Introduction
Oral neoplasms account for approximately 6% of all cancers in dogs.1 Despite their relatively low prevalence, they have the potential to be severely detrimental to a dog’s quality of life when they go unnoticed at early stages. The first signs of local disease often include oral bleeding, inappetence, or an overt mass effect/facial swelling. Depending on the tumor type and grade, the patient may also present with signs of locoregional or distant metastasis at the time of diagnosis, which vastly affects the median survival time and, inherently, the treatment paradigm.1 For most oral tumors, surgery is an intrinsic part of treatment planning. However, up to 37% (171 of 459) of patients will experience complications, with 38% (96 of 253) of those considered major or catastrophic.2 Further, in advanced-stage disease, surgery alone does not result in durable remission, even with the addition of adjuvant therapy (radiation, chemotherapy, immunotherapy) when indicated.3–17
However, early detection and intervention can result in long-term remission4,8,18,19 and may prevent the need for high-morbidity treatments. In fact, even for oral malignant melanoma (OMM), survival is reported at up to 874 and 818 days for stages I and II disease, respectively, with surgery exclusively.4 Moreover, for tumors that carry a less aggressive biological behavior, such as oral squamous cell carcinoma (OSCC), long-term remission can be achieved in early-stage disease when curative surgery with clean margins is still feasible.8,18,19 By extrapolating oral surgery patient data, veterinarians can be made aware of emerging patterns in canine populations and devise the most appropriate treatment plan.
On the basis of previous epidemiology and histopathological studies of dogs, 37% (151 of 403) of oral cavity biopsies are malignant neoplasms and 34% (138 of 403) are odontogenic.20 Oral malignant melanoma (18.8% [39 of 208] to 39.7% [209 of 526]), OSCC (13.9% [29 of 208] to 15.0% [79 of 526]), fibrosarcoma (FSA; 6.73% [14 of 208] to 11.2% [59 of 526]), and canine acanthomatous ameloblastoma (CAA; 3.85% [8 of 208] to 9.70% [51 of 526]) were consistently among the most common tumors and ranked in this order of frequency.21,22 Yet due to the intrinsic limitations of pathological studies, the true prevalence of disease cannot be inferred and the clinical and environmental risk factors cannot be appropriately evaluated.
In humans, there are not only genetic risks for oral cancer but also well-documented extrinsic risk factors, such as drinking, smoking, severe periodontal disease, and human papillomavirus.23 To date, there has been no canine-specific retrospective study published of oral neoplasia that has analyzed both intrinsic and extrinsic risk factors over multiple decades. The only similar study,24 published 4 years ago, reviewed 2 decades of data and did not account for environmental factors and annual occurrences. Also, since the study utilized a national database, the different diagnostic procedures at various veterinary institutions likely created confounding variables. Understanding incidence data and underlying risk factors associated with oral neoplasms in dogs is crucial for making timely diagnoses and effectively managing the condition—ideally by preventative strategies. Identifying underlying causes will allow veterinarians to employ risk stratification, utilize targeted screening and preventative protocols, and guide pet owners in making informed decisions about their dog’s healthcare. We hypothesized that increasing age, distinct pure breeds, poor air quality, and severe periodontal disease would represent risk factors for oral neoplasia in dogs.
Methods
Data collection: oral tumor cases
The University of California-Davis Veterinary Medical Teaching Hospital (VMTH) patient data were bulk extracted from the Veterinary Medical and Administrative Computing System (VMACS) from January 1985 to April 2024. Specific permutations of keywords for the tumor location and tumor type were selected to return the desired output. The search criteria for locations were mandible and maxilla (rostral, caudal, and rostral/caudal), palate, lingual/tongue, buccal/cheek, lip, and tonsil. Tumor type keywords were SCC/squamous, OMM/melanoma, CAA/ameloblastoma, MLO/multilobar, POF/fibroma, FSA/fibrosarcoma, OSA/osteosarcoma, sarcoma, carcinoma, and epulis. When a tumor pathology had an abbreviation, a pipe symbol (|) was included between the 2 words to denote an OR command. An asterisk (*) was used as a wildcard command to account for multiple spellings of a word. Individual patient files were then checked manually to verify the tumor type and location. Primary extraoral tumors (nasal, retrobulbar, and cutaneous head) were excluded, regardless of whether there was a secondary extension into the oral cavity. Oral masses without a diagnosis were similarly excluded.
Collected data from the electronic medical record included patient ID, breed, sex, weight, date of birth/age, date of intake, tumor diagnosis, and location. Pit Bull Terriers and Staffordshire Bull Terriers were grouped together as pit bull–type dogs. The patient’s county location was stratified with the use of the zip code of the patient’s home address. A local database of median air quality index (AQI) data available online through the Environmental Protection Agency25 was categorized on the basis of California zip codes and the years (1980 to 2023). According to the Environmental Protection Agency, the AQI is an estimation of air pollution based on 5 major particles: O3, PM2.5/PM10, CO, SO2, and NO2. An AQI below 100 is considered good to moderate, an AQI between 100 and 200 is considered unhealthy, and an AQI above 200 is considered very unhealthy or hazardous.25
Data collection: general patient population
General patient population data for age, sex, and breed were collated by collecting data on all dogs seen at the VMTH between January 1, 1984, and April 1, 2024, utilizing the PatientsStats function on VMACS. Because PatientsStats did not output the home addresses, another general patient population of patients that presented to the hospital during the same time frame was compiled from all dogs that received bloodwork. The zip codes from these patients were subsequently extracted for AQI analysis.
Periodontal disease evaluation
For periodontal disease status, a subset of CT scans from VMACS were randomly evaluated for periodontal disease status by board-certified dentists and oral surgeons (SG and MSR). Prior to scan selection and randomization, all patients that had a head CT were collated and then stratified into small dogs (< 15 kg), medium dogs (15 to 25 kg), and large dogs (> 25 kg) to ensure that each group was represented equally in the analysis. Power analysis was performed to select the minimum number of CT scans required for review to ensure we could accurately calculate the OR between the cancer and periodontal disease exposure. As no canine-specific literature exists, the OR (3.2 to 3.3) was based on human research of the association between periodontal disease and oral cancer.26,27 For power analysis, we used the incidence of oral cancer detected at the University of California-Davis (rounded to 0.1) and the reported approximately 80% periodontal prevalence among companion dogs.28 A total of 171 patients (57/weight group) in each group (oral cancer vs no oral cancer) resulted in 90% power to detect a relative risk (RR) of 3 and 80% power to detect an RR of 2.5.
As dental charts were not available for review, CT scans were categorized as no alveolar bone loss and then periodontal disease stages 2 to 429 based on the most severe tooth in the oral cavity. Matched (total number of patients and patient weight) scans from patients with brain tumors and no oral tumors were also reviewed and scored for comparison. After initial evaluation, we also corrected for alveolar bone loss confined to the incisors only, as it has been shown that CT scans may overestimate incisor bone loss30 compared to dental radiographs. Thus, when the alveolar bone loss was only documented in the incisors and normal alveolar bone structure was detected in all other regions, we classified these cases as no alveolar bone loss and termed these scans after incisor correction.
Data processing
Several Google Sheets (Google LLC) functions were used to aid in data processing. The CONCAT function was used to combine the patient histories and diagnoses into a single cell, while the REGEXMAX function was used to confirm the presence of the keywords in that string. The function, VLOOKUP, was used to convert zip codes into California counties and search the local AQI database on the basis of the year of the patient intake and the county of the owner’s home address. IFNA and IFERROR were used to skip zip codes not found in the AQI database. The YEAR function was used to retrieve the year of the first visit, and the DAYS function was used to determine the number of days between the date of birth and intake at the VMTH. IF statements were used to format data for charts when comparing multiple subcategories. RAND() and sort sheet_by_column_(A–Z) were used to randomize rows for periodontal disease assessment. Conditional formatting was used to identify blank cells, irregularly high values, and misspellings of words. An artificial Google Sheets extension was also used to rapidly review breed abbreviations and return the breeds’ full names.
A high-speed computer (Quadro RTX 3000; NVIDIA) was required to analyze the general patient population. To compensate for puppies not yet at skeletal/dental maturity, patients younger than 2 years were excluded from weight evaluation and periodontal disease review in the general patient population. The recommended minimum age for screening was calculated by evaluating the 95% CI and recording the lower bounds of the CI.
Statistical analysis
Logistic regression was used to assess the marginal risk for developing oral cancer separately for each independent variable: breed, age, sex, intact status, and AQI. Age and AQI were treated as continuous independent variables with a possible nonlinear relationship with cancer development. Breeds with fewer than 30 oral tumor patients were excluded from the breed data analysis to avoid sampling error. Additionally, due to the large number of breeds available, only the top 20 breeds (by sample size) were considered for modeling. Case status was identified by type/pathology, allowing for different “case” outcomes for the logistic regression. The top 3 pathologies (by sample size) were analyzed further as separate cases, leading to the 4 following possible case types: OMM, OSCC, CAA, and overall (any pathology).
Logistic regression results were interpreted as RR for cancer development using the equivalence of the OR under either retrospective or prospective sampling, and the low prevalence of cancer risk in the population implies that the OR closely approximates the RR. The association between tumor location and pathology was assessed with the use of likelihood ratio tests under the null hypothesis of each location being equally likely for each pathology. Multiple comparisons were adjusted with the Holm procedure. A Fisher exact test was used to assess the effects of periodontal disease on oral tumor risk. All statistical analyses were performed in R, version 2024 (R Foundation for Statistical Computing). Statistical significance was defined as P < .05.
Results
Patients
An initial search in the VMACS resulted in 17,810 instances of oral disease or neoplasia between 1985 and 2024 from a total of 256,802 dogs. After accounting for duplicate patient IDs, 2,693 patients were identified as potential candidates. After a manual review of the medical records, 1,011 patients were excluded for primary extraoral (eg, primarily nasal) tumors with extension into the oral cavity. This left 1,682 oral tumor patients for analysis. Of these 1,682 patients, there were 1,710 tumors in total since a subset of patients presented with 2 separate tumor pathologies. For the 28 patients with 2 oral neoplasms, each tumor was treated separately during the pathology analysis.
The most common tumor types were OMM (29% [492 of 1,710]), OSCC (19% [316 of 1,710]), CAA (11% [184 of 1,710]), FSA (9% [150 of 1,710]), and osteosarcoma (OSA; 7% [117 of 1,710]). The mean incidence of oral tumors presenting to the veterinary teaching hospital over the past 38 years was 4.59 cases/1,000 patients (Figure 1). Oral tumor incidence increased each year by approximately 0.141/1,000 patients seen until 2019, then declined until 2022, when it started to increase again. The average annual increase of the incidences for OMM, OSCC, and CAA were 0.0417, 0.0236, and 0.0204/1,000 patients, respectively.
The annual incidence of oral neoplasia is shown as the number of oral tumor cases per 1,000 patients between the years of 1985 and 2023 for tumors overall and the 3 most common oral pathologies. CAA = Canine acanthomatous ameloblastoma. OMM = Oral malignant melanoma. OSCC = Oral squamous cell carcinoma.
Citation: Journal of the American Veterinary Medical Association 263, 3; 10.2460/javma.24.09.0594
Patient signalment
One hundred eight unique breeds were identified with oral tumors during the study period. The 8 most common breeds with oral tumors included Labrador Retrievers, Golden Retrievers, German Shepherd Dogs, Australian Shepherds, pit bull–type dogs, Rottweilers, Poodles, and Cocker Spaniels (Table 1). Golden Retrievers, Labrador Retrievers, and Australian Shepherds were significantly at risk for oral cancer development compared to the general patient population (Supplementary Figure S1; Supplementary Table S1). An omnibus test also revealed that the 20 most common breeds for oral tumors (the 8 aforementioned breeds, Australian Cattle Dogs, Beagles, Border Collies, Boxers, Chihuahuas, Chow Chows, Doberman Pinschers, Great Danes, Schnauzers, Shetland Sheepdogs, and Shih Tzus) were overrepresented (P < .001) compared to the general patient population.
The 8 most common pure dog breeds and mixed/other breeds diagnosed with oral tumors from 1985 to 2024 are listed with their proportion and collapsed ORs. The 2 most common tumor pathologies are noted for each breed.
| Breed | Cases (n = 1,682) | General population (n = 253,438) | OR | Common pathologies |
|---|---|---|---|---|
| Labrador Retriever | 14.1% (n = 238) | 2.22% (n = 5,627) | 7.26 | OMM: 26% (n = 61) |
| OSCC: 20% (n = 46) | ||||
| Golden Retriever | 11.2% (n = 118) | 1.00% (n = 2,535) | 12.5 | OMM: 28% (n = 161) |
| FSA: 23% (n = 44) | ||||
| German Shepherd Dog | 4.02% (n = 68) | 1.18% (n = 2,979) | 3.54 | OSCC: 28% (n = 20) |
| OMM: 15% (n = 10) | ||||
| Australian Shepherd | 2.73% (n = 46) | 0.445% (n = 1,127) | 6.29 | OSCC: 24% (n = 11) |
| CAA: 22% (n = 10) | ||||
| Pit bull–type dogs | 2.68% (n = 45) | 0.995% (n = 2,521) | 2.74 | CAA: 24% (n = 11) |
| OSA: 18% (n = 8) | ||||
| Rottweiler | 2.68% (n = 45) | 0.764% (n = 1,936) | 3.57 | OMM: 42% (n = 19) |
| OSA: 24% (n = 11) | ||||
| Poodle | 2.32% (n = 39) | 0.752% (n = 1,907) | 3.13 | OMM: 41% (n = 16) |
| OSCC: 23% (n = 9) | ||||
| Cocker Spaniel | 2.02% (n = 34) | 0.520% (n = 1,317) | 3.95 | OMM: 38% (n = 13) |
| OSCC: 17% (n = 6) | ||||
| Mix | 14.2% (n = 239) | 81.3% (n = 206,083) | 0.038 | OMM: 44% (n = 104) |
| OSCC: 20% (n = 48) | ||||
| Other | 44.0% (n = 740) | 11.5% (n = 29,088) | 6.06 | OMM: 28% (n = 205) |
| OSCC: 21% (n = 154) |
CAA = Canine acanthomatous ameloblastoma. FSA = Fibrosarcoma. OMM = Oral malignant melanoma. OSA = Osteosarcoma. OSCC = Oral squamous cell carcinoma.
Oral malignant melanoma was most commonly diagnosed in mixed breeds (44% [104 of 239]), Rottweilers (42% [19 of 45]), Poodles (41% [16 of 39]), Cocker Spaniels (38% [13 of 34]), Golden Retrievers (28% [53 of 188]), and Labrador Retrievers (26% [60 of 238]). Squamous cell carcinoma was most commonly diagnosed in German Shepherd Dogs (28% [20 of 68]), Poodles (23% [9 of 39]), Labrador Retrievers (20% [47 of 238]), and mixed-breed dogs (20% [48 of 239]). Canine acanthomatous ameloblastoma was commonly diagnosed in pit bull–type dogs (24% [11 of 45]). Fibrosarcoma and OSA were most common in Golden Retrievers (23% [44 of 188]) and Rottweilers (24% [11 of 45]), respectively. No statistical analysis was performed to evaluate the risk of each breed for specific tumor pathologies.
Patients with oral neoplasms were significantly (P < .001) older than the average hospital population, with the median age of oral tumor patients being 9.66 years (IQR, 7.32 to 11.6 years; n = 1,682) versus 5 years (IQR, 2 to 9 years; 241,945) for the general patient population (Figure 2). Age had an OR of 1.13 (P < .001), indicating the risk of oral tumors increased by 13% for each increasing year of age up until 13.6 years.
The distribution of ages for all oral tumor cases and the general patient population from 1985 to 2024 is reported. The y-axis represents the proportion of total oral tumor cases and general population that was found in each age range. VMTH = University of California-Davis Veterinary Medical Teaching Hospital.
Citation: Journal of the American Veterinary Medical Association 263, 3; 10.2460/javma.24.09.0594
There was minimal variation in age by tumor type, with the median age being 10.8 years (IQR, 8.73 to 12.7 years), 10.0 years (IQR, 7.58 to 11.7 years), 9.05 years (IQR, 6.94 to 10.6 years), 8.44 years (IQR, 5.87 to 10.5 years), and 9.78 years (IQR, 7.35 to 11.0 years) for OMM, OSCC, CAA, FSA, and OSA, respectively. The OR of age for OMM was 1.30 (P < .001), for OSCC was 1.14 (P < .001), and for CAA was 1.35 (P < .001).
Heavier dogs were significantly (P < .001) more likely to be diagnosed with oral tumors compared to the general patient population. However, the median weight of the oral tumor group was 27.7 kg (IQR, 15.1 to 35.6 kg; n = 1,626), while the general patient population was 22.7 kg (IQR, 9 to 33 kg; 90,491), so this finding is not likely to be clinically impactful.
There was a nearly equal distribution of male (51.8% [871 of 1,710]) and female (48.2% [811 of 1,710]) patients presenting with oral tumors. In particular, 45.0% (770 of 1,710) were spayed females, 2.40% (41 of 1,710) were intact females, 42.4% (725 of 1,710) were neutered males, and 8.54% were (146 of 1,710) intact males. Moreover, the effect of intact status, regardless of sex, was significant (P < .001), with spaying/neutering increasing the risk of cancer by 259%. Specifically, spayed dogs had the greatest OR (5.51; P < .001), followed by neutered dogs (OR, 2.34; P < .001).
This pattern was conserved among some, but not all, of the common tumor types. Specifically, being female was significantly protective for OMM (OR, 0.484; P < .001), OSCC (OR, 0.659; P < .05), and CAA (OR, 0.517; P < .05); these findings ignored intact status. For OMM, the OR of cancer for spayed females compared to intact females was 6.78 (P < .001) and for neutered males compared to intact males was 2.35 (P < .001). Collectively, the OR developing OMM in a sterilized patient was 3.99 (P < .001). For OSCC, the overall OR of developing cancer in a sterilized patient was 3.02 (P < .001); this was more significant regarding females (OR spayed vs intact female, 5.94; P < .001) than males (OR neutered vs intact male, 1.54; P < .05). For CAA, the OR of developing cancer in a sterilized patient was 4.61 (P < .001). This was conserved in both female (OR of spayed to intact female, 7.82; P < .001) and male (OR neutered to intact male, 2.72; P < .001) patients.
Tumor localization
There was a pattern between tumor location and tumor type (Figure 3). Oral malignant melanoma was most commonly diagnosed on the lips/commissures (25.6% [126 of 492]) and was least common on the tonsils (0.8% [4 of 492]). Oral squamous cell carcinoma was most commonly diagnosed on the rostral mandible (17.9% [60 of 316]) and was least common on the lip/commissures (7.3% [23 of 316]). Canine acanthomatous ameloblastoma was most commonly diagnosed on the rostral mandible (45.4% [88 of 194]) and was absent on the tonsils, cheek, lips, and tongue. Each oral cavity region had a unique profile of pathology incidences (Supplementary Figure S1); OMM, OSCC, CAA, FSA, OSA, and plasmacytoma were found to be significantly overrepresented in at least 1 area.
The oral location distribution of the 3 most common oral neoplasms (CAA, OMM, OSCC) is displayed. The percentages represent the percent of a given tumor type diagnosed in the defined location. When a certain tumor type was overrepresented (P < .05) in a location compared to the other tumor types, it is marked with an asterisk. N/A = Not applicable.
Citation: Journal of the American Veterinary Medical Association 263, 3; 10.2460/javma.24.09.0594
Extrinsic risk factors
In evaluating risk factors for disease, no significant correlation was found between the home AQI (Figure 4) and overall oral tumor occurrence (P = .299). This was conserved for OMM (P = .463), OSCC (P = .516), and CAA (P = .162). The median AQI for all oral tumor patients was 50 (IQR, 41 to 77; n = 1,578). The AQI for the general patient population was 49 (IQR, 41 to 77; n = 100,271). Although certain zip codes had a significantly higher relative incidence of oral tumors (P < .001; χ = 1,863; df = 1,632), a choropleth map for California zip codes did not reveal any apparent geographical patterns for either OSCC or all oral tumor patients other than their close proximity to Davis.
A histogram of home air quality index (AQI; 1985 to 2024) for the oral tumor cases and the general patient population showing the proportion of patients that fell in each AQI range. The AQI increments are 10 units.
Citation: Journal of the American Veterinary Medical Association 263, 3; 10.2460/javma.24.09.0594
Fifty-seven percent (57.3% [98 of 171]) of patients with oral neoplasia and 70.2% (120 of 171) of patients in the general patient population had signs of alveolar bone loss on CT scans. Following incisor correction, 42.7% (73 of 171) of patients with oral tumors and 41.5% (71 of 171) of brain tumor patients had evidence of periodontal diseases (Table 2). The OR for periodontal disease was 0.571 before incisor correction and 1.05 after incisor correction, and the overall difference between the oral tumor cases and the general patient population was insignificant (P > .05) regardless of incisor correction whether periodontal disease was treated as binary (alveolar bone loss versus periodontal disease stages II to IV).
Periodontal disease prevalence in dogs with and without oral tumors was evaluated on CT by board-certified dentists (SG and MSR). After initial evaluation, we corrected for alveolar bone loss confined to the incisors only, as it has been shown that CT scans may overestimate incisor bone loss30 compared to dental radiographs. Thus, when the alveolar bone loss was only documented in the incisors and normal alveolar bone structure was detected in all other regions, we classified these cases as no alveolar bone loss and termed these scans after incisor correction. The general patient population consisted of dogs imaged for brain tumors.
| Before incisor correction | After incisor correction | |||||
|---|---|---|---|---|---|---|
| S | M | L | S | M | L | |
| Stage | Oral tumor | |||||
| 0–I | 17.5% (10/57) | 47.4% (27/57) | 63.2% (36/57)*** | 21.1% (12/57)* | 78.9% (45/57) | 71.9% (41/57) |
| II | 24.6% (14/57) | 42.1% (24/57) | 35.1% (20/57)** | 24.6% (14/57) | 17.5% (10/57) | 26.3% (15/57) |
| III | 22.8% (13/57) | 10.5% (6/57) | 0.00% (0/57)** | 19.3% (11/57) | 3.51% (2/57) | 0.00% (0/57) |
| IV | 35.1% (20/57) | 0.00% (0/57) | 1.75% (1/57) | 35.1% (20/57) | 0.00% (0/57) | 1.75% (1/57) |
| Stage | General population | |||||
| 0–I | 14.0% (8/57) | 50.9% (29/57) | 24.6% (14/57) | 42.1% (24/57) | 78.9% (45/57) | 54.4% (31/57) |
| II | 29.8% (17/57) | 40.4% (23/57) | 61.4% (35/57) | 12.3% (7/57) | 15.8% (9/57) | 38.6% (22/57) |
| III | 31.6% (18/57) | 8.77% (5/57) | 14.0% (8/57) | 24.6% (14/57) | 5.26% (3/57) | 7.02% (4/57) |
| IV | 24.6% (14/57) | 0.00% (0/57) | 0.00% (0/57) | 21.1% (12/57) | 0.00% (0/57) | 0.00% (0/57) |
L = Large (> 25 kg). M = Medium (15 to 25 kg). S = Small (< 15 kg).
*.01 < P ≤ .05.
**.001 < P ≤ .01.
***P ≤ .001.
When periodontal disease stages were treated as continuous, small dogs were significantly (P < .05) more likely to be diagnosed with stage II and stage IV disease after incisor correction in both the oral tumor cases and the general patient population. Of note, the distribution of no alveolar bone loss diseases after incisor correction was uneven between the small and large classes of the oral tumor cases and the general patient population (P < .05). Yet the distribution of periodontal disease stages for small dogs was also uneven (P < .05): there were more small dogs with advanced periodontal disease and oral tumors compared to small dogs with advanced periodontal disease and brain tumors. The association between periodontal disease and oral tumors was significant in large dogs before incisor correction (P < .001) and small dogs after incisor correction (P < .05).
Discussion
This was the largest and most comprehensive study to date on the prevalence and risk factors for oral neoplasia at a single institution. This study found that the incidence of oral tumors was 4.59 cases/1,000 (4.73 before COVID-19) and has been increasing by an average of 0.141 cases/1,000 (0.219 before COVID-19) annually for the past 4 decades. There were distinct signalment features that put dogs at an increased risk of oral tumors. Conversely, AQI and geographical location did not noticeably impact oral tumor predisposition. Periodontal disease may play a role, but further evaluation is required.
This study revealed an incidence of 4.59 cases/1,000 dogs, which was higher than a similar study24 at another teaching hospital, which revealed an annual incidence of 3.4 cases/1,000 dogs. This was likely due to differences in hospital populations and clientele that presented for treatment. However, it may also represent a geographical difference in predisposition between local and national data. The most common oral tumor type presented at the veterinary hospital alternated every few years between OMM and OSCC until 2002, at which point OMM consistently remained the most frequently diagnosed pathology. During 1985 to 1986, 1987 to 1993, and 1995 to 1996, there was not a single diagnosis of CAA. However, with terminology changes for odontogenic tumors,31 veterinarians at the teaching hospital may not have adopted this term and instead opted to use acanthomatous epulis. To the authors’ knowledge, there is no previous publication that discussed changes in annual occurrence for oral neoplasms, nor has there been a study that analyzed oral pathological incidences for a single population.
It was found that the median age for all tumors was 9.65 years (IQR, 7.29 to 11.6 years), which was close for each specific tumor type. This is unsurprising given the body of literature that reports oral tumors in geriatric populations for OMM,20,32,33 OSCC,20,32,33 CAA,20,33 FSA,20,32,33 and osteosarcoma.20,33 Of note, OSCC has also been reported in the juvenile population19; within our dataset, 3.48% (11 of 316) of OSCC were juvenile cases. Collectively, juvenile OSCC appears to be rare. On the basis of this dataset, the optimal age to begin routine dedicated screening for oral tumors is 9 years; screening earlier may impose financial burdens or uncover nonsignificant incidental findings, and screenings later may lead to considerably missed diagnoses.
Labrador Retrievers, Golden Retrievers, and Australian Shepherds were overrepresented in this study and have also been shown in other studies to be predisposed to cancer in oral22 and nonoral sites.33 Since decades of inbreeding have decreased their genetic fitness, certain breeds likely have a higher genetic risk of oral tumors.34,35 Consequently, mixed breeds usually have higher genetic diversities and will be better equipped to repair critical DNA mutations before they can facilitate tumorigenesis.36 Our study reinforced previous findings, and unlike many historical studies, it compared the proportion of breeds with tumors to the patient population, ensuring that overrepresented breeds do not solely represent regional breed preferences.
The higher proportion of oral tumors in sterilized dogs was surprising. This may be due to the fact that these dogs are older (and thus at higher risk) rather than indicating that sterilization (or the lack of hormones) increases the risk of developing oral tumors. Yet the effect of hormones cannot be totally excluded, as premature neutering has been linked to an increased risk of osteosarcoma.37 Further work examining the relationship between hormones and cancer development is warranted. This is especially true, as there was a distinct difference between female and male ORs and the effect of sterilization in each of these patient cohorts. Oral malignant melanoma has historically been shown to be predisposed in females38; thus, this may indicate an at-risk population.
Additionally, certain tumor pathologies were more likely to originate in specific regions. Compared to other tumor histologies, OMM and OSCC were overrepresented in all locations, demonstrating their higher prevalence in dogs. Canine acanthomatous ameloblastoma was disproportionately diagnosed on the rostral mandible, rostral maxilla, caudal mandible, and rostral to caudal mandible. Other studies39,40 also found that CAA was more common on the mandible (70% [48 of 68] to 72% [191 of 263]) compared to the maxilla, and it was diagnosed more frequently on the rostral mandible (41% [27 of 68] to 51% [135 of 263]) as opposed to the caudal mandible. Our results do not directly match historical data, as we compared the risk of different tumor histologies for each location instead of comparing the distribution of location within a particular tumor type.
The large overlap of AQI distributions between the oral tumor cases and the general patient population demonstrates that environmental factors like air quality and soil contaminants are unlikely to increase the risk of developing oral neoplasms in dogs. This is surprising given the literature linking poor air quality to oral cancer occurrence in humans. Higher concentrations of PM2.5 have been correlated (P < .05) with oral cancer in 366,597 nonsmoking Taiwanese men older than 40 years.41 Moreover, based on a clinical study42 of 43 Chinese children, the proinflammatory response from PM2.5 air pollutants predisposes people to oral cancer at a molecular level. Similarly, a lack of clustering on a choropleth map shows that geographical location does not appear to significantly impact the rates of oral tumors within the state of California. However, there is an inherent bias, as patients from certain regions may not present to a tertiary care facility for oral cancer treatment due to financial constraints.
It was identified that periodontal disease was not a clear risk factor for oral cancer in dogs. This poor association was surprising given the extensive research connecting periodontal disease to an increased oral cancer risk in humans. However, when looking at each size group separately and evaluating images before and after incisor correction, differences between the oral tumor and the general patient population were noted. Specifically, the number of large dogs before incisor correction (stages 0 to III) and small dogs after incisor correction (stage 0 to I) from the oral tumor group was significantly (P < .05) higher than the general patient population. This may represent a potential risk factor that may be extrapolated in a larger patient population. As our power analysis was based on human literature, there is a chance that significant differences were missed. Further, small breeds with brain tumors were primarily brachycephalic, so they may have falsely elevated the degree of periodontal disease.
There were distinct limitations to this research, such as its retrospective nature and its focus on a single hospital population. Although the searches performed in the VMACS were extensive, they may not have captured all of the oral tumor patients seen. Clinicians occasionally misspell diagnoses when charting, so several patients may have been overlooked. Also, advances in veterinary diagnostics throughout the past few decades and the reclassification of tumor nomenclature (ie, CAA) further complicated histopathological analysis. We also specifically searched for common oral types and excluded rare tumors from our analysis. It is also worth noting that this data may not represent the general canine population because low-income owners are less likely to pursue specialty referrals. Additionally, this study did not analyze home-environmental factors such as diets and owner smoking, as has been done for feline OSCC,43–45 since there was no feasible way to consistently discern this information. Furthermore, even though the AQI formula accounted for several air pollutants, there may be other contaminants in the drinking water that contributed to oral tumor development.
Despite the limitations, this was the largest and most extensive study to date that has evaluated canine oral tumor clinical prevalence and quantified signalment, location, and environmental factors. The recommended age to begin screening for oral tumors is 9.14 years; however, if a patient is younger and already receiving a dental cleaning, it is worthwhile to conduct a thorough oral examination for lesions, ensuring to include the sublingual and tonsil region, which is often overlooked. Labrador Retrievers, Golden Retrievers, and Australian Shepherds have the highest risk for oral tumors, so these dogs should be screened biannually, regardless of age. Additionally, AQI, geographical location, and periodontal status did not clearly impact tumor occurrence and are not modifiable risk factors.
Knowing the risk factors for oral neoplasms will facilitate enhanced prognostic insight and more efficient resource allocation; this will improve clinicians’ understanding of the distribution of oral cancer within their patients and allow them to devote more time and resources to screening at-risk dogs. By diagnosing oral tumors at early stages, they can help prevent high-morbidity interventions and improve patient outcomes.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
The authors thank Christ Bradt, DVM, MS, for answering questions about the Veterinary Medical and Administrative Computing System; Satoshi Shinkawa, BS, for providing data analysis advice; and Chrisoula Skouritakis, PhD, for rendering Figure 3. The authors also acknowledge the University of California-Davis veterinary team for diagnosing these patients and the Marcu Laboratory for providing access to their high-speed computers.
Disclosures
GPTforWork, an AI extension, was used to convert breed abbreviations to full breed names. However, its outputs were manually verified. Grammarly was also used sparingly to correct sentence structure when drafting this manuscript. No other AI-assisted technologies were utilized for data collection or manuscript writing.
Funding
The authors have nothing to disclose.
References
- 2.↑
Cray M, Selmic LE, Kindra C, et al. Analysis of risk factors associated with complications following mandibulectomy and maxillectomy in dogs. J Am Vet Med Assoc. 2021;259(3):265-274. doi:10.2460/javma.259.3.265
- 3.↑
Boria PA, Murry DJ, Bennett PF, et al. Evaluation of cisplatin combined with piroxicam for the treatment of oral malignant melanoma and oral squamous cell carcinoma in dogs. J Am Vet Med Assoc. 2004;224(3):388-394. doi:10.2460/javma.2004.224.388
- 4.↑
Tuohy JL, Selmic LE, Worley DR, Ehrhart NP, Withrow SJ. Outcome following curative-intent surgery for oral melanoma in dogs: 70 cases (1998-2011). J Am Vet Med Assoc. 2014;245(11):1266-1273. doi:10.2460/javma.245.11.1266
- 5.
Mestrinho LA, Bernardo E, Niza MMR, Lloret A, Buracco P. Neoadjuvant chemoradiotherapy and surgery as treatment for oral maxillary squamous cell carcinoma in a dog. Aust Vet J. 2012;90(7):264-268. doi:10.1111/j.1751-0813.2012.00928.x
- 6.
Gardner H, Fidel J, Haldorson G, Dernell W, Wheeler B. Canine oral fibrosarcomas: a retrospective analysis of 65 cases (1998-2010). Vet Comp Oncol. 2015;13(1):40-47. doi:10.1111/vco.12017
- 7.
Polton G, Borrego JF, Clemente-Vicario F, et al. Melanoma of the dog and cat: consensus and guidelines. Front Vet Sci. 2024;11:1359426. doi:10.3389/fvets.2024.1359426
- 8.↑
Fulton AJ, Nemec A, Murphy BG, Kass PH, Verstraete FJ. Risk factors associated with survival in dogs with nontonsillar oral squamous cell carcinoma 31 cases (1990-2010). J Am Vet Med Assoc. 2013;243(5):696-702. doi:10.2460/javma.243.5.696
- 9.
Kawabe M, Mori T, Ito Y, et al. Outcomes of dogs undergoing radiotherapy for treatment of oral malignant melanoma: 111 cases (2006-2012). J Am Vet Med Assoc. 2015;247(10):1146-1153. doi:10.2460/javma.247.10.1146
- 10.
Culp WT, Ehrhart N, Withrow SJ, et al. Results of surgical excision and evaluation of factors associated with survival time in dogs with lingual neoplasia: 97 cases (1995-2008). J Am Vet Med Assoc. 2013;242(10):1392-1397. doi:10.2460/javma.242.10.1392
- 11.
Riggs J, Adams VJ, Hermer JV, Dobson JM, Murphy S, Ladlow JF. Outcomes following surgical excision or surgical excision combined with adjunctive, hypofractionated radiotherapy in dogs with oral squamous cell carcinoma or fibrosarcoma. J Am Vet Med Assoc. 2018;253(1):73-83. doi:10.2460/javma.253.1.73
- 12.
Pazzi P, Steenkamp G, Rixon AJ. Treatment of canine oral melanomas: a critical review of the literature. Vet Sci. 2022;9(5):196. doi:10.3390/vetsci9050196
- 13.
Giuliano A, Dobson J. Prospective clinical trial of masitinib mesylate treatment for advanced stage III and IV canine malignant melanoma. J Small Anim Pract. 2020;61(3):190-194. doi:10.1111/jsap.13111
- 14.
Myoung H, Kim MJ, Lee JH, Ok YJ, Paeng JY, Yun PY. Correlation of proliferative markers (Ki-67 and PCNA) with survival and lymph node metastasis in oral squamous cell carcinoma: a clinical and histopathological analysis of 113 patients. Int J Oral Maxillofac Surg. 2006;35(11):1005-1010. doi:10.1016/j.ijom.2006.07.016
- 15.
Selmic LE, Lafferty MH, Kamstock DA, et al. Outcome and prognostic factors for osteosarcoma of the maxilla, mandible, or calvarium in dogs: 183 cases (1986-2012). J Am Vet Med Assoc. 2014;245(8):930-938. doi:10.2460/javma.245.8.930
- 16.
Farcas N, Arzi B, Verstraete FJM. Oral and maxillofacial osteosarcoma in dogs: a review. Vet Comp Oncol. 2014;12(3):169-180. doi:10.1111/j.1476-5829.2012.00352.x
- 17.↑
Martano M, Iussich S, Morello E, Buracco P. Canine oral fibrosarcoma: changes in prognosis over the last 30 years? Vet J. 2018;241:1-7. doi:10.1016/j.tvjl.2018.09.005
- 18.↑
Grimes JA, Mestrinho LA, Berg J, et al. Histologic evaluation of mandibular and medial retropharyngeal lymph nodes during staging of oral malignant melanoma and squamous cell carcinoma in dogs. J Am Vet Med Assoc. 2019;254(8):938-943. doi:10.2460/javma.254.8.938
- 19.↑
Sharma S, Boston SE, Skinner OT, et al. Survival time of juvenile dogs with oral squamous cell carcinoma treated with surgery alone: a Veterinary Society of Surgical Oncology retrospective study. Vet Surg. 2021;50(4):740-747. doi:10.1111/vsu.13625
- 20.↑
Wingo K. Histopathologic diagnoses from biopsies of the oral cavity in 403 dogs and 73 cats. J Vet Dent. 2018;35(1):7-17. doi:10.1177/0898756418759760
- 21.↑
Mikiewicz M, Paździor-Czapula K, Gesek M, Lemishevskyi V, Otrocka-Domagała I. Canine and feline oral cavity tumours and tumour-like lesions: a retrospective study of 486 cases (2015-2017). J Comp Pathol. 2019;172:80-87. doi:10.1016/j.jcpa.2019.09.007
- 22.↑
Satthathum C, Srisampane S, Jariyarangsrirattana P, Anusorn P, Sattasathuchana P, Thengchaisri N. Characteristics of canine oral tumors: insights into prevalence, types, and lesion distribution. J Adv Vet Anim Res. 2023;10(3):554-562. doi:10.5455/javar.2023.j709
- 23.↑
Akinkugbe AA, Garcia DT, Brickhouse TH, Mosavel M. Lifestyle risk factor related disparities in oral cancer examination in the US: a population-based cross-sectional study. BMC Public Health. 2020;20(1):153. doi:10.1186/s12889-020-8247-2
- 24.↑
Cray M, Selmic LE, Ruple A. Demographics of dogs and cats with oral tumors presenting to teaching hospitals: 1996-2017. J Vet Sci. 2020;21(5):e70. doi:10.4142/jvs.2020.21.e70
- 25.↑
Air quality index report. 2023. US Environmental Protection Agency. Accessed October 15, 2024. https://www.epa.gov/outdoor-air-quality-data/air-quality-index-report
- 26.↑
Komlós G, Csurgay K, Horváth F, Pelyhe L, Németh Z. Periodontitis as a risk for oral cancer: a case-control study. BMC Oral Health. 2021;21(1):640. doi:10.1186/s12903-021-01998-y
- 27.↑
Ye L, Jiang Y, Liu W, Tao H. Correlation between periodontal disease and oral cancer risk: a meta-analysis. J Cancer Res Ther. 2016;12(suppl):C237-C240. doi:10.4103/0973-1482.200746
- 28.↑
Wallis C, Holcombe LJ. A review of the frequency and impact of periodontal disease in dogs. J Small Anim Pract. 2020;61(9):529-540. doi:10.1111/jsap.13218
- 29.↑
Lobprise HB, Stepaniuk K. Oral surgery-periodontal surgery. In: Wiggs’s Veterinary Dentistry: Principles and Practice. John Wiley & Sons Inc; 2019:193-228. doi:10.1002/9781118816219.ch10
- 30.↑
Campbell RD, Peralta S, Fiani N, Scrivani PV. Comparing intraoral radiography and computed tomography for detecting radiographic signs of periodontitis and endodontic disease in dogs: an agreement study. Front Vet Sci. 2016;3:68. doi:10.3389/fvets.2016.00068
- 31.↑
Gardner DG, Baker DC. The relationship of the canine acanthomatous epulis to ameloblastoma. J Comp Pathol. 1993;108(1):47-55. doi:10.1016/S0021-9975(08)80226-5
- 32.↑
Brodey RS. A clinical and pathologic study of 130 neoplasms of the mouth and pharynx in the dog. Am J Vet Res. 1960;21(84):787-812.
- 33.↑
Kosovsky JK, Matthiesen DT, Marretta SM, Patnaik AK. Results of partial mandibulectomy for the treatment of oral tumors in 142 dogs. Vet Surg. 1991;20(6):397-401. doi:10.1111/j.1532-950X.1991.tb00346.x
- 34.↑
Grüntzig K, Graf R, Boo G, et al. Swiss canine cancer registry 1955-2008: occurrence of the most common tumour diagnoses and influence of age, breed, body size, sex and neutering status on tumour development. J Comp Pathol. 2016;155(2-3):156-170. doi:10.1016/j.jcpa.2016.05.011
- 35.↑
Dobson JM. Breed-predispositions to cancer in pedigree dogs. ISRN Vet Sci. 2013;2013:941275. doi:10.1155/2013/941275
- 36.↑
Ruple A, MacLean E, Snyder-Mackler N, Creevy KE, Promislow D. Dog models of aging. Annu Rev Anim Biosci. 2022;10(1):419-439. doi:10.1146/annurev-animal-051021-080937
- 37.↑
Cooley DM, Beranek BC, Schlittler DL, Glickman NW, Glickman LT, Waters DJ. Endogenous gonadal hormone exposure and bone sarcoma risk. Cancer Epidemiol Biomarkers Prev. 2002;11(11):1434-1440. doi:10.1158/1055-9965.EPI-02-0386
- 38.↑
Boston SE, Lu X, Culp WT, et al. Efficacy of systemic adjuvant therapies administered to dogs after excision of oral malignant melanomas: 151 cases (2001-2012). J Am Vet Med Assoc. 2014;245(4):401-407. doi:10.2460/javma.245.4.401
- 39.↑
Fiani N, Verstraete FJ, Kass PH, Cox DP. Clinicopathologic characterization of odontogenic tumors and focal fibrous hyperplasia in dogs: 152 cases (1995-2005). J Am Vet Med Assoc. 2011;238(4):495-500. doi:10.2460/javma.238.4.495
- 40.↑
Goldschmidt SL, Bell CM, Hetzel S, Soukup J. Clinical characterization of canine acanthomatous ameloblastoma (CAA) in 263 dogs and the influence of postsurgical histopathological margin on local recurrence. J Vet Dent. 2017;34(4):241-247. doi:10.1177/0898756417734312
- 41.↑
Chu YH, Kao SW, Tantoh DM, Ko PC, Lan SJ, Liaw YP. Association between fine particulate matter and oral cancer among Taiwanese men. J Investig Med. 2019;67(1):34-38. doi:10.1136/jim-2016-000263
- 42.↑
He L, Norris C, Cui X, et al. Oral cavity response to air pollutant exposure and association with pulmonary inflammation and symptoms in asthmatic children. Environ Res. 2022;206:112275. doi:10.1016/j.envres.2021.112275
- 43.↑
Snyder LA, Bertone ER, Jakowski RM, Dooner MS, Jennings-Ritchie J, Moore AS. p53 expression and environmental tobacco smoke exposure in feline oral squamous cell carcinoma. Vet Pathol. 2004;41(3):209-214. doi:10.1354/vp.41-3-209
- 44.
Bertone ER, Snyder LA, Moore AS. Environmental and lifestyle risk factors for oral squamous cell carcinoma in domestic cats. J Vet Intern Med. 2003;17(4):557-562. doi:10.1111/j.1939-1676.2003.tb02478.x
- 45.↑
Zaccone R, Renzi A, Chalfon C, et al. Environmental risk factors for the development of oral squamous cell carcinoma in cats. J Vet Intern Med. 2022;36(4):1398-1408. doi:10.1111/jvim.16372
