Response of Spontaneous Oral Tumors in Canine Cancer Patients Treated with Stereotactic Body Radiation Therapy (SBRT)

Abstract

The objective of this study is describe outcome and toxicity for dogs with oral tumors, specifically oral malignant melanoma (OMM), squamous cell carcinoma (SCC), and soft tissue sarcoma (STS) after stereotactic body radiation therapy (SBRT). A single institution retrospective study was conducted. Outcomes were analyzed using Kaplan-Meier analysis and Cox proportional hazard analysis. Treatment responses at different time points were evaluated with Pearson’s Chi-squared test to identify prognostic factors. Acute and late toxicities were recorded according to VRTOG criteria and were analyzed to identify risk factors. Adverse events other than acute and late toxicities were recorded. A total of 98 patients met the inclusion criteria (OMM n = 37; SCC n = 18; STS n = 43). The SBRT prescription was 1–6 fractions, with a total dose range of 12–40 Gy. Local progression-free survival (PFS) for OMM, SCC, and STS was 187, 253, and 161 days, respectively. Overall PFS was 152 days and median survival time (MST) was 270 days, with no statistical difference between tumor types. The presence of lymph node metastasis and the use of elective nodal irradiation (ENI) were associated with shorted PFS and MST. Severe acute toxicities to organs at risk affected 10/85 (11.8%) of patients. Osteoradionecrosis and oronasal fistula formation occurred in 23/81 (28.4%) of patients and was significantly associated with tumor type (SCC, P = 0.006). SBRT can be offered as a treatment option for oral tumors in dogs. Toxicities were common and warrant risk factor considerations and adjustments to current SBRT protocols.

INTRODUCTION

The oral cavity represents a common site of head and neck cancer in canine patients (1–3). The most prevalent tumor types include oral malignant melanoma (OMM), squamous cell carcinoma (SCC), and soft tissue sarcoma (STS) (2–5). These malignancies can arise from maxillary or mandibular gingival mucosa, lip, tongue, soft and hard palate, tonsils, and oropharynx (2, 4). Due to significant biological behavior overlap, shared genomic aberrations and molecular drivers (6), and ability to receive similar state-of-the-art medical care, dogs with spontaneously occurring oral SCC have been proposed as a predictive animal model for human HPV-negative head and neck squamous cell carcinoma (HNSCC) (6–9). Convergence of genomic changes and shared clinical and histopathological features have also led to the proposal of canine oral melanoma as a preclinical comparative oncology model for human mucosal melanoma (10–13). Human head and neck soft tissue sarcomas (HNSTS) are a rare entity accounting for 1% of all head and neck cancers (HNC) (14–17), with a minority of these tumors affecting the oral cavity (15) and no comparative translational model identified. There is a growing interest in the emerging role and potential of the dog as a translational model for cancer research to provide insight into the natural biology and response to therapy of common and uncommon human tumors, including HNC (18–20).

Radiation treatment plays a fundamental role in the local management of head and neck tumors, for both curative-intent and palliative-intent treatment (3, 21–24). Aggressive surgery followed by adjuvant radiation therapy for high-risk cases is generally considered standard-of-care for curative-intent treatment in both human and veterinary medicine (3, 22, 25–28). If surgical excision is deemed not possible, declined or if postsurgical local recurrence is noted, the use of conventionally fractionated and hypofractionated radiation treatment has been described in canine OMM (29–39), SCC (38, 40–46) as well as STS (38, 47–49).

Stereotactic body radiation therapy (SBRT) is an emerging approach for the treatment of oral tumors in the macroscopic disease setting (27, 50–52). This technique involves a highly targeted and precise delivery of high doses of radiation in a few (1–5) fractions on a daily or every other day schedule and is characterized by a sharp dose drop-off to spare normal tissues from significant radiation exposure (27). Its use has been described for local control treatment of various head and neck cancers in companion animals (53–60). In human medicine, it has been applied for recurrent, previously treated head and neck cancer, oligometastatic disease, as a neoadjuvant treatment in combination with immunotherapy as well as primary therapy in elderly or poorly performing patients (50, 52, 61, 62). Currently, there is limited information regarding toxicity and adverse events after the SBRT approach for oral tumors in a treatment-naïve setting in both human (50, 52, 62) and veterinary medicine (58, 59). Severe toxicity is reported in head and neck patients undergoing re-irradiation with an SBRT protocol (63–65). The objective of this study is to describe therapeutic outcomes and normal tissue toxicity for dogs with macroscopic oral tumors, specifically OMM, SCC, and STS, after SBRT with the secondary aim of providing baseline clinical data for the treatment of canine oral tumors with SBRT for future canine comparative oncology model for translational radiation research. The data provided in this study could be useful in designing canine oral tumor clinical trials to investigate experimental therapeutic approaches, such as immunotherapy, that may be used in conjunction with SBRT to treat head and neck cancers.

MATERIALS AND METHODS

Case Selection and Medical Record Review

A single institution retrospective study of dogs with macroscopic OMM, SCC, or STS treated with SBRT was conducted. Inclusion criteria for the study were client-owned canine patients diagnosed with oral tumors via cytology and/or histopathology, treated daily or every other day with external beam SBRT protocols at the authors’ institution between January 2009 and August 2022. Exclusion criteria included carcinoma of the salivary gland, primary osteosarcoma, primary temporomandibular joint location, microscopic disease, and concurrent immunotherapy.

All patients were evaluated with a comprehensive cancer team approach and offered single-modality or multimodality treatment options as deemed appropriate by the attending clinicians. For local tumor control of each case, board-certified radiation and surgical oncologists evaluated patient candidacy for surgical excision with or without adjuvant therapy or radiotherapy alone. Typically, fractionated radiation therapy, SBRT, and palliative radiation therapy options would be discussed with owners for each case, at the discretion of the attending radiation oncologist. Treatment selection was based on client decision; however, due to the retrospective nature of the study, details regarding how treatments were elected were not consistently recorded within the medical records. Medical records were reviewed to collect patient demographics including breed, sex, age at the time of diagnosis, weight, clinical signs associated with the primary oral tumor, date of start of the clinical signs and date of diagnosis, any tumor types other than oral tumors present at the time of treatment. Previous medical and surgical management of the oral tumor, including the outcome of therapy prior to presentation and oral medications at the time of radiation therapy, were also recorded. Gross tumor-specific characteristics included: specific tumor location (maxilla, mandible, soft palate, oropharynx, tongue) and tumor extension [rostral, caudal oropharyngeal, or a combination of these (Fig. 1)], pretreatment tumor size based on physical examination, caliper measurement of the longest tumor diameter and cross-sectional imaging based electronic measurement, as well as specific characteristics identified on cytology and/or histopathology reports.

FIG. 1.

Specific tumor location amongst the total population (%) and between tumor types (n) and specific anatomical extension amongst the total population. Created with Biorender.com. Abbreviations: OMM, oral malignant melanoma; SCC, squamous cell carcinoma; STS, soft tissue sarcoma; PM4, fourth premolar; M2/3, 2nd and 3rd molar; R-C, rostral-caudal; C-O, caudal-oropharyngeal; R-C-O, rostral-caudal-oropharyngeal.

Staging and Diagnostic Testing

The stage of the oral tumor was determined based on the TNM staging system (66). Data recorded from staging diagnostics before radiation therapy included: complete blood count and biochemistry within one month of the first anesthesia event, computed tomography (CT) of the primary tumor with radiation planning, thoracic imaging with three-view thoracic radiographs and/or thoracic CT. The results of cytologic interrogation and/or the findings on biopsy following extirpation of any locoregional lymph node were recorded, when available. CT simulations for radiation therapy planning were performed as previously described (57, 67). Dogs were positioned in sternal, dorsal, or lateral recumbency, with forelimbs positioned caudally.

Tumor stage was determined based on retrospective evaluation of CT scans interpreted by American College of Veterinary Radiology board-certified radiologists. Primary tumor characteristics recorded included largest tumor diameter, presence of bony lysis, nasal invasion, and tumor extension (rostral, caudal oropharyngeal or a combination of these). To describe the anatomical extension of the oral masses, tumors were classified as: rostral if they were described to extend cranially to the 4th premolar (PM4); caudal if they extended from PM4 to the 2nd molar (M2) for maxillary tumors or the 3rd molar (M3) for mandibular tumors; oropharyngeal if they were located caudal to M2/3 (see Fig. 1), or a combination of the above. The presence of abnormal regional lymph node abnormalities in terms of size or contrast-enhancing pattern, as interpreted by the board-certified radiologist, was recorded. If soft tissue pulmonary nodules were noted on thoracic radiographs or CT scan, they were assumed to be distant metastatic disease and were not confirmed by cytology or histology.

Radiation Planning, Delivery, and Plan Parameters

Pre- and post-contrast CT scans were acquired at 2 mm spacing. They were imported and utilized for inverse treatment planning with the Varian Eclipse treatment planning system versions 8.0, 11.0, and 15.5 (Varian Medical Systems, Inc. Palo Alto, CA). The gross tumor volume (GTV) was identified and contoured. Clinical target volume (CTV) expansions were applied at the discretion of the radiation oncologist and typically included 2–15 mm of nasal cavity in dogs with intranasal cavity tumor invasion. A planning target volume (PTV) incorporated a 2–3 mm isotropic expansion from the GTV (or CTV, when applicable) limited within the body contouring, to account for daily set-up error associated with positioning. Locoregional lymph node inclusion was inconsistent and was based on clinical and imaging-based concerns at the discretion of the attending radiation oncologist. In this study, a patient with one or more cytologically confirmed metastatic lymph node(s) that were included in the radiation treatment, was identified as having metastatic lymph node radiation treatment (mLN RT). A patient with any lymph node included in the radiation treatment that was not sampled or not cytologically confirmed to be metastatic, was considered as having elective nodal irradiation (ENI). A patient having a lymph node that was not sampled, but had high clinical concern for metastasis in the medical record based on imaging findings (enlargement, abnormal contrast enhancement) was further described as having suspect metastatic LN RT (smLN RT). Gross tumor volume of the lymph node(s) was contoured separately (GTVn), with a nodal planned target volume (PTVn) of 2–5 mm isotropic expansion from the GTVn or 4–9 mm asymmetric expansion for mandibular and retropharyngeal lymph nodes as previously described (68) for cases after 2012.

The organs at risk (OAR) included skin, pharynx/trachea, esophagus, spinal cord, eyes, lenses, and brain. The normal tissue constraints for the OAR were adopted and modified for canine patients from a previous report of human OAR constraints (69). Mucosal sparing over GTV volume was also retrospectively recorded, based on the presence of target volume structures contoured out of the skin/mucosa OAR.

SBRT plans were created using inverse planning, with coplanar or noncoplanar 6 MV modulated static radiation beams or Volumetric Modulated Arc Therapy (VMAT). Beams were modulated using a 120 multileaf collimator and sliding-window technique. At the time of development, radiation therapy plans were assessed based on the intent to deliver 100% of the radiation prescription to 99% of the GTV, 99% of CTV, and 95% of the PTV. Varian’s AAA dose calculation (version 15.6.) was used for dose predictions and optimization was performed using Varian’s Photon Optimizer (version 15.6.06). PTV-less structures were created when indicated to remove normal OAR structures (eyes, lenses, haired skin, full or partial oral mucosal surface, esophagus, pharynx/trachea, brain) from the PTV and utilized in plan optimization and evaluation. The oral mucosa was contoured as a continuum single structure with the haired skin and the area overlying the tumor was treated at full thickness for tumor disease extension at the discretion of the radiation oncologist, in light of the physical exam and imaging interpretation. An American College of Veterinary Radiology board-certified veterinary radiation oncologist evaluated and approved all SBRT plans. Quality assurance was performed as previously described (57).

All patients were treated under general anesthesia. Protocols for anesthesia were similar to those previously described at the authors’ institution (57). Daily patient positioning was verified by using orthogonal kV–kV images or on-board kV cone-beam CT (CBCT) images, based on the attending clinician’s preferences, and positioning errors were corrected with a four degrees-of-freedom couch. A Varian Trilogy linear accelerator (Varian Medical Systems, Inc.) was used for treatment delivery.

Radiation treatment plans were later assessed in a retrospective manner with the following data collected: GTV, PTV with corresponding less structures when available, volumes and doses to OARs, dose to 99% of the primary GTV (pGTV_99%), dose to 99% of the primary CTV (pCTV_99%, when applicable), dose to 95% of the primary PTV (pPTV_95%), near minimum (D_98%) and near maximum (D_2%), mean and median dose (D_50%) to target volumes (70, 71). D_min, D_max, D_mean, and D_50% doses, as well as volume at specific constraint points were recorded for OARs (69, 70). The intent of mucosal sparing over the GTV volume was also retrospectively recorded, based on the presence of target volume structures contoured out of the skin/mucosa OAR. Written informed consent for treatment was obtained from the owners.

Response and Clinical Follow-up

Information collected in the follow-up period included resolution of previous clinical complaints or development of new clinical signs, development of acute and late tissue toxicities, response to therapy based on imaging and/or physical exam findings, dates and results of follow-up staging via imaging, date, and cause of death. If not available, additional follow-up information was obtained via electronic communication with referring veterinarians. The medical management in the follow-up period was at the discretion of the veterinary team managing the case at the time of each visit. Follow-up anti-inflammatory therapy, chemotherapy, immunotherapy, and radiation therapy in the previously treated field were recorded. Surgical intervention in the radiation field, time from the end of radiation treatment to the surgical procedure, and associated complications were also recorded. The authors intended to assess treatment response according to RECIST criteria (72) utilizing physical exam, caliper and cross-sectional imaging measurements, when available. If no tumor measurements were available, no progressive disease was noted and if the patient’s signs were subjectively improved on the medical records, the tumor response was classified as clinical benefit (CB). For this study, the clinical benefit rate (CBR) was defined as complete response (CR) + partial response (PR) + stable disease (SD) + CB.

Normal tissue toxicity grading was recorded based on medical records review or retrospectively assigned according to the Veterinary Radiation Therapy Oncology Group (VRTOG v1.0) morbidity scoring scheme for acute and late radiation effects (73), with acute effects being defined as signs occurring within 90 days after treatment, and late effects being defined 90 days after treatment. Toxicities were determined based on medical record review and graded retrospectively based on available information. Adverse events, defined as sequelae not fitting VRTOG criteria within the radiation field, were also recorded.

The cause of death was determined based on the review of the medical record documentation, which included physical examination and clinical assessment by the attending veterinarian(s), diagnostic tests, and/or necropsy findings, if available. If the cause of death was unknown or for quality-of-life reasons, then these deaths were attributed to the oral tumor.

Statistical Analysis

Patient outcomes were reported as local progression-free survival (L-PFS), overall progression-free survival (PFS), and overall survival time (OST), as calculated via Kaplan-Meier curve analysis. Local PFS from the start of SBRT to local progressive disease or death (censor = lost to follow-up or local treatment other than fistula repair at previous treated site, without evidence of local tumor progression). Overall PFS were calculated from the start of irradiation to the time of progression or death (censor = alive without progression, lost to follow-up). OST was calculated from the start of irradiation to the time of death of any reason (censor = alive or lost to follow-up). Survival outcomes of different tumor types were compared using the Log Rank test. Univariate Cox proportional hazard models were used to estimate the hazard ratios of risk factors for local PFS, overall PFS, and OST. The analysis was performed for all dogs and individual groups divided by tumor type (OMM, SCC, and STS groups). Specific variables included were pre-irradiation longest tumor diameter (LTD), T3 vs. T2/T1, presence of bony lysis, lymph node metastasis, lung metastasis, total radiation dose (Gy), pGTV and pPTV volume, pGTV_99%, pGTV_95%, elective nodal irradiation, development of grade 3 acute or late VRTOG toxicity of any organ at risk. Specific tumor subtype (amelanotic tumor for OMM, tonsillar location for SCC, hi-low fibrosarcoma for STS group) was also analyzed on the Univariate Cox proportional hazard model. For OMM specifically, the impact on outcome after SBRT of the Oncept^™ melanoma vaccine (Merial, Duluth, GA), a bacterial plasmid DNA vaccine encoding the tumor-targeted antigen human tyrosinase (huTYR pDNA), was also included in the analysis. Multivariate Cox proportional hazard analysis was then performed by including variables with P < 0.2 in the univariate analysis. For L-PFS, dosimetry parameters such as pGTV_99%, pGTV_98%, pGTV_95%, pGTV_50%, pGTV_2%, pPTV_98%, pPTV_95%, pPTV_50% were also separately analyzed by fraction number for OMM (1 and 3 fractions) and STS (3 and 5 fractions). Local response based on RECIST criteria was recorded at 0–3 months, 3–6 months, 6–12 months, and 12–18 months postirradiation. Pearson’s Chi-squared and Fisher’s exact test (if sample size too small) were used to analyze association between response at different time points and tumor type, primary tumor size (T stage, pre-irradiation LTD), tumor extension (rostral, caudal, oropharyngeal, rostrocaudal, caudal-oropharyngeal, and rostro-caudal-oropharyngeal), radiation treatment (total dose, dose/fraction). For toxicity analysis, multiple variables were analyzed to identify prognostic factors associated with acute grade 2 and 3 toxicity (Y/N) and late grade 3 bone toxicity (Y/N). Pearson’s Chi-squared test was used to identify categorical variables associated with high-grade toxicity. Continuous predictors were evaluated by the Mann-Whitney test. Factors evaluated included tumor type, patient age, weight, primary tumor size (pre-RT LTD, T stage, pGTV and pPTV volume), radiation treatment (total dose, dose/fraction, number of fractions), target dose coverage (pGTV_2%, pGTV_99%, pGTV_98%, pGTV_50%, pPTV_98%, pPTV_95%, pPTV_50%), intra-oral full thickness mucosal dose, bony lysis and nasal cavity invasion. The dosimetry data listed above was analyzed overall by absolute dose and separated into subgroups by fraction number. P values, 0.05 were considered statistically significant. All statistical analyses were performed using commercial software Prism v8.4.3 (GraphPad Software, San Diego, CA). The association of elective nodal irradiation (ENI; Y/N) with the development of lymph node progressive disease (LN PD; Y/N) was analyzed with Fisher’s exact test. The impact of ENI on time to LN PD was analyzed with Kaplan-Meier curve analysis and Log Rank test, accounting for the development of new LN PD as an event and censoring patients that did not have recorded lymph node progression. Shapiro-Wilk test was used to test for the normality of the data. A one-way ANOVA followed by Student’s t-test was used to compare body weight and age between tumor type groups.

RESULTS

Patient Demographics

A total of 98 patients met the inclusion criteria for this study. Patient demographics and pre-irradiation tumors are summarized in Table 1. Patients with STS were younger than OMM (Student’s t-test: P = 0.0009) and SCC (P = 0.0036) and had significantly higher body weight compared to OMM (P = 0.0003) and SCC (P = 0.0024). The tumor was incidentally identified during a dental cleaning procedure in 15/98 (15.3%) patients. For 83/98 (84.7%) of dogs with clinical signs, the median time of clinical sign duration before definitive diagnosis of an oral mass was 12 days (range: 0–224). Clinical signs that were reported to be associated with the tumor included: oral mass effect (n = 69, 70.4%), oral pain (n = 17, 17.3%), upper respiratory signs such as sneezing, nasal discharge, stertor, snoring, etc. (n = 15, 15.3%), oral bleeding (n = 14, 14.3%), halitosis (n = 11, 11.2%), hyporexia or anorexia (n = 8, 8.2%), facial swelling (n = 7, 7.1%), sialorrhea (n = 6, 6.1%), hacking or cough (n = 4, 4.1%), exophthalmos (n = 4, 4.1%), lethargy (n = 3, 3.1%), teeth chattering (n = 2, 2%), dysphagia (n = 2, 2%), epistaxis (2, 2%), ptosis (n = 2, 2%), unilateral muscle atrophy (n = 2, 2%), ocular discharge (n = 2, 2%), third eyelid prolapse (n = 1, 1%). Specific tumor anatomical location and extension by tumor type are summarized in Fig. 1.

TABLE 1.

Patient Demographics and Pre-Irradiated Tumor-Directed Therapies

	OMM	SCC	STS
Total patients	N = 37 (37.8%)	N = 18 (18.4%)	N = 43 (43.9%)
Age (years)	10.5 ± 3.3 (1.6–18.7)	10.6 ± 2.7 (7–15)	7.9 ± 3.3 (1.6–15.6)
Weight (Kg)	20 ± 10.8 (3.1–41.3)	19 ± 12.1 (4–40.7)	29.9 ± 12.2 (5.9–56.2)
Sex	FS = 18 (48.6%), MC = 17 (45.9%), MI = 2 (5.4%)	FS = 10 (55.6%), MC = 8 (44.4%)	FS = 15 (34.9%), MC = 24 (55.8%), MI = 3 (7%), FI = 1 (2.3%)
Breeds	MBD (11), Labrador (4), Cocker Spaniel (3), Chow (3), Eskimo (2), Golden Retriever (2), Pug (2), Scottish Terrier (2), Australian Heeler, GSD, Irish Terrier, Keeshond, Maltese, Pomeranian, Rottweiler, Staffordshire Terrier	Labrador (2), Australian Heeler, Australian Shepherd, Bichon, Boxer, Brittany Spaniel, Cairn Terrier, Cocker Spaniel, Eskimo, French Bulldog, GSD, MBD, Pointer, Rat Terrier, Shetland Sheepdog, Shih Tzu, Standard Poodle	MBD (13), Golden Retriever (11), Labrador (6), GSD (2), Husky (2), Bernese Mountain Dog, Bloodhound, Boxer, Cairn Terrier, Irish Setter, Kuvasz, Shih Tzu, Staffordshire Terrier, WHWT
Pre-RT NSAIDs	24/37 (64.9%)	10/18 (55.6%)	28/43 (65.1%)
Pre-RT antibiotics	8/37 (21.6%)	5/18 (27.8%)	20/43 (46.5%)
Pre-RT surgery	11/37 (29.7%)	6/18 (33.3%)	17/43 (39.5%)
Surgical procedures	Cytoreductive surgery (8), maxillectomy (2), mandibulectomy (1)	Cytoreductive surgery (3), hemi-mandibulectomy (2), dental extractions with biopsy (1)	Cytoreductive surgery (8), maxillectomy (6), dental extractions with biopsy (3)
Pre-irradiation chemo	2/37 (5.4%)	3/18 (16.7%)	5/43 (11.6%)
Chemotherapy protocols	TOC (1), IMA (1), CIS (il) + VINC (il) +, 5FU (il) + TOC (1)	CARBO (2), CTX (m) (1)	CTX (m) (3), DOX (2)

Notes. Age and weight data are reported in: mean ± SD (range).

Abbreviations: FS, female spayed; MC, male castrated; MI, male intact; FI, female intact; MBD, mixed breed dog; GSD, German Shepherd Dog; WHWT, West Island White Terrier; RT, radiation therapy; LTD, longest tumor diameter; NSAIDs, non-steroidal anti-inflammatories; 5FU, 5-fluorouracil; CARBO, carboplatin; (il), intralesional; CIS, cisplatin; DOX, doxorubicin; CTX (m), metronomic cyclophosphamide; IMA, imatinib; TOC, toceranib, VINC, vincristine.

Tumor diagnosis was confirmed via biopsy and histopathology in 34/37 (91.9%) of OMM, in 16/18 (88.9%) of SCC and 40/43 (93%) of STS, with the remainder of patients having a cytology-based diagnosis. A total of 34/98 (34.7%) patients underwent some form of cytoreductive surgery (see Table 1), with microscopic disease achieved in 24 cases total prior to mass recurrence. One patient with an STS diagnosis underwent partial maxillectomy (narrow excision) followed by full course of fractionated radiation (3 Gy × 18 fractions) prior to recurrence. A total of 5/37 (13.5%) patients with melanoma received the Oncept^™ melanoma vaccine before presenting for radiation treatment. Prescription of one or more antibiotics, including amoxicillin/clavulanic acid, clindamycin, enrofloxacin, chloramphenicol, and doxycycline, for the management of oral clinical signs was recorded in a total of 33/98 (33.7%) of patients. Non-steroidal anti-inflammatories (NSAIDs) such as carprofen, piroxicam, meloxicam, deracoxib, firocoxib, grapiprant, or aspirin, were prescribed by the referring veterinarians in a total of 62/98 (63.3%) of cases. One patient with melanoma was prescribed prednisolone as an anti-inflammatory.

Staging and Imaging Findings

All dogs had a head/neck pre- and post-contrast CT scan of the skull with radiation planning. One dog had a whole-body fluorine 18 fluorodeoxyglucose (¹⁸F-FDG) Positron Emission Tomography-Computed Tomography (PET-CT). Four dogs were set up in the supine position, one in the lateral position, and the remainder in sternal recumbency. All patients (100%) had thoracic staging with either thoracic radiographs, thoracic CT, or both. Sampling of locoregional lymph nodes, as well as inclusion in the radiation therapy treatment, was dependent on the attending clinician’s discretion. Sampling via fine needle aspirate or excisional biopsy of at least one locoregional lymph node was recorded in a total of 81/98 (82.7%) of patients. Specific advanced imaging findings and staging are summarized in Table 2.

TABLE 2.

Staging and Advanced Imaging Findings

	OMM	SCC	STS
Pre-irradiation LTD (cm)	3.6 (1.1–11.3)	4.5 (1.5–12.5)	6 (1.2–15)
Bony lysis	21/37 (56.7%)	15/18 (83.3%)	40/43 (93%)
Nasal cavity invasion	13/37 (35.1%)	7/18 (38.8%)	24/43 (55.8%)
Lymphadenopathy on CT	22/37 (59.4%)	14/18 (77.8%)	32/43 (74.4%)
Lymph node sampling	32/37 (86.4%)	17/18 (94.4%)	32/43 (74.4%)
Confirmed nodal metastasis	8/37 (21.6%)	5/18 (27.7%)	5/43 (11.6%)
Diagnosis method	C (7); H (1)	C (4); H (1)	C (4); H (1)
Pre-irradiation thoracic radiographs	29/37 (78.4%)	16/18 (88.8%)	32/43 (74.4%)
Pre-irradiation thoracic CT	23/37 (62.1%)	8/18 (44.4%)	21/43 (48.8%)
Pulmonary nodules	3/37 (8.1%)	0/18 (0%)	3/43 (6.9%)
Stage I	3 (8.1%)	2 (11.1%)	1 (2.3%)
Stage II	13 (35.2%)	4 (22.2%)	5 (11.6%)
Stage III	18 (48.6%)	12 (66.7%)	34 (79.1%)
Stage IV	3 (8.1%)	0 (0%)	3 (7%)

Notes. Pre-irradiated LTD reported in median (range).

Abbreviations: C, cytology; H, histology.

Radiation Treatment, planning, and dosimetry profile

The first fraction of SBRT was administered in a median of 26 days (range: 1–552) after definitive diagnosis for OMM, a median of 41 days (range 6–158) for SCC, and a median of 29 days (range: 0–854) for STS. A total of 22 (22.4%) patients (OMM n = 20, STS n = 2) were prescribed a single fraction (dose range 12–22 Gy), 57 patients (OMM n = 16, SCC n = 15, STS n = 26) were prescribed a three-fraction protocol (dose range 8–10 Gy/fraction), and 18 patients (OMM n = 1, SCC n = 2, STS n = 15) were prescribed a five fraction protocol (dose range: 6–8 Gy/fraction), while one dog received 5 Gy 36 fractions treatment (Table 3). Tumor type was significantly associated with the number of fractions prescribed (Fisher’s exact test: P = 0.00000004) (Supplementary Table S1;² https://doi.org/10.1667/RADE-24-00079.1.S1). Static beam IMRT (6–18 beams) or VMAT (1–3 partial or full arcs) were used for planning in 48 and 50 cases, respectively.

TABLE 3.

Summary of Different Radiation Schedules

Single fraction	Total dose (Gy)	EQD2* (Gy)	Treatment duration (days)	Total patients	OMM	SCC	STS
22 Gy × 1	22	58.7	1	1	1	0	0
20 Gy × 1	20	50	1	6	5	0	1
18 Gy × 1	18	42	1	8	7	0	1
15 Gy × 1	15	31.2	1	6	6	0	0
12 Gy × 1	12	22	1	1	1	0	0
				22	20	0	2
Three fractions
10 Gy × 3	30	50	3 (3–7)	47	10	12	25
9 Gy × 3	27	42.7	5 (3–5)	8	5	3	0
8 Gy × 3	24	36	3	2	1	0	1
				57	16	15	26
Five fractions
8 Gy × 5	40	60	5 (5–7)	3	0	0	3
7 Gy × 5	35	49.6	7 (5–7)	10	1	0	9
6 Gy × 5**	30	40	7 (7–8)	5	0	2	3
				18	1	2	15
Six fractions
5 Gy × 6	30	37.5	8	1	0	1	0
				98	37	18	43

Notes. Treatment duration reported in median (range).

^*Assumed alpha:beta of 10 for all oral tumors.

^**One patient received a single fraction of 6 Gy and did not complete protocol, thus was not included in treatment duration analysis.

Lymph nodes were included in the radiation treatment at the discretion of the attending radiation oncologist in a total of 39/98 (39.8%) patients. A total of 5/39 (12.8%) had metastatic lymph node radiation treatment (mLN RT) only; 21/39 (53.8%) had elective nodal irradiation (ENI) only without any confirmed lymph node metastasis; 13/39 (33.3%) patients had a combination of mLN RT and ENI. For the ENI only group, a three-fraction protocol was administered to 15/21 (71.4%) of cases, with a median dose of 25 Gy (range: 18–30 Gy), and single fraction protocol was administered in the remaining 6/21 (28.6%) cases, with a median dose of 16 Gy (range:13–17 Gy). These doses represented a median of 91.6% of the primary tumor prescription (range: 66.6–100%). In 19/21 (91%) of ENI cases, bilateral mandibular and medial retropharyngeal lymph nodes were included, one case had bilateral mandibular lymph nodes included and one case had a single right retropharyngeal lymph node included.

Radiation treatment dosimetry data is summarized in Supplementary Table S1 (https://doi.org/10.1667/RADE-24-00079.1.S1). The intended prescription was not achieved in all cases due to considerations for the dose constraints for organs at risk (OAR). Institutional constraints used for OAR treated with an SBRT protocol, as well as delivered dose to the OAR and reported side effects are summarized in Supplementary Table S2.

Treatment Response, Toxicity, and Clinical Follow-up

The median follow-up period post-radiation was 255 days (range: 7–1,896 days), with a total of 18/98 (18.3%) patients that were lost to follow-up. Three patients were still alive at the time of data collection, at days 451, 550, and 808 postirradiation. After SBRT, clinical improvement was documented in 35/37 (94.5%) of OMM, 17/18 (94.4%) of SCC, and 34/43 (79%) of STS cases, with a median time to improvement of 18 days (range: 4–86), 17 days (range: 2–153) and 26 days (range: 2–119), respectively. Follow-up with at least one CT scan was performed in 34/98 (34.7%) patients. In patients where RECIST criteria were applicable and a primary tumor size measurement was recorded based on physical exam or electronic caliper measurement, the maximum local response recorded for OMM cases was CR in 16/37 (43.2%), PR in 9/37 (24.3%), SD in 3/37 (8.1%) and PD in 2/37 (5.4%), at a median time of 64 days (range: 10–325 days); the maximum local response for SCC cases was CR was recorded in 9/18 (50%), PR in 2/18 (11.1%), SD in 3/18 (16.6%) and PD in 1/18 (5.5%), at a median time of 88 days (range: 24–155); the maximum local response recorded for STS cases was CR in 6/43 (13.9%), PR in 12/43 (27.9%), SD in 4/43 (9.3%) of cases and PD in 2/43 (4.6%), at a median time of 59 days (range: 10–742 days). Dogs that achieved a CR in ≤3 months had a significantly smaller primary disease longest tumor diameter at presentation (Mann-Whitney: P = 0.002), with a median of 3.5 cm (range: 1.1–7.4) compared to a median of 5 cm (range: 1.2–15 cm) for the cases that did not achieve a CR. Overall, there was no significant difference in response between tumor types; however, tumor factors such as T stage and histologic type were significantly associated with the achievement of complete responses and development of local progression within specific time points, as summarized in Fig. 2. At 0–3 months, dogs with T1/T2 stage were more likely (Pearson’s Chi-squared test: P = 0.0006) to achieve a CR compared to stage T3. Patients with OMM (P = 0.005) and SCC (P = 0.003) were more likely to achieve a CR within the 3–6 months period compared to STS cases. Finally, patients with OMM (P = 0.006) and STS (P = 0.014) were associated with a higher likelihood of having local PD at the 6–12 months postirradiation compared to SCC cases.

FIG. 2.

Timeline summarizing local response of oral tumors to SBRT. Chi-square and Fisher’s exact test were used to analyze the association between response and tumor type, primary tumor size (T stage, pre-RT LTD), tumor extension (rostral, caudal, oropharyngeal, rostral-caudal, caudal-oropharyngeal, rostral-caudal-oropharyngeal), radiation treatment (total dose, dose/fraction). Abbreviations: CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; CB, clinical benefit: RECIST criteria cannot be applied; N/A, not applicable: patient diseased or tumor underwent local treatment post-RT; N/D, patient alive, response not documented; LTD, longest tumor diameter; fx, fraction; CBR, Clinical benefit rate (CR + PR + SD + CB).

Acute toxicities affecting skin, oral mucosa, eyes, and/or CNS occurred in 42/85 patients (49.4%) with physical recheck exams within 4 weeks postirradiation. VRTOG grade 3 acute toxicities affected 10/85 (11.8%) of patients. Amongst these 10 cases, oral mucosa was affected in 8 dogs (ulceration, necrosis), eyes in 2 dogs (keratoconjunctivitis sicca progressing to corneal ulceration), and skin in 1 patient (necrosis).

Late toxicities affecting skin, bone, joint, and/or eyes were documented in 35/81 (43.2%) of cases alive and evaluated after 3 months, with VRTOG grade 3 toxicity affecting 24/81 (29.7%) of patients. Amongst these twenty-four cases, bone was affected in twenty-three dogs (osteoradionecrosis, oronasal fistula), skin in three dogs (necrosis, necrotizing cheilitis, cutaneous fistula), and eye (blindness) in one dog. A summary of events attributable to radiation-induced side effects scored via VRTOG criteria is available in Table 4.

TABLE 4.

Summary of VRTOG Acute and Late Toxicities Recorded

Acute VRTOG score	Mucosa	Skin	Eye	CNS
Grade 0	56/85 (65.9%)	71/85 (83.5%)	78/85 (91.8%)	83/85 (97.7%)
Grade 1	7/85 (8.2%)	11/85 (12.9%)	0/85 (0%)	2/85 (2.3%)
Grade 2	14/85 (16.4%)	2/85 (2.4%)	5/85 (5.9%)	0/85 (0%)
Grade 3	8/85 (9.5%)	1/85 (1.2%)	2/85 (2.3%)	0/85 (0%)
Late VRTOG score	Bone	Skin	Eye	Joint
Grade 0	58/81 (71.6%)	68/81 (83.9%)	79/81 (97.4%)	80/81 (98.7%)
Grade 1	0/81 (0%)	8/81 (9.8%)	0/81 (0%)	0/81 (0%)
Grade 2	0/81 (0%)	2/81 (2.4%)	1/81 (1.3%)	1/81 (1.3%)
Grade 3	23/81 (28.4%)	3/81 (3.7%)	1/81 (1.3%)	0/81 (0%)

For single fraction protocols, the volume of skin/mucosa at 16.3 Gy was significantly associated with the development of moderate to severe (VRTOG grade 2 or 3) acute mucosal side effects (Mann-Whitney test: P = 0.035), dogs developing mucositis had a median volume of 1.9 cc at 16.3 Gy (range: 0.1–10.9 cc). None of the other factors included in the analysis were found to be associated with the development of VRTOG grade 2 or 3 acute mucosal toxicity. A total of 9/18 (50%) of SCC experienced osteonecrosis and/or fistula formation, compared to 10/37 (27%) of OMM and 4/43 (9.3%) of STS. There was a significant difference amongst tumor types for the development of grade 3 late bone toxicity (P = 0.0062), with SCC having a higher incidence compared to STS (P = 0.004). Toxicity risk factor analysis is summarized in Table 5.

TABLE 5a.

Association between Categorical and Continuous Variables for Acute Mucosal and Late Bone Toxicity

	VRTOG acute mucosal toxicity (grade 2 or 3) n = 85	VRTOG late bone toxicity (grade 3) n = 78
Categorical variables
OMM vs. SCC vs. STS	0.71	0.0062**
OMM vs. SCC	–	0.20
OMM vs. STS	–	0.069
SCC vs. STS	–	0.0042**
T1/T2 vs. T3	0.60	0.61
Bony lysis Y/N	>0.99	0.56
Nasal invasion Y/N	0.80	0.13
No. of fractions 1 vs. 3 vs. 5	0.96	0.55
CTV Y/N	0.30	0.48
Mucosal sparing Y/N	0.20	0.14
Continuous variables
Pre-irradiation LTD	0.91	0.38
Age	0.63	0.15
Weight	0.93	0.066
Dose/fraction	0.91	0.93
Total dose	0.52	0.15
pGTV volume	0.90	0.64
pPTV volume	0.82	0.63
pGTV 99% dose	0.44	0.34
pPTV 95% dose	0.39	0.18
Intraoral breakthrough	0.27	0.98

Adverse events that represented sequelae not fitting VRTOG criteria (73) 1.0 were recorded. Oral pain in the acute phase postirradiation was documented in 42/85 (49.4%) of patients. Adverse events associated with the oral cavity include: persistent halitosis (n = 11), tumor infection/abscessation (n = 6), chronic mucosal defect at >3 months postirradiation (n = 5), tooth root or bone exposure (n = 4), self-trauma to the oral cavity (n = 3), tooth fracture (n = 1), tooth discoloration and necrosis (n = 1), gum recession and pigmentation change (n = 1), dysphagia (n = 1), oral bleed (n = 1), pharyngeal necrosis documented 93 days postirradiation (n = 1). One dog with oral sarcoma had a biopsy one year postirradiation due to chronic irritation of the oral mucosa in the radiation field and the incisional biopsy described granulation tissue with multifocal perivascular lymphoplasmacytic inflammation. Upper respiratory signs (sneezing, mucopurulent to bloody nasal discharge, and odor) consistent with chronic rhinitis were noted in 12/85 (14.1%) of dogs. One of these patients with intranasal tumor extension had histopathological confirmation of chronic active necro-suppurative rhinitis with intralesional bacteria on incisional biopsy 42 days postirradiation. Epistaxis was reported in 4/85 (4.7%) dogs with intranasal cavity disease; of these, two dogs were noted to have progressive local intraoral disease and one of these patients required a transfusion 139 days postirradiation. Hyperkeratosis and dryness of the nasal planum were reported in three patients. Periorbital swelling was documented in one dog 138 days postirradiation. Acute tumor necrosis was recorded in 32/85 (37.6%), with oronasal fistula formation in < 3 months due to tumor response in three (3.5%) patients. Antibiotics in the postirradiation period were prescribed in 51/85 (60%) of patients for oral health purposes.

Within 4 weeks after SBRT, a total of 77 dogs were prescribed an NSAID, and 3 dogs were prescribed a steroid as an anti-inflammatory therapy. Systemic chemotherapy was administered to a total of 37 patients. A summary of tumor-directed therapies divided by tumor type can be found in Table 6.

TABLE 5b.

Association between dosimetry parameter variables for Acute Mucosal and Late Bone Toxicity

	Single fraction	Three fractions	Five fractions
VRTOG acute mucosal toxicity (grade 2 or 3) n = 85 evaluable
Intraoral breakthrough	0.84	0.46	0.10
Max dose to skin/mucosa	0.25	0.82	0.17
skin/mucosa volume at constraint	0.035*	0.78	0.12
pGTV 2%	0.071	0.17	0.66
pGTV 99%	0.75	0.36	0.61
pGTV 98%	0.42	0.31	0.54
pGTV 50%	0.14	0.79	0.79
pPTV 98%	0.42	0.20	0.88
pPTV 95%	0.34	0.18	0.95
pPTV 50%	0.18	0.15	0.74
VRTOG late bone toxicity (grade 3) n = 78 evaluable
Intraoral breakthrough	0.90	0.44	0.35
Max dose to skin/mucosa	0.42	0.91	0.79
skin/mucosa volume at constraint	0.15	0.74	0.77
pGTV 2%	0.53	0.47	0.99
pGTV 99%	0.30	0.65	0.90
pGTV 98%	0.53	0.78	0.79
pGTV 50%	0.43	0.44	0.99
pPTV 98%	0.37	0.33	0.79
pPTV 95%	0.32	0.32	0.64
pPTV 50%	0.44	0.19	0.97

Surgery in the radiation-treated site was documented in a total of 20 patients, with a median time to the first recorded procedure of 145 days (range: 3–643). Major complications were associated with fistula repair (failure in 7/8 cases), dental extractions, and enucleation procedures, while cytoreductive surgeries, biopsies, and debridement procedures without dental extractions were not documented to have significant complications (see Table 7). Mandibulectomy was performed in 3/9 of patients with mandibular primary tumor location due to tumor progression (n = 2) and osteoradionecrosis (n = 1) and one patient with a rostral sublingual location that experienced mandibular fracture two months postirradiation. Reirradiation was performed in 11 dogs at a median time of 224 days (range: 168–857). Ten patients had treatment in the macroscopic disease setting after local tumor recurrence (SBRT n = 6, weekly hypofractionated n = 4) and one patient underwent fractionated irradiation after mandibulectomy and concerns for incomplete tumor excision. Of the re-irradiated patients, 3/11 (27.2%) were reported to develop VRTOG grade 3 bone toxicity.

TABLE 6.

Summary of Tumor-Directed Therapies that Were Given to Patients after SBRT, Divided by Tumor Type

	OMM	SCC	STS
Postirradiation NSAID	30/37 (81%)	13/18 (72.2%)	34/43 (79%)
Postirradiation steroids	1/37 (2.7%)	1/18 (5.5%)	1/43 (2.3%)
Postirradiation vaccine	17/37 (45.9%)	n/a	n/a
Surgery to irradiated site	6/37 (16.2%)	6/18 (33.3%)	8/43 (18.6%)
Surgical interventions (no. of procedures)	Partial mandibulectomy (2), cytoreductive surgery (1), exenteration (1), silicone button placement (1), cryotherapy (2) and laser/cautery (3), biopsy (1)	Fistula repair (5), dental extractions (2), enucleation (2), mandibulectomy (2), biopsy (1)	Fistula repair (2), cytoreductive surgery (11), dental extractions (3), enucleation (1), biopsy (1)
Lymph node extirpation	2/37 (5.4%)	1/18 (5.5%)	3/43 (6.9%)
Postirradiation chemotherapy	15/37 (40.5%)	7/18 (38.8%)	15/43 (34.8%)
Chemotherapy protocols	CARBO (8), TOC (4), CTX (m) (3), TMZ (1), MAS (1), DTIC (1), 5FU (il) (1)	CARBO (6), DOX (1), ZOL (1), CTX (m) (1), CHL (m) (1)	CTX (m) (7), DOX (5), CARBO (4), TOC (3), VBL (2), MTX (1), CCNU (1), DAS (1)
Radiation course no. 2	5/37 (13.5%)	1/18 (5.5%)	5/43 (11.6%)
Radiation protocol	SBRT 18 Gy × 1 (1), SBRT 10 Gy × 3 (1), weekly 6 Gy × 6 (2), weekly (unknown) × 3 (1)	IMRT 2.7 Gy × 20 to microscopic disease	SBRT 6 Gy × 5 (2), SBRT 7 Gy × 5 (2), weekly 8 Gy × 3 (1)

Abbreviations. CARBO, carboplatin; TOC, toceranib; CTX (m), metronomic cyclophosphamide; TMZ, temozolomide; MAS, masitinib; DTIC, dacarbazine; 5FU, 5-fluorouracil; (il), intralesional; DOX, doxorubicin; ZOL, zoledronate; CHL (m), metronomic chlorambucil; VBL, vinblastine; MTX, mitoxantrone; CCNU, lomustine; DAS, dasatinib.

Local progression was recorded in 45/98 (45.9%) of patients, with no significant difference in incidence amongst tumor type groups (Chi-squared: P = 0.49). The SCC group was found to have a significantly longer local PFS when compared to OMM (P = 0.008) and STS (P = 0.046). There was no difference in overall PFS and OST between tumor types. Survival data is summarized in Fig. 3. Dogs that had a CR in the first 3 months had a significantly improved L-PFS (245 days; P = 0.033), overall PFS (157 days; P = 0.015), and OST (529 days; P = 0.011) when compared to the ones that did not achieve a CR in the first 3 months (L-PFS: 182 days; overall PFS 148 days; OST: 261 days). Dosimetry data analyzed by tumor type and fraction number was not found to be significantly associated with L-PFS.

FIG. 3.

Survival analysis using Kaplan-Meier curves by tumor types. Panel a: Local progression-free survival (L-PFS). Panel b: Overall progression-free survival (PFS). Panel c: Overall survival time (OST). Panel d: Disease-specific survival time (DSS).

Progressive metastatic disease after SBRT occurred more frequently in the OMM group (22/37, 59.4%) when compared to SCC (5/18, 27.7%) and STS (13/43, 30%), and this was statistically significant (Chi-squared: P = 0.01). For OMM cases, development of new metastatic disease was identified in locoregional lymph nodes (5/37, 13.5%), skin (3/37, 8.1%), and tonsils (1/37, 2.7%). New pulmonary nodules were documented in 12/37 (27.9%) and metastatic involvement of the CNS was suspected in 4/37 (10.8%) patients. For SCC cases metastasis post-SBRT was identified in the lymph node in 2/18 (11.1%), lungs in 2/18 (11.1%), spleen in 2/18 (11.1%), and liver and glossal arch/lingual in one patient each (5.5%). In STS cases, new metastatic disease was documented in lymph nodes (6/43, 13.9%), distant periorbital and subcutaneous neck region (1/43, 2.3%), and lungs in 4/43 (9.3%) of patients during follow up period. A total of 52.6% (50/95) of patients were euthanized due to decline in quality of life attributed to disease progression and amongst these, 29.4% (28/95) was due to local oral disease. Radiation therapy-associated toxicity for the oral tumor was documented to be the sole reason for euthanasia in two patients and contributed to the decision of euthanasia in an additional nine patients. In 30.5% (29/95) of cases, the reason for euthanasia was not specified (see summary in Table 8).

TABLE 7.

Summary of Complications for Surgeries in the Oral Tumor Radiation Field

Surgical procedure	No. of complications	Details
Mandibulectomy	1/4	Dehiscence, drift
Fistula repair	7/8	Failure, dehiscence, seroma
Dental extractions	3/5	Fistula
Enucleation or exenteration	2/4	Bleeding, abscessation, fistula
Biopsy or cytoreductive surgery	0/15	–
Cryotherapy and cautery/laser	1/5	Necrosis

The univariate and multivariate Cox Proportional Hazard analysis data analysis for L-PFS, overall PFS, and OST for all tumors and divided by tumor types are summarized in Supplementary Table S3 (https://doi.org/10.1667/RADE-24-00079.1.S1). Specific tumor subtypes within the different tumor type groups did not affect outcomes in this study. No difference in outcomes was noted with the use of Oncept^™ melanoma vaccine after SBRT in OMM cases. On multivariate analysis, OMM cases were at increased risk for shorter overall PFS if they had lymph node metastasis at the time of treatment (HR: 3.089, CI 95% 1.322 to 6.904, P = 0.006) and chemotherapy postirradiation (HR: 2.350; CI 95%: 1.322 to 6.904, P = 0.029). PFS for dogs with lymph node metastasis was 89 days (range: 10–529) compared to 161 days (range: 12–1499) for dogs with no lymph node irradiation. When considering all tumors for OST on multivariate Cox proportional hazard analysis, dogs with lymph node metastasis (HR: 2.315, CI 95% 1.225 to 4.256, P = 0.008) and CTV use (HR: 2.103, CI 95%: 1.315 to 3.916, P = 0.023) were significantly associated with shorter overall survival time. The use of ENI with or without presence of confirmed lymph node metastasis negatively affected overall PFS (HR: 1.827, CI 95% 1.104 to 2.983, P = 0.017) and OST (HR: 2.287, CI 95% 1.315 to 3.916, P = 0.002). When dogs with confirmed lymph node metastasis radiation treatment with or without the combination with ENI (mLN RT ± ENI – OST: 168 days, range: 14–529) were separated from patients without confirmed metastasis treated with ENI (ENI only), the ENI only group had a significantly shorter OST (276 days, range: 24–527; P = 0.033) compared to dogs with no lymph node irradiation (No LN RT – OST: 347 days, range: 30–1896) (see Fig. 4b). Only 5 dogs had confirmed lymph node metastasis irradiation without ENI, therefore they were grouped with the mLN + ENI. When outcomes were analyzed by tumor type, the ENI only group did not have a significantly improved or worse outcome compared to no LN RT in OMM (PFS P = 0.60, OST: P = 0.055), SCC (PFS: P = 0.35, OST: P = 0.44) or STS (PFS: P = 0.25; OST: P = 0.29).

FIG. 4.

Impact of nodal irradiation on PFS and OST. A total of 5 dogs had confirmed lymph node metastasis irradiation without elective nodal irradiation (ENI) and were grouped with the mLN + ENI. Panels a and b: Dogs were divided into three groups. If lymph nodes were included, but not confirmed to be metastatic with sampling, they were categorized as ENI. Panels c and d: Dogs in the above ENI only group were further divided into ENI when lymph nodes were sampled and had low clinical concern for undetected metastasis, and smLN PNRT when lymph nodes were not sampled and there was concern for metastasis on advanced imaging and medical record review. Abbreviations: ENI, elective nodal irradiation; mLN RT, metastatic lymph node radiation therapy; smLN, suspected metastatic lymph node.

The ENI only group was further analyzed with Kaplan-Meier curves. 11/21 (52.3%) cases were OMM, 4/21 (19%) were SCC, and 6/21 (28.5%) were STS. FNA and cytology of at least one lymph node were pursued in 16/21 (76.1%) of cases, revealing no evidence of metastatic disease. In the remainder of cases, no lymph node sampling was obtained, however bilateral mandibular and retropharyngeal lymph nodes were included based on CT imaging and clinical concern for metastasis. When cases were further selected for lymph node sampling indicating no metastasis and low clinical suspicion (n = 16), median overall PFS was not significantly different (PFS 175 days, range: 1–527; P = 0.35) compared to the no LN RT group (PFS 157 days, range: 30–1,499) (see Fig. 4c) and OST was also not significantly different (MST 277 days, range 7–527; P = 0.14) compared to the no LN RT group (see Fig. 4d). When evaluating the use of any ENI with or without confirmed LN metastasis, this did not improve the incidence of LN PD (P = 0.53) or the time to LN PD (P = 0.69) (see Fig. 5).

FIG. 5.

Summary of the use of ENI in the canine population with oral tumors treated with SBRT and outcomes associated with regional lymph node disease progression. The incidence of new nodal progressive disease post-SBRT was compared using Fisher’s exact test (P = 0.53). The time to nodal progression between no ENI and ENI use was compared using Kaplan-Meier curve analysis and Log-Rank test (P = 0.82). Created with Biorender.com. Abbreviations: ENI, elective nodal irradiation; mLN RT, metastatic lymph node radiation therapy; smLN, suspected metastatic lymph node; LN PD, regional lymph node progressive disease.

DISCUSSION

This study represents a novel study describing safety and efficacy of SBRT for the treatment of primary oral mucosal melanoma, carcinoma, and soft tissue sarcoma tumors in canine patients. Local tumor control, defined as clinical benefit rate (CR + PR + SD + CB) was achieved for a majority of dogs for the 3 months (87.9%) and 6 months after SBRT (59.1%). A minority of cases were alive and had long-term local tumor control over one year (16.3%). Smaller tumors (median LTD ≥3.5 cm) had a higher chance of achieving CR within three months, and dogs who achieved CR had significantly improved L-PFS (245 vs. 182 days), overall PFS (157 vs. 148 days) and OST (529 vs 261 days). In this study, there was a variety of protocols and a wide range of dosimetric data, and the number of fraction(s) prescribed was associated with tumor type, thus we were unable to determine if the number of fraction or dosimetric parameters were best for achieving tumor control for the different tumor types. There was a significant difference in clinical results between tumor types only for L-PFS, where SCC had a longer tumor control compared to OMM and STS, which tended to have local progression within 6 months.

For OMM, the median PFS (148 days) and median OST (270 days) were comparable to those of other previously published studies of hypofractionated radiation therapy in the macroscopic disease setting which reported a PFS range of 150–205 days (29–31) and a MST range of 171–401 days (29–35). The MST (332 days) for SCC was similar to the MST of 355 days (41) and 365 days (40) previously reported for hypofractionated protocols in the gross disease setting. Finally, for STS cases, the PFS (157 days) follows a similar trend to a previously published study with a PFS of 180 days (47). Factors to consider in the results of our study include the fact that 70/98 (71.4%) of cases were stage III or IV, meaning they had locally advanced cancers (Fig. 2B) and/or had developed metastatic disease at the time of treatment (Table 2). These factors have been shown to negatively affect outcomes after irradiation in some studies across OMM (29, 30, 35, 38), SCC (38, 42), and STS (38, 49) cases of the oral cavity, with the patient population of these studies having mostly compatible advanced stage (III-IV) at presentation across tumor types (OMM, 31.6–59%; SCC 69.7–75.8%; STS 69.7–73.8%).

Melanoma cases tended to have an increased metastatic rate, as expected (5). Lymph node metastasis negatively affected PFS on multivariate analysis. Chemotherapy administration negatively affected OMM outcomes for L-PFS and PFS, but not OST. We suspect that this result of chemotherapy being negatively associated with PFS may be attributed to case selection, as clinicians may have recommended chemotherapy for cases with more advanced stages at presentation. Administration of the Oncept^™ vaccine did not affect outcome. Overall, the role of systemic agents in the treatment of canine OMM remains to be clarified in veterinary oncology (29, 39). No significant prognostic factors were identified for L-PFS, overall PFS, or OST for SCC and STS cases specifically.

When evaluating dogs by anatomical location and extension, there was no statistically significant difference in outcome. However, when separating cases that were confined to the rostral oral cavity, the OST was 461 days (range: 24–1,896; P = 0.13) compared to 261 days (range: 4–1,499) for a caudal extension, which could be considered a clinically relevant finding in our canine population.

When analyzing all tumors, the use of a CTV in planning to account for microscopic disease extension inside the nasal cavity was negatively associated with OST on multivariate analysis. Although nasal cavity invasion was not identified as negatively affecting outcome, we hypothesized that this finding may be due to case selection for tumors with more aggressive local invasion features and/or clinical concern.

The presence of confirmed lymph node metastasis was identified as a negative prognostic factor, with a significantly shorter PFS and OST. Additionally, in this study the use of ENI was found to negatively affect PFS and OST on initial consideration with univariate and multivariate Cox Proportional hazard analysis, however, there is a significant confounding component in these results due to the use of this approach in patients with presence of at least one regional lymph node metastasis confirmation or in cases where advanced imaging findings raised significant clinical concerns. The use of prophylactic nodal radiation therapy has been described in veterinary radiation oncology for different head and neck cancers, with the goal of improving outcomes by preventing locoregional progression in patients at high risk for undetected micrometastasis (29, 74, 75). It has been empirically pursued in canine OMM without definitive evidence of improved outcome (29). Unlike human medicine (76–78), there are no examples of consensus statements on the management of oral tumors or of standardized locoregional lymph node inclusion with radiation therapy for companion animals. In this case population, the use of ENI did not improve the chance of development or the time to development of progressive regional nodal metastatic disease. The use of ENI only in cases that lacked confirmed lymph node metastasis negatively affected outcomes (OST: 276 days) compared to dogs that did not have any lymph node irradiation (OST: 347 days). When ENI cases with low clinical concern based on cytology sampling and imaging were reviewed separately, outcomes were not significantly different than dogs with no LN RT. In this canine population, there was incomplete or inconsistent regional lymph node sampling and additional limitations in drawing definitive conclusions on the ENI approach with this study population, due to inconsistent inclusion in the treatment plan and relative difference in dose prescribed to electively irradiated lymph nodes. However, these results should encourage a pause before considering the benefit irradiation of large lymph node beds with SBRT approaches, given the possible increased morbidity for patients [e.g., soft tissue necrosis, dysphagia, and carotid blowout syndrome reported in humans (52)], technical complications, planning and treatment time, without improved outcome. Additionally, particular attention should be given when combining SBRT protocols with immunotherapy, given the emerging evidence regarding the negative effects of ENI with radiation treatment protocols on the systemic and local immune responses, specifically effects associated with CD4 and CD8 T cells, in preclinical models and clinical patients, including mice, canine and humans (79–81).

It is important to keep in mind when evaluating the lymph node status of this retrospective canine cohort for outcome, that lymph node size, imaging findings, and cytology sampling results have not been shown to be always predictive of lymph node status of canine head and neck tumors (82–87) and that the regional lymph node may not always represent the draining sentinel lymph node (86, 88–90). Further efforts should be made towards complete staging with lymph node sampling via cytology and/or histopathology and incorporation of sentinel lymph node mapping approaches for lymph node inclusion in the radiation field to provide the best opportunity for translational research.

In this population, 10/85 (11.8%) patients experienced high-grade acute effects to the oral mucosa, skin and/or eyes and 24/81 (29.7%) experienced high-grade late effects to bone, skin and/or eye. Late toxicity incidence is above the quoted risk of 2–20% of developing osteoradionecrosis as late complications for treatment of primary head and neck cancers with SBRT by human institutions (52). In veterinary radiation oncology, normal tissue tolerance for organs at risk (OAR) and associated recommendations for normal tissue constraints when generating an SBRT treatment plan and protocol are still being defined. For comparison, the incidence of high-grade late bone cancer is reported to be 0–21.4% for fractionated protocols (38, 45, 47, 48) and 0–14.2% for palliative hypofractionated protocols. The incidence of high-grade acute and late toxicity results revealed from this retrospective study indicates that adjustments to SBRT protocols for canine oral tumors with respect to OARs should be considered.

Osteoradionecrosis and fistula formation occurred more frequently with SCC cases, with half of the cases being affected. No other tumor characteristics, such as bony lysis on CT, nasal cavity invasion, or planning information was found to be predictive, however, there was variety in contouring and planning approach, as well as dose prescription, which could affect these results. Due to the biological behavior of malignant oral tumors, which commonly present with bony lysis (3) as described in this study, sparing bone by physical avoidance while maintaining tumor coverage appears challenging. Fractionation could play a role in the goal of reducing osteoradionecrosis incidence, however, this was not possible to accurately assess in this cohort due to bias in the prescription of dose number based on tumor type.

In addition to the impact on quality of life of companion animal patients, the concern for the high incidence of severe bone complications post-SBRT is also relevant because fistula repair in the radiation field failed in most cases in this study. The mandible is generally considered a primary site of radiation-induced osteonecrosis due to the low blood perfusion compared to the other osseous structures in the oral cavity (22, 91). Mandibulectomy as salvage therapy was pursued in 4/9 mandibular cases where there was local recurrence and/or osteoradionecrosis formation, with minor complications, most likely because of the possibility of performing a resection outside of the radiation field. Xerostomia is a common radiation-induced side effect experienced by human patients undergoing radiotherapy of the oral cavity and neck region (92–94), however, this was not found to be a significant problem reported in this canine patient cohort. This is likely attributed to the highly targeted approach of SBRT and salivary gland anatomical differences in terms of number and location between dogs (95, 96) and humans (97), however lack of clinical recognition in canine patients is also possible.

In this study, toxicity was recorded based on VRTOG criteria v1.0 (73) published at the time of patient treatment and data collection. A proportion of the described sequelae that did not fit such criteria, fall into the more recently published radiation-induced toxicity criteria by VRTOG v2.0 (98). Additionally, there could be an underestimation of side effects recorded due to a lack of consistent rechecks and/or full oral exams recorded at the time of acute and late toxicity development.

Oral pain was a common concern in the follow-up period, despite the fact that many patients (81.6%) were prescribed oral anti-inflammatories postirradiation, of which the majority were NSAIDs. This opens the question of the ideal pain control regimen after irradiation to the oral cavity in companion animals and whether this should include routine use of neuropathic pain medications such as gabapentin, opioids, such as transdermal fentanyl patches, and topical rinses (99, 100). Another common concern that emerged from our study was the frequent prescription of antibiotics for oral health purposes in the follow-up period. Oral health management can have a major impact on quality of life of human patients after radiation therapy for the head and neck (91, 100). Our finding raises the concern for the quality of life of canine patients related to the development of opportunistic infections and chronic inflammation after SBRT as well as the concern for judicious antimicrobial use and antibiotic-resistance development. Oral tumors are by natural biological behavior in direct contact with the oral cavity microbiome and acute tumor necrosis after SBRT protocols (37.6% in this study) could increase the risk for secondary infections which may be challenging to manage. SBRT can affect local tumor microenvironment and systemic immune response (101–103), and this raises the consideration that this approach may also affect interactions between host immunity and oronasal microbiome.

Limitations

This was a retrospective study, which included a variety of radiation protocols and heterogeneity in lymph node treatment. There was a significant variability in prescription and fractionation, target inclusion, and target dose coverage within the authors’ institution. Similar heterogeneity in protocols has been described for SBRT of HNC amongst experienced human centers (52). Another limitation was the lack of complete regional lymph node sampling in all cases, which could have affected staging, treatment, and prognosis for cases. There was a lack of consistent follow-up physical exam evaluation and imaging evaluation at different time points, which could have affected the recording of tumor responses and toxicity. These limitations can affect the generalizability of our findings and reproducibility in subsequent canine cohorts. Future studies should focus on achieving standardization in contouring, target inclusion, and dose prescription, ideally in a prospective, randomized controlled clinical trial with consistent follow-up for toxicity evaluation and 3D imaging-based response assessments.

CONCLUSIONS

SBRT can be offered as a treatment option for oral tumors in dogs. Toxicities were common and warranted risk factor consideration as well as appropriate veterinary and client education. In this study, it is unclear which risk factors predispose dogs to severe late bone toxicity, aside from SCC tumor type. Dose intensity and/or fractionation adjustments should be considered to help decrease the incidence of osteoradionecrosis and fistula formation. Optimal radiation protocols for local tumor control of macroscopic oral tumors and improved toxicity management should be investigated with further studies for veterinary oncology patients. Surgical intervention in the oral cavity after SBRT should be approached being mindful of the extension of the radiation field and the high risk of complications, specifically for dental extraction and fistula repair attempts, and warrants a multidisciplinary team discussion. This study can provide the basis for the design of future prospective clinical trials investigating the risks vs. benefits of the use of elective lymph node irradiation in SBRT protocols, with or without the combination of immunotherapy. Hopefully, veterinary and human medicine can work together to help investigate new treatment approaches with the goal of improving outcomes for both human and companion animal patients affected by HNC. Given their potential to develop naturally occurring tumors in the presence of an intact immune system, their potential for similar comorbidities as cancer patients, and their ability to receive similar state-of-the-art medical and radiation therapy care to their human counterparts (8, 12, 19, 20), dogs could be a representative and valuable spontaneous oral cancer model for future translational research studies.

TABLE 8.

Summary of Reasons for Euthanasia in 95 Patients

	Disease associated
	local and metastatic	of these, local failure	of these, disease + radiation toxicity	Radiation toxicity associated alone	Non-tumor or radiation associated	Not documented
OMM (n = 36)	24/36 (66.6%)	12/36 (33.3%)	4/36 (11.1%)	0 (0%)	4/36 (11.1%)	8/36 (22.3%)
SCC (n = 17)	5/17 (29.4%)	4/17 (23.5%)	3/17 (17.6%)	1/17 (5.8%)	4/17 (23.5%)	7/17 (41.3%)
STS (n = 42)	21/42 (50%)	12/42 (28.5%)	2/42 (4.8%)	1/42 (2.4%)	7/42 (16.6%)	13/42 (31%)
Total	50/95 (52.6%)	28/95 (29.5%)	9/95 (9.5%)	2/95 (2.1%)	15/95 (15.8%)	28/95 (29.5%)