Evaluation of the use of chemotherapy and other prognostic variables for surgically excised canine thyroid carcinoma with and without metastasis

Abstract

This retrospective study compared the efficacy of surgery alone versus surgery in combination with chemotherapy in the treatment of canine thyroid carcinoma; potential prognostic factors were evaluated. Forty-four dogs with biopsy-confirmed thyroid carcinoma met the inclusion criteria. Twenty-eight dogs were treated with surgery alone and 16 with surgery and chemotherapy. The median survival of dogs treated with surgery and chemotherapy was 518 d, which was not statistically different from that of the dogs treated with surgery alone. The number of thyroid lobes removed at surgery was prognostic with respect to survival. Despite an overall metastatic rate of 48%, the addition of chemotherapy to surgical excision did not improve survival; however, this finding may be due to inadequate power to demonstrate a difference.

Introduction

Thyroid tumors constitute 1.2% to 3.8% of all tumors in the dog (1). Approximately 90% of clinically detectable thyroid tumors are carcinomas (2). Detectable metastases are present at the time of diagnosis in more than 35% of these cases (3–7). Metastases are generally located in the lungs and in regional lymph nodes (1). Size of the primary tumor and degree of invasiveness into normal tissues correlate with the presence of metastasis and with prognosis (3,8).

The use of surgery alone in freely movable non-metastatic thyroid tumors is associated with a median survival time over 36 mo (9). Radiation and chemotherapy are also employed for thyroid carcinoma treatment. Cisplatin chemotherapy alone in advanced cases resulted in a median survival duration of 98 d in one report (8). Doxorubicin alone has also shown efficacy in some cases (4,10). Radiation therapy has been used to treat fixed thyroid carcinomas with median survival times of 22 to 24 mo (5,11), and 1- and 3-year progression free survival rates of 80% and 72%, respectively (12).

Despite the apparent success of radiation therapy and surgery in the management of localized thyroid carcinomas, up to 80% of canine patients develop metastasis during the course of their disease (2–7). The objectives of this study were, therefore, to retrospectively evaluate the use of chemotherapy in addition to surgery in the curative intent management of canine thyroid carcinoma and to identify possible prognostic variables for the disease.

Materials and methods

The medical records of dogs that were presented to the University of Illinois Veterinary Teaching Hospital and the Université de Montréal Centre Hospitalier Universitaire Vétérinaire for thyroid tumors between 1980 and 2004 were retrospectively evaluated. Patients with biopsy-confirmed thyroid carcinoma were eligible for entry into the study. Information collected from the medical records included age, breed, sex, body weight; the presence, nature, and duration of clinical signs; and histological diagnosis. When available, information regarding tumor size and tumor invasion into normal tissues was collected. Tumor invasion was determined by physical palpation, and defined as a fixed (invasive) or a mobile (non-invasive) tumor. Staging was performed according to the World Health Organization (WHO) TNM classification system for tumors of the thyroid gland in dogs (13) (T0 — No evidence of tumor, T1 — Tumor ≤ 2 cm maximum diameter, T2 — Tumor 2 to 5 cm maximum diameter, T3 — Tumor ≥ 5 cm maximum diameter). Staging evaluation and minimum data base were recorded, including complete blood (cell) count, serum chemistry panel, urinalysis, thoracic radiographs, regional lymph node evaluation (cytology or histopathology), and advanced imaging using computed tomography (CT), magnetic resonance imaging (MRI), and/or nuclear scintigraphy (Tc-pertechnetate). Thyroid gland function was assessed by 1 or more of the following tests: total T3 and T4, free T4, and TSH to determine the functional endocrine status of the tumor.

The type of treatment administered was based on the clinician’s preference or limited by the owner’s prerogative; this was retrieved from medical records. Patients were classified as having received surgery alone or in combination with chemotherapy. Information with respect to surgery included unilateral or bilateral lobe removal, the status of the surgical margins on histopathology, and the incidence and type of surgical complications. The type of surgery performed was based on the surgeon’s preference and classified as cytoreductive if complete excision was not attempted, simple excision if excision of the gland only was performed, and radical excision if surrounding tissues such as the larynx, esophagus, and carotid were also resected. Whether the surgery was performed at the referral institution or at the referring veterinarian’s practice was also noted when available. Information with respect to chemotherapy included the agent used, the dose, frequency and number of doses received, and whether clinical response and/or adverse effects were observed.

Endpoints were response to therapy and overall survival. A complete response was defined as a complete resolution of measurable disease, a partial response was defined as a decrease in measurable tumor volume of ≥ 50%, stable disease as < 50% decrease but < 25% increase in measurable tumor volume with no new lesions detected, and progressive disease was defined as an increase of > 25% in measurable tumor volume. Duration of survival was defined as the time from the start of treatment (surgery) to the patient’s death. Survival information was obtained from the medical record or telephone conversation with the referring veterinarian and/or the owners. Dogs were censored if they were still alive at the time of analysis, if they were lost to follow-up, or if they died or were euthanized for unknown or unrelated causes. Age, sex, tumor size, type and invasiveness, the duration of clinical signs, the presence of metastasis at onset, and the use of chemotherapy were analyzed for prognostic significance with respect to survival. With respect to surgical treatment, the margin status, whether the tumor was fixed or mobile, the number of lobes removed, the occurrence of surgical complications, whether the surgery was performed at the referring veterinarian’s clinic, the type of surgery performed (cytoreductive, radical, simple excision), and the use of postoperative chemotherapy, the chemotherapy drug administered, the number of doses administered, and whether a response was seen were also assessed for prognostic significance with respect to survival.

For statistical analysis, 2 treatment groups were identified: surgery alone, or surgery with chemotherapy. Survival curves were generated for each group and compared using the Kaplan-Meier product limit test. Prognostic variables were first assessed by univariate analysis using the Kaplan-Meier product limit test and then by multivariate analysis using the Cox proportional hazards test. We also evaluated the potential association of certain categorical variables using the exact chi-square test. Variables tested included: development of metastasis and chemotherapy; time to metastasis and chemotherapy; duration of clinical signs and metastasis at diagnosis; number of lobes removed and post-operative complication; tumor invasiveness and postoperative complication; margin status and post-operative complication; type of surgery and post-operative complication. Statistical analyses were carried out with SAS version 9.1 (SAS, Cary, North Carolina, USA). The level of statistical significance was set at P < 0.05.

Results

Clinical presentation

Forty-four cases met the inclusion criteria. Various breeds were represented, with mixed breed being the most common and accounting for 8 of the 44 cases. Golden retrievers, Labrador retrievers, and Afghan hounds accounted for 3 each; beagles, English bulldogs, and German shepherd dogs accounted for 2 each. The male-to-female ratio was 0.6 with 27 female and 17 males. The median and mean ages at presentation were 10 and 9.3 y, respectively, with an age range of 1 to 14 y. The median and mean weights of dogs in this population were 26.4 and 26.2 kg, respectively, with a range of 5.5 to 50.9 kg. The presenting complaint was the presence of a cervical mass in 37 cases, with 4 dogs presenting for respiratory problems (labored breathing, respiratory distress, dyspnea, excessive panting) and 1 with digestive signs (difficulty swallowing). In 1 dog the mass was an incidental finding on routine physical examination. In 1 dog the presenting complaint was not recorded. The median and mean durations of clinical signs before diagnosis were 42 and 98 d, respectively, with a range of 2 to 730 d.

Diagnosis and staging

Thirty cases were diagnosed from biopsy material as having follicular thyroid carcinomas, 6 had solid carcinomas, 4 had carcinomas, 2 had medullary carcinomas, and 1 each had anaplastic carcinoma and adenocarcinoma. The tumors were classified as fixed in 11 cases, mobile in 14 cases and undetermined in 19 cases. According to the WHO classification (13), 20 dogs were locally staged with T3 tumors (9 were treated with surgery and chemotherapy and 11 with surgery only), 18 with T2 tumors (3 were treated with surgery and chemotherapy and 15 were treated with surgery only), 2 with T1 tumors (1 in each treatment group) and tumor size was not recorded in 4 dogs (3 were treated with surgery and chemotherapy and 1 was treated with surgery only). Metastasis was present at the time of diagnosis in 15 cases and eventually developed in another 6 dogs, for an overall metastatic rate of 48%. Metastases at presentation were noted in the lungs in 6 dogs, regional lymph nodes in 4 dogs, cranial mediastinum in 1 dog, lung and regional lymph nodes in 2 dogs, and lung and liver in 1 dog. An additional dog was reported to have metastasis but the location was not recorded. Six dogs with metastasis were treated with a combination of surgery and chemotherapy and 9 with surgery only. A functional tumor was diagnosed in 6 of the 34 cases (17.6%) that had thyroid hormone levels evaluated. Dogs determined to be hyperthyroid had a mean total T4 value of 98.6 nmol/L with a range of 66.5 to 167 nmol/L.

Treatment and outcome

Surgery was performed as the sole treatment in 28 dogs and 16 were treated with a combination of surgery and chemotherapy. Twenty-four of the 42 patients for which information was available had a unilateral thyroidectomy. The overall surgical complication rate was 35.7%. The complication rate for unilateral thyroidectomy was 29.2% (7/24) and 50% (9/18) for bilateral. The development of transient hypocalcemia in 6 dogs was the most common complication, followed by hypothyroidism in 3 dogs and the development of seroma at the surgical site in 3 dogs. One dog had laryngeal dysfunction, 1 dog had a temporary upper airway obstruction, and 1 dog developed aspiration pneumonia. No dogs died of their complications. Surgery was performed at the referring institution in 8 dogs. Four cytoreductive procedures were performed; all of these were on dogs with invasive tumors (2 fixed). One dog had a radical procedure (removal of part of the eosophagus) and was euthanized 12 d after surgery at the owner’s request due to the poor prognosis associated with the diagnosis. Margins were defined as clean on the histopathology report in 22 of the 44 dogs.

With respect to chemotherapy, 8 dogs received platinum agents as part of their initial protocol. Five dogs were treated with carboplatin alone. Four received 300 mg/m² and 1 that weighed 5.5 kg was treated at 166 mg/m². One dog was treated with a combination of carboplatin (300 mg/m²) and gemcitabine (120 mg/m²) and 2 dogs with a combination of cisplatin (50 mg/m²) and doxorubicin (15 mg/m²). Eight dogs received doxorubicin as a single agent at 30 mg/m². An average of 4.4, 3-week cycles of chemotherapy were administered, with a range of 1 to 13 cycles. One complete response was noted in a dog with lung metastasis receiving doxorubicin and that dog was still in complete remission at the last visit 362 d later. Three partial remissions were noted. One dog that received carboplatin alone had a partial remission of 480 d. One dog treated with a combination of gemcitabin and carboplatin had a partial remission of 396 d. One dog treated with doxorubicin had a partial remission of 41 d. One dog had stable disease for 365 d while treated with doxorubicin. One dog did not respond to the combination cisplatin and doxorubicin. Eleven dogs receiving chemotherapy could not be evaluated for response as they were treated in a microscopic disease setting. Information on complication from chemotherapy was not available.

Of the 44 cases, survival information was available for 35: 13 in the surgery and chemotherapy group and 22 in the surgery only group. The overall median survival duration was 518 d. Fifteen cases were censored. Seven dogs were still alive, 3 were lost to follow-up, 3 were euthanized for unrelated disease, and 2 had unknown cause of death or euthanasia. The cause of death was identified as due to the primary tumor in 3 dogs, 2 of which were in the surgery only group, and secondary to metastasis in 3 dogs, 2 were of which were in the surgery and chemotherapy group. Two dogs were euthanized because of progression of pre-existing metastatic disease (1 in each treatment group) and one because of the development of metastasis (surgery and chemotherapy treatment group). Two dogs were euthanized because of the primary tumor and the development of metastasis and both were in the surgery only treatment group. The remaining 12 dogs were documented as having died as a result of their cancers but there was no indication whether death was due to local disease or dissemination.

None of the prognostic factors that were assessed were significant. The duration of clinical signs prior to diagnosis approached significance at a P = 0.06. The median survival time for patients displaying tumor-related clinical signs for less than 35 d was 1677 d compared with 391 d for dogs presenting tumor-related clinical signs for greater than 35 d (Figure 1). Statistical significance was not retained upon multivariate analysis with P = 0.22.

Figure 1.

Survival curve illustrating the relationship between survival and the duration of clinical signs. Dogs with clinical signs prior to presentation for 35 d or less (n = 16) had a median survival of 1677 d compared with 391 d for dogs with clinical signs for greater than 35 d (n = 14); P = 0.06.

Median survival time for the surgery alone group was 510 d compared with 518 d for the surgery and chemotherapy group. This difference was not statistically significant with P = 0.46. Several factors were assessed for prognostic importance in the surgery treated group. The only factor that had a statistically significant impact on survival was the number of lobes removed (P = 0.004). The median survival for patients having only 1 lobe removed was 1462 d compared with 365 d for patients undergoing bilateral thyroidectomy (Figure 2). The statistical significance was retained upon multivariate analysis and revealed a risk ratio of death from disease of 9.1 for patients having undergone bilateral thyroidectomy (P = 0.02). There was no association between the number of lobes removed and the presence of post-operative complications (P = 0.32). There was also no association between the presence of post-operative complication and margin status (P = 0.16), fixed tumors (P = 0.68) or debulking surgery (P = 26). With respect to chemotherapy, no variable tested (number of cycles, doxorubicin versus platinum and response to chemotherapy) had a significant impact on survival. There was no association with chemotherapy administration and the development of metastasis (P = 0.18) or with time to metastasis (P = 0.89). Overall median time to metastasis was 347.5 d. It was impossible to compare this variable between the 2 treatment groups due to the small numbers.

Figure 2.

Survival curve illustrating the relationship between survival and the number of lobes removed. Dogs subjected to the removal of 1 lobe of the thyroid gland (n = 21) had a median survival of 1462 d compared with 365 d for dogs that had both lobes removed (n = 12); P = 0.004.

The possibility of a case selection bias between the surgery group and the surgery and chemotherapy group was reviewed, with regard to the presence of metastasis at diagnosis; there was a similar distribution between the 2 groups, with a P-value of 1. We also looked at the possibility of a bias with respect to duration of clinical signs and again found similar distributions within the treatment groups, with P = 1. There was no association between the duration of clinical signs and the presence of metastasis (P = 0.08).

Discussion

The patient demographics, clinical presentation, and tumor type distribution for this study population were similar to what has been previously published for canine thyroid carcinomas (2,3,9,11,12). Most cases were determined to have T2 or T3 tumors with 44.0% (11/25) being fixed. Thirty-four percent presented with metastasis, mostly located in the lungs as is consistent with previous literature (1–4). In this study, 17.6% of dogs were identified as having functional tumors (making thyroid hormone in excess). This is also consistent with previous reports (6,14).

The median survival time for patients with mobile tumors is comparable to that of previous publications with 20.5 mo median survival times reported (9). The median survival times for patients with fixed tumors was not reached with survivals ranging from 5 to 1541 d. However, 75% of dogs with fixed tumors (6/8) survived at least 1 y. This is in contrast with previous literature which reports that only 25% of dogs with fixed tumors survived at least 1 y (7). Only 1 of the 6 dogs with fixed tumors surviving at least 1 y was treated with a combination of surgery and chemotherapy. This likely reflects the important role of local treatment in the management of thyroid carcinomas as is supported by numerous publications on the role of radiation therapy in fixed thyroid carcinomas (5,11,12,15).

The number of lobes removed was a prognostic factor, with patients having bilateral thyroidectomy experiencing shorter survivals. Interestingly this was not related to the fixed or mobile nature of the tumor or with the occurrence of surgical complications. The rationale for using bilateral thyroidectomy or unilateral thyroidectomy was not provided in the medical records. However, it may be that dogs undergoing bilateral thyroidectomy had more aggressive disease involving both lobes.

The median survival time for dogs with metastasis at diagnosis (n = 15) was 506 d, which is surprisingly high. Fifty-eight percent of the dogs with metastasis at presentation were treated with surgery alone. This likely reflects the important role of local treatment even in metastatic thyroid carcinoma. It may also reflect the slow progression of the natural disease. Unfortunately, due to the small numbers of dogs in our study, the median time to metastasis could not be compared between treatment groups.

The duration of clinical signs prior to presentation may also be an important determinant of survival. It has been suggested that with the more frequent use of thyroid ultrasonography for a variety of medical conditions, incidental thyroid carcinomas may be diagnosed before they are clinically palpable, allowing for a better prognosis (2).

As with any retrospective study design, there are several limitations to our study. Treatment was based on the clinician and owner preference, and this creates a bias. Staging was not standardized throughout the study population, which created missing data, limiting the power of our statistical analysis. The power to detect differences in survival with respect to treatment was also limited and it remains possible that chemotherapy does indeed have a positive impact on survival of dogs with thyroid carcinoma. Thus, one must exercise caution when drawing conclusions. Based on the data generated from this study, it is estimated that 131 dogs would need to be included in each group in order to detect a statistical difference of 40% in survival 80% of the times using a 5% level of significance. Nonetheless this study allows for the evaluation of a significant subset of patients treated with surgery and identifies new potential prognostic factors that merit further investigation in a prospective setting.

In conclusion, there was no significant difference between the surgery-only treated group and the surgery and chemotherapy group with respect to survival. The use of surgery in combination with chemotherapy in the metastatic setting did not seem to significantly increase survival at least for the chemotherapy agents used in this study. Further studies are warranted to prove this in a randomized prospective study and to determine the impact of chemotherapy on the time to metastasis.