• Open Access

Canine Intraspinal Meningiomas: Imaging Features, Histopathologic Classification, and Long-Term Outcome in 34 Dogs


  • Presented in part as an abstract at the 24th Annual American College of Veterinary Internal Medicine Forum, Louisville, KY, May 25–28, 2006.

Corresponding author: Dr Beverly K. Sturges, DVM, Department of Surgical and Radiological Sciences, University of California, Davis, CA 95616-8747; e-mail: bksturges@vmth.ucdavis.edu.


Background: Meningioma is the most common primary intraspinal nervous system tumor in dogs. Clinical findings, clinicopathologic data, and treatment of these tumors have been reported sporadically, but little information is available regarding cerebrospinal fluid (CSF) analysis, histologic tumor grade, or efficacy of radiation therapy as an adjunct to cytoreductive surgery.

Animals: Dogs with histologically confirmed intraspinal meningiomas (n = 34).

Methods: A retrospective study of dogs with intraspinal meningiomas between 1984 and 2006 was carried out. Signalment, historical information, physical examination, clinicopathologic data, radiation therapy protocols, surgery reports, and all available images were reviewed. All tumors were histologically classified and graded as defined by the international World Health Organization classification scheme for central nervous system tumors.

Results: Intraspinal mengiomas in dogs are most common in the cervical spinal cord but can be found throughout the neuraxis. Location is correlated with histologic grade, with grade I tumors more likely to be in the cervical region than grade II tumors. Myelography generally shows an intradural extramedullary compressive lesion. On magnetic resonance imaging, the masses are strongly and uniformly contrast enhancing and a dural tail often is present. CSF analysis usually shows increased protein concentration with mild to moderate mixed pleocytosis. Surgical resection is an effective means of improving neurologic status, and adjunctive radiation therapy may lead to an improved outcome.

Conclusions and Clinical Importance: Biopsy is necessary for definitive diagnosis, but imaging and CSF analysis can suggest a diagnosis of meningioma. Treatment of meningiomas with surgery and radiation therapy can result in a fair to excellent prognosis.

Meningioma is the most common primary central nervous system (CNS) neoplasm affecting the spinal cord of dogs.1–8 Typically, intraspinal meningiomas cause a chronic, progressive myelopathy with mild to moderate spinal pain.5,9–14 A tentative diagnosis of intraspinal tumor may be made after advanced imaging and cerebrospinal fluid (CSF) analysis, but histologic evaluation of tumor type is essential for definitive diagnosis. Results of CSF analysis usually are within normal limits,7,8,13 but mild to moderately increased protein concentration with or without pleocytosis also has been reported.9,11,13,15,16 On myelography, spinal cord meningiomas usually are discrete extramedullary intradural masses.5−8,10–18 Reports of computed tomography (CT) imaging are scarce, but contrast-enhancing mass lesions have been described.14,17 Typical magnetic resonance imaging (MRI) characteristics include a contrast-enhancing mass with a broad-based dural attachment and variable signal intensity on precontrast T1- and T2-weighted images.4,6,14,19

Most histopathologic characteristics of canine meningiomas are strikingly similar to those in people.20–22 The international World Health Organization (WHO) classification system for human meningiomas provides specific criteria for their histopathologic classification and grade.23 Tumor grading is an important diagnostic criterion in human intracranial meningiomas and has both predictive and prognostic value.21 Most human spinal meningiomas are grade I24–26 but higher grade tumors (II and III) are directly correlated with a worse prognosis.27,28 Currently, the WHO classification system for canine meningiomas20 and several textbooks29–32 do not include all of the histologic subtypes that occur nor do they provide any guidelines or criteria for grading malignancy potential. Although several reports have described various histologic subtypes in canine spinal cord meningiomas, tumor grading is rarely used.33,34 A recent review of 112 intracranial canine meningiomas has applied the human WHO classification and grading system to these tumors and clearly outlines the advantages of and purpose for using this system compared with the current WHO system for dogs.22 Reported treatments for canine intraspinal meningiomas consist of medical management and cytoreductive surgery with or without radiation therapy.2,5,9–19,35,36 After surgery, survival times vary from 4 to 47 months.2,5,9–19 The effect of adjunctive radiation therapy on survival time is unclear, because there are few reports describing its use.10,17–19

The goals of this retrospective study were to describe the clinical, imaging, and histologic characteristics of canine primary spinal cord meningiomas, to evaluate surgical intervention and radiation therapy as treatment options, and to examine any correlation between tumor grade and clinical outcome.

Materials and Methods

Medical records of dogs with a histopathologic diagnosis of intraspinal meningioma at the Veterinary Medical Teaching Hospital (VMTH), University of California, Davis, between 1984 and 2006 were reviewed. The following information was recorded from the patient records: age at diagnosis, breed, weight, presenting signs, prior medical history, and physical and neurologic examination findings. Abnormal findings on any of the following diagnostic tests also were recorded: CBC, serum biochemical profile, urinalysis, thoracic radiographs, and abdominal radiographs or abdominal ultrasound. Results of CSF analysis, including red blood cell count, total nucleated cell count (TNCC), and differential count and protein concentration, were recorded.

All available imaging studies were reviewed by a board-certified radiologist (R.E.P.) and board-certified neurologist (B.K.S.), and a consensus interpretation was reached. On myelograms, CT images, and MR images, tumors were categorized according to their location longitudinally along the neuraxis as well as transversely within the vertebral column according to their relationship with the dura mater (extradural, intradural/extramedullary, or intramedullary). Additionally, vertebral column radiographs were reviewed for any associated structural abnormalities. Myelograms were performed by lumbar or cisternal injection of iopamidala or iohexol.b Using a 4 mm slice thickness, CTc images were reviewed for the presence or absence of contrast enhancement when IV contrast agentd was administered. MR images were obtained with either a 0.34 Te imaging system or a 1.5 Tf imaging system, and T1-weighted (T1W) and T2-weighted (T2W) spin echo pulse sequences were acquired with settings considered optimal for imaging the spinal cord. Additional T1W transverse and sagittal images were acquired after administration of a contrast mediumg IV. MR images were evaluated for the character of contrast enhancement (strength and homogeneity), signal intensity relative to spinal cord gray matter, and the presence or absence of a dural tail. Dural tail was defined as at least 1-slice thickness of meningeal enhancement beyond the tumor on postcontrast T1W images in the dorsal plane, sagittal plane, or both.

Each tumor was histologically classified and graded by a board-certified pathologist (R.J.H.) using the criteria defined by the latest human WHO international histologic classification of CNS tumors.37 Tumors were graded as benign (grade I), atypical (grade II), or anaplastic (grade III). All grade I tumors were subtyped histologically. Features that characterized grade II tumors included the presence of either >4 mitotic figures/10 high-power fields (HPF; area of 0.16 mm2) or exhibiting at least three of the following features: necrosis, small cells, increased cellularity, nuclear atypia, or sheet-like growth patterns.37,38 The chordoid subtype was assigned to grade II. The anaplastic grade III tumor had >20 mitotic figures/10 HPF and contained malignant cytologic features, including necrosis.37

Radiation therapy was performed with a Cobalt 60 teletherapy unit (2 dogs) or a 4 MV linear accelerator (5 dogs) and started a median of 10 days after cytoreductive surgery (range, 7–16 days). The median prescribed dose was 45 Gy (range, 45–48 Gy) with a median prescribed dose per fraction of 3 Gy (range, 2.5–4 Gy). All dogs were treated with parallel opposing fields. Three dogs had treatment plans calculated with the dose prescribed to the level of the spinal cord and 4 dogs had computer plans done based on postoperative CT images. All dogs received prednisone at anti-inflammatory dosages during and for a variable period of time after radiation therapy.

All available information concerning treatment protocol utilized in each dog was reviewed. When follow-up information was not available from the medical record, owners, referring veterinarians, or both were contacted for information regarding the dog's clinical progression and survival time.

The Kaplan-Meier method of survival function estimation was used to determine mean survival for the 4x combinations of treatment modality and grade of tumor. The χ2 test of independence was used to test for significant differences in distribution of tumors longitudinally along the neuraxis according to grade as well as breed distribution relative to hospital population. Statistical computations were done with computer softwareh and significance was defined as P≤ .05.



Thirty-four dogs (20 females and 14 males) had a histologically confirmed diagnosis of intraspinal meningioma. Median age at the time of diagnosis was 9.1 years (range, 4–15 years) and median weight was 30 kg (range, 3–47 kg). Fifteen breeds were represented, including 6 Boxers, 5 Golden Retrievers, 5 Labrador Retrievers, 3 German Shepherd dogs, 2 Shetland Sheep dogs, and 1 each of 10 other pure bred dogs. In addition, 3 mixed breed dogs were included. Compared with the breed distribution of cases admitted to the VMTH over these years, both Boxers and Golden Retrievers were overrepresented.

Clinical Presentation

Information regarding clinical signs was available for 32/34 dogs. The median time between the initial clinical signs reported by the owner and presentation to the VMTH was 98 days (range, 8 days to 3 years). Twenty-eight of 32 dogs (88%) had clinical signs of neurologic dysfunction for at least 1 month before presentation. The most common complaints included ataxia and paresis of ≥1 limbs (24/32), apparent spinal pain (16/32), lameness (7/32), or urinary or fecal incontinence (5/32). Abnormalities found on neurologic examination consisted predominantly of ataxia and paresis (26/32) and apparent spinal pain (17/32) reflecting the neuroanatomic localization of the tumor.

Diagnostic Testing

Results of CBC, serum biochemistry, and urinalyses were reviewed. The most common abnormalities included mild increases in liver enzyme activity (16/30 dogs), mild nonregenerative anemia (4/30 dogs), and evidence of urinary tract infection (6/25 dogs). Thoracic radiographs and abdominal radiographs or abdominal ultrasound were normal except in 3 dogs that had masses in the heart base, esophagus, and thyroid, respectively.

CSF analysis was performed on 29 samples from 26 dogs. The mean protein concentration in the CSF of all 26 dogs (based on lumbar puncture when both lumbar and cisternal samples were available) was 212 mg/dL (range, 19–836; reference range <35 mg/dL). The mean WBC count for all CSF analyses was 11 cells/μL (range, 1–47; normal <4/μL). Ten of the dogs had cell counts within the reference range. Neoplastic cells were not seen in any sample. CSF protein concentrations and TNCC were significantly lower in dogs with cervical tumors (mean protein = 98 mg/dL; mean cell count = 4/μL) versus lumbar tumors (mean protein, 158 mg/dL; mean TNCC, 24/μL; P= .044 for protein levels and P= .001 for TNCC). There was no significant difference in protein concentrations (P= .12) or TNCC (P= .80) in CSF taken from cisternal versus lumbar sites. Also, the CSF protein concentrations (P= .21) and cell counts (P= .53) in grade I versus grade II tumors were not significantly different.

In 12/25 dogs where 2-view vertebral column radiographs were completed, there was subjective widening of the vertebral canal at the site of the tumor (Fig 1a). Additional findings included spina bifida at C5-6 in a dog with a meningioma at C6-7 and generalized osteopenia in 2 dogs. Myelography clearly defined the location of the tumor longitudinally along the neuraxis in 23/24 dogs. The 1 dog in which myelography did not define the tumor had a tumor at C1, where there was poor contrast filling. Thirteen tumors were categorized as intradural/extramedullary (Fig 1b), five as extradural, two as intramedullary and, in 3 dogs, the transverse location of the tumor with respect to the dura mater could not be determined. Localization of the tumor obtained at surgery or necropsy was in agreement with localization along the neuraxis in all cases; however, prediction of location with respect to the dura was accurate only in 19/24 dogs. All meningiomas had an intradural component at surgery or necropsy.

Figure 1.

 Cervical spine of a dog with a meningioma at C2. (A) Plain lateral spinal radiograph showing widening of the vertebral canal at C2. (B) myelogram showing an intradural-extramedullary contrast pattern.

CT scanning was completed after myelography in 7 dogs. In 5 of these dogs, all of which had an intradural/extramedullary mass according to myelography, the classification of the tumor on the CT images was consistent with the myelographic findings and also agreed with surgical or histopathologic findings. In the other 2 dogs, both myelography and CT imaging were inconsistent with the ultimate transverse location of the tumor. IV contrast was administered in 2 dogs and there was contrast enhancement of the mass in both.

A discrete mass causing marked spinal cord compression (Fig 2a) was present on MR images of 15 dogs undergoing MR imaging. Precontrast T1W images (12/15 dogs) showed mild to moderate hyperintensity of the tumor relative to the spinal cord; the tumors in the remaining 3 dogs were hypointense. All tumors for which precontrast and postcontrast images were available (12/15) were contrast enhancing. The masses on the postcontrast images were strongly hyperintense in 3 dogs for which no T1W precontrast images were available, consistent with contrast enhancement. A dural tail was identified in 12/15 dogs. T2W images available in 10 dogs showed mild to marked hyperintensity of the tumor relative to the spinal cord in 8 dogs and isointensity in 2 dogs. In 8 dogs, MRI was done after myelography (Fig 2a and b). In 1 dog in which myelography failed to locate a lesion along the neuraxis, MRI demonstrated a mass at C1. In six of the remaining 7 dogs, transverse location agreed among myelography, MRI, and surgical or necropsy findings. In 1 dog, MRI indicated an intramedullary location whereas myelography indicated an extradural location and necropsy identified an extramedullary mass that was both extradural and intradural.

Figure 2.

 Cervical spine of the same dog as in Fig 1 with a meningioma at C2. (a, b) Sagittal and transverse T1W postcontrast MR images showing large uniformly contrast-enhancing mass with dural tail. (c, d) Sagittal and transverse T1W postcontrast MR images obtained immediately following surgical debulking of the tumor. Contrast enhancement ventral to the spinal cord may represent residual tumor (arrow).

Postoperative MRI studies were done in 4 dogs (mean, 15 months postoperatively; range, 6–26 months) to determine the cause of neurologic deterioration (Fig 2c and d). In 2 dogs, neither of which had received radiation therapy, regrowth was apparent after 6 and 26 months, respectively, and a discrete contrast enhancing mass with similar imaging characteristics to the original lesion was seen. In 1 dog that had undergone radiation therapy 18 months before (Fig 3), a small discrete mass at the original site of the tumor was seen. In addition, there was marked hyperintensity in the spinal cord on T2W images caudal to the original tumor. In 1 dog that was 9 months postsurgery (without radiation), there was diffuse hyperintensity of the spinal cord on T2W images at the tumor site and contrast enhancement on T1W images. However, there was no evidence of a discrete mass as would be expected with tumor regrowth.

Figure 3.

 Sagittal T2W MR image of the same dog as shown in Fig 1 and 2. The image was obtained 18 months following radiation therapy with the recurrence of clinical signs. There is recurrence of the tumor at the initial site (C2) as well as diffuse hyperintensity of cervical spinal cord in the radiated field.

Pathologic Examination

Of the 34 intraspinal meningiomas, 18 were classified histologically as grade I tumors, of which 10 were meningothelial, 7 were transitional, and 1 was microcystic. Fifteen tumors were classified as grade II, of which 12 were atypical and 3 were of the chordoid type. One tumor was classified as grade III (malignant). There was a significant difference (P=.023) in the mean age of diagnosis in dogs with grade II tumors (8.3 years) versus dogs with grade I tumors (10.3 years).

In 23/34 dogs (68%) meningiomas were located in the cervical region, with 20 of these at the level of, or cranial to, C3 vertebrae (Fig 4). Eight meningiomas (24%) were located within the lumbar spinal canal, two in the thoracic spinal canal, and 1 tumor spanned multiple segments in the cervical and thoracic spinal canal. Of the 23 meningiomas located in the cervical region, 7 were grade II (4 atypical and 3 chordoid). The remaining 16 tumors were grade I. Both tumors in the thoracic region were grade II (atypical) meningiomas. Of the lumbar tumors 5/8 were grade II (atypical), 1/8 was a grade III, and 2/8 were grade I. Grade II tumors were statistically more likely to be found in the thoracolumbar region and grade I tumors were more likely to be found in the cervical region (P= .049). No correlation between the imaging characteristics of the tumor and grade was found.

Figure 4.

 Number and grade of meningiomas according to location along the neuraxis.

Treatment and Outcome

Dogs were divided into 3 treatment/outcome categories: (1) dogs that were euthanized with no surgery attempted (10/34) with a median time to euthanasia of 1 day (range, 0–21 days); (2) dogs that underwent surgery but died or were euthanized intraoperatively or perioperatively (6/34) so that the benefits of surgery could not be evaluated (median time to euthanasia, 0.5 days; range, 0–27 days); and (3) dogs that underwent surgery with or without postoperative radiation therapy and survived at least 3 months (18/34). All dogs in category 2 had tumors in the cranial cervical spinal cord. All dogs in category 3 improved clinically after surgery and were normal or nearly normal neurologically except for the dog with a grade III tumor. This dog had surgery without radiation therapy and was euthanized after 3 months as a result of neurologic deterioration. The outcome in the remaining 17 dogs in this category was further examined with respect to grade of tumor and whether or not adjunctive radiation therapy was administered.

Ten dogs in category 3 with grade I or II tumors had surgery with no radiation therapy (two of these dogs were reoperated for tumor recurrence at 12 and 25 months, respectively, after the original surgery). Eight of these dogs (8/10) were eventually euthanized due to recurrence and progression of neurologic signs (mean survival time; 19 months). One dog was alive 29 months after surgery. Two dogs were euthanized after 14 and 16 months, respectively, for unrelated reasons.

Of the 7 dogs in category 3 completing radiation therapy after surgery, three were euthanized for recurrence and progression of neurologic signs at 18, 27, and 36 months postoperatively. One dog was alive at the time of writing (41 months after surgery) and 2 dogs died of other causes after 72 and 78 months, with no apparent neurologic deterioration. One dog was lost to follow up after 4 months. Using neurologic deterioration (defined as progression to the point of inability to walk or to time of 2nd surgery) as an endpoint, there was a significantly longer time to neurologic deterioration in dogs that received radiation therapy compared with those that did not (P= .015). Of the 7 dogs completing radiation therapy, 3 had grade I tumors and 4 had grade II tumors. Of the 11 dogs that did not undergo radiation therapy, 6 had grade I tumors, 3 had grade II tumors, and 1 had a grade III tumor.


Sixty-one dogs with histologically confirmed meningiomas affecting the spinal cord have been reported in the past 32 years.2–19,33,36,39 Analysis of previous cases and those included in the present report provides interesting comparisons. Boxers were the breed most overrepresented in this study and were the most frequently affected breed in previous studies (4 of the 36 cases for which breed was reported).5,7,9–12,14–19,33,35,36,39,40 This may be a reflection of breed popularity or may represent the propensity of this breed to develop neoplasms. A search of pathology reports at our institution over the years included in this study for neoplasia affecting the vertebral column, spinal cord, or both indicated that about 14% of such dogs had primary spinal cord neoplasia (eg, ependymomas, gliomas, spinal nephroblastomas, meningiomas). Of these, approximately 65% were meningiomas.

In this study, mean age at the time of diagnosis was 9.1 years (range, 4–15 years), similar to the mean age of 8.6 years (range, 0.4–13 years) of the 51 dogs for which age was reported in the literature. Younger dogs were significantly more likely to have grade II tumors, and grade II tumors were significantly more likely to be thoracolumbar in location.

Increased CSF protein concentration was a consistent finding in this study. Dogs with meningiomas in the lumbar spinal cord had significantly higher protein concentrations and cell counts than did dogs with cervical tumors, most likely reflecting a local increase of cells and protein in the vicinity of the tumor, because most CSF samples were collected from the lumbar site. CSF findings in this study differ from those reported in the literature.7–13,15,16 In the 19 dogs with spinal meningiomas for which CSF analyses were reported, 11 dogs had normal CSF, 5 dogs had a normal cell count associated with an increased protein concentration, and 3 dogs had an increase in both cell count and protein concentration.7,8,13

Findings on spinal radiographs and myelograms in this study were similar to those reported previously.5–12,16–18 Myelography was accurate at locating the tumor along the neuraxis and typically showed an intradural/extramedullary pattern with a minority of extradural, intramedullary, or indeterminate patterns. In general, CT was no more successful than the original myelogram at characterizing tumor location with respect to the dura.

MRI characteristics of spinal tumors reported here with regard to contrast enhancement and dural tails were very similar to those previously reported.4,19 T2W images and precontrast T1W images both were more likely to be hyperintense than reported in previous studies. Differences in MRI protocol may account for these variations in intensity.

Meningiomas occurred primarily in the cranial cervical and lumbar regions of the spine in this study. Previous studies showed a similar distribution of spinal meningiomas. Of the 53 reported cases for which location was mentioned,2–5,7–9,11–14,17–19,33,36,39 36/53 (68%) were cervical, 2/53 (4%) were thoracic, 12/53 (23%) were lumbar, and 3/53 (6%) were multifocal. The frequency of tumor occurrence longitudinally along the neuraxis in this study was statistically associated with tumor grade, suggesting a different etiology or pathogenesis for the different grades of tumors. One of the dogs in this study had a grade I meningioma extending over several spinal cord segments in the cervical and thoracic spine. This highly unusual locational variant has been reported in 1 other dog, a 5-month-old Rhodesian Ridgeback puppy.39

In contrast to dogs, the majority of human intraspinal meningiomas occur in the thoracic region.27 A reason for this predilection has not been established, but in humans arachnoid cap cells (found within arachnoid proliferations) are believed to be the progenitor cell for meningiomas and arachnoid proliferations are most common in the thoracic spine.41 The distribution of arachnoid proliferations in the canine spinal meninges is unknown.

In this study, grade I and grade II tumors occurred with similar frequency, but there was only 1 grade III tumor. There have been very few other reports of graded spinal meningiomas. One study reported that 2/4 meningiomas were malignant (grade III)33 whereas another reported 2/2 grade I tumors.34

Surgery resulted in neurologic improvement in 17/24 (71%) of dogs. The 7 dogs for which surgery was ineffective may be divided into 4 categories. In 3 dogs, euthanasia was necessitated by complications during or immediately after surgery. All 3 dogs had meningiomas in the C1-3 region of the spinal cord. This increased mortality may be due in part to the highly variable vascular and vertebral anatomy of the region as well as the close proximity to the brain stem. In 2 dogs, surgery was done to define a mass lesion incompletely characterized by myelography. These dogs were euthanized intraoperatively based on the presence of the tumor. Had MRI been available, they would most likely have been euthanized, with surgery not attempted. In 1 dog, the meningioma was a grade III intramedullary tumor. This dog's clinical signs did not improve after surgical resection, and the dog was euthanized because of neurologic deterioration 90 days postoperatively. In 1 dog, euthanasia was done 27 days after surgery because of septicemia that apparently was not related to the surgical procedure.

Excluding the 2 dogs for which surgery was exploratory in nature, the probability of complications associated with surgery was 3/22 (14%). The probability of complications for dogs with tumors at C1-C3 was 3/15 (20%) and was 0/7 for dogs with tumors caudal to C3. The probability of neurologic improvement after surgery (excluding the 2 dogs for which the surgery was exploratory and the dog that died of septicemia 1 month later) was 17/21 (81%). The mean reported survival time in the literature after cytoreductive surgery of spinal meningiomas is 15.6 months (15.3 months in dogs that received radiation therapy and 15.8 months in dogs that did not).2,5,10,11,14,17–19,35 These times should be interpreted with caution because the time of follow-up was short in several reports.

The results of this study suggest an improved outcome in dogs that received postoperative radiation versus those that did not. Among the dogs that did not receive radiation therapy, the longest survival times (4 and 3.5 years) were in dogs that required a 2nd surgery, and both of these dogs eventually were euthanized because of severe progression of neurologic signs attributed to regrowth of the tumor. The dog with the 3rd longest survival time (still alive after 27 months) has been paraplegic with urinary and fecal incontinence for more than a year, and has thus lived well beyond the point when many owners would have chosen euthanasia. Among the 7 dogs that received radiation therapy, the 2 dogs with the longest survival times (6 and 6.5 years) did not experience neurologic decline before dying of other causes. A 3rd dog has survived 41 months with excellent neurologic status and no deterioration. When taking these factors into account, and with neurologic deterioration rather than survival as an endpoint, there was a significant difference between dogs that received radiation therapy and those that did not (P= .015). Three of the 7 irradiated dogs declined neurologically after 18–36 months. Radiation therapy can cause delayed damage to neural tissue,42 and this may have caused or contributed to their decline.

Meningiomas also affect the spinal cord in humans, but in contrast to those in dogs, the vast majority are benign (WHO grade I) and well encapsulated.43 Surgical resection is the treatment of choice. In contrast to dogs, incidence of recurrence postoperatively is low (3–15%)27 and, for this reason, radiation therapy is rarely necessary. It is sometimes undertaken with incomplete resection when there has been recurrence or when surgery is not possible because of poor general health.25,44 Limited data indicate that radiation therapy is beneficial in these cases.25,27,44

There was no obvious correlation between outcome and grade of the tumor in this study, but the small, nonrandomized population makes it impossible rule out such a correlation. Uncontrolled factors include surgical technique, initial size of the tumor, and decision to pursue radiation therapy. One potentially important confounding factor is that almost all grade I tumors are in the cervical region, whereas grade II tumors are more evenly spread out along the neuraxis. Animals with tumors in the thoracolumbar region are likely to have an inherent bias toward increased longevity because paraparetic animals are much easier for owners to manage than tetraparetic animals. No animal in our series with a cervical tumor had a 2nd surgery, and 1 animal with a lumbar tumor has been kept alive long past the point of any spinal cord function, which would not be an option in a dog with a cervical tumor. At this point, we do not know whether the clearly defined and highly predictive human grading scheme is applicable to dogs. In order to determine the prognostic value of this systematic grading scheme, both grade and clinical outcome of many more meningiomas would need to be evaluated, ideally in a controlled, prospective study.

In summary, most meningiomas in the spinal cord in dogs are grade I or II, with younger dogs more likely to develop grade II meningiomas. Grade I meningiomas have a strong predilection for the cervical spinal cord, whereas grade II tumors are more evenly distributed between the cervical and lumbar spinal cord. CSF analysis almost always demonstrates increased protein concentrations and often shows mild, mixed pleocytosis. Myelography is an effective method of identifying the location of meningiomas along the neuroaxis and often shows an intradural extramedullary contrast pattern. Meningiomas have a characteristic appearance on MRI with homogeneous contrast enhancement, iso- to hyperintensity on T1W and T2W images, and frequently a dural tail. Dogs that have grade I or II meningiomas generally improve clinically after surgery and may remain clinically improved for months to years postoperatively. Radiation therapy improves outcome after surgery and appears to prevent recurrence of disease in some dogs. There is no apparent correlation between grade of tumor and imaging characteristics or long-term outcome; however, larger numbers of dogs would need to be studied to rule out such a correlation.


aIsovue 200, Bracco Dx, Princeton, NJ

bOmnipaque, Winthrop-Breon Laboratories, New York, NY

cGE 9800 or GE HiSpeed FX/I, Milwaulkee, WI

dConray 400 Mallinckrodt Inc, St Louis, MO

eResonex 5000, Resonex Inc, Sunnyvale, CA

fGeneral Electric Signa, Milwaukee, WI

gMagnevist, 469.01 mg gadopentatate per mL, Berlex Laboratories, Wayne, NJ

hEgret Statistical Software V 2.0.1 1999, StatXact v 8.0, Cytel Software Corporation, Cambridge, MA