male Golden Retriever with hydrocephalus secondary to a medullary lesion. Images were acquired prior to and 4 weeks following surgical treatment for the hydrocephalus, and the dog was euthanased following the second imaging session. The MR images demonstrated a medullary lesion with patchy but predominantly hyperintense signal with both T1- and T2-weighting, within which small areas of low signal were scattered. There was little edema associated with this lesion and no enhancement with gadolinium. Postmortem examination revealed the medullary mass to be a dermoid cyst. Several small nodular lesions were identified within the central nervous system on the magnetic resonance images whose origin was uncertain on postmortem examination. Veterinary Radiology & Ultrasound, Vol. 40, No. 1, 1998, p p 23-26. Key words: dermoid cyst, magnetic resonance imaging, dog, hydrocephalus,ventriculoperitonealshunt.

D

formed using a 0.5 Tesla superconducting magnet.* Transverse T1-weighted images and T2-weighted images in the transverse, sagittal, and dorsal planes were acquired. Following intravenous administration of 0.1 mmolkg bodyweight gadodiamide contrast medium? T1-weighted images were acquired in the transversem, sagittal, and dorsal planes. The T1-weighted scan sequences used a repetition time (TR) of 600 ms and an echo time (TE) of 24 ms, and the TZweighted scans a TR and a TE of 200 ms and 80 ms, respectively. The field of view used was 220 mm, and the slice thicknesses were 7 mm for the transverse scans and 6 mm for the dorsal plane and sagittal scans. The imaging matrices were 256 x 256 for the T1-weighted scans and 256 x 192 for the T2-weighted scans. There was ventriculomegaly affecting the lateral and third ventricles and the mesencephalic aqueduct. A small nodule was present adjacent to the dorsal wall of the left lateral ventricle together with two similar nodules in the brainstem. One of these nodules was associated with the mesencephalic aqueduct, and the other was adjacent to the right lateral aspect of the brainstem, possibly in the subarachnoid space (Fig. 1). With T1-weighting, these nodules were dramatically hyperintense, and with T2-weighting, they had equal amounts of very high and absent signal, thought to be caused by a chemical shift artifact. Immediately caudal to the brainstem nodules and in the region of the fourth ventricle, was an irregular, roughly spherical mass approximately 2-cm in diameter (Fig. 2). This gave a patchy, but predominantly hyperintense, signal with both

neoplastic lesions with a complex cyst wall containing both epidemoid tissue and adnexa.' They have been found associated with the dermal, ocular, and central nervous systems.2-6 The most common central nervous system manifestation in domestic species is of a dermoid sinus connecting with the surface of the skin and extending to affect the spinal ~ o r d . ~ Primary -~ intracranial dermoid cysts are rare in man,','' and only two have been reported in the dog."312 This report describes the magnetic resonance (MR) findings in a 7-year-old golden retriever that presented with clinical signs attributable to hydrocephalus. ERMOID CYSTS ARE

Case Report A 7-year-old, male, neutered golden retriever was presented at the Queen's Veterinary School Hospital for investigation of episodes of hindlimb ataxia associated with aggressive behavioral change. The dog had progressive behavioral changes for a period of 7 months with reluctance to exercise, hindlimb ataxia, and bumping into obstacles. There was a left-sided head tilt. A degree of sinoatrial block associated with a wandering pacemaker was also detected, but, the results of cardiac ultrasonographic and radiographic examination were within normal limits. Magnetic resonance imaging (MRI) of the brain was perFrom the Queen's Veterinary School Hospital, University of Cambridge, Madingley Road, Cambridge, CB3 OES, UK, and the Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk, CB8 7UU, UK. Address correspondence and reprint requests to Mike Targett, MA, VetMB, PhD, MRCVS, Queen's Veterinary School Hospital, Madingley Road, Cambridge, CB3 OES, UK. Received December 8, 1997; accepted for publication April 21, 1998.

*Surrey Medical Imaging Systems Ltd., Guildford, UK. TOMNISCAN; Nycomed (UK) Ltd., Birmingham, UK.

23

24

TARGETT ET

A

B

AL

1998

there was a dramatic reduction in ventricular dilation, with increased thickness of the cerebral cortex, although this was partly offset by a marked increase in the size of the subarachnoid space, presumably attributable to CSF leaking around the shunting tube. The position of the shunt tubing was verified (Fig. 3 ) . The ventricular nodule seemed unchanged, but the brainstem mass had enlarged slightly, and the previously identified brainstem nodules were less obvious. Again, no contrast enhancement or secondary edema was evident. At this stage, the dog was euthanized. Grossly, there was a 2-cm diameter cystic lesion ventral to the fourth ventricle within the medulla in the midline. This lesion contained hair. The ventricular lesion was represented by a small, fluid-filled, thin-walled cyst on the ventricular wall. The lateral ventricles were grossly enlarged with marked thinning on the cortex. The ventriculoperitoneal shunt contained cerebrospinal fluid along its length, and a fibrin plug was present distally. Histologically, the brainstem lesion was a medullary derA

FIG.1. Initial transverse T1 (A) and T2 (B) weighted images in which the ventriculomegaly, the nodular lesion in the roof of the left lateral ventricle and the nodular lesions within the medulla associated with the mesencephalic aqueduct and larger medullary lesion are apparent.

T1- and T2-weighting, within which small areas of low signal were scattered. None of the lesions was characterized by contrast enhancement on T1-weighted images or was seen to be associated with edema on T2-weighted images. The hydrocephalus was assumed to be obstructive, caused by a mass within or adjacent to the fourth ventricle. The complex nature of the mass, and in particular, its high signal intensity on T1-weighted images, was thought to be indicative of a hemorrhagic lesion, whether benign or malignant. A ventriculoperitoneal shunt13 was used to treat the hydrocephalus. The MR study was repeated several weeks later, by which time the MR signal from hemorrhage would be expected to have changed. A ventricular CSF sample taken at the time of surgery was normal. Following shunt placement, the dog made a good recovery with resolution of the head tilt and return to normal exercise tolerance and behavior. Four weeks after shunt placement, the dog started to circle to the left. On repeat MRI examination at this time,

FIG. 2. Initial sagittal T1 (A) weighted image and a T2 (B) weighted image acquired 4 weeks following insertion of the ventriculoperitoneal shunt. The roughly spherical mass in the region of the fourth ventricle is apparent.

VOL. 40, No. 1

MRI

OF

DERMOID CYSTIN

FIG.3. Sagittal T2-weighted image acquired 4 weeks following insertion of the ventriculoperitoneal shunt. The ventriculoperitoneal shunt can be clearly seen.

A

DOG

25

FIG. 4. Photomicrograph of the wall of the medullary dermoid cyst. Note the stratified squamous epithelial lining (E) of the cyst cavity (C)and underlying glandular adnexa containing sebaceous glands (S) and dilated sweat glands (D) (van Gieson, scale-bar 100 I*).

magnetic properties of melanin but will appear isointense or hypointense on T2-~eighting,’~.’~ because of shortening of the TZrelaxation time. Hyperintensity caused by slowly moid cyst (Fig. 4). This was associated with Wallerian deflowing blood may be detected by the use of flow-sensitive generation in the underlying medulla. The cystic area on the imaging technique^.'^ Pure fat lipomata usually can be difdorsal roof of the left lateral ventricle was histopathologiferentiated from dermoids by their uniform signal characcally normal. teristics, because dermoids typically contain areas of low signal, but a mainly lipomatous dermoid cannot easily be Discussion differentiated from a lipoma.I8 Other tumors that contain fat In humans, intracranial dermoid cysts are characterized in as well as other such debris as epidermoid cysts and terMR images by mixed, mainly high signal on both T1- and atoma are hard to separate from dermoids. T2-weighting because of the presence of variable amounts If dermoid cysts rupture, droplets of fatty material may of fat.14-17 Areas of low signal are attributed to the presence become distributed through the ventricular system and subof hair. In one report, hair within a dermoid cyst was even arachnoid spaces and appear as small, hyperintense noddetected as curvilinear inclusions within the mass. l4 Small u l e ~ . ’ ~ , ’Sulcal ~ * ~ widening ~ , ~ ~ by fat is said to be pathogareas of calcification, glandular tissue, and sweat secretion nomic of a ruptured intracranial d e r m ~ i d Detection .~~ of may also contribute to areas of low signal intensity.14316318 rupture of dermoid cysts has a bearing on prognosis, beContrast enhancement does not occur,17219and the masses cause of the resultant aseptic meningitis and other side efare not associated with surrounding edema. l 5 Other disease f e c t ~ . ~Three-dimensional ~ , ~ ~ , ~ ~ chemical-shift-selective processes that may produce hyperintense lesions on TI gradient echo imaging has been reported as a means of weighted images include hemorrhage,14915,’ 8*20-22 other fatdetecting even small amounts of free lipid in cerebrospinal containing tumors,14’18’22s23 m e l a n ~ m a t a , ~and ~ - ~slowly ~ fluid by exploiting the different resonant frequency of fat flowing bl00d.l~Hemorrhage may occur on its own or relative to water-containing tissue to produce images in within tumors and creates high intensity signal because of which fat appears hyperintense, and the signal from brain the paramagnetic properties of blood breakdown products, and cerebrospinal fluid is greatly reduced.22It is postulated notably methemaglobin, making it hard to differentiate vithat the three hyperintense nodules visible in this dog’s sually from lipomatous lesions on T1-weighted scans.21 brain were created by extruded fatty material corresponding However, T2-weighted studies may show lower signal, beto the ventricular cystic lesion seen at postmortem examicause intracellular methemaglobin is hyp~intense,’~ and nation. calculation of T1- and TZ-relaxation times shows that these In humans, diagnosis of dermoid cysts is made more are usually longer in areas of old hemorrhage.21 Differencertain using a combination of CT and MRI, because on tiation of fat and methemoglobin may also be made using CT the mass appears hypoattenuating, typical of fat, chemical shift imaging or fat suppression techniques such as usually allowing the possibility of hemorrhage to be exSTIR (short T1 inversion recovery).18 Metastatic melacluded. 14-18320 Pure lipomata appear uniformly hypodense; nomata in the brain appear hyperintense on TI -weighted whereas, dermoids, epidermoids, and teratoma produce a scans because of a combination of hemorrhage and the paramore patchy appearance often with associated calcifica-

26

TARGETT ET AL

tion.14,15,17,19 CT may also show the presence of fat in the subarachnoid space in the case of ruptured dermoid cysts,26 but the perspicuity of MRI for droplets of fatty material is better.I7 The resolution of the clinical signs following reversal of the hydrocephalus is consistent with a secondary hydrocephalus. The second episode of circling, however, may have been produced directly by the medullary lesion. The

1998

dramatic resolution of the hydrocephalus seen on the second MRI examination confirms the effectiveness of the shunting technique as a short-term diagnostic tool and the usefulness of MRI for noninvasive monitoring of this condition. Dermoid cyst should be considered as a differential diagnosis for lesions having high signal intensity in T1weighted images but do not enhance following gadolinium administration.

REFERENCES 1. Lantos PL, Vandenberg SR, Kleihues P. Tumours of the nervous system. In: Graham DI, Lantos PL. Greenfield’s neuropathology, 6th ed. London: Arnold, 1997. 2. Marks SL, Harari J, Dernell WS. Dermoid sinus in a Rhodesian Ridgeback. J Small Anim Pract 1993;34:356-358. 3. Lee Gross T, Ihrke PJ, Walder EJ. Veterinary dermatopathology. St Louis: Mosby, 1992. 4. Slatter D. Fundamentals of veterinary ophthalmology, 2nd ed. Philadelphia: WB Saunders, 1990. 5. Summers BA, Cummings JF, de Lahunta A. Veterinary neuropathology. St. Louis, MO: Mosby, 1995. 6. Walter GF, Kleinert R. Dysontogenetic brain tumours-Proposal for an improved classification. Neuropath App Neurobiol 1987; 13:273287. 7. Henderson JP, Pearson GR, Smerdon TN. Dermoid cyst of the spinal cord associated with ataxia in a cat. J Small Anim Pract 1993;34402404. 8. Fatone G, Brunetti A, Larnanga F, Potena A. Dermoid sinus and spinal malformations in a Yorkshire Terrier: Diagnosis and follow-up. J Small Anim Pract 1995;36:178-180. 9. Selcer EA, Helman RG, Selcer RR. Dermoid sinus in a Shih Tzu and a Boxer. J Am Hosp Assoc 1984;20:634-636. 10. Miller NR, Epstein MH. Giant intracranial dermoid cyst: Case report and review of the literature on intracranial dermoids and epidermoids. Can J Neurol Sci 1975;2:127-134. 11, Howard-Martin M, Bowles MH. Intracranial dermoid cyst in a dog. J Am Vet Med Assoc 1988;192:215-216. 12. Mcgarth JT. Neurologic examination of the dog: With clinicopathologic observations, 2nd ed. London: Henry Kimpton, 1960. 13. Clemmens RM. Ventriculoperitoneal shunt for hydrocephalus. In: Bojrab MJ. Current techniques in small animal surgery, 2nd ed. Philadelphia: Lea & Febinger, 1983. 14. Markus H, Kendall BE. MRI of a dermoid cyst containing hair. Neuroradiology 1993;35:256-257. 15. Gormley WB, Tomecek FJ, Qureshi N, Malik GM. Craniocerebral

epidermoid and dermoid tumours: A review of 32 cases. Acta Neurochir (Wein) 1994;128: 115-121. 16. Erdem G, T o p p M, Topaloglu H, Bertan V, Arikan U. Dermoid tumor with persistently low CSF glucose and unusual CT and MRI findings. Pediatr Neurol 1994;10:75-77. 17. Zimmerman JR, Raila FA, Russell WA, Smith RR. Case report: CT and MRI of intracranial dermoid causing headache. J Mississippi State Med Assoc 1996;37:509-512. 18. Behari S, Benerji D, Gupta RK, Agarwal P, Chhabra DK. Problems in differentiating intradural lipoma from dermoid on magnetic resonance imaging. Australas Radio1 1997;41:196-198. 19. Rubin G, Scienza R, Pasqualin A, Rosta L, Da Pian R. Craniocerebral epidermoids and dermoids: A review of 44 cases. Acta Neurochir 1989;97:1-16. 20. Davidson HD, Ouchi T, Steiner RE. NMR imaging of congenital intracranial germinal layer neoplasma. Neuroradiology 1985;27:301-303. 21. Dooms G, Hricak H, Sollitto RA, Higgins CB. Lipomatous tumors and tumors with fatty component: MR imaging potential and comparison of MR and CT results. Radiology 1985;157:479483. 22. Nagele T, Klose U, Grodd W, Opitz W, Schick F, Steinbrecher A, Voigt K. Three-dimensional chemical shift-selective MRI of a ruptured intracranial dermoid cyst. Neuroradiology 1996;38:572-574. 23. Bradshaw JR. Magnetic resonance imaging of intracranial tumows-Review. Clin MRI 1992;2:77-85. 24. Hasso AN, Bell SA, Tadmor R. Intracranial vascular tumors. Neuroimaging Clin North Am 1994;4:4:849-870. 25. Keller MA, Smith RML. In: Kucharczyk W. MRI: Central nervous system, 1st ed. London: Mosby-Wolfe. 26. Wilms G, Casselman J, Demaerel Ph, Plets C, De Haene I, Baert AL. CT and MRI of ruptured intracranial dermoids. Neuroradiology 1991;33:149-151. 27. Hahn FJ, Ong E, McComb RD, Mawk JR, Leibrock LG. MR imaging of ruptured intracranial dermoid. J Comput Assist Tomogr 1986;10:888-892.