Survival Data for Canine Oral Extramedullary Plasmacytomas: A Retrospective Analysis (1996–2006)

Introduction

Extramedullary plasmacytomas (EMP) consist of neoplastic plasma cells that do not arise from the bone or bone marrow. These tumors are functional derivations of B-cell lymphocytes that produce a single clonal immunoglobulin secretion (i.e., either lambda or kappa light chains, or a single heavy chain of the immunoglobulin molecule).¹ In the past, EMP were often mistaken for other types of neoplasia, including reticulum cell sarcomas, undifferentiated round cell tumors, or neuroendocrine tumors.^2,3 However, increased awareness of this neoplastic process has led to improved diagnostic accuracy.² Recent reports state that EMP comprise approximately 2.5% of all neoplasms in dogs,⁴ usually among middle-aged to older (9 to 10 years of age) dogs.^2,3,5

In people, 80% to 90% of EMP are localized to the head and neck region,^1,10 and they are most commonly found in the nasopharyngeal/sinonasal region.¹ In dogs, EMP can develop from either mucocutaneous or mucosal tissue, and they have been reported in a variety of locations, including the oral cavity, trachea, esophagus, stomach, colon, and skin.^2–9 The veterinary literature suggests that 22% to 28% of EMP in dogs are located within the oral cavity.^2,5,11 However, EMP are often not mentioned in the literature, because they are not among the most common oral tumors of dogs (e.g., melanoma, squamous cell carcinoma, and fibrosarcoma).¹²

Half of all human EMP cases present with a circulating hyperglobulinemia, and one in three tumors will evolve into multiple myeloma.¹³ While multiple myeloma is a documented neoplasm in veterinary medicine, repeated studies have not shown an obvious correlation between EMP and the development of multiple myeloma in dogs.^3,5 Indeed, EMP are most often independent of multiple myeloma and provide no evidence of metastasis (as once believed).^2,3,5 Unlike multiple myeloma, EMP have historically carried a much more favorable prognosis. These tumors tend to be locally invasive, often involving bony destruction, but having a low rate of metastasis. Consequently, complete surgical removal is often curative.^2,5

Materials and Methods

The Texas A&M University Veterinary Medical Teaching Hospital (TAMU VMTH) database was searched for all cases of canine oral plasmacytoma seen from 1996 to mid-2006. Oral plasmacytic gingivitis, reticulum cell sarcomas, and undifferentiated round cell tumors were also identified and evaluated for possible inclusion. For incidence data, the database was also searched for all cases of canine plasmacytomas, as well as all oral tumors within the same time period. Oral tumors were defined as neoplastic processes involving the oral mucosa, gingiva, tongue, lips, odontogenic tissue, mandible, or maxilla. All cases were evaluated by a single pathologist for confirmation as oral plasmacytoma.

Pertinent data were abstracted from medical records, and follow-up information was obtained via telephone conversations with either the referring veterinarian or the owner. Survival time was defined as the time from histological confirmation until death. “Time to recurrence” was defined as the amount of time for regrowth of the tumor after initial surgical cytoreduction, radiation therapy, or systemic treatment (including chemotherapy).

Survival data were evaluated using Kaplan-Meier graphs produced via a computerized statistical program.^a The relationship between various prognostic factors and survival time was evaluated using two-sample survival models. This modeling approach includes four nonparametric statistical tests, including the Gehan-Wilcoxon, Cox-Mantel, Logrank, and Peto-Wilcoxon tests.^a

Age and tumor size were calculated as dichotomous variables. Age was categorized as ≤8 years, and tumor size was dichotomized as either >4 cm³ or ≤4 cm³. Both values were derived using the median of all data.

Results

A search of medical records identified 302 oral tumors diagnosed between January 1996 and July 2006. The most common oral cancers were melanoma (n=93, 30.7%), squamous cell carcinoma (n=55, 18.2%), osteosarcoma (n=27, 8.9%), and fibrosarcoma (n=27, 8.9%). The epulides comprised 18% of oral masses (n=55). Other oral tumors included lymphoma (n=13), chondrosarcoma (n=1), hemangiosarcoma (n=7), fibroma (n=3), neurofibrosarcoma (n=3), and hamartoma (n=2).

Twenty cases of oral plasmacytoma, undifferentiated round cell neoplasm, or severe plasmacytic gingivitis were identified. One case was removed because of lack of followup, and three additional cases (one with plasmacytic gingivitis and two suggestive of plasmacytomas with inconclusive immunohistochemistries) were removed after pathology reevaluation. The remaining 16 confirmed cases of oral EMP represented 5.2% of all oral tumors.

The database was also searched for all plasmacytomas in dogs (regardless of tumor location), and 56 plasmacytomas were found over the same 10-year time frame. While these records were not evaluated, the tumors were not coded in conjunction with multiple myeloma, suggesting a plasmacytic cell neoplasia independent of multiple myeloma. Oral EMP represented 28.5% (n=16) of all EMP diagnosed at TAMU VMTH over this 10-year period.

No obvious sex predilection was found for development of oral EMP, which were diagnosed in two intact and six spayed females, and in one intact and seven neutered males [see Table]. Breeds consisted of three golden retrievers, three Yorkshire terriers, two Cairn terriers, one boxer, one Airedale terrier, one Scottish terrier, one Labrador retriever, one cocker spaniel, and three mixed-breed dogs. The median age of all dogs at the time of diagnosis was 8.0 years (range 1.5 to 17 years).

Diagnostic three-view thoracic radiography was performed in 14 of the 16 dogs, with no evidence of pulmonary or intrathoracic lymph node metastasis. Mandibular lymph node aspirates were collected in five of the 16 dogs. No aspirate suggested metastasis of the oral EMP, but one node was positive for metastasis of a concurrent oral melanoma. In four cases, the retropharyngeal lymph node was the regional lymph node of interest, but it was not clinically accessible.

Complete blood counts and serum biochemical evaluations were performed on all 16 dogs. Two dogs had elevated globulin concentrations of 6.6 g/dL and 4.4 g/dL (reference range 1.7 to 3.8 g/dL). The dog with the globulin level of 6.6 g/dL also had active concurrent systemic illness (i.e., coccidiomycosis and ehrlichiosis). Alkaline phosphatase concentration was modestly elevated, ranging from 262 to 665 U/L (reference range 24 to 147 U/L) in four of the 16 dogs. No other clinically significant laboratory abnormalities were found. No bone marrow aspirations were performed in any dog. Ten dogs had mild to severe dental tartar, with varying gingivitis noted on their initial physical examinations.

Masses were found on the tongue in four (25%) dogs, originating from the rostral aspect of the mandible in seven (46%) dogs, on the maxilla/hard palate in three (18.75%) dogs, and on the buccal surface of the upper lip in one (6.2%) dog. Two dogs (case nos. 1, 3) had more than one mass, with all masses arising from the same anatomical tissue. In one dog, soft-tissue damage to the maxilla, mandible, and hard palate was extensive, making it impossible to determine the original location of the mass.

No clinical signs related to an oral mass were reported in seven of the 16 dogs. The chief complaint was bleeding from the mouth for as long as 6 weeks in six dogs. Two dogs had a decrease in appetite, and one dog was pruritic at the site of tumor growth.

Accurate measurements of the EMP were available for 12 of the 16 dogs. The median size of the plasmacytoma was 4.12 cm³ (range 0.16 to 90 cm³). No statistical difference in survival was found to be associated with tumors > or <4.0 cm³ at the time of diagnosis.

Three dogs did not receive any treatment for their oral EMP. Seven of 16 dogs had prior cytoreductive surgery performed by the referring veterinarian. All seven had histological evidence of neoplastic cells at the surgical margins, as well as regrowth of the tumor at a median recurrence time of 50 days (range 23 to 690 days). Two of these seven dogs received no further treatment; one received systemic chemotherapy consisting of prednisone and melphalan^b; and four were treated with more aggressive surgery (three partial mandibulectomies, one partial maxillectomy) after referral. The six remaining dogs received aggressive surgical removal at the referral center (two partial mandibulectomies, one partial maxillectomy, and three local mass removals from the tongue) as their first treatment.

Of the 10 dogs that had aggressive surgical removals at a referral center (six as a first surgery and four as a second treatment), four had tumor infiltration at the surgical margins. One of these four underwent full-course radiation treatment; one began treatment with prednisone and melphalan; and one received both prednisone/melphalan therapy and palliative radiation (case no. 14). The fourth dog had no further treatment.

Five dogs were alive at the time of this writing. Median survival time for the 11 dogs that died was 474 days (range 37 to 2906 days) [see Figure]. Eight of the 11 dogs were euthanized for reasons other than uncontrolled tumor growth. One dog that was euthanized for severe immune-mediated thrombocytopenia and hemolytic anemia (i.e., Evan’s syndrome) was found to have infiltrative plasmacytosis in multiple organs (spleen, liver, pancreas, and lung) at necropsy. It was unclear if this plasmacytosis represented metastasis or antigenic stimulation of the autoimmune disease. No other dog had any indication of metastasis at time of death.

The median survival time for the four dogs that had either no treatment (case nos. 3, 5) or incomplete resections without adjuvant therapy (case nos. 8, 10) was 90 days (range 37 to 830 days). Two of these four dogs were euthanized because of uncontrolled tumor proliferation and/or progression of clinical signs secondary to the oral mass.

No statistically significant relationship was seen between overall survival and variables of age, tumor size, or presence of infiltrative surgical margins. Furthermore, none of these factors were significant prognostic indicators for death caused by tumor progression.

Histopathology typically revealed a poorly circumscribed mass composed of closely packed sheets and cords of neoplastic round cells with an occasional faint packeting arrangement, as well as varying degrees of peripheral infiltration. A moderate amount of cytoplasm that was typically amphophilic was seen, although basophilic or eosinophilic cytoplasm was also recorded. Nuclei were round to oval, pleomorphic, and sometimes eccentrically placed. Occasional multinucleated cells were typically scattered throughout the mass. The mitotic index ranged from 10.0 to 60.0 per ten high-power fields (400×). In four cases, a moderate amount of mature lymphocytes was within the periphery of the primary tumor. Amyloid deposits were not found in any specimen. These histopathology findings were consistent with those of previous reports.^2,14

Discussion

The incidence of EMP appears to be increasing within veterinary medicine.⁵ Previous reports suggest that this is largely due to nomenclature changes, so that reticulum cell sarcomas and other poorly differentiated tumors are now histologically categorized as plasmacytomas.^2,3,5 The emerging science of immunohistochemistry has also led to fewer reports of undifferentiated round cell neoplasia,^2,3 although positive immunohistochemical results were not inclusion criteria in the current study.

Monoclonal light-chain reactivity helps differentiate plasma cell tumors from other round cell variants. Specifically, 97% to 100% of all EMP are positive for lambda light chain reactivity.^11,14 Mb-1 (CD79a) and HM57 (anti-Mb-1) are other immunological markers that can be used to further confirm plasmacytoma. Plasma cells stain positive for Mb-1 because of their evolution from B-cell lymphocytes.¹⁵ The Cyclin D1 marker has been linked to multiple myeloma, but not to extramedullary plasma cell tumors. The presence of Cyclin D1 within EMP may suggest an exceptionally aggressive primary tumor or a metastatic lesion of multiple myeloma.¹⁴

Unknown antigenic stimuli—possibly gingivitis and periodontal disease—have been speculated to be responsible for the increasing incidence of EMP.¹⁵ This hypothesis is supported by the identification of mature T cells and dendritic cells within oral EMP, suggesting both neoplastic and inflammatory components.¹⁵ Approximately 62.5% of dogs in the current study had some level of dental disease, which is consistent with the above hypothesis. Data were not sufficient to evaluate the dental health of the other six cases. Tumors from four of the cases had moderate infiltration of mature lymphocytes, which may suggest a T-cell response to antigenic stimulation; however, this was not confirmed via immunohistochemistry.

Findings from the current study indicated that 28.5% of all EMP presented to the authors’ referral center were located within the oral cavity. This is similar to previous studies in which an oral prevalence of 22% to 28% was found.^2,5,11 Plasmacytomas represented 5.2% of all oral tumors, making EMP the sixth most common oral neoplasm in dogs presented to TAMU VMTH.

An unexpected finding of this study was the decreased incidence of oral fibrosarcoma. Osteosarcoma and fibrosarcoma each comprised approximately 9% of all oral tumors in this study. The incidence of oral osteosarcoma is consistent with that in previous reports, but the incidence of fibrosarcomas in the authors’ study is significantly lower than that previously reported (13.3% to 16%).^12,16,17

A 12:1 male:female ratio has been suggested by some EMP studies in people.¹⁸ The authors found no such sex predilection, which is consistent with numerous other veterinary studies.^2,4,5,14 The 8-year median age of dogs enrolled in the authors’ study is consistent with that in previous reports.^2,3–5,14 A possible correlation between male gender and EMP of the tongue has been suggested in the literature, ⁵ but the four EMP of the tongue in the authors’ study were split equally between males and females. Golden retrievers and Yorkshire terriers were identified as the most common breeds in this study (with three cases each), and both have been reported as common breeds for the development of EMP.^3–5,8,14

The presence of multiple masses did not appear to adversely affect survival times, which were 430 and 830 days in the two affected cases of this study. At the time of diagnosis, none of the 16 dogs had evidence of metastatic disease. This supports previous evidence that EMP are locally invasive but rarely metastatic.^2–9,14,15 However, the EMP of the tongue were not completely staged because of the difficulty in accessing the regional retropharyngeal lymph nodes. It is unclear if the one dog that was found to have disseminated plasmacytosis at necropsy developed the oral EMP coincidentally with, or as a metastasis from, an undiagnosed primary tumor. While not previously reported, it is possible that the plasmacytosis was part of the inflammatory response to an aggressive form of Evan’s syndrome. This lack of metastasis continues to suggest that EMP are independent from multiple myeloma.

Regional lymph nodes show evidence of metastasis in 25% of EMP cases in people, and almost half of the human patients have a concurrent hyperglobulinemia.^13,18 By contrast, serum globulin levels were within normal limits in 14 of 16 dogs in the current study. One of the two dogs with elevated globulins had coccidiomycosis and ehrlichiosis, which might account for antigenic stimulation. This dog responded to appropriate treatment (antifungals and doxycycline), but follow-up globulin levels were not obtained. The authors thought that the slightly elevated globulin level (4.4 g/dL) in the second dog was clinically insignificant and most likely attributable to chronic antigenic stimulation. Again, follow-up protein electrophoresis and globulin levels were not obtained.

Many of the dogs in this study underwent mandibulectomies or maxillectomies for boney infiltration of the primary EMP, suggesting local invasion but no systemic metastasis. This is consistent with previous reports.^2,5 Regrowth of the original tumor occurred in all instances in which surgical removal was incomplete and adjuvant therapy was not done.

In two recent studies, a correlation between histopathological variation and survival was not demonstrated.^11,14 In these studies, the histopathology of EMP was divided into five or six different categories. Conclusions reached after both studies indicated that no morphological variations were helpful prognostic indicators. These findings confirm the previous study by Baer et al., who found no correlation between a devised grading scheme and clinical prognosis.³

A large percentage of the EMP in the two above studies originated in the oral cavity.^4,14 Amyloid was reported in 12.5% of these EMP cases,¹⁴ although no amyloid was identified in any of the 16 dogs in the authors’ study. The presence or absence of amyloid within the tumor architecture appears to have no clinical significance in human medicine. ^14,19,20

Survival time of veterinary patients with oral EMP is based primarily on adequate surgical removal.^2–9 In the authors’ study, all dogs euthanized for tumor progression had incomplete surgical removal as the only treatment. Adjuvant treatment options (e.g., full-course radiation or melphalan) appeared to improve survival time in dogs with residual tumor. Radiation therapy is the treatment of choice in people because of the disfiguration associated with aggressive surgical removal.^1,18 The usual high-grade nature of round cell tumors indicates a propensity to respond to radiation treatment.^8,9

In this study, survival time was not significantly associated with age, tumor size, or tumor location. None of these variables was a significant prognostic indicator of death caused by tumor progression. Previous reports indicate histopathology is clinically irrelevant with regard to oral EMP.^11,14

The six most common oral neoplasms (including the epulides and EMP) represent nearly 90% of all oral tumors.^12,16,17 All of these tumors carry a risk for local invasion and rapid regrowth after incomplete surgical excision. This suggests that imaging studies, such as computed tomography and dental radiography, would be useful before surgical removal in order to accurately gauge the level of tumor infiltration regardless of underlying pathology. Such modalities may prevent conservative removal and subsequent regrowth. Furthermore, adjuvant radiation or chemotherapy with prednisone and melphalan may adequately control microscopic disease. Prospective studies are needed to accurately evaluate this hypothesis.

Conclusion

Oral EMP were identified in 16 dogs at TAMU VMTH over a 10-year period. Median survival time was 474 days. These tumors displayed local infiltration but are unlikely to metastasize. Consequently, complete surgical excision appears to be curative. The increased incidence of this tumor type suggests that EMP should be considered as a differential for locally aggressive oral tumors. No obvious correlation exists between EMP in dogs and multiple myeloma, which appears to be a distinct neoplasm.