Outcome of intensity-modulated radiation therapy-based stereotactic radiation therapy for treatment of canine nasal carcinomas

1 INTRODUCTION

Canine nasal tumors represent approximately 1% of all canine neoplasms and carcinomas are identified most frequently in this location.^1-4 Common histologic subtypes include adenocarcinoma, squamous cell carcinoma, and undifferentiated carcinoma.^3-5 Definitive radiation therapy is the standard of care treatment for canine nasal tumors, with conventionally fractionated radiation therapy (CFRT) used predominately for the past 35 years.¹ Studies evaluating cumulative doses of 4200-6300 centigray (cGy) in 9-21 fractions as sole treatment of nasal tumors report median survival time (MST) ranging from 8.9 to 19.7 months.^5-11 With the advent of more precise radiation delivery techniques, such as IMRT, reported survival times are comparable to those previously reached with megavoltage and cobalt radiation, with less severe toxicity to adjacent structures.^{7, 8}

Stereotactic radiation therapy (SRT) delivers definitive dosing over an accelerated time period, reducing patient hospital time and increasing convenience.^13-15 This type of external beam radiation therapy uses image guidance to precisely deliver high doses of radiation over a small fraction number.¹⁵ Two studies are published evaluating IMRT-based SRT for dogs with nasal tumors. Glasser et al evaluated three 800-1000 cGy fractions for treatment of 19 dogs with nasal tumors and found an overall MST of 399 days and a carcinoma-specific MST of 332 days. This study reported minimal acute and late toxicity, comparable to previously published IMRT-based CFRT.^{7, 13} Gieger and Nolan evaluated the use of 3000 cGy delivered in three daily fractions in 29 dogs with nasal tumors, including 20 with carcinomas. This study reported minimal toxicity and a median progression-free survival (PFS) of 354 days (150-558 days). In this population, the MST was 586 days (448-724 days), which is at the upper range of what has been reported for radiation therapy as sole treatment for nasal tumors.¹² Although tumor type was not found to affect outcome, carcinoma-specific PFS and MST were not provided.¹²

We hypothesized that SRT for canine nasal carcinomas would produce similar survival outcomes (PFS and MST) to those previously reported for patients with all nasal tumor types treated with SRT, with predominantly grade 1-2 acute and late toxicities.

2 METHODS

This was a retrospective, observational study. Medical records of dogs treated for nasal carcinoma with SRT at the University of Florida Small Animal Hospital (UF SAH) from March 2013 to December 2016 were compiled. Patients were included if they received curative intent IMRT-based SRT for histopathologically confirmed nasal cavity carcinoma. Within this time frame, IMRT-based CFRT, SRT, and palliative-intent dose schemes were offered based on patients’ factors. Staging (complete blood count, chemistry panel, urinalysis, bilateral mandibular lymph node aspirates, thoracic CT or radiographs, and abdominal CT or ultrasound) was recommended but not required. Recheck examinations were recommended 2 and 4 weeks after SRT and then every 3-6 months. Patients were excluded if they received prior radiation therapy or surgery for nasal carcinoma. Patients were excluded if they received prior radiation therapy or surgery for nasal carcinoma. Although previous reports indicate that nasal squamous cell carcinoma may be more radioresistant than other carcinomas in this location, this was not an exclusion criteria for this study as both studies evaluating SRT for canine nasal tumors each include one of these patients.^{12, 13, 16}

Data were retrieved and recorded by two clinicians (SF and KS) and included patient information (breed, age, sex, weight, and presence/type of clinical signs), tumor descriptors (sampling method, diagnosis, and modified Adams’ tumor stage⁵), treatment dates, radiation dosimetry information, adjuvant medical therapy (nonsteroidal anti-inflammatory drugs [NSAIDs] and chemotherapy), and outcome (toxicity, imaging, clinical/image-based response, and survival). Radiation toxicity was retrospectively graded according to the Veterinary Radiation Therapy Oncology Group from information in medical records and through telephone interviews with owners and primary veterinarians.¹⁷ Acute toxicity was defined as occurring fewer than 90 days from initiation of SRT and late toxicity was defined as occurring after 90 days from start of SRT.^{7, 12} Results of staging tests were recorded when performed. Long-term follow-up and death information was obtained from hospital records or telephone contact with primary veterinarians and owners. Computed tomography-based response was classified using the Veterinary Cooperative Oncology Group response evaluation criteria.¹⁸

Radiation dosimetry data included: minimum, mean, maximum gross tumor volume (GTV) and planning target volume (PTV); minimum and maximum dose to GTV and PTV; dose at 98% (D98), 95% (D95), and/or 2% (D2) of volume of GTV and PTV; normal brain volume (cm³) that received more than 2500 cGy and more than 1500 cGy; contralateral eye and ipsilateral eye tissue volume and percentage of structure volume that received more than 1200 cGy. D98 approximates minimum dose to target. D2 approximates maximum dose to target.

Homogeneity index (HI), conformity index (CI), and gradient index (GI) were evaluated. HI evaluates degree of dose distribution homogeneity in the PTV and is defined as HI = max isodose/prescription isodose (PI).^{19, 20} Conformity index indicates degree of conformity of dose distribution within the PTV.²⁰ It is defined as (volume of target covered by PI)/(PI volume × target volume). GI indicates degree of dose fall-off outside PTV.²¹ GI is defined as body volume receiving 50% prescription dose/body volume receiving 100% prescription dose. The gradient measure was equal to the 50% isodose line.

The biologically effective dose (BED) was calculated using α/β = 10 (BED₁₀) for acute responding tissue and α/β = 3 (BED₃) for late responding tissue. Equivalent dose in 200 cGy/fraction (EQD2) was calculated to assess 200 cGy equivalent dose for late responding normal tissue based on the linear-quadratic model. Although the linear-quadratic model may be inappropriate to predict biological response for SRT, BED and EQD2 were used in our study due to the lack of a superior model.

3 STATISTICAL ANALYSIS

Statistical analyses were performed by one author with postgraduate training in statistics (KS) and checked by one author with graduate- and undergraduate-level coursework training in statistics (SF), using commercially available software (SPSS Version 25, IBM Corp., Armonk, NY). Survival time (ST) was defined as time from start of SRT until death. Dogs were censored if alive at time of analysis, at time of loss to follow-up, or at time of second radiation treatment. Progression-free survival was defined as time from start of SRT until tumor recurrence/metastasis. Local recurrence was confirmed by CT or suspected if clinical signs progressed in absence of CT confirmation. Patients were censored from PFS at time of loss to follow-up, death, or if alive at time of analysis if without image-based or clinical evidence of disease progression.

Several factors were evaluated for potential effects on PFS and ST. These included age, body weight, sex (male/female), days to complete SRT (3/>3), modified Adams’ stage (Stage 1-2, Stage 3, and Stage 4), epistaxis, facial deformity, adjuvant NSAID, adjuvant chemotherapy, clinical response, CT-based response (complete response [CR] or partial response [PR]¹⁸), BED₁₀, D2 GTV, and GTV volume. For continuous variables, median was determined and values were categorized as above or below the mean for further analysis. Unless specified, categorical variables were evaluated on presence or absence.

Continuous data were assessed for normality by the Shapiro-Wilk test. The Kaplan-Meier (KM) product-limit method was used to generate PFS and ST. The log-rank test was used to compare KM curves in evaluation of potential risk factors. A P-value of <.05 was considered significant.

4 RESULTS

Twenty dogs received SRT for nasal carcinomas within the study period. One dog was excluded due to prior radiation treatment. Two dogs with severe comorbidities were excluded due to the palliative intent of their SRT. Seventeen dogs met inclusion criteria.

5 DISCUSSION

This study evaluates the use of SRT for canine nasal carcinoma. This population of dogs reached a median PFS of 359 days and an MST of 563 days. Although the nature of this study precludes direct comparison to other studies evaluating IMRT-based SRT for dogs with nasal tumors, survival in our population appears similar to what has been previously reported (PFS 354 days and MST 339 and 586 days).^{12, 13} Fifteen dogs (88.2%) experienced a clinical benefit, which is comparable to the clinical response rates (94.7% and 100%) reported in the two prior studies investigating IMRT-based SRT for all nasal tumors.^{12, 13}

Of the factors evaluated for impact on PFS and ST, only clinical response was found to have a significant impact on survival time. Based on our review of the literature, this has not been previously reported for dogs receiving IMRT-based radiation for nasal tumors but has been documented for orthovoltage radiation treatment of this disease.²⁶ Improvement in clinical signs demonstrated that these patients benefitted from radiation. For a portion of dogs, clinical improvement could have been the result of tumor response to treatment, though without follow-up CT and histopathology, radiation mitigation of local inflammation cannot be excluded as a contributor.

On follow-up CT, six of 10 dogs obtained a PR and one experienced a CR. Computed tomography-based response has previously been correlated with improved survival among dogs undergoing definitive radiation for nasal tumors.²⁷ Within this population, clinical response conferred a survival benefit, but image-based response did not, which likely reflects the limitations of our study, including small patient number and lack of standardized follow-up. The majority of patients underwent CT imaging following radiation therapy (59%); however, timing of imaging was not uniform among patients. In the study by Thrall et al, CT scans were performed every 3 months following orthovoltage radiation treatment for canine nasal tumors and maximal imaging response most frequently occurred 3-6 months posttreatment.²⁷ In our population, median time to recheck CT was 8 months. Some patients underwent CT to investigate clinical suspicion of progressive disease. Our image-based response data may underestimate true rate of response among this population.

One drawback of CT-based response assessment for nasal tumors is the inability to differentiate between residual disease, dormant tumor, fibrosis, and rhinitis. This may explain why patients achieving subtotal resolution of soft tissue density material in the nasal passage (PR) can experience prolonged progression-free and overall survivals following radiation for nasal tumors, as demonstrated in our population. Histopathology is needed to distinguish fibrosis or rhinitis from residual tumor, which is rarely performed. For one patient, biopsy via rhinotomy following SRT identified rhinitis without evidence of carcinoma in the evaluated tissue. If this sample were representative, it would indicate a histologic CR versus the PR designated from imaging criteria. Unfortunately, this procedure also likely contributed to fistula formation at the rhinotomy site and patient morbidity and mortality.

Among patients evaluated for acute toxicity, toxicity was mild, and no patients experienced grade 3 acute toxicity. Toxicity could have been underreported due to lack of follow-up for a portion of the population in the appropriate time frame.

Late toxicity in our population was higher than anticipated based on previously published SRT- and IMRT-based protocols for treatment of nasal tumors in dogs. Distinguishing between late toxicity and tumor progression can be challenging without imaging or necropsy. In this location following radiation therapy, neurologic, ocular, and nasal signs can be caused by either tumor progression or radiation toxicity. In some patients, duration and progression of signs can help distinguish between radiation-induced change and tumor recurrence. For example, return of nasal discharge, epistaxis, or sneezing due to radiation-induced nasal cavity changes can be persistent but often respond to symptomatic therapy (antibiotics and steroids) and is not expected to progress to include other abnormalities (eg, facial deformity). Other signs (mentation change and seizures) can have a similar clinical presentation and outcome from radiation toxicity or tumor progression, and cause may be indistinguishable without advanced imaging or histopathology. Unfortunately, imaging and necropsy were not obtained for any patients experiencing potential severe late toxicity, which is a limitation of this study.

In this study, 50% of patients developed possible grade 1-3 late ocular toxicity. Among these dogs, five of the eight patients experienced grade 1 toxicity (KCS), and three of the eight dogs developed possible grade 3 toxicity (blindness). One dog developed progressive, bilateral blindness followed by hearing loss starting at 13 years of age, 912 days following SRT. One dog developed acute-onset bilateral blindness along with mentation change, circling, and seizures 200 days following treatment. One patient developed acute vision loss 469 days after SRT, and further follow-up was recommended but not obtained. Lawrence et al reported a late toxicity rate of 25% with no dogs developing blindness following IMRT-based CFRT treatment for nasal tumors.⁹ In that prospective study, serial ophthalmologic examinations were used to help characterize ocular changes, which would have been helpful in our population. Lacking serial follow-up, we evaluated doses of radiation delivered to the eyes and optic nerves for patients that developed blindness, with only one patient receiving radiation approaching a dose expected to cause toxicity. With doses delivered to the eyes and optic nerves among these patients in addition to their clinical presentations, we suspect that some developed blindness due to age-related disease, other central nervous system (CNS) disease (including radiation toxicity to brain), or potentially tumor progression rather than radiation toxicity to ocular structures.

The two prior studies evaluating SRT for nasal tumors report oronasal or nasocutaneous fistula formation following radiation. Gieger and Nolan and Glasser et al report two dogs (6.8%) and one dog (5.2%), respectively, with fistula formation likely attributable to radiation, which is similar to what occurred in our population (5.8%).^{12, 13}

Three of our patients developed potential grade 3 CNS toxicity (seizures). No patients had seizures at presentation, and unfortunately, none had imaging performed after seizures developed. When evaluating dose to brain achieved in these patients, two received doses at or exceeding the recommendation of full prescription dose to <1.1 cm³ brain established in Griffin et al.²⁵ One patient received a dose well below this cutoff, and it is considered less likely that this patient's signs can be attributed to toxicity rather than tumor progression. Hunley et al reported no late CNS toxicity in patients receiving fractionated IMRT for nasal tumors, which is lower than what we observed.⁷

Given the higher rates of potential ocular and CNS toxicity, more precise planning or increased fractionation should be considered. Utilization of concurrent MRI for more precise planning may reduce both late ocular and CNS toxicity. Many nasal tumor patients have soft tissue density extension into the frontal sinus. Due to the possibility of tumor extension, some radiation oncologists include any soft tissue density within the frontal sinus in the GTV, even if noncontrast-enhancing, to avoid tumor under-treatment. Radiation delivery to nontumor tissue could increase dose exposure to adjacent normal brain. In the retrobulbar space or sphenoid sinus, poor visualization of the optic nerves with CT creates an additional challenge. Magnetic resonance imaging could aid in delineating optic nerves in the retrobulbar space, which could reduce exposure to normal CNS tissues, and potentially reduce toxicity. One pilot study evaluating MRI in conjunction with CT in dogs with nasal tumors demonstrated improvement in evaluation of margins at the soft tissue-tumor interface and identification of meningeal enhancement.²⁷ All patients had a greater than 10% difference (11.2-32.1%) in calculated tumor volume with combined CT/MRI compared with CT alone.²⁸ Magnetic resonance imaging may become more valuable in treatment planning and should be considered to improve tumor contouring and patient outcomes.

The main limitations of this study are its retrospective nature, lack of uniform follow-up, lack of imaging or pathology for patients experiencing potential severe late radiation effects, and small patient number. Many patients had regular follow-up at our hospital; however, timing of imaging and toxicity assessment was not standardized. Most patients had at least some follow-up via phone interviews months to years after treatment. Retrospective grading of toxicity is subject to recall bias, which likely resulted in under-reporting of toxicity, particularly lower grade toxicity. Adjunctive therapy (NSAIDs and chemotherapy) was not standardized, and the small patient number may have prevented detection of a difference in outcome among patients where one exists. At least one dog experienced clinical benefit from chemotherapy following confirmed progressive disease. Other dogs in this population may have experienced a survival benefit from adjuvant treatment that was not detected.

In conclusion, we found that dogs with nasal carcinoma treated with SRT experienced progression-free and overall survivals comparable to previous reports for dogs receiving SRT for all nasal tumors. We identified clinical response as a positive prognostic factor for survival. A higher than anticipated rate of possible late toxicity occurred in our population, and we intend to further investigate the use of concurrent MRI-limited scans in future patients to improve tumor contouring within the cranial vault and retrobulbar spaces, hopefully maximizing response durations with reduced toxicity to CNS structures.

LIST OF AUTHOR CONTRIBUTIONS

CONFLICT OF INTEREST

The authors declare no conflict of interest.