Background

Mast cell tumours (MCTs) arise from malignantly transformed mast cells. In dogs, they are the most common cutaneous malignancy, accounting for 11–18% of cutaneous neoplasms [1,2,3,4,5,6]. Mast cell tumours are highly variable in location, behaviour and appearance. Typically occurring in the dermis and subcutaneous tissue, MCTs can occur as solitary or multiple lesions. They may infiltrate surrounding tissues and show widespread disseminated metastasis [7,8,9,10]. Location of tumours has been associated with biological behaviour, with tumours arising from the mucosa/mucocutaneous junctions, muzzle and inguinal regions potentially exhibiting more aggressive behaviour [11,12,13,14]. Uncommon extracutaneous sites of primary tumours include the conjunctiva, salivary glands, nasal cavity, nasopharynx, larynx, oral cavity, gastrointestinal tract, urethra and vertebral column [14,15,16,17,18,19,20].

As primary or metastatic tumours, MCTs affecting the vertebral column are uncommon in the veterinary literature. Polyostotic MCT metastasis affecting the vertebrae has been reported once previously [15]. The aim of this case series is twofold, to describe the clinical presentation of two dogs with cutaneous MCTs and one with subcutaneous MCT subsequently diagnosed with metastatic disease causing polyostotic lesions, and to describe the MRI features of metastatic osseous MCTs.

Case presentation

Case 1

An 8-year-old male neutered, small crossbreed dog was referred for further assessment and treatment of a metastatic low-grade (Kiupel), grade II (Patnaik) cutaneous MCT. The dog had presented to the primary practitioner (PP) a year prior for an excisional biopsy of a 5.0 × 4.6 mm, right sided, raised cutaneous upper lip mass. The patient was otherwise well, and the caregiver had no other pertinent health concerns. Histopathology confirmed a grade II (Patnaik), low-grade (Kiupel) MCT with a mitotic count (MC) of 1/10 high power fields (HPF), Ki-67 testing was not performed. Complete excision was achieved with a narrow deep margin of 0.7 mm. Seven months later, right mandibular lymphadenomegaly was noted and excisional biopsy was performed. Histopathology was consistent with MCT metastasis. Fine needle aspirates (FNAs) of the left mandibular lymph node, liver, and spleen did not show evidence of metastasis. Adjuvant Lomustine (70 mg/m2 per os every 4 weeks [Bova]) chemotherapy was initiated one month later. Three months after starting chemotherapy, a subcutaneous nodule was noted in the area of the previously excised right mandibular lymph node, cytology of which confirmed lymph node MCT metastasis. Lomustine was halted and the treatment was changed to toceranib phosphate (3.25 mg/kg per os every 48 h [Palladia; Zoetis UK limited]). The patient was then referred for further treatment and assessment, one year after the initial presentation to the PP.

At the time of assessment by the referral clinician (European Specialist in Small Animal Oncology), the patient was bright and alert and with normal vital parameters. Right-sided mandibular and left-sided cervical masses were noted, suspected to represent mandibular and medial retropharyngeal lymph nodes effaced by mast cell metastasis, respectively. In-house cytology of the mandibular mass was performed and consistent with MCT. Cytology of the cervical mass was not performed. A computed tomography (CT) of the head, thorax and abdomen revealed right mandibular and bilateral medial retropharyngeal lymphadenomegaly. There was heterogeneous enhancement of the liver following contrast administration; however, no evidence of hepatic nodules or lesions. There was no evidence of pulmonary metastasis at the time of the scan.

Splenic cytology showed reactive lymphoid hyperplasia, intramedullary haematopoiesis and small numbers of individual mast cells. Hepatic cytology showed vacuolar hepatopathy and mild mastocytosis with low numbers of individualised mast cells without overt atypia.

Right and left mandibular and retropharyngeal lymphadenectomy and surgical scar revision of the upper lip were performed. The right retropharyngeal and mandibular complex was invasive and adherent to the right carotid, vagosympathetic trunk, laryngeal nerve and right mandibular salivary gland. It was, therefore, suspected that residual disease remained following surgery. Histopathology revealed overtly metastatic MCT within the left and right retropharyngeal and right mandibular lymph nodes, classified as HN3 by the Weishaar et al. grading system for histopathology of nodal mast cells [21]. No neoplastic tissue was identified in the excised scar tissue.

A radiotherapy planning CT showed a right cervical subcutaneous nodule at the recent surgery site, most likely consistent with lymphoid tissue. FNAs of this lesion confirmed the presence of residual MCT tissue. Chemotherapy and radiotherapy were initiated two and five weeks, respectively, after surgical removal of the right retropharyngeal and mandibular complex. Vinblastine chemotherapy (2mg/m2 intravenous once weekly for four weeks, followed by once a fortnight until week 12), and prednisolone (40 mg/ m2 per os every 24 h for two weeks, before a reduction to 20mg/m2 per os every 48 h). The patient was maintained on gabapentin (10 mg/kg per os every 12 h [Milpham]), paracetamol (20 mg/kg per os every 12 h [Bova UK]) and amantadine (12.5 mg/kg per os every 24 h [Bova UK]). The patient received five consecutive daily 5 Gy radiotherapy treatments (20 Gy total dose) to the postsurgical site after bilateral removal of mandibular and retropharyngeal lymph nodes. The chemotherapy protocol was halted at week six due to a deterioration in clinical signs.

One week after radiotherapy completion, the patient presented with reluctance to exercise. Neurological examination by a European Specialist in Small Animal Neurology revealed discomfort on palpation of the thoracolumbar region but no neurological deficits. Gabapentin administration was increased to 10 mg/kg every 8 h. The patient was reviewed one week later, and neurological examination documented thoracolumbar kyphosis with ambulatory paraparesis. Postural reactions were absent in the pelvic limbs but normal in the thoracic limbs. Pelvic limb withdrawal reflexes were reduced bilaterally. The tail and anal tone were normal. A neuroanatomical localisation to the L4-S1 spinal cord segments was suspected. C-reactive protein was elevated at 139.56 mg/L (0–10 mg/L), and haematology revealed neutropenia 2.80 × 109/L (3-11.8 109/L) and leukopenia 3.46 × 109/L (6-17 × 109/L). The manual platelet count was moderately reduced at 32 × 109/L.

Case 2

A 15-year-old female neutered springer spaniel was referred for investigation and treatment of ambulatory paraparesis. The patient presented to the PP ten months prior for the removal of a mass within the right caudal mammary gland. Histopathological assessment of the mass and regional lymph node indicated a 2 cm subcutaneous MCT (MC of >10/10 HPF) and invasion of the lymph node by sheets of hyperchromatic, intermediately differentiated mast cells. The lymph node was classified as HN2 by the Weishaar et al. grading system [21]. Lateral and deep tumour-free margins of 1 cm were achieved. The mass was c-KIT mutation negative.

Seven months after the MCT removal, a 3 cm subcutaneous mass was noted in the right groin. Cytology of the mass was consistent with MCT. Abdominal ultrasound showed enlarged, rounded, hypoechoic abdominal lymph nodes and liver nodules. Cytology of the spleen, hepatic nodules and right medial iliac lymph node were consistent with MCT metastasis.

Prednisolone and Vinblastine chemotherapy was initiated, as described in case 1. Two months after initiating chemotherapy, the dog represented to the PP with ambulatory paraparesis. The patient was referred for further investigations and assessment ten months after initial presentation for a subcutaneous MCT.

At the time of assessment by the referral clinician (European Specialist in Small Animal Neurology), the patient was non-ambulatory paraparetic with increased muscle tone of the thoracic limbs. Postural reactions were normal in the thoracic limbs and absent in the pelvic limbs. Withdrawal reflexes were decreased in the pelvic limbs, and normal in the thoracic limbs. Patellar reflexes were normal. Cranial nerves examination was normal. Pain was suspected on palpation of the cervical spine. The neuroanatomical localisation was to the T3-L3 spinal cord segments, but a lesion affecting the cervical area (i.e. muscles, articular processes, bone, meninges, nerve roots or intervertebral discs) could not be ruled out due to the presence of cervical discomfort. The patient’s vital parameters were within normal limits.

Thoracic radiographs did not reveal overt evidence of intrathoracic lymphadenopathy or pulmonary nodules.

Case 3

A 13-year-old male neutered Nova Scotia Duck Toiling Retriever was referred for removal of a cutaneous MCT affecting the mid dorsum, just caudal to the scapulae. Six months prior to presentation the patient had undergone excision of a recurrent low grade (Kiupel) MCT affecting the left shoulder, which had previously been removed eight years prior. Histopathology of the most recent MCT revealed a 20 × 20 mm high grade (Kiupel) cutaneous MCT (MC 15/10 HPF) with narrow margins of < 1 mm deep and lateral. Abdominal ultrasound showed nodular hepatopathy and splenopathy. Hepatic and splenic cytology was consistent with mild vacuolar degeneration of the liver and mild extramedullary haematopoiesis of the spleen. Ultrasonic assessment of the axillary and inguinal lymph nodes did not show evidence of pathological changes.

Vinblastine and prednisolone chemotherapy was initiated as described in case 1. The patient completed the 12-week course of chemotherapy, but did not present for restaging. The patient was then re-referred seven months after completion of chemotherapy with a one-week history of non-ambulatory paraparesis. On general examination the patient had a large subcutaneous mass noted on the right thoracic wall within the axillary region, cytology of which showed spindle cell proliferation with atypia. On neurological exam the patient was non-ambulatory paraparetic. Postural reactions were normal in all four limbs when the patient was adequately supported. Patellar and withdrawal reflexes were normal in the pelvic limbs. Withdrawal reflex was subjectively reduced in the right thoracic limb and normal in the left thoracic limb. Cutaneous trunci reflex was absent caudal to the cranial thoracic region. Cranial nerves examination was unremarkable. The patient was painful on palpation of the cranial thoracic region. The neuroanatomical localisation was to the T3-L3 spinal cord segments. The decreased withdrawal noted in the right thoracic limb was attributed to the large mass lesion, however involvement of C6-T2 spinal cord segments or right brachial plexus could not be ruled out. Haematology, biochemistry and electrolytes were unremarkable.

Magnetic resonance imaging findings

All cases underwent MRI of the vertebral column. Magnetic resonance imaging findings for each case are summarised in Table 1. Following MRI all patients were humanely euthanised as requested by the caregivers. In all patients the MRI studies identified multiple polyostotic vertebral lesions. These varied between multifocal well-defined nodules to ill-defined patches affecting multiple vertebrae (Figs. 1, 2 and 3) as well as other bones including iliae, ribs, sternebrae and scapula. All vertebral lesions were T2W and T1W mildly hyperintense to isointense to the spinal cord and STIR hyperintense. T1W post contrast images were acquired in all cases. All vertebral lesions showed moderate homogeneous contrast enhancement. The vertebral shape was preserved in all cases, but all showed vertebral cortical osteolysis. In all cases, several of the osseous nodules and masses extended beyond the cortical margins of the vertebrae into the extradural space as well as the perivertebral soft tissues. This resulted in spinal cord compression ranging from moderate to severe. In addition to these lesions centred on the vertebral column, additional lesions of similar signal intensity were noted in the medullary cavity of the iliac bones (all cases), ribs (all cases), sternebrae (cases 1 and 2), and scapula (case 2). Where included, these osseous lesions showed a similar signal intensity to that of the vertebral lesions. Splenic nodules and masses were noted in all cases, with variable signal intensity (see Table 1), as well as intra-abdominal lymphadenomegaly, liver nodules and peritoneal as well as retroperitoneal free fluid (case 2). A large intermuscular thoracic wall mass was also seen in case 3.

Table 1 Magnetic resonance imaging findings by case. T2W – T2 weighted, STIR – short Tau inversion recovery, T1W – T1 weighted
Full size table
Fig. 1
figure 1

Case 1 - Mid sagittal images of the thoracolumbar spine of patient 1, in a T2-weighting (A), Short Tau Inversion Recovery (B), T1-weighting (C) and T1-weighting post contrast (D) sequences. Multifocal variably defined nodules are visible in multiple vertebral bodies and in the vertebral lamina of L4 (arrows). Both extradural and paraspinal lesions are noted (arrowheads)

Full size image
Fig. 2
figure 2

Case 2- mid sagittal T2-weighted image of the lumbar spine (A). This shows multiple nodules in the vertebral bodies (arrows), that are T2W isointense to the spinal cord. Several splenic nodules (*) and a markedly enlarged medial iliac lymph node (**) are also visible. The transverse images in T2-weighting (B), T1-weighting (C) and T1-weighting post contrast (D) show the nodule in the vertebral body of L1 vertebra (arrow)

Full size image
Fig. 3
figure 3

Case 3 - Dorsal Short Tau Inversion Recovery (A) and transverse T2-weighted (B), T1-weighted (C), and T1-weighted post contrast (D) images of the T5 vertebral lesion. The intraosseous lesion (arrows) causes bone lysis (arrowheads) and extends into the extradural space (*) and paravertebral soft tissues (**)

Full size image

Outcome

All cases were humanely euthanised either under anaesthetic (case 2 and case 3) or the day following their MRI (case 1). Post-mortem samples of the vertebrae and extradural masses were obtained in all cases, FNAs of the spleen and liver was also performed in cases 1 and 2.

Histopathological and cytopathological findings

In cases 1 and 2, neoplastic round cells were present within the assessed vertebrae, infiltrating intertrabecular spaces and effacing haematopoietic tissue (Fig. 4). In case 1, the same population of cells extended into the adjacent skeletal muscle and extradural space of the spinal canal. MC was 49/10 HPF in case 1 and 5/10 HPF in case 2.

Fig. 4
figure 4

Case 2 - Histopathology of T6 vertebrae. (A) T6 vertebrae and associated extradural mass 2.5x, a neoplastic extradural mass is seen within the spinal canal (blue arrow), adipose tissue is densely infiltrated by a neoplastic round cell population; in multifocal to confluent areas the intertrabecular spaces of the vertebrae are filled with the same population of neoplastic round cells (red arrow). (B) T6 vertebrae 40x, filling the intertrabecular spaces is monomorphic population of neoplastic round cells. Cells have indistinct cell borders and contain moderate amounts of clear cytoplasm. Nuclei are round to oval with stippled chromatin and generally inconspicuous nucleoli. There is mild anisocytosis and anisokaryosis

Full size image

In case 3, histopathology of the T5 vertebrae showed a focal, moderate lymphoplasmacytic synovitis. The fibrous connective tissue was focally markedly infiltrated with macrophages, lymphocytes and plasma cells arranged around a small amount of crystalline material. Some larger round cells were present within this infiltrate. These round cells exhibited no cellular atypia and may have represented neoplastic cells or macrophages. In this region the number of these cells was too small to definitively determine their origin.

In cases 1 and 2 sections of the spinal cord corresponding to the affected vertebrae were found to be histologically unremarkable. The spinal cord was not assessed in case 3.

Extradural masses in all cases displayed a neoplastic round cell population. The epidural mass in case 3, had a MC of 30/10 HPF. An additional intermuscular axillary mass was examined in case 3 which possessed a similar population of neoplastic cells, this had a MC of 7/10 HPF.

In all cases, neoplastic cells were round to oval with indistinct cell borders with moderate amounts of pale cytoplasm, varying between amphophilic, basophilic or clear. Nuclei were round to oval with stippled chromatin. Multinucleated cells were not observed in any samples. Occasional pleomorphic nuclei were present in case 3. Anisokaryosis and anisocytosis was mild in cases 1 and 2 and marked in case 3.

In case 1, no overt positive staining of mast cell granules was observed with Toluidine Blue stain (Fig. 5). In all cases, neoplastic cells exhibited staining for CD117. In case 1, staining was consistent with cKit staining pattern III, in cases 2 and 3 staining was consistent with pattern II (Fig. 6). In all cases CD3 staining was negative, excluding T-cell lymphoma. In case 1, CD20 staining was negative and in case 2 CD79A staining was negative, excluding B-cell lymphoma.

Postmortem splenic cytology was consistent with mast cell neoplasia in cases 1 and 2. In case 1, hepatic cytology displayed mast cell infiltration consistent with metastasis. In case 2, cytology of a hepatic nodule and right mandibular lymph node was consistent with mast cell tumour metastasis. Splenic and hepatic sampling was not performed in case 3. Lung cytology was obtained in case 2 and did not show evidence of mast cell neoplasia. In case 1, cerebrospinal fluid (CSF) collected postmortem showed mild mononuclear pleocytosis.

Fig. 5
figure 5

Case 1 – Skeletal muscle adjacent to the spine, 20x. (A) Neoplastic cells are negative in Toluidine Blue stain, (B) neoplastic cells are negative for CD3, (C) neoplastic cells are negative for CD20, (D) Neoplastic cells are positive for CD117 with cKit staining pattern III, confirming mast cell tumour

Full size image
Fig. 6
figure 6

Case 3 – Extradural mass at the level of T5 vertebrae, 40x. (A) Neoplastic cells are positive for CD117 with cKit staining pattern II, confirming mast cell tumour. (B) A few small reactive lymphocytes are positive for CD3

Full size image

Discussion

This case series describes three dogs diagnosed with mast cell tumours (MCTs), all of which developed polyostotic vertebral lesions. Cases 1 and 2 in this series were euthanised approximately one year following initial diagnosis, with confirmed metastasis to the local lymph nodes, liver, spleen and vertebral bodies. Case 3 was euthanised within a year of initiating chemotherapy for a high grade MCT, with the development of a new thoracic wall mass, an extradural mass, vertebral polyostotic lesions and suspected splenic metastasis.

The most common differential diagnoses for polyostotic aggressive lesions of the vertebrae include round cell tumours (e.g. MCT, multiple myeloma, plasma cell tumours, lymphoma and histiocytic sarcoma), and metastasis of distant solid tumours (e.g. carcinoma), much less likely mesenchymal tumours if the distribution of the lesion is multifocal rather than segmental [15, 22,23,24,25,26,27,28,29]. In a retrospective study of MRI features in 60 dogs with tumours affecting the vertebral column, preservation of vertebral shape, homogenous contrast enhancement and lesions centring on the bone were more commonly shown to be associated with round cell neoplasms [22]. These findings are therefore similar to those reported in this case series. The T2W and T1W iso to mild hyperintensity to the spinal cord is also similar to the reported signal intensity of vertebral multiple myeloma [26] and similar to the T2W and STIR signal reported with vertebral lymphoma [30].

Round cell vertebral tumours in dogs include MCT, lymphoma, multiple myeloma, plasmacytomas and histiocytic sarcomas [16, 22, 31, 32]. Bone invasion by MCTs is rarely reported in dogs [15, 20, 33,34,35]. Osteolysis and associated pathological vertebral fractures have been described in people with systemic mastocytosis [36]. A MCT involving the deep cervical soft tissues, with focal osteolysis of the calvarium and brain parenchyma invasion, has been documented [35]. Another published report describes a disseminated MCT that infiltrated the sphenoid bones and resulted in blindness; splenic and jejunal lymph node metastasis was confirmed via cytology but no primary tumour was identified [34]. Local infiltration of the stifle joint and suspected invasion of the adjacent bone, resulting in polyostotic lesions of the tibial plateau and distal patella, has been reported in the case of a subcutaneous MCT [33]. Only one of these cases describes a disseminated MCT affecting distant sites [34]. In previous reports, bone involvement has more commonly comprised of lysis, infiltration or invasion of bone local to the tumour, [33, 35] rather than metastasis from a distant site as seen in this case series. When polyostotic lesions have been reported previously, they have arisen through local infiltration from neighbouring bones [33]. This contrasts with what is seen in this case series, where MCT metastasis is seen as multiple focal lesions affecting different vertebrae.

The majority of previously reported cases of MCT spinal involvement have described extradural masses causing spinal cord compression without vertebral involvement. The first published report of a spinal MCT describes a presumed primary extradural MCT, causing compression of the spinal cord. The mass invaded the dura mater and adjacent connective tissue without vertebral involvement [16]. Invasion of the dura mater was not noted in this case series. A suspected primary extradural MCT extending from the paravertebral musculature into the vertebral canal and causing spinal cord compression has also been reported [37]. Complete staging was not performed, therefore, it is not possible to confirm whether this was truly a primary lesion. Metastatic extradural MCT have been reported previously. In all but one of these cases, lesions resulted in spinal cord compression without vertebral involvement [15, 38, 39]. A case series by Moore et al., describes four cases of spinal MCTs [15]. A primary focal extradural MCT was described with no involvement of the vertebra. The three other cases in the Moore et al., case series were metastatic tumours, two of which were focal extradural lesions. The final case was a metastatic MCT causing polyostotic lesions within the vertebral bodies of L4 and L5, the sacrum and the right ileum. No lesions extended beyond the vertebral bodies or showed expansion into adjacent soft tissues; no focal extradural lesion was noted in this case. On MRI, the lesions were T2W slightly to moderately hyperintense, T1W hypointense and markedly STIR hyperintense. The T2W and T1W signal intensities reported here differ to those in our case series, where the lesions were T1W and T2W isointense to mildly hyperintense to the spinal cord.

In cases 1 and 2 of our series, histopathology confirmed infiltration of the cancellous bone of the vertebrae by a neoplastic mast cell population, extending into adjacent skeletal muscles and extradural space. Case 3 presented with a focal extradural mass at the level of T4 to T6 vertebrae and a second mass surrounding the head of the fifth rib, causing lysis and extending into hypaxial musculature. Histopathology of the T5 vertebral body did not show clear evidence of mast cell infiltration. Magnetic resonance imaging did not identify lesions within this vertebra. Although it is not possible to confirm, it is reasonable to consider that neoplastic cells may be present within the other vertebrae affected by the polyostotic lesions in this case.

Extraspinal lesions were present in all cases, including masses and nodules affecting the spleen (all cases) and liver (cases 2 and 3). Metastasis to the spleen and liver was confirmed in cases 1 and 2, postmortem sampling was not performed in case 3. A large thoracic wall mass was visible in case 3 and postmortem sampling confirmed the diagnosis of MCT. Abdominal and thoracic lymphadenomegaly was also present in case 2. Lesions in these cases were significant, highlighting the importance of assessing extraspinal structures on MRI; of note, none of the cases showed isolated vertebral lesions, and all had evidence of multicentric spread of their MCTs.

This series highlights the potential for MCTs to cause polyostotic vertebral metastases, a presentation not widely recognised in the literature. Metastasis should be considered for patients with previously diagnosed MCTs presenting with spinal hyperaesthesia.