A retrospective study of chemotherapeutic effect without wide-margin surgery or radiation therapy in dogs with oral malignant melanoma

Abstract

Background

Effective treatment for canine oral malignant melanoma (e.g., curative-intent surgery) may not be feasible or radiation therapy may be unavailable. However, chemotherapy is usually an option, and more information is needed regarding its use without adequate local treatments.

Objective

Our objective was to investigate the efficacy of chemotherapy in canine oral malignant melanoma without adequate local control, using carboplatin with dose reduction in small-breed dogs and metronomic chemotherapy.

Animals and procedure

Client-owned dogs with histopathologically diagnosed oral malignant melanoma were retrospectively enrolled from 2016 to 2022. The chemotherapy protocol in each case was determined by the attending clinician.

Results

Thirteen dogs were included. The median progression-free interval of all 13 dogs was 42 d (14 to 953 d). The median overall survival time of dogs with chemotherapy as their only systemic treatment was 181 d (50 to 960 d; n = 11). The median dosage of carboplatin was 250 mg/m². Response to treatment and clinical stage were significant prognostic factors.

Conclusion and clinical relevance

As chemotherapy provided a median survival of 6 mo, it could be considered when adequate local control is infeasible. Earlier clinical stages or achievement of at least stable disease during chemotherapy may indicate better survival in dogs.

Introduction

Melanoma, the most common malignant tumor in the canine oral cavity, has a relatively poor prognosis due to its strong propensity for local invasion and distant metastasis (1). Locoregional control is considered the most effective treatment for canine oral malignant melanoma (OMM). Considering the metastatic rate of up to 80%, multimodality management with systemic control is necessary. However, systemic maximum-tolerated- dose (MTD) chemotherapy achieved only a 28 to 37% response rate in the case of macroscopic disease (2,3). There are limited reports of targeted therapy with masitinib or toceranib (4,5) and a need for additional evidence. There has been progress in immunotherapy against malignant melanoma in both humans and dogs with approval of the DNA vaccine Oncept (6). However, Oncept treatment outcomes were inconsistent (7–11). There have been various clinical trials of immunotherapies in canine OMM (12–14), including an immune checkpoint blockade that significantly prolonged survival in those with end-stage disease (13). Despite a better prognosis, the cost may be prohibitive for most owners. Further, as neither a dedicated animal radiation treatment facility nor immunotherapy is widely available in many countries, chemotherapy is still considered a primary or adjuvant treatment in canine OMM.

The most common chemotherapy used in canine OMM is MTD carboplatin, a platinum drug that affects cell cycles by producing DNA crosslinks. In the study that investigated the effectiveness of carboplatin in macroscopic OMM (3), the response rate was 28%. The carboplatin dose used in that research was 300 to 350 mg/m² and was a significant prognostic factor. However, there was a higher incidence of adverse effects in small dogs in that study, indicating that the chemotherapy dose may have not been optimized (3).

A few studies utilized chemotherapy in the adjuvant setting (15,16). Following radiation therapy, MTD carboplatin did not seem to affect survival. Median survival of 9 to 12 mo was reported in studies that combined either low-dose concurrent or adjuvant carboplatin (17–19). In dogs that underwent surgery followed by chemotherapy or other systemic treatments, no survival benefit was detected by Boston et al (15), although Dank et al reported a 7-month progression-free survival that was not different from that for surgery plus radiation therapy and chemotherapy (16). After curative-intent surgery, Tuohy et al reported that adjuvant therapy, including MTD and metronomic chemotherapy (MC) and an immunotherapeutic vaccine, was related to a poorer prognosis (20). However, those studies could not be directly compared due to differences in inclusion criteria, clinical stages, and treatment combinations, plus potential selection bias. In addition to carboplatin, 1 study discussed the concept that combining radiation therapy and temozolomide could achieve significantly longer progression-free survival (21).

Metronomic chemotherapy, which utilizes oral chemotherapeutic drugs at a lower dose but higher frequency regimen, as well as an NSAID, inhibits tumor angiogenesis, modulates host immunity and cancer stem cells, and induces tumor dormancy (22). The most common chemotherapy drug used in canine MC was cyclophosphamide, an alkylating agent. Chlorambucil was an alternative agent that had a similar effect to cyclophosphamide without hemorrhagic cystitis complication (23). In canine soft tissue sarcoma, metronomic cyclophosphamide chemotherapy significantly prolonged disease-free interval (24) and reduced the regulatory T cell amount and tumor microvessel density (25). In a canine OMM xenograft mouse model, low-dose cyclophosphamide and piroxicam suppressed tumor growth and reduced cancer proliferation index, microvessel density, and vascular endothelial growth factor (26). As a subset of the oral tumors, Milevoj et al reported 4 dogs with malignant melanoma receiving metronomic cyclophosphamide in a palliative setting, and 3 dogs had progressive disease in 30 d (27). However, other clinical evidence of MC in canine OMM is limited.

In some areas where radiation therapy and Oncept are unavailable, chemotherapy with or without surgery is still a main treatment for canine OMM. In Taiwan, small- to medium-sized dogs are predominant, necessitating a dose reduction in chemotherapy to reduce severe adverse events. However, efficacy may be compromised.

Objectives of this study were: i) to determine effects of either MTD or MC against canine OMM in the absence of wide-margin surgery or radiation therapy; ii) due to the largely small-breed canine population in Taiwan, to determine effects of lower-dosage carboplatin; and iii) to determine prognostic factors related to chemotherapeutic outcomes in dogs with OMM.

Materials and methods

Inclusion criteria and medical records review

Medical records of dogs with histopathological diagnoses of OMM in the National Taiwan University Veterinary Hospital Animal Cancer Treatment Center (Taipei, Taiwan) from 2016 to 2022 were retrospectively reviewed. Dogs that received chemotherapy as their primary or adjuvant treatment, either MTD or MC, were included. However, dogs that received adequate local control (e.g., wide-margin excision and/or radiation therapy, for primary tumor and regional lymph nodes) were excluded. In addition, dogs that received other systemic treatments (e.g., other chemotherapy, targeted therapy, or immunotherapy) before chemotherapy were also excluded. Dogs that received other treatments after failing chemotherapy were only analyzed for progression-free interval (PFI) and not included in the overall survival analysis.

We obtained the following information: signalment, clinical stage, tumor location, tumor size, bone invasion status, regional lymph node status, diagnostic imaging, histopathological features (lymphovascular invasion and mitotic count), whether primary or recurrent tumor at presentation, whether macroscopic or microscopic disease before chemotherapy, chemotherapy protocol, side effects and best response of chemotherapy, whether local progression or distant metastasis, other treatments after disease progression, and cause of death.

Clinical stage estimation was based on the World Health Organization TNM staging system for dogs with oral melanoma. Stage I: primary tumor size < 2 cm, no metastatic evidence; Stage II: primary tumor size 2 to 4 cm, no metastatic evidence; Stage III: primary tumor size 2 to 4 cm with metastatic regional lymph nodes, or primary tumor size > 4 cm; and Stage IV: if pulmonary or distant metastasis was detected, irrespective of tumor size or regional lymph node status.

Tumor size was recorded based on caliper measurement or imaging. Regional lymph nodes were examined via physical examination and cytology or histopathology. If no lymphadenopathy was noted, an ipsilateral mandibular lymph node was surveyed. Pulmonary metastasis was detected with 3-view thoracic radiographs and/or computed tomography.

Chemotherapy

Each dog’s body weight was recorded and body surface area was calculated. Chemotherapy regimen and follow-up protocol were determined by the attending veterinarian. Tumor response evaluation was based on the response evaluation criteria for solid tumors in dogs consensus document from the Veterinary Cooperative Oncology Group (28). If the best response was stable disease, the duration should be at least 4 wk. Adverse event evaluation was based on the Veterinary Cooperative Oncology Group Common Terminology Criteria for Adverse Events (29). Physical examination, blood work (CBC and chemistry), tumor size measurement, and history were done before chemotherapy and at each revisit.

Statistical analyses

Progression-free interval was defined as the time interval from start of chemotherapy to development of local progression or regional or distant metastasis. Overall survival time (OST) was defined as the time interval from diagnosis to death (for any reason). If an animal was lost to follow-up without evidence of tumor progression, it was censored. Intervals from chemotherapy to local progression and to distant metastasis were recorded. The PFI was calculated in all animals. Those that had chemotherapy as their sole treatment, with or without a previous conservative surgery, were allowed for OST analysis. Kaplan-Meier curves were used for survival depiction and a log-rank test was applied for univariable analysis of factors obtained from medical records. All analyses were done with GraphPad Prism software (RRID:SCR_002798, Version 9.4.0; GraphPad Software, San Diego, California, USA) and P-values < 0.05 were considered statistically significant.

Results

Animal characteristics

Thirteen dogs were enrolled in this retrospective study. The median age was 13 y (range: 8 to 15 y), and the median body weight was 9.1 kg (range: 1.5 to 20 kg). There were 5 miniature schnauzers, 2 Shibas, and 1 each of dachshund, miniature poodle, Maltese, Scottish terrier, beagle, and mixed-breed dog. Four dogs were spayed females, 5 were intact males, and 4 were castrated males. All dogs had histopathological diagnoses of OMM. Six dogs had primary tumors and 7 had tumors that had recurred. Thirty-one percent of dogs had Stage I/II disease (3 with Stage I, 1 with Stage II) and 69% had Stage III/IV disease (7 with Stage III, 2 with Stage IV). The median tumor maximum diameter was 2.8 cm. Six dogs had metastatic tumor cells in their regional lymph nodes (histopathological diagnosis in 2, cytological in 4), 6 dogs had no evidence of lymph node metastasis (histopathological diagnosis in 1, cytological in 5), and lymph nodes were not evaluated in 1 dog. Five dogs had maxillary melanomas, 7 had mandibular tumors, and 1 had tonsil melanoma with lymph node and facial skin metastasis. Three dogs had bone involvement, 5 did not, and the others did not have this information in their medical records. Ten dogs received chemotherapy with macroscopic disease and 3 with microscopic tumor cells. The median interval from diagnosis to start of chemotherapy was 15 d (range: 3 to 134 d).

Chemotherapy protocols and side effects

Eight dogs had MTD chemotherapy, and all received carboplatin (Table 1). The median carboplatin dosage was 250 mg/m², ranging from 250 to 300 mg/m², in a 3-week interval. The median number of carboplatin injections was 2, ranging from 1 to 3. Carboplatin (Kemocarb, Paraplatin) was administrated IV in 20 to 30 min. Maropitant was administered through the same IV route before carboplatin injection, depending on the clinician’s assessment. Five dogs had gastrointestinal adverse events, including altered appetite, vomiting, and diarrhea; all were Grades 1 or 2. Three dogs became lethargic (all were Grade 1). One dog had a Grade 2 elevated ALT but recovered following receiving liver-protecting medications. One dog had a Grade 1 elevated creatinine and self-recovered after 1 wk. One dog had partial remission of the enlarged metastatic lymph node (response of the primary tumor was not evaluated as the mass was located at the caudal soft palate region). Five dogs had stable disease for at least 4 wk and 2 had tumor progression.

Table 1.

Descriptive data for the 13 dogs with oral malignant melanoma enrolled in a chemotherapy trial.

Dog	Stage	Pre-chemotherapy tumor	Mitotic index	Chemotherapy	Response	PFI (d)	OST (d)	Outcome
1	III	Microscopic	30/10 HPF	Carboplatin; 250 mg/m2, 3 cycles	SD	169	345	Pulmonary metastasis
2	II	Macroscopic	8 to 10/10 HPF	Carboplatin; 250 mg/m2, 2 cycles	SD	44	203	Local progression
3	I	Microscopic	0 to 2/HPF	Carboplatin; 250 mg/m2, 3 cycles	PF	PF	Alive	Follow-up: > 960 d without progression
4	III	Macroscopic	0 to 2/HPF	Carboplatin; 250 mg/m2, 1 cycle	PR	42	50	Local progression
5	IV	Macroscopic	NA	Carboplatin; 300 mg/m2, 1 cycle	NA	21	60	NA; died 21 d after chemotherapy
6	IV	Macroscopic	5 to 8/HPF	Carboplatin; 300 mg/m2, 3 cycles	SD	31	103	Pulmonary metastasis progression
7	IV	Macroscopic; metastatic LN and lung	5 to 7/HPF	Carboplatin; 300 mg/m2, 1 cycle	PD	14	181	Pulmonary metastasis progression; anorexia
8	I	Macroscopic	2 to 3/HPF	Metronomic CYC; 10 mg/m2, qd	SD	70	238	Local progression
9	I	Macroscopic	Rare	Metronomic CLB; 6.25 mg/m2, q2d to tiw	SD	700	935	Pneumorrhagia; no tumor progression
10	III	Macroscopic	24/10 HPF	Metronomic CLB; 4.65 mg/m2, q2d	PD	21	50	Local progression
11	III	Macroscopic; metastatic LN	1 to 2/HPF	Metronomic CLB; 4.65 mg/m2, q2d	PD	28	180	Local progression
12	III	Macroscopic	Rare	Carboplatin; 250 and 300 mg/m2	PD	28	116	Local progression
13	III	Microscopic	4/10 HPF	Metronomic CYC; 10 mg/m2, q2d	PF	54	272	Local progression

CLB — Chlorambucil; CYC — Cyclophosphamide; HPF — High-power field; LN — Lymph node; NA — Not assessed; OST — Overall survival time; PD — Progressive disease; PF — Progression free; PFI — Progression-free interval; PR — Partial remission; SD — Stable disease.

Five dogs received MC. Two had cyclophosphamide (Endoxan), 10 mg/m² q24h and 10 mg/m² q2d for each dog, along with piroxicam. The other 3 dogs had chlorambucil (Leukeran): 2 dogs received 4.5 mg/m² q2d, 4.65 mg/m² q2d, along with piroxicam. The 3rd dog received 6.25 mg/m² q2d, then was switched to 6.25 mg/m² on a Mon-Wed-Fri protocol, without NSAIDs, due to chronic kidney disease. Treatment duration ranged from 14 to 928 d. One dog developed Grade 1 thrombocytopenia after long-term use (897 d) of chlorambucil but recovered after protocol adjustment from 6.25 mg/m² q2d to 6.25 mg/m² Mon-Wed-Fri. One dog had Grade 1 neutropenia after 4 wk of chlorambucil and tumor progression was noted; therefore, MC was discontinued and no additional management for neutropenia was needed. Three dogs maintained stable disease for 54, 70, and 928 d, respectively, whereas the other 2 dogs were defined as having progressive disease within 4 wk after treatment initiation. For all 13 dogs, the overall response rate to chemotherapy was 8% (1/13, partial remission); for the MTD group, the response rate to carboplatin was 12.5% (1/8, partial remission).

Survival and prognostic factors

Ten dogs died due to tumor-related causes, including 7 deaths due to local disease progression. One dog received metronomic chlorambucil against the OMM for 928 d and died from cardiopulmonary disease. Another dog received 3 carboplatin injections and remained in a microscopic progression-free status; the owner requested regular follow-up and the dog survived > 960 d. The overall median PFI for all 13 dogs was 42 d (range: 14 to 953 d). The median overall survival for the 11 dogs that had chemotherapy as their only systemic treatment, with or without a previous conservative surgery, was 181 d (range: 50 to 960 d). Time to distant metastasis was not recorded because < 1/2 of the dogs developed pulmonary metastasis, except for the dogs with initial Stage IV.

Clinical stage was significantly related to both PFI and OST (Figure 1 A, PFI: 385 d for dogs with Stages I and II and 28 d for dogs with Stages III and IV, P = 0.023; Figure 1 B, OST: 586.5 d for dogs with Stages I and II and 103 d for dogs with Stages III and IV, P = 0.0252). Oral tumor size was divided into 2 groups, with a median tumor size of 2.8 cm used as a cutoff. Dogs with a tumor < 2.8 cm had longer PFI (169 versus 29.5 d; P = 0.008) and OST (142 versus 350 d; P = 0.037) than dogs with a tumor ≥2.8 cm (Table 2).

Figure 1.

Kaplan-Meier curves of progression-free interval (A; n = 13) and overall survival time (B; n = 11) in dogs with earlier (Stage I/II) and later (Stage III/IV) clinical stages of disease. Median progression-free interval was 385 d in dogs with Stages I and II and was 28 d in dogs with Stages III and IV; P = 0.023. Median overall survival time was 586.5 d in dogs with Stages I and II and was 103 d in dogs with Stages III and IV; P = 0.0252.

Table 2.

Summary of the univariable prognostic factor analysis in 13 dogs with oral malignant melanoma and receiving chemotherapy.

Signalment
Age	Median: 13 y (8 to 15 y)
Body weight	Median: 9.1 kg (1.5 to 20 kg)
Breed	Schnauzer beagle, Scottish terrier, mixed-breed dog) (n = 5), Shiba (n = 2), others (n = 6: dachshund, miniature poodle, Maltese,
Sex	Male castrated (n = 4), female spayed (n = 4), male intact (n = 5)
Tumor information	n	PFI (d)	P-value	n	OST (d)	P-value
Primary or recurred
Primary	6	56	0.17	6	170.5	0.497
Recurred	7	28		5	181
Location
Maxilla	5	42	0.379	5	60	0.678
Mandible	7	44		5	203
Median tumor size (cm)
≥ 2.8	6	29.5	0.008	4	142	0.037
< 2.8	5	169		5	350
Bone involvement
Yes	3	21	0.153	3	181	0.455
No	4	42.5		2	226.5
Regional lymph node
Nonmetastatic	6	62	0.105	5	238	0.058
Metastatic	6	29.5		5	103
Tumor stage
Early (Stage I/II)	4	385	0.023	4	586.5	0.0252
Late (Stage III/IV)	9	28		7	103
Residue disease
Macroscopic	10	29.5	0.071	9	180	0.076
Microscopic	3	169		2	655
Chemotherapy
Chemotherapy type
MTD carboplatin	8	36.5	0.879	7	181	0.964
MC CYC or CLB	5	54		4	209
Response (in 4 wk)
PD	4	24.5	0.0004	2	115.5	0.15
Non-PD	8	62		8	220.5
Side effectsa
Yes	8	37.5	0.345	7	203	0.058
No	5	42		4	55
Survival
Median PFI (n = 13)	42 d
Median OST (n = 11)	181 d
Response rate (n = 13)	8%
Response rate (MTD group, n = 8)	12.5%

CLB — Chlorambucil; CYC — Cyclophosphamide; MC — Metronomic chemotherapy; MTD — Maximum-tolerated-dose; OST — Overall survival time; PD — Progressive disease; PFI — Progression-free interval.

^aSide effects included gastrointestinal and hematological events; all were Grades 1 to 2.

P-values < 0.05 are shown in bold font.

Treatment response was significantly related to PFI. Dogs with progressive disease had a median PFI of 24.5 d, compared to 62 d for those without progression (P = 0.0004; Figure 2). Residual tumor of macroscopic or microscopic disease (PFI: 29.5 versus 169 d; P = 0.07) and lymph node status with or without metastasis (PFI: 29.5 versus 62 d; P = 0.1) did not alter survival. Due to the insufficient case number, no multivariable analysis was done.

Figure 2.

Kaplan-Meier curves of progression-free interval of all 13 dogs with different tumor responses to chemotherapy. Median progression-free interval was 24.5 d in dogs with progressive disease (PD) during chemotherapy, but was 62 d in dogs that achieved at least stable disease; P = 0.0004.

There were no significant relationships between survival and tumor location, bone involvement, primary or non-primary tumor status, receiving MTD or MC, or chemotherapeutic adverse events. Histopathological features including mitotic index and lymphovascular invasion were not analyzed due to inconsistent recording.

Discussion

This retrospective study reported an 8% response rate of chemotherapy against canine OMM. However, the duration was short, with a median PFI of only 42 d. For dogs that received chemotherapy as their only systemic treatment, median PFI and OST were 42 and 181 d, respectively. Survival in the current study was not comparable to that in previous studies (2,3), nor was the 8% response rate. Although the results of those studies could not be directly compared due to differences in patient populations and treatment protocols, the majority of dogs in the current study did not receive curative-intent treatment plans before chemotherapy, either because of an owner’s decision (e.g., declining wide-margin surgery) or some force majeure (e.g., lack of radiation therapy availability in Taiwan since 2020).

Wide-margin excision and radiation therapy can slow progression of canine OMM and prolong patient survival (16,17,19,20,30,31). In the current study, 1 dog had a wide-margin surgery for the primary tumor without removing the regional lymph node; however, both lymph node and pulmonary metastasis were detected 3 mo later. Carboplatin was initiated as soon as the Stage IV disease was diagnosed, but the dog had no response to chemotherapy and died within 3 mo after the Stage IV disease diagnosis.

Regarding the chemotherapeutic effect, Brockley et al (2) reported a 37% response rate (3/8, all partial remission) in dogs with OMM that received carboplatin alone, with median survival of responders significantly longer than that of non-responders (978 versus 147 d). Further, Rassnick et al (3) reported a 28% response rate for carboplatin in dogs with measurable melanomas. In the current study, the overall response rate for chemotherapy was 8%, and 12.5% in the carboplatin group, with neither approaching previous results. One major difference between current and previous studies was the doses of chemotherapy agents used. In our study, the median dose of MTD carboplatin was 250 mg/m² and 12.5 mg/kg. In contrast, induction doses of 300 and 300 to 350 mg/m² were applied in the Brockley et al (2) and Rassnick et al (3) studies, respectively, and a higher dose of 15.1 mg/kg was significantly related to a better response compared to 12.6 mg/kg (on a mg/kg basis) (3). Therefore, the lower response rate in the current study was partially attributed to the lower carboplatin dose.

Regarding MC, 3 dogs received chlorambucil and 2 received cyclophosphamide. The cyclophosphamide dosage used for the 2 dogs was not as standard as previous research suggested, in which 12.5 to 15 mg/m² q24h proved effective in immune regulation and anti-angiogenesis (22,25). However, considering drug compound and owner-related issues, the clinicians prescribed cyclophosphamide at 10 mg/m² q24h and q2d. Due to the low body weight and drug compound problem, the dose of chlorambucil prescribed for the other 3 dogs was lower than suggested (4 mg/m² q24h). Although no study has addressed metronomic chlorambucil dosage and immune regulatory effects, a prospective study using 4 mg/m² q24h reported an 11% response rate, and 47% of the dogs achieved stable disease with only limited toxicities (23). Perhaps both the MTD carboplatin and MC drug dosages were less effective, resulting in a suboptimal antitumor response. In addition to the 8% response rate, when we designated dogs according to the presence or absence of progressive disease during chemotherapy, the clinical benefit rate was 67%. Although the clinical benefit is more commonly used in evaluating targeted therapy, immunotherapy, and MC, instead of MTD chemotherapeutic effect, dogs with progression had a significantly shorter PFI (24.5 versus 62 d).

The impetus for a lower carboplatin dose (median: 250 mg/m², 12.5 mg/kg) was that most animals in our study were small-sized dogs (< 15 kg or even < 10 kg). Although Rassnick et al reported that carboplatin dosage was positively related to the response of melanoma, a higher dose also led to a higher adverse events rate (3). It was recommended that a reduced dose of carboplatin be used in smaller-sized dogs (32). In addition, most owners in Taiwan stated that they did not want side effects to impair their dog’s quality of life. Although a lower carboplatin dose may not generate a high response rate, we must consider a lower dose (e.g., 250 mg/m²) to avoid severe adverse events. There were only 5 gastrointestinal and 4 hematologic adverse events and all were Grades 1 to 2, which seemed slight compared to previous studies (2,3,16). Although not a significant difference, the median OST was extended in dogs that had adverse events. Thus, further dose escalation could be considered to possibly generate a higher response rate and a prolonged survival if the initial treatment was well-tolerated.

In univariable analysis, oral tumor size, clinical stage, and response to chemotherapy were significant prognostic factors predicting progression-free survival. The tumor size and clinical stage were significantly related to OST.

Clinical stage was analyzed for prognosis in several studies but did not have a significant effect in chemotherapy-related studies (2,3,15–18). In the current study, clinical stage had significant relationships with both PFI and OST. Based on a previous study (19), dogs were sub-grouped into Stages I/II and Stages III/IV groups due to low case number. Four of 13 dogs were Stages I/II (31%). With lower disease stage, tumor size was smaller with neither regional nor distant metastasis, enabling higher chemotherapeutic efficiency either by killing tumor cells directly or slowing tumor progression. Among the 4 Stages I/II dogs, 1 had microscopic disease and survived > 900 d after 3 doses of carboplatin injection. The other 3 had macroscopic disease. For the 9 Stages III/IV dogs (69%), median PFI and OST were 28 and 103 d, respectively. A higher percentage of dogs at advanced stages may explain the poorer survival and lower response rate in the current study. In previous studies, tumor size was also a significant prognostic factor but mainly correlated with clinical stage (1). Larger tumor volume may affect chemotherapeutic drug delivery and patient quality of life. Although dogs with oral tumors > 2.8 cm had a poorer prognosis, the significance cannot be defined without multivariable analysis.

Tumor location was not a significant prognostic factor in the current study, perhaps due to a lack of curative-intent surgeries. Conversely, because most of the dogs were smaller dogs, even a 2- to 3-centimeter tumor may have affected the owner’s desire to pursue surgical management and rendered the case non-operable. As the effect of the surgery was not standardized, factors that might affect surgery may have been less consequential.

Receiving MTD or MC also revealed no significant survival difference. To the authors’ knowledge, there has been no specific study of MC in canine oral melanoma. In the research on MC in canine oral tumors (27), 4 dogs were diagnosed with melanoma, and 3 of the 4 achieved short stable disease in 21 and 30 d.

Macroscopic disease was a negative prognostic factor in a study of dogs receiving 6 Gy × 6 radiation therapy via univariable analysis. It remained significantly prognostic in dogs with Stages I/II disease via multivariable analysis (19). In the current study, only 3 dogs had microscopic disease and the median PFI was 169 d, whereas the remaining dogs with macroscopic disease had a median PFI of 29.5 d. The median OST was also longer in those with microscopic disease. Although the case number was low and survival relevance was nonsignificant, perhaps chemotherapy was more efficient with a smaller tumor volume.

Obvious limitations of the current study were its retrospective nature and small population. Although all dogs had examinations of clinical stage, not all cases had a definitive diagnosis; e.g., lymph node evaluation through cytology instead of histopathology and pulmonary metastasis evaluation through chest radiography instead of computed tomography. There was inconsistency between the cytological results of lymph nodes versus histopathology (33). No standard cytological diagnostic criteria exist for metastatic melanoma tumor cells in a lymph node. Perhaps some dogs had an early lymph node metastasis that was not detected by fine-needle aspiration, resulting in underestimation of their clinical stage in the current study. Conversely, due to the low case number of dogs at Stages I/II, additional, larger-scale investigations are warranted to investigate whether chemotherapy could provide a survival benefit for dogs at Stages I/II without radiation therapy or immunotherapy.

In the present study, treatment protocols were not standardized. For dogs that had their primary tumors excised at private veterinary hospitals and were referred after tumor recurrence, surgery approaches were not recorded. Carboplatin administration was relatively conforming, but the metronomic protocols differed, including the use of cyclophosphamide or chlorambucil and the prescribed dosages and intervals. In addition, histopathological factors, including mitotic index and lymphovascular invasion, which were significant prognostic factors according to a recent consensus paper (34), were not consistently recorded and some slides were not available for review. The prognostic cutoff value of mitotic index is 4 per 10 high-power fields (34). In the current study, 6 of 13 dogs had a mitotic index > 4 per 10 high-power fields (PFI: 14 to 169 d), whereas the 2 long-term survivors each had a low mitotic index (Dog 3: 0 to 2/high-power-field; Dog 9: rare).

In conclusion, the current study indicated that, although there were rare long-term survivors, chemotherapy only provided relief and temporary survival benefit. Without adequate local control, dogs with early clinical stage disease and maintaining stable disease or progression-free status during chemotherapy may have better survival. Optimization of MTD carboplatin and MC dosages for small dogs should be further considered. New research on novel or multimodal treatments for OMM may provide better outcomes. Finally, chemotherapy may still be a worthy treatment for managing canine oral melanoma. CVJ