Effect of radiotherapy on freedom from seizures in dogs with brain tumors

Abstract Background Seizures are a common presenting sign in dogs with brain tumors. Hypothesis/Objectives To investigate the effect of radiotherapy on freedom from brain tumor‐associated seizures and survival time in dogs. Animals Thirty‐two client‐owned dogs with brain tumor‐associated seizures; 18 received medical treatment and radiotherapy, 14 received medical treatment alone. Methods Multicenter retrospective study. Baseline characteristics (seizure semiology, magnetic resonance imaging [MRI] characteristics, and treatment) and duration of seizure freedom were recorded for the 2 treatment groups. Duration of seizure freedom between groups was compared (log‐rank test) using Cox's proportional hazard analysis, with baseline characteristics entered as covariates. Results The duration of seizure freedom and survival time were significantly longer in the radiotherapy group (P < .001), with a mean of 24 months (95% confidence interval [CI], 14.3‐33.8) versus 1.7 months in the control group (95% CI, 0.5‐2.9) and a mean of 34.6 months (95% CI: 25.2‐44.1) versus 6.2 months in the control group (95% CI, 2.6‐9.7) respectively. Baseline characteristics were not associated with duration of seizure freedom after the start of treatment. In the radiotherapy group, 5 dogs were euthanized during the study period because of causes other than seizures. In the control group, recurrence of seizures was observed before death in all dogs. Conclusions and Clinical Importance A longer period of seizure freedom and longer survival time was observed in dogs with brain tumors after radiotherapy compared to medical treatment only. The pathophysiological mechanisms of epileptogenesis and the effect of radiation therapy on seizure control are unclear to date. Further prospective studies are needed.


| INTRODUCTION
Seizures are a common presenting clinical manifestation in dogs with brain tumors 1 and are a major concern of the owners of affected dogs because they are perceived to greatly affect the animal's quality of life.
In veterinary medicine, control of seizure activity in patients with brain tumors is mainly based on the use of antiepileptic drugs (AED) and steroidal anti-inflammatory drugs. Other treatment modalities reported in conjunction with medical treatment include surgical resection, chemotherapy or radiotherapy. 2,3 In human patients, there has been interest in investigating the role of surgical resection, chemotherapy and radiotherapy in seizure control. 4,5 Radiation therapy improves seizure control in humans, but no similar studies are available in dogs. 5 Our aim was to investigate the effect of radiotherapy on freedom from seizures and survival time in dogs with brain tumors. We also investigated whether the clinical presentation of seizures, magnetic resonance imaging (MRI) characteristics of the lesions and drug dosage (AED and corticosteroids) would have confounded this outcome. Study and control dogs were included if they fulfilled the following inclusion criteria: (a) clinical features of seizures and (b) MRI features of a brain tumor. All dogs underwent MRI of the brain and only dogs in which both a board-certified radiologist and neurologist had described a brain tumor as the most likely differential diagnosis were included. The presumed diagnosis of brain tumor on MRI was based on lesion topography (ie, number of lesions, shape of the lesion, and presence of mass effect, dural tail [thickening of the dura adjacent to intracranial pathology on contrast enhanced T1-W images] or contrast enhancement). 6 Histopathological data were analyzed when available. Dogs were excluded if they underwent additional chemotherapy or brain surgery. The minimum follow-up period for both groups was 12 months or until death or euthanasia before reaching the end of the 12-month follow-up period.

| Radiotherapy and medical group
Dogs that received whole brain radiotherapy and medical treatment formed the radiotherapy group and dogs that received medical treatment alone formed the medical group. The radiotherapy course (using 6 MV X-rays or 6  to the base plane of the skull) was calculated to determine a relative ratio of the tumor to brain size. Any mass effect was calculated as the largest deviation from midline in transverse sections. All measurements on MRI were made using a computer software program (Visbion) by one of 2 of the authors (SM and JHR) on T1-W postcontrast images or T2-W images, if no contrast uptake was observed.
A tentative diagnosis of brain tumor type was made based on the MRI findings. Definitive diagnosis by histopathological analysis was recorded, if available.
After radiotherapy, the owners were contacted at the end of 2012 and, according a standard questionnaire, asked whether the animal's seizures were controlled (defined as <1 seizure per month) before treatment and to grade the animal's seizure frequency at 3, 6, and 12 months after treatment as part of clinical follow-up. The owners T A B L E 1 Comparison of the baseline characteristics of the study population between both groups (radiotherapy and medical) were not made aware of the purpose of the study. When available, additional records about seizure activity of the patient before and after initiation of treatment were reviewed. If the animal had been euthanized, the owner was asked whether euthanasia was related to seizure activity. The time when the animal was euthanized was recorded. For each dog, seizure freedom was calculated, defined as the duration of time that the dog remained seizure-free either since the start of radiotherapy for the radiotherapy group or since the start of the drug treatment for the medical group. Corticosteroid and AED dosages during the course of medical treatment and radiotherapy also were recorded (Table 1).
To identify errant variables (those nonrelated to treatment) that might predispose patients to achieve seizure freedom, data from both groups were compared, including patient signalment, clinical presentation (seizure characteristics), drug choice and dosage and MRI findings (Table 2). Baseline characteristics of the 2 groups also were compared (Table 1).

| Statistical analysis
The radiotherapy and medical treatment groups were compared for seizure freedom and survival time, using a Kaplan-Meier Sur- presentation, with single seizure episodes (P = .001) being overrepresented in the radiotherapy group (Table 1). However, Cox hazard proportional analysis (see later) did not identify any variables cofounding the outcome (seizure freedom).  Note that time 0 is defined as the start of radiotherapy for the radiotherapy group or the start of the drug therapy for the medical group F I G U R E 2 Kaplan-Meier survival curves, comparing the cumulative percentages of dogs in the 2 groups. Dogs of the radiotherapy group (red) had a longer survival time as compared to dogs of the medical group (blue), following the start of therapy. Note that time 0 is defined as the start of radiotherapy for the radiotherapy group or the start of the drug therapy for the medical group less severe seizure forms (changing from generalized to focal) or only 1 or 2 episodes after the course of radiotherapy (Figure 2).

| Comparison of seizure freedom and survival time between radiotherapy and medical groups
In the medical group, all the dogs had died or been euthanized by the end of the study, 71% (10/14) because of seizure activity and 29% (4/14) because of the development of other neurological signs.

| Confounding variables
A Cox proportional analysis was used to explore whether signalment, clinical presentation, MRI characteristics or drug dosage could have confounded the time of seizure freedom after radiotherapy. None of the explored characteristics was found to be associated in a statistically significant manner with time of seizure freedom after intervention (Table 2). Only group (radiotherapy versus medical) was significantly different (P = .001).

| DISCUSSION
To our knowledge, this study is the first to investigate the potential of radiotherapy to prolong seizure freedom in dogs with brain tumors, and suggests that radiotherapy decreases seizure frequency in dogs. Similar findings have been observed in humans patients with brain tumors treated with radiotherapy. 5 This information is crucial for the owner's decision making when choosing whether to pursue further treatment or not, because often recurrence of seizures is a common reason for euthanasia 7 in these patients. Interestingly, in our study 70% of the dogs in the medical group were euthanized because of seizure activity rather than other neurological impairment, which could indicate the importance of seizure freedom on the perceived quality of life of the dog for the owners. Similar findings have been described previously. 7 The significant difference between seizure-free duration in the radiotherapy group as compared to the control group allows radiation therapy to be presented to owners as a potential aid in seizure control in dogs with brain tumors. However, because ours was a retrospective study, prospective studies are warranted to confirm this observation.
The reason for the increased duration of seizure freedom after radiotherapy observed in our study remains unclear. In a study of humans with gliomas 5 a significant difference in seizure freedom was achieved at 12 months in 32%-38% of patients. Only the duration of seizures before radiotherapy influenced outcome: Patients with a seizure history of >1 year before radiotherapy had a worse outcome. 5 No significant correlation was found between patients with seizure reduction and reduction of tumor size after radiotherapy. 5 In our study, tumor size after radiotherapy was unknown, because the owners declined repeated MRI. Based on extrapolation from studies in humans, we speculate that the role of radiotherapy in seizure control is not related directly to tumor size, and it is postulated that another mechanism likely plays a role, particularly in those patients with no tumor size reduction. 5 Epileptogenesis in patients with brain tumors still is not well understood. Tumors in the frontal, temporal, and parietal lobes and in the olfactory bulb have been associated with an increased incidence of seizures. 1 The epileptogenic zone often is not restricted to the tumor area. Perilesional edema also is implicated in the mechanism of epileptogenesis., 1,8 In fact, in approximately one-third of the human patients, the epileptogenic zone is not within the tumor itself, but in the surrounding brain tissue. 9 Seizures in humans generally are associated with low-grade tumors, suggesting that slow rate growth might allow time for development of functional (rather than structural) changes to occur in the tumor cells and surrounding tissue. 10 The use of radiation therapy, although aiming to destroy the neoplastic cells, also damages surrounding brain tissue. Because the peritumoral area is the most likely site for the seizure focus, the effect of radiotherapy on the myelin and thus electrical conduction, and also on altered neurons, conceivably could eliminate the functional deficits. 11 This effect could lead to eliminating the source of the hyperexcitability, and therefore the seizures, without a marked decrease in lesion size. Further studies would be necessary to validate this hypothesis.
Our study had a number of limitations. Only the control group had histologically confirmed tumors. None of the dogs in the radiotherapy group that died had necropsy performed, thus the exact histological origin of the presumed tumor on MRI remains unknown. It is known that MRI has a sensitivity of approximately 90% and specificity of approximately 87% for correct identification of brain neoplasia. 10 It also has been shown, however, that cerebrovascular incidents can be mistaken for brain tumors. 6  Seizure free time 6-9 mo 2 (11%) 2 (14%) Seizure free time 9-12 mo 2 (11%) 0 (0%) Seizure free time > 12 mo 11 (61%) 0 (0%) Seizure free by end of study 8 (44.4%) 0 (0%) the exception of seizure presentation, with dogs in the radiotherapy group being presented mainly with single seizures. None of the variables recorded, however, had a significant association with freedom from seizures, with the exception of treatment modality. Other unrecorded characteristics related to the geographic differences between populations, such as diet and environment, could have an influence. The time between presenting seizure activity and the collection of information from the owners may have led to some recall bias.
It would have been useful to evaluate the efficacy of radiation therapy in decreasing seizure frequency by comparing seizure frequency before and after treatment between the 2 groups. Doing so was not possible in our study because intervention (both medical or radiotherapy) was started soon after the onset of seizures and diagnosis. A prospective study could further explore this question.
In conclusion, in our study a longer period of seizure freedom was observed in dogs with brain tumors after radiotherapy compared with those that received medical treatment alone. These findings encourage performance of a prospective multicenter study to confirm our results. Such information might help owners in their decision whether to pursue radiotherapy as a treatment option for animals with suspected brain tumors. Further prospective case-controlled, blinded, randomized studies are needed in humans and dogs to investigate underlying pathophysiological mechanisms, because tumor size and shrinkage and severity of edema do not seem to be the only reasons for the decrease in seizure frequency.