The Prognostic Value of Lymph Node Metastasis with Grade 2 MCTs in Dogs: 55 Cases (2001–2010)
This study evaluates a series of dogs diagnosed with grade 2 cutaneous mast cell tumors (MCTs) with concurrent lymph node (LN) metastasis. All dogs had surgical excision of the primary tumor. The presence of metastasis was confirmed with either histopathology (n = 35) or cytology (n = 20). There was no significant difference in survival times (STs) between dogs with and without LN metastasis. Median survival time (MST) was not reached at 65.9 mo. LN palpation was a poor predictor of metastasis (sensitivity, .71; specificity, .54). Tumor location was the only prognostic factor for survival in this series of dogs. ST was greater for dogs that had removal of their metastatic LN. This study suggests that in dogs with grade 2 MCTs, outcome may not be affected by the presence of LN metastasis; however, removal of the metastatic LN may prolong survival.
Introduction
Mast cell tumors (MCTs) are the most commonly diagnosed malignant cutaneous tumors in dogs, representing up to 21% of all skin tumors.1–4 Most MCTs fall into the intermediate-grade (grade 2) category of the Patnaik histologic grading scheme, which is currently the most commonly used method of predicting MCT behavior.5 The biologic behavior of grade 2 MCTs can be variable. Although many will be well controlled with surgical excision, a proportion will metastasize. Metastasis most commonly involves the regional draining lymph nodes (LNs); the spleen and liver; and, less commonly, the bone marrow, lungs, kidneys, and heart.6 Previously reported metastatic rates for completely excised grade 2 tumors are estimated at 5–22%.7–9 Metastatic rates for subcutaneous MCTs were lower at 4% in recent literature, but those tumors were not graded.10
Complete excision of cutaneous tumors with adjunctive chemotherapy has been recommended for stage II (modified World Health Organization scheme) disease, but the role of LN metastasis in the prognosis of grade 2 tumors is not clearly defined.1,3 In one study, regional LNs were positive for metastatic spread in 45% of cases of incompletely excised grade 2 MCTs.9 Although only 11 dogs were evaluated, that study suggested that the presence of metastasis was not prognostic for overall survival. In contrast, regional LN metastasis has been shown to be a negative prognostic factor for oral/perioral and muzzle MCT.11,12 Chaffin and Thrall (2002) reported that radiation therapy (RT) of a solitary MCT and the metastatic LN resulted in a median disease-free survival comparable to that in dogs with stage 0 disease. The majority of the tumors in that study were grade 2.13 With grade 3 MCTs, it was recently reported that LN status alone significantly affected survival.14
In addition to histologic grade, literature has focused on finding other prognostic factors in dogs with MCTs.4,12,15–18 Most evaluate survival time (ST), but not factors associated with a higher rate of metastasis. Tumor location, such as MCTs of the muzzle, is the only clinical factor known by the authors associated with a higher rate of metastasis.12 For subcutaneous MCTs, the only reported risk factor for decreased time to metastasis in multivariable analysis was mitotic index (MI), but again, those tumors were ungraded. The rate of metastasis for tumors having a MI > 4 was 53.89 times greater than for a MI of 0.10 Specifically, for grade 2 MCTs, it would be helpful to focus on clinical stage with respect to metastasis in the regional draining LN and its role in prognosis. That information can be used to direct the use of adjunctive therapy and to determine if LN extirpation should be routinely recommended.
The purpose of this study was to retrospectively evaluate dogs with surgically excised grade 2 MCTs both with and without regional metastasis to the draining LN and to determine if nodal metastasis and/or removal of the metastatic LN affected prognosis. This study also evaluated if a palpably enlarged LN was predictive for metastasis and if there were any associations between LN metastasis and either patient factors or tumor location.
Materials and Methods
Tumor, LN, and Patient Evaluation
Medical records from client-owned dogs with a diagnosis of cutaneous grade 2 MCT at Red Bank Veterinary Hospital between January 2001 and October 2010 were evaluated retrospectively. Inclusion criteria were dogs with a histologically confirmed grade 2 MCT, either histologic or cytologic LN evaluation within 1 mo of diagnosis, and either histologic or cytologic confirmation of LN metastasis. Primary tumor classification was determined by one of several board-certified pathologists during the study perioda. Surgical margins were recorded as complete if surgical borders were separated from tumor tissue by ≥ 2 mm.8,9 LN classification (either positive or negative for metastasis) was determined via histopathology by one of several board-certified pathologists and, in some cases, through in-house cytology. For in-house cytologic analysis, a positive LN was confirmed if the predominant cell type (i.e., > 50%) were mast cells, if several aggregates of mast cells were observed, and/or if poorly differentiated mast cells were seen. A separate group of 35 dogs with histologically confirmed grade 2 MCTs without either histologic or cytologic evidence of LN metastasis was identified during the same time period and used as a control group.
Medical records in both groups were reviewed for breed, sex, weight, age at diagnosis, date of diagnosis, location of the primary tumor (i.e., extremity, mucocutaneous, prepucial, other), and regional LN sampled. Results of LN palpation either at the time of its aspiration or extirpation were recorded as either normal or enlarged. Additional tests for staging each patient were variable but included complete blood counts, serum biochemical analyses, abdominal ultrasounds, and thoracic radiographs.
Therapy
Records were reviewed for adjunctive therapy and were categorized as chemotherapy and RT, chemotherapy alone, RT alone, steroids alone, or none. Chemotherapy was recommended at the discretion of the attending oncologist. Protocols were comparable and varied minimally depending on clinician. A 14 wk protocol included IV vinblastineb (2 mg/m2) either alone or with oral lomustinec (60–80 mg/m2) or cyclophosphamided (200 mg/m2). Vinblastine was given once weekly if alone or every third wk if alternating with either lomustine or cyclophosphamide. Dogs treated with chemotherapy were examined regularly for local recurrence and chemotherapy side effects. Dogs that received oral prednisone as a sole agent received a dose of .5–1 mg/kg per os q 12–24 hr.
RT was recommended in cases of incomplete excision where additional surgery was not pursued. If the regional LN was not excised for histopathology, it was included in the radiation field if palpable. The standard RT protocol consisted of 16 daily fractions to a total of 48 gray (3 gray/fraction) at the primary tumor site and regional LN (if present).
Outcome
Medical records were evaluated for all follow-up information, including re-examinations and/or communications with owners. Additional follow-up data regarding either cause of death or euthanasia if deceased and date of last examination (mo and yr) if alive were obtained through both medical records and communication with the owners via telephone. The information obtained from telephone conversations were recorded on standardized questionnaire sheets. Cause of death was defined as either MCT-related (i.e., local disease, distant metastasis) or MCT-unrelated disease.
Statistical Analysis
Quantitative descriptive data for metric variables were presented as either median (range) or mean ± standard deviation for normally distributed data. Tests of association between categorical variables and metric variables used tests of either two medians or two means, as appropriate. Associations between categorical variables were tested using Pearson’s χ2 test of independence or Fisher exact test. Both groups were tested for similarity on the following factors: age, weight, neuter status, palpably enlarged node, removal of node, tumor location, complete excision of primary tumor, and adjunctive therapy. Due to the small number of cases, tumor locations were categorized as extremity, mucocutaneous, and other for statistical analysis. Similarly, adjunctive therapies were categorized as chemotherapy and/or RT, none, and steroid only. Curves for ST were generated using the Kaplan-Meier product-limit method. Dogs were censored in the analysis if they were alive at the time of statistical analysis, lost to follow-up, or died of causes unrelated to their MCT. Variables examined for predictors of ST in the univariate analysis included sex, tumor location, LN status, and removal of LN. Adjunctive therapy was evaluated for effect on ST in each of the groups separately using Kaplan-Meier stratified analysis and again in multivariate analysis. Multivariate analysis used Cox regression, following the backward elimination and assumption assessment guidelines in Hosmer et al. (2008) and briefly described below.19 For the multivariate analysis, tumor location had too few mucocutaneous cases to yield meaningful results and was therefore categorized as either extremity or other. Variables determined to be significant in the univariate analysis as well as LN status and possible confounding variables (e.g., therapy) were entered first. The P values of the Wald statistic of each variable and as the P value of the partial likelihood ratio test for the overall model were used to eliminate variables from the model one at a time and then used to confirm that the deleted variable was not significant. At each step, an assessment of whether the removed variable was a confounder was made. If it was, it was added back into the model. LN status was always the last predictor to be removed. For all univariate analyses, a value of P < .05 was considered significant. For multivariate analysis, a value of P = .1 was used to remove a variable from the model. All analyses were performed using commercial softwaree.
Results
There were 581 dogs with a histopathologic diagnosis of grade 2 MCT during the study period and 55 dogs met the inclusion criteria. Thirty-five dogs were included in the control group. Six dogs had a history of multiple MCTs. All dogs were categorized as having stage 0, I, or II disease.3
There were 47 spayed females, 41 neutered males, and 2 intact males. Median body weight at time of diagnosis was 30.2 kg (range, 3–58.3 kg) and median age was 7.25 yr (range, 1–15 yr). There was no significant difference in the number of spayed females or neutered males (P = .294), age (P = .355), or body weight (P = .869) between dogs with or without LN metastasis.
Physical exam findings for LN palpation were available for all 90 dogs. Fifty-five dogs (61.1%) had an enlarged regional LN. Dogs with palpably enlarged LNs were more likely to have metastasis (P = .017). The sensitivity and specificity of LN enlargement as a predictive value for the spread of MCT disease was .71 (95% confidence interval [CI], .569–.819) and .54 (95% CI, .369–.708), respectively. Metastatic LNs that were palpably enlarged were more likely to be excised (60%) compared with normal size nodes (40%).
Ninety LNs were sampled in 90 dogs. Distribution of LNs included 42 popliteal, 18 prescapular, 13 submandibular, 9 inguinal, 5 axillary, 2 sublumbar, and 1 sternal. The incidence of LN metastasis (55 dogs) was 61.1%. Of the dogs with confirmed nodal metastasis, 35 were diagnosed by histopathology and 20 by cytology. Of the dogs without nodal metastasis, 21 were diagnosed by histopathology and 14 by cytology.
Margin evaluation was available for 89 cases. Complete excision was ultimately achieved in 53 cases (59.6%). Nineteen of those dogs had undergone additional surgery to achieve clean margins after an incomplete excision of the primary tumor. Tumor locations in both groups included 46 on the extremities (51.1%), 9 mucocutaneous junctions (10%; i.e., lip, nasal planum, or perineum), 3 prepucial (3.3%), and 32 on other locations (35.6%), including the head, neck, or trunk. LN metastasis was seen in 50% of tumors on the extremity, 66.6% of tumors at mucocutaneous junctions, and 74.2% of tumors in other locations. All three prepucial tumors had LN metastasis. There was no effect of tumor location on presence of LN metastasis (P = .08).
In dogs without LN metastasis, 1 had both chemotherapy and RT (2.9%), 6 had chemotherapy alone (17.1%), 11 had RT (31.4%), and 3 received only oral steroids (8.6%). In dogs with LN metastasis, 9 had both chemotherapy and RT (16.4%), 35 had chemotherapy alone (63.6%), 2 had RT (3.6%), and 3 received only oral steroids (5.5%). Chemotherapy and/or RT was used in significantly higher percentages in node-positive dogs (83%) than node-negative dogs (51%; P = .001).
Forty-one dogs were deceased at the time of analysis. According to medical records, 13 dogs (31.7%) died due to MCT-related disease. Two died due to systemic involvement (spleen and liver), 10 had local disease (i.e., recurrence, edema, lameness, pain), and 1 had both local and systemic disease. Twenty-seven dogs (65.8%) died due to unrelated disease. The cause of death for one dog was unknown. For 47 dogs that were alive at the time of telephone follow-up, 45 dogs (95.7%) had been examined by a veterinarian within the preceding 12 mo. Two dogs were last examined 13 mo and 22 mo before conclusion of the study. Owners were available by telephone for 87 dogs (96.7%).
The primary measure of outcome was survival. Seventy-seven cases were censored from analysis. Forty-seven dogs were still alive, 27 died due to unrelated causes, and 3 were lost to follow-up. Overall median survival time (MST) was not reached at 65.9 mo (range, 2–65.9 mo). Survival rates (1–4 yr) for LN status are reported in Table 1. There was no significant difference in MST between dogs with and without LN metastasis (P = .133; Figure 1). In both groups, MST was not reached (range, 7.03–65.9 mo for LN-negative dogs and 2–63.9 mo for LN-positive dogs).
LN, lymph node, MCT, mast cell tumor.
Citation: Journal of the American Animal Hospital Association 50, 2; 10.5326/JAAHA-MS-5997
Outcome was affected by removal of the metastatic LN (P = .007; Figure 2, Table 1). For dogs with LN metastasis, MST was 43 mo for those with LN aspiration and cytology (range, 2–47.01 mo), whereas MST was not reached for those with LN removal and histopathology (range, 5.95–63.9 mo).
Citation: Journal of the American Animal Hospital Association 50, 2; 10.5326/JAAHA-MS-5997
When stratified by adjunctive therapy, there was no significant difference in ST between LN-negative and LN-positive dogs (P = .357 with therapy and P = .218 without). When stratified by LN status, there was no significant difference in ST between dogs that did and did not receive adjunctive therapy (P = .847 for LN-positive and P = .681 for LN-negative dogs).
Outcome was affected by location of the primary tumor (P = .015). Nonetheless, MST was not reached for any location (range, 7.03–65.9 mo for extremities, 2–47 mo for mucocutaneous sites, and 5.02–63.9 mo for other sites).
LN status was not a significant predictor of survival in the multivariate model (P = .641). Multivariate analysis of ST did not display confounding interactions between predictor variables (e.g., LN status and adjunctive therapy). The only significant multivariate prognostic factor was tumor location (P = .008; hazard ratio, 5.8, 95% CI = 1.57–21.6 for other locations referenced to extremities).
Discussion
Regional LN status was not a significant predictor of ST in this group of dogs with surgically excised grade 2 MCTs. The lack of prognostic value found in this series of dogs is in contrast to grade 3 MCT where LN status alone significantly affects prognosis.14 Therapy aimed at the primary tumor with grade 2 tumors has been recommended.3,20 Nevertheless, removal of the regional LN may be indicated. Indications for LN extirpation include local morbidity associated with an enlarging node, staging, and/or removal of a potential source for further metastasis. There is little information on whether LN status and/or adjunctive therapy for the metastatic node prolongs survival with grade 2 MCT. With grade 3 MCT, treatment of the LN improved survival only in univariate analysis.14
Dogs with removal of the metastatic LN in this study had a MST significantly greater than dogs with cytologically diagnosed metastasis (i.e., LN was not removed). This may be because excising the metastatic node decreases regional tumor burden and therefore eliminates the LN as a reservoir for mast cell proliferation. Previously, cytoreductive surgery followed by RT of the primary site and metastatic node as well as oral prednisone prolonged survival in dogs with grade 2 tumors.13
Although the most reliable tool for definitive diagnosis is histopathology, an error made in the diagnosis of metastasis could have led to improved outcomes for the dogs included in this study with removal of their LN. This is unlikely, however, because effacement of LN architecture with mast cells in a dog with a confirmed MCT could not be explained with other disease. Excisional biopsy of the LN is the current standard for evaluating LN metastasis.21 Until further studies evaluating the influence of LN extirpation on prognosis of grade 2 MCTs, the study authors recommend excising the regional node at the time of tumor excision despite its size, especially if it has cytologic evidence of metastatic spread. If surgery is not feasible due to either location or other clinical factors, then RT of a metastatic regional node is recommended. A metastatic LN with grade 2 tumors should not preclude curative intent therapy with surgery and/or RT.
Approximately 71% of dogs with LN metastasis had enlarged LNs on physical examination, which was significantly greater than dogs with either a normal or reactive LN. Sixteen LNs that were normal in size were positive for metastatic disease; however, the sensitivity and specificity of LN size as a predictor of metastasis were 71% and 54%, respectively. This is comparable to oral melanoma, where a study evaluating the association between LN size and metastasis in dogs found a sensitivity and specificity of 70% and 51%, respectively.22 This study’s findings indicate that LN palpation remains an important aspect in evaluating patients with MCTs; however, a proportion of positive nodes will be normal in size. In dogs with cutaneous MCTs, the authors recommend evaluating all regional LNs either with cytology or histopathology as part of staging.
The study authors found no significant difference for primary tumor location or other clinical factors between dogs with and without LN metastasis. A previous study evaluating MCTs of the muzzle showed a 63.6% incidence of regional metastasis for grade 2 tumors, which is higher than reported rates at other cutaneous sites.12 There were only four dogs with tumors of the muzzle area in this study population, including three at a mucocutaneous junction. With a larger number of tumors in those areas, the association between tumor location and LN metastasis may have reached significance.
The only prognostic factor in this group of dogs was location of the primary tumor. Because of small sample sizes, tumors of the head, neck, and trunk were grouped together. Consistent with previous literature, tumors of mucocutaneous sites had a worse outcome compared with other locations, but MST was still not reached, which was likely due to the small group of nine cases.23,24 Overall, dogs with tumors of an extremity lived longer than other locations. Tumors at a location other than an extremity had a 5.8 times greater risk of dying.
Unlike previous literature, the diagnosis of LN metastasis was based on histopathology in the majority of LNs in this series of dogs. A smaller proportion of LNs was diagnosed with cytology, which could be considered a limitation. A recent study reported criteria for the cytologic diagnosis of LN metastasis and suggested the clinical usefulness and relevance of cytology for staging.25 Criteria for ”probable metastasis” and ”certain metastasis” yielded similar STs. The study authors used similar criteria for in-house LN aspirates, including finding either effacement with mast cells, several aggregates of mast cells, and/or poorly differentiated mast cells. A study by Langenbach et al. (2001) concluded that cytologic evaluation provides a high sensitivity and specificity for diagnosis of metastasis for solid tumors.21 Fine-needle aspiration of regional LNs, whenever possible, should be performed routinely during initial patient assessment.
This study is limited by its retrospective nature and small sample size. Case-control in a retrospective setting can be challenging; however, the two groups of dogs were not significantly different except in LN status, proportion with enlarged LNs, and proportion receiving adjunctive therapy. The latter is the most important limitation when comparing the two groups. Therapeutic differences could be a confounding factor to the survival results. To address this, the authors statistically compared the use of adjunctive therapy (i.e., with or without) within each LN status group and ST was not affected by treatment. In other words, dogs with or without LN metastasis had no significant differences in survival whether they received additional therapy after surgery. Although those therapies overlap for regional control (LN), combining chemotherapy and RT in the analysis was not ideal because one addresses distant metastasis and the other local disease. Indeed, chemotherapy was used more commonly in the LN-positive group and therefore could have led to similar STs to LN-negative dogs. Due to that limitation, the results should be interpreted with caution.
The incidence of stage II disease (61%) with grade 2 MCTs in this study group was higher than previously reported and is likely due to a biased population including only those with sampled LNs. In one report, 21.6% had nodal metastasis with grade 2 MCTs.25 Difficulties comparing metastatic rates among the literature include populations with different histologic grades, those with undefined grades, and/or small subpopulations with LN cytology.8–10,12,15,25,26 Reason for present bias may be because dogs with lymphadenopathy had extirpation more commonly performed than those with normal peripheral LNs on physical examination. Location of the primary tumor (e.g., trunk) may have an influence as well because a draining LN may not be easily included in the surgical field. Finally, LN extirpation without knowing tumor grade may not always be recommended in clinical cases due to additional anesthetic time and morbidity. All of those factors likely contributed to a smaller population of negative LNs. A final limitation of this study is that the diagnosis of the primary tumor was made by one of several pathologists. It has previously been shown that there is discordance between histologic grades assigned by different pathologists in up to 50% of MCTs.27 More recently in a multi-institutional study with 28 surgical pathologists, concordance was < 64% for cutaneous grade 2 MCTs using the Patnaik grading system.28 Interestingly, that study proposed a two-tier histologic grading system, which provided a 96.8% consistency and an improved prediction of survival. That system was not in use during the study period. Only three dogs in the current study developed distant metastasis, and STs in this study group were much longer than reported for grade 3 tumors, making inclusion of high-grade MCT unlikely.14,29
Conclusion
This study indicates that in dogs with a surgically excised grade 2 MCT, survival may not be affected by the presence of regional LN metastasis. This should be interpreted with caution due to differences in adjunctive therapy between groups. However, dogs with a metastatic LN lived longer if it was removed. Given the differences in STs between dogs with and without removal of the metastatic LN, extirpation of the draining node is recommended in cases where it is clinically feasible.
Kaplan-Meier curve depicting median survival time (MST) for dogs with and without lymph node (LN) metastasis. LN, lymph node.
Kaplan-Meier curve depicting effect of removal of a metastatic LN (n = 35) on MST in all LN-positive dogs (n = 55). LN, lymph node.
Contributor Notes
