Evaluation of adjuvant carboplatin chemotherapy in the management of surgically excised anal sac apocrine gland adenocarcinoma in dogs

Authors


Abstract

There is no widely accepted standard of care for canine anal sac apocrine gland adenocarcinoma (ASAGAC). Surgery alone is inadequate in many cases, but the benefit of adjuvant chemotherapy is not well established. The primary objective of this retrospective study was to evaluate the role of carboplatin chemotherapy in the post-operative management of ASAGAC. Seventy-four dogs with naturally occurring ASAGAC underwent surgery. Forty-four dogs received adjuvant carboplatin and 30 did not. Median overall survival (OS) was 703 days. Median time to progression (TTP) was 384 days. Only primary tumour size and lymph node metastasis at diagnosis significantly impacted the outcome. Differences in OS and TTP, between the dogs that received adjuvant carboplatin and those that did not, failed to reach statistical significance. Treatment of progressive disease, whilst not limited to chemotherapy, significantly prolonged the survival. This study shows that adjuvant carboplatin chemotherapy is well tolerated and may have a role in the management of dogs with ASAGAC.

Introduction

Anal sac apocrine gland adenocarcinoma (ASAGAC) is a malignant tumour arising from the secretory epithelium of the walls of the bilateral anal sacs. ASAGAC accounts for only 2% of all canine skin tumours and 17% of all perianal tumours in this species.[1, 2]

ASAGAC displays aggressive biologic behaviour. It is both locally invasive and highly metastatic, with a reported overall metastatic rate of between 40 and 96%.[1, 3-10] Metastasis to the regional lymph nodes is frequently reported early in the course of the disease,[5-7, 9, 10] whilst metastasis to distant sites, such as the lungs, liver, spleen and bone, is typically observed months to years later.[1, 3-7, 9]

In addition to the morbidity associated with the tumour's locally aggressive behaviour, systemic clinical signs may occur as a result of paraneoplastic hypercalcaemia.[4-8, 11-13] The hypercalcaemia associated with ASAGAC occurs because of the synthesis of parathyroid-hormone-related peptide (PTH-rp) and secretion by the tumour, and the protein's downstream endocrine sequelae.[11, 12] Paraneoplastic hypercalcaemia is reported in approximately 25% of cases of ASAGAC, although documented incidences are as high as 53%, and whilst there exists some discordance in the literature,[5, 6, 8] hypercalcaemia has been associated with decreased survival time.[4, 7, 13]

Multiple treatment modalities have been utilized in the management of ASAGAC, including surgery, radiation therapy, chemotherapy, electrochemotherapy and various combinations thereof.[5, 6, 8, 9, 13-16] Currently, there is no widely accepted standard of care for canine ASAGAC, other than to recommend surgical resection of the primary tumour and to consider adjuvant therapy. Surgery alone is often inadequate for controlling either local or metastatic disease.[4, 5, 7] With surgery alone, local recurrence occurs in 20–50% of dogs with ASAGAC,[4, 5] and metastasis occurs in 40–96%.[1, 3-10] The longest reported OS times were achieved in dogs that received multimodal therapy, consisting of a combination of surgery, radiation therapy and chemotherapy.[6]

Numerous chemotherapeutic agents have been investigated for both the adjuvant and palliative management of ASAGAC, including the platinum compounds carboplatin and cisplatin, actinomycin-D, the synthetic anthracycline/acenedione mitoxantrone, melphalan, and more recently the tyrosine kinase inhibitor toceranib phosphate (Palladia™).[5, 6, 8, 9, 14, 16] Despite this, the role of adjuvant chemotherapy in the treatment of ASAGAC remains poorly defined. Undoubtedly, this is a consequence of inconsistent treatment, diagnostic and monitoring regimens, study population heterogeneity and resultant under-powering of statistical analyses and conclusions.

The primary objective of this study was to retrospectively evaluate the role of adjuvant carboplatin in the management of surgically excised ASAGAC in a relatively homogenous population of dogs with consistent diagnostic, treatment and monitoring regimens. The authors' hypothesis was that adjuvant carboplatin chemotherapy would improve time to progression (TTP) and/or OS time in dogs with AGASAC relative to dogs treated with surgery alone. The secondary objective of this study was to evaluate prognostic factors that may be used to guide treatment recommendations.

Materials and methods

Case selection and data collection

Cases with a definitive ante-mortem diagnosis of ASAGAC were identified from a search of the medical records of the University of Wisconsin-Madison Veterinary Medical Teaching Hospital (UW-M VMTH), spanning a 10-year period, between September 2001 (9/10/2001) and December 2011 (12/6/2011).

Inclusion criteria applied to the cases required a histopathologically confirmed diagnosis of ASAGAC for which surgery was the primary treatment option employed, as well as complete staging, consisting of history, physical examination including per rectum digital examination, hematologic, biochemical and urine analyses, thoracic radiography and abdominal ultrasound. For those cases receiving adjuvant carboplatin chemotherapy, dogs were included on an intent-to-treat basis. This institution's standard protocol consisted of an intended four doses of carboplatin to be administered at 300 mg m−2 intravenously every 3 weeks. A 10–20% dose reduction is used for dogs weighing less than 10 kg and a 20% dose reduction is used for dogs developing any grade IV toxicity.

Cases were excluded if surgery was not the initial treatment option employed, if radiation therapy was incorporated in the initial treatment plan, or if the adjuvant chemotherapy protocol administered was not the aforementioned standard protocol. Cases suspected of having extra-nodal metastatic disease identified during staging were excluded from the study. Cases were also excluded if inadequate information was available, specifically a lack of detail regarding the diagnostic, staging and monitoring tests, and a duration of follow-up of less than 30 days.

Information retrieved from the medical records included signallment, presenting clinical signs, relevant physical examination findings, results of clinicopathologic testing, including hematologic, biochemical and urine analyses, radiographic, ultrasonographic and advanced imaging [computed tomography (CT)] findings, a surgical description, the details of adjunctive chemotherapy, and any rescue and/or palliative treatments instituted, as well as outcomes. The referring veterinarian(s) and/or owner(s) were contacted for follow-up where additional details were required.

Toxicities were graded according to the Veterinary Co-operative Oncology Group's common terminology criteria for adverse events (VCOG-CTCAE) following chemotherapy or biological antineoplastic therapy in dogs and cats v1.1.[16]

Statistical analysis

For the purposes of fulfilling the study's primary objective, cases were divided into two subgroups; those dogs that underwent surgery alone and those that underwent surgery followed by adjuvant chemotherapy.

Descriptive and comparative statistics were performed for the overall study population, as well as the two main subgroups. Fisher's exact test was used to compare categorical non-numerical data, such as gender, between groups. The Wilcoxon rank-sum test was used to compare continuous numerical data, such as age and weight, between groups. Results were considered significant at P < 0.05.

Two measures of outcome were evaluated. TTP was defined as the time (in days) from the date of surgery until the date of confirmed disease progression (either local recurrence or metastatic disease). OS was defined as the time (in days) from the date of surgery until the date of death as a result of ASAGAC or another cause. Data was considered right-censored at the time of last veterinary contact if the dogs remained alive at the conclusion of the study period, or if they were ultimately lost to follow-up. Kaplan-Meier survival curves were estimated for both outcomes, along with median survival times and confidence intervals (CIs). OS and TTP were compared between the study's two main subgroups using the log-rank test. Results were considered significant at P < 0.05.

The effect of variables, including signallment, primary tumour characteristics, presence or absence of clinical signs at the time of diagnosis, presence or absence of hypercalcaemia at the time of diagnosis, presence or absence of metastasis at the time of diagnosis, the location and extent of metastasis, the administration of adjuvant chemotherapy and the administration of various rescue therapies at the time of progression, on OS and TTP was evaluated using Cox proportional hazard models. Variables were first fit singly, and those that were significant were then used as a potential predictor set for a backward stepwise procedure. Adjuvant chemotherapy was retained at all steps so that its contribution could be evaluated in the presence of other potentially confounding variables. All other variables that had P < 0.05 were considered significant and retained in the model.

All statistical analyses were performed utilizing two commercially available software packages Microsoft® Excel® for Mac 2011 (version 14.2.5) and R for Mac OS X GUI 2012 (version 2.15.2).

Results

Patient population

One hundred and twenty-four cases of ASAGAC were treated surgically during the study's time period. Of these cases, 74 dogs fulfilled the study's inclusion criteria. Thirty dogs (40.5%) received adjuvant carboplatin chemotherapy, whilst 44 dogs (59.5%) did not receive adjuvant chemotherapy. Nineteen dogs were excluded because of the primary use of non-carboplatin chemotherapy protocols. Ten dogs were excluded because radiation therapy was incorporated into the initial treatment regimen. In six cases, surgery was not the primary treatment modality employed. In only one case, the presence of diffuse pulmonary metastatic disease precluded inclusion. Fourteen cases were excluded as a result of inadequate historical, diagnostic, staging or monitoring data, or duration of follow-up not exceeding 30 days.

The median age of all the dogs in the study population was 9.71 years (range: 6.06–16.15 years). Forty-eight dogs (65%) were castrated males and 26 were spayed females (35%). There were no intact dogs in the study population. Twenty-nine different breeds were represented in the study population. Labradors and their crosses predominated (n = 17), followed by golden retrievers (n = 7), German shepherds (n = 7), Siberian huskies (n = 6), Springer spaniels (n = 5), dachshunds (n =3), poodles (n = 3), cocker spaniels (n = 2), beagles (n = 2), border collies (n = 2), Rottweilers (n = 2), and one each of 18 other breeds. The median weight for the dogs was 28.8 kg (range: 2.4–51.0 kg). The demographic characteristics were similar between the study's two main subgroups (Table 1).

Table 1. Demographic characteristics for the overall study population and both major subgroups; dogs with ASAGAC that underwent surgery alone and dogs with ASAGAC that underwent surgery followed by adjuvant carboplatin chemotherapy
Parameter evaluatedOverall study population(n = 74)Surgery alone (n = 30)Surgery with adjuvant chemotherapy(n = 44)Significance(P-value)
Age (years)
Median9.7110.179.45P = 0.126
Range6.06–16.156.10–16.156.06–14.49 
Gender
Males48 (65%)23 (77%)25 (57%)P = 0.090
Females26 (35%)7 (23%)19 (43%) 
Weight (kg)
Median28.831.228.6P = 0.222
Range2.4–51.08.2–51.02.4–48.0 
Hypercalcaemia12 (16%)2 (7%)10 (23%)P = 0.107
Mild312P = 0.697
Moderate514 
Severe404 

Diagnostics and staging

All dogs underwent complete staging, consisting of history, physical examination incorporating per rectum digital examination, hematologic, biochemical and urine analyses, thoracic radiography and abdominal ultrasound.

In 43% of the overall study population (n = 32/74), with similar percentages in both major subgroups (surgery alone n = 14/30; adjuvant chemotherapy n = 18/44), the disease was incidentally identified during a routine clinical examination or as part of the work-up of an unrelated problem. Most common clinical signs reported for the remaining dogs were similar to those previously reported for dogs with ASAGAC, and included scooting or apparent perineal irritation (n = 12), tenesmus (n = 11), polyuria and polydipsia (n = 9) and a noticeable perianal mass or swelling (n = 7). Less frequently reported clinical signs included a dull demeanour (n = 6), inappetence to anorexia (n = 3), hind limb lameness (n = 3), diarrhoea or inappropriate defecation (n = 2), and dysuria, stranguria or inappropriate urination (n = 1). The proportion of dogs displaying clinical signs was not statistically different between both major subgroups (P = 0.809). The presence or absence of specific clinical signs was not statistically analysed, given small numbers.

Of the overall study population, 12 dogs (16%) were hypercalcaemic at the time of diagnosis. Of these, three dogs (25%) were classified as having mild hypercalcaemia, five (42%) as having moderate hypercalcaemia and four (33%) with severe hypercalcaemia. The degree of hypercalcaemia was classified according to previously published ranges of the University of California-Davis School of Veterinary Medicine Endocrinology Laboratory.[17] The proportion of hypercalcaemic dogs did not statistically differ between the surgery alone and adjuvant chemotherapy subgroups (n = 2/30, 7%; n = 10/44, 23%) (P = 0.107), nor did the proportion of dogs in each of the extent categories (P = 0.697).

Except for the presence of hypercalcaemia, there were no notable or repeatable hematologic or biochemical abnormalities identified at the time of diagnosis neither for the overall study population nor for either of the major subgroups. Most panels were considered unremarkable. Six dogs had mildly elevated liver enzymes, three dogs were mildly anaemic and two dogs had a mild thrombocytosis. One dog was diagnosed with International Renal Interest Society (IRIS) stage II chronic kidney disease.

Nineteen dogs (26%) had caudal abdominal lymphadenopathy as determined by abdominal imaging at the time of diagnosis. The proportion of dogs with caudal abdominal lymphadenopathy did not statistically differ for the surgery alone and adjuvant chemotherapy groups (n = 6/31, 20%; n = 13/44, 30%) (P = 0.424). Of the 19 dogs with enlarged lymph nodes identified on imaging, only five had metastasis confirmed by a pre-surgical fine needle aspirate and cytology, but all dogs with suspected lymph node metastasis underwent surgical lymph node extirpation and subsequent histopathologic evaluation for confirmation. Enlarged lymph nodes on imaging were invariably metastatic.

No repeatable abnormalities were revealed by thoracic radiography at the time of diagnosis. One dog had sternal lymph node enlargement evident on thoracic radiographs. This dog did have extensive mid-to-caudal abdominal lymph node enlargement evident on both abdominal ultrasonography and subsequent CT of the abdomen, which was cytologically diagnosed as multifocal metastasis. Cytologic evaluation of a CT-guided fine needle aspirate sample from the sternal lymph node also confirmed the metastasis to that site. This dog was one of only two that received advanced imaging (CT). In the other dog, CT of the abdominal cavity was intended to aid surgical planning and merely confirmed the presence of caudal abdominal lymphadenopathy.

Metastasis

The overall metastatic rate at the time of diagnosis was approximately 26% (n = 19/74), and did not statistically differ for the surgery alone and adjuvant chemotherapy groups, approximately 23% (n = 6/30) and 30% (n = 13/44), respectively (P = 0.424). Eighteen dogs (24%) had solely caudal abdominal lymph node metastases, and one dog had metastatic extension to the sternal lymph node.

Surgery

All dogs underwent anal sacculectomy as primary treatment. All 19 (25%) dogs with caudal abdominal lymphadenopathy had concurrent lymph node extirpation performed. The one dog with sternal lymphadenopathy also had a concurrent thoracotomy, reducing the disease burden to microscopic post-operatively. Lymph node extirpation was not performed if the intra-abdominal lymph nodes were clinically assessed as normal based on ultrasound imaging by a board-certified radiologist. The proportion of dogs that underwent lymph node extirpation was similar for both major subgroups (surgery alone n = 6/30, 20%; adjuvant chemotherapy n = 13/44, 30%) (P = 0.424).

The study population's overall reported surgical complication rate was 15% (n = 11/74). The surgical complication rate once again was not different for each of the major subgroups (surgery alone n = 4/30, 13%; adjuvant chemotherapy n = 7/44, 16%) (P = 1.0). These comprised peri-operative hemorrhage (n = 4), surgical site seroma formation (n = 3), surgical site dehiscence (n = 3) or herniation (n = 1). One dog, however, underwent a pull-through procedure because of dehiscence, which resulted in unacceptable post-procedural fecal incontinence and ultimately euthanasia. There was an increased complication rate in those dogs that had concurrent caudal abdominal lymph node extirpation (n = 8/19, 42%) compared with those that did not (n = 3/55, 5%) (P = <0.005).

Primary tumour characteristics and histopathology

Of the overall study population, 42 dogs (57%) had left-sided masses, 29 dogs (40%) had right-sided masses and three dogs (4%) had bilateral disease. It was found that two dogs had bilateral anal sacculectomy based on per rectum digital examination and suspicion of bilateral disease, although neoplasia was only identified histopathologically in the left anal sac in both dogs. Three dogs had bilateral anal sacculectomy performed because this was the surgeon's preference, despite disease being clinically suspected and ultimately histopathologically identified on only one side. These proportions were not statistically different between the surgery alone and adjuvant chemotherapy subgroups (P = 0.765). The median size of the mass(es) at the time of diagnosis was 2.10 cm (range: 0.3–9.0 cm) and there was no significant difference between the two major subgroups (surgery alone median: 2.20 cm, range: 0.8–9.0 cm; adjuvant chemotherapy median: 2.05 cm, range: 0.3–6.80 cm) (P = 0.611).

Surgical margins were variably described as incomplete or complete. However, for those cases in which the margins were described as ‘complete’, scrutiny of the histopathology report typically revealed statements such as ‘1–2 mm complete margins’ or ‘completely removed with narrow (unspecified) margins’. Given these histopathologic descriptions, what is known about the surgical technique of anal sacculectomy and the institution's clinical assumption that anal sacculectomy for removal of ASAGAC is invariably incomplete when the measurable extent of the microscopic margins is critically evaluated; this parameter was not statistically analysed.

Chemotherapy

For those dogs included in the surgery with adjuvant chemotherapy group, all dogs received at least one dose of carboplatin administered intravenously within 33 days of surgery. Median time to commencing chemotherapy was 17 days (range: 11–33 days). Median number of doses administered was four (range: 2–4 doses). The median dose administered was 300 mg m−2 (range: 240–300 mg m−2).

Forty dogs (91%) completed the planned four doses of carboplatin, with minimal chemotherapy-associated adverse events. The overall rate of chemotherapy-associated adverse events was 30% (n = 13/44), whilst the percentage of dogs that experienced dose-limiting side effects was 18% (n = 8/44). Reported side effects included hematologic abnormalities, both neutropenia and thrombocytopenia, and gastrointestinal upsets, varying from solely inappetence to vomiting and/or diarrhea. Three dogs (7%) developed side effect(s) resulting in cessation of therapy, including two grade IV neutropenias and associated illness, and one grade III neutropenia in a dog also presenting with neurologic signs. One dog ceased chemotherapy because of progressive disease.

Outcomes

Forty-six dogs (62%) died or were euthanized during the study period, whilst the remaining 28 dogs (38%) were censored either alive (n = 17) or lost to follow-up (n = 11). Thirty dogs (41%) died or were euthanized as a result of progressive disease, and 15 dogs (20%) died or were euthanized as a result of unrelated illness or trauma. The cause of death was not recorded for one dog in the surgery with adjuvant chemotherapy group. Of the two major subgroups, 20 (67%) in the surgery alone and 26 dogs (59%) in the adjuvant chemotherapy groups, respectively, died during the study period. This difference was not statistically significant (P = 0.217). The apparent differences in cause of death or censoring between the two major subgroups, where 10 (33%) and 20 dogs (45%) in the surgery alone and adjuvant chemotherapy groups, respectively, died as a result of progressive disease, 10 (33%) and 5 dogs (11%), respectively, died of unrelated illness or trauma and 7 (23%) and 4 dogs (9%) were lost to follow-up, were also not statistically significant (P = 0.086).

Forty-one dogs (55%) developed progressive disease during the study's follow-up period. Of the two major subgroups, 15 dogs (50%) in the surgery alone and 26 dogs (59%) in the adjuvant chemotherapy groups, respectively, developed progressive disease. This difference was not statistically significant (P = 0.638). Manifestations of progressive disease included local recurrence (n = 18, 44%), locoregional lymph node metastasis (n = 27, 66%) and pulmonary progression (n = 7, 17%). With regard to locoregional recurrence or metastasis, 9 dogs (22%) developed local recurrence, 16 dogs (39%) developed locoregional lymph node metastasis and 5 dogs (12%) a combination of both. For dogs developing distant metastasis, one dog (2.5%) had pulmonary metastasis only, one dog (2.5%) both pulmonary metastasis and local recurrence, two dogs (5%) both pulmonary and locoregional lymph node metastasis and two dogs (5%) all three. One dog (2.5%) developed locoregional lymph node metastasis and pelvic osteolytic lesions consistent with bone metastasis. One dog (2.5%) developed renal metastasis in addition to locoregional lymph node and pulmonary metastasis. In three dogs (7%), the site of metastasis was not recorded. The proportion of dogs with locoregional lymph node metastasis or pulmonary metastasis did not differ between the two subgroups (P = 0.457, P = 0.972, respectively). Interestingly, there was a statistically significant difference in the proportion of dogs that developed local recurrence between the two subgroups, with dogs that received carboplatin being less likely to develop local recurrence (surgery alone n = 10/15, 67%; adjuvant chemotherapy n = 8/26, 31%) (P = 0.016).

Not all dogs that developed progressive disease died (n = 6), and despite having progressive disease three dogs died of unrelated causes (n = 3). Treatment for progressive disease was pursued in 28 cases (n = 28/41, 68%). Importantly, treatment for progression was far more likely to be undertaken in the group of dogs that initially received adjunctive chemotherapy (P = 0.001). Treatment for progressive disease included various combinations of repeated surgery, radiation therapy and chemotherapy, in addition to palliative measures, such as analgesia. Of the dogs with progressive disease, those that received treatment had a significantly longer OS (P < 0.005) compared with those that did not, with a median survival time from progression to death of 374 days for those dogs that received treatment compared with 47 days for those that did not.

Median OS was 703 days (95% CI: 594–894) with a range 5–2402 days (Fig. 1). Median overall TTP was 384 days (95% CI: 280–745) with a range 58–745 days in those dogs with confirmed disease progression (Fig. 2). The median OS and TTP in those dogs that received adjuvant chemotherapy were 723 days (95% CI: 620–1116) (range: 28–1721 days) and 382 days (95% CI: 287 to NA) (range: 58–745 days), respectively. The median OS and TTP in those dogs that underwent only excisional surgery were 581 days (95% CI: 394–948) (range: 5–2402 days) and 402 days (95% CI: 243 to NA) (range: 91–469 days), respectively. These differences were not statistically significant (OS P = 0.174; TTP P = 0.370).

Figure 1.

Estimated Kaplan-Meier survival curve for overall survival times in 74 dogs with ASAGAC treated with surgery +/− adjuvant carboplatin chemotherapy.

Figure 2.

Estimated Kaplan-Meier survival curve for time to progression in 74 dogs with ASAGAC treated with surgery +/− adjuvant carboplatin chemotherapy.

Several clinical variables were identified as related to outcome on univariate analysis, including the presence of clinical signs at diagnosis (Table 2). However, only size of the primary tumour (OS P = <0.021; TTP P = 0.050) and presence of metastasis at the time of diagnosis (OS P = 0.042; TTP P = 0.018) were significant in the multivariate model (Figs 3 and 4). The median OS and TTP in those dogs that did not have metastasis were 761 days (95% CI: 703–1116) (range: 14–2402 days) and 594 days (95% CI: 346 to NA) (range: 68–745 days), respectively. The median OS and TTP in those dogs that did have metastasis were 448 days (95% CI: 270 to 885) (range: 5–946 days) and 134 days (95% CI: 103 to NA) (range: 58–476 days), respectively.

Table 2. Results of the univariate analysis with respect to overall survival (OS) and time to progression (TTP)
 OSTTP
Variableexp(coef)aP-valueexp(coef)aP-value
  1. aExponential (coefficient) is defined as the multiplied relative risk of an event (death or progression) occurring for a unit change in the variable.
  2. *Statistically significant at P = <0.05.
Age1.11130.1421.01590.833
Gender0.81260.5161.33020.420
Weight1.00980.4131.00070.956
Clinical signs1.98850.036*1.71040.106
Hypercalcaemia1.85220.1491.49000.342
Metastasis (lymph node)3.02450.001*3.8686<0.001*
Side1.27670.5941.50370.453
Size1.2984<0.001*1.23780.007*
Lymph node extirpation3.02450.001*3.8686<0.001*
Complications3.4844<0.001*11.5647<0.001*
Chemotherapy0.74150.3241.01630.960
Start days1.00270.9331.03680.251
No of treatments0.60780.1930.90330.783
Toxicity0.71750.4850.74210.501
Dose reductions0.63560.4160.66150.448
Progression2.85550.004*NANA
Local progression0.46510.047*0.47820.034*
Lymph node progression1.18060.6901.26250.520
Pulmonary progression1.31020.5391.76320.187
Progression elsewhere2.14530.3061.51090.579
Treatment for progression0.40430.020*1.14610.697
Figure 3.

Estimated Kaplan-Meier survival curves for overall survival times in dogs with ASAGAC diagnosed with (dashed) or without (solid) metastasis at the time of initial diagnosis. The median overall survival time of 448 days in dogs with metastasis at the time of initial diagnosis was significantly shorter (P = 0.042) than the 761-day median overall survival time in dogs that did not have metastasis.

Figure 4.

Estimated Kaplan-Meier survival curves for time to progression in dogs with ASAGAC, with (dashed) or without (solid) metastasis at the time of initial diagnosis. The median time to progression of 134 days in dogs with metastasis at the time of initial diagnosis was significantly shorter (P = 0.018) than the 448-day median time to progression in dogs that did not have metastasis at the time of initial diagnosis.

Lymph node extirpation and the related issue of post-operative complications were both significantly associated with poorer outcomes. Dogs that underwent lymph node extirpation had a median OS and TTP of 448 and 134 days, respectively (OS P = 0.001; TTP P = <0.001), whilst dogs that did not undergo lymph node extirpation had a median OS and TTP of 764 and 594 days, respectively (P = <0.001; P = <0.001).

The administration of chemotherapy had no statistically detectable impact on OS or TTP in dogs with ASAGAC, although the group that did not have metastasis at the time of diagnosis and did receive adjuvant chemotherapy had the longest median TTP and OS (Figs 5 and 6).

Figure 5.

Estimated Kaplan-Meier survival curves for overall survival times in dogs with ASAGAC, with or without metastasis at the time of initial diagnosis, and that did or did not receive adjuvant carboplatin chemotherapy. Dogs that did not have metastasis at the time of diagnosis and received adjuvant carboplatin chemotherapy (dotted) had the longest median survival time of 952 days, whilst dogs that did have metastasis at the time of initial diagnosis and did not receive adjuvant carboplatin chemotherapy (dashed) had the shortest median survival time of 448 days.

Figure 6.

Estimated Kaplan-Meier survival curves for time to progression in dogs with ASAGAC, with or without metastasis at the time of initial diagnosis, and that did or did not receive adjuvant carboplatin chemotherapy. Dogs that did not have metastasis at the time of diagnosis and received adjuvant carboplatin chemotherapy (dotted) had the longest median time to progression of 745 days, whilst dogs that did have metastasis at the time of initial diagnosis and did receive adjuvant carboplatin chemotherapy (alternating dashed & dotted) had the shortest median time to progression of 113 days.

Discussion

The primary objective of this study was to investigate the role of carboplatin chemotherapy in the microscopic disease setting following anal sacculectomy +/− metastatic lymph node extirpation. Survival analyses, however, failed to unequivocally prove the authors' primary hypothesis; that adjuvant carboplatin chemotherapy would improve measures of outcome in dogs with AGASAC. Median OS in the study population was, however, longer than earlier reports of between 6 and 11 months[4, 5] and more comparable to the recent studies of Williams et al.[7] and Turek et al.[6] Median TTP was also no longer considered in this study compared to previous reports.[5, 7] Although comparison of outcomes between dissimilar studies is suboptimal, the prolonged values reported in this study might reflect improvements in the diagnosis, monitoring and treatment of canine ASAGAC over time, as well as a change in society's attitude towards dog ownership and veterinary care. The significant impact of treatment on survival time, even in the face of progressive disease, supports this theory, and is an important information for veterinarian and owners.

The currently recommended adjunctive chemotherapy protocol employed at UW-M VMTH and utilized in this study consists of four doses of carboplatin administrated intravenously every 3 weeks. The use of carboplatin is based on the previously documented efficacy (33%) of the platinum agents in macroscopic ASAGAC.[5] Generally, chemotherapy is considered more effective against microscopic disease. It is interesting, therefore, that the administration of platinum agents for metastatic AGASAC in the microscopic disease setting has failed to improve survival compared with surgery alone.[7, 9] In both the aforementioned studies, the population numbers were small, and the chemotherapy protocols were not fully described and/or standardized. In Williams et al.'s[7] retrospective evaluation of 113 dogs with ASAGAC, either carboplatin or cisplatin was administered to 28 dogs, but the details of the protocols were not described.[7] In Polton and Brearley's retrospective evaluation of 80 dogs with ASAGAC, approximately half the dogs received carboplatin administered in the adjuvant microscopic setting, the neoadjuvant peri-operative setting and the palliative macroscopic setting, but once again the details of the protocols were not elaborated further.[9] To the authors' knowledge, this is the first study investigating the efficacy of a standardized carboplatin protocol as adjunctive treatment in the management of surgically excised AGASAC and comparing measures of outcomes between the dogs receiving adjunctive chemotherapy and a contemporaneous cohort of dogs, with similar clinical characteristics, that did not receive adjunctive chemotherapy.

Carboplatin was well tolerated by the dogs in this study. Despite an overall adverse event rate of 18%, most were mild, with only two dogs developing grade IV toxicities (neutropenia), and 91% of dogs completing the standard protocol. The incidence of dose-limiting side effects was 18%, which is higher than the incidence of dose-limiting side effect(s) for carboplatin when used as adjunctive therapy for osteosarcoma as previously reported by Saam et al., but seemingly lower than that reported by Phillips et al.[17, 18] During the study period, hemograms were routinely assessed for 7–10 days after the treatment. More recent studies suggest that hemograms obtained on days 14 and 21 may provide more information about treatment-induced myelosuppression relative to hemograms assessed at 7–10 days.[17] Consequently, the incidence of neutropenia and thrombocytopenia may have been underestimated in this study.

Breed, body weight and age of the dogs in this study were similar to previous reports of dogs with ASAGAC.[1, 3-5, 7, 10] However, 65% of the dogs in this study were castrated males. This seems in contrast to the older studies in which females were overrepresented,[1, 3, 4] and even to more recent studies in which the proportions of males and females were approximately equal.[5, 7, 10, 19] However, our study population was not compared with all hospital admissions during the time period, and males may have exceeded females, thereby explaining this anomaly.

ASAGAC was incidentally identified in 43% of the cases in this study. This is a much higher percentage of cases compared with previous publications, in which although 50% of dogs presented with no clinical signs obviously suggestive of anal sac disease, only 7% were diagnosed in dogs on routine examination with no owner complaints of any clinical signs attributable to either primary or metastatic ASAGAC.[5] This may be a result of an increasing awareness of the disease by veterinarians, as well as individuals peripherally involved in canine care, such as dog groomers.

Only 16% of the dogs in this study were hypercalcaemic at the time of diagnosis. Again, this is in contrast to previous studies in which the reported incidences of paraneoplastic hypercalcaemia were higher, ranging between 25 and 53%.[4-8] The lower incidence of hypercalcaemia in this study may be due to the higher proportion of dogs diagnosed incidentally, implying minimal tumour burdens and/or less advanced disease. Although there has been no previously reported association between tumour size and hypercalcaemia, this might not apply with incidentally identified disease, which perhaps infers an even smaller tumour burden. Selection bias may also have influenced this parameter, as it is possible that the owners of a proportion of dogs diagnosed with AGASAC during the study's time period may have elected not to pursue surgery at the outset. All hypercalcaemic dogs became eucalcaemic following surgery, irrespective of adjuvant chemotherapy, which supports the previous recommendation to perform excision of the mass and affected lymph nodes, even in a palliative setting, in order to control the pathophysiologic sequelae associated with hypercalcaemia.[11-13] Hypercalcaemia was ultimately not associated with measures of outcome in this study. This issue remains unresolved in the literature.[5-8, 13]

The metastatic rate reported in this study is similar to that reported previously.[1, 3-10] Moreover, the pattern of progression observed in this study, characterized by early local lymph node metastasis and later extension to distant sites, is also similar to that described previously.[1, 3-7, 9, 10] It was interesting that dogs that received adjuvant carboplatin were less likely to develop local recurrence, but a prospective randomized clinically controlled trial would be required to confirm the role of chemotherapy in inhibiting the local disease recurrence.

In this study, all dogs with suspected lymph node metastases had lymph node extirpation performed concurrently with anal sacculectomy. This study did not address the potential for microscopic metastases in normal-appearing lymph nodes and the clinical implications thereof, as lymph nodes appearing ultrasonographically normal were not investigated further. The impact of microscopic metastases is yet to be investigated in dogs with ASAGAC. Previous studies have shown that despite grossly apparent metastases to the iliac lymph nodes, dogs with ASAGAC can have long-term survival when surgery incorporates excision of the metastatic lesion(s).[9, 13, 20, 21] Lymphadenectomy also assists in controlling the paraneoplastic hypercalcaemia associated with ASAGAC.[13, 20, 21] These prior surgical reports had a low complication rate, even in dogs in which pubic osteotomy was performed.[13, 20, 21] Pubic osteotomy was not necessary in any of the dogs in this study. The overall surgical complication rate was low in this study, but there was an increased complication rate in those dogs that underwent lymph node extirpation compared with those that did not. However, most complications were minor and readily managed. In agreement with previous reports,[13, 20, 21] therefore, the authors recommend excision of metastatic lymph nodes.

This study did not analyse the impact of incomplete or apparently complete surgical margins on measures of outcome. The lack of assessment is due to our institution operating under the clinical assumption that margins for surgical excision of ASAGAC are invariably incomplete, when the measurable extent of the microscopic margins are critically evaluated.[7] The anal sacs are cutaneous diverticula of the stratified squamous epithelium originating at the anocutaneous junction, in between the internal and external anal sphincters, into which adjacent apocrine glands empty their secretions.[6] Given the complex anatomy of the region, anal sacculectomy is a necessarily marginal procedure in order to avoid undesirable post-operative adverse events, such as fecal incontinence.

This study possesses several limitations. A limitation inherent to all retrospective studies is the lack of case standardization and deficits in medical record keeping. In this study, the investigators endeavoured to standardize the study population, diagnostics, treatment and monitoring, such that there were no statistically appreciable differences between the study's two main subgroups. Cox regression models theoretically account for the statistical impact of intertwining clinical variables, but they cannot completely eliminate the bias. Given that the study's two main subgroups were neither randomized nor blinded, there did exist the potential for selection and client biases. Individual clients may be more inclined to treat aggressively, more likely to pursue treatment in the setting of progressive disease and less likely to euthanatize early. Such client bias can confound the analysis of benefit conferred by therapeutic interventions, and in this study this may be particularly applicable to the apparent benefit of treating progressive disease. A randomized, blinded and clinically controlled study would be necessary to investigate this clinical question further. Selection bias can also confound statistical analyses, and may be one reason why adjuvant chemotherapy did not ultimately confer a statistically significant survival advantage.

It is argued that OS is not a valid endpoint for veterinary studies given the aforementioned client bias, and that measures of outcome such as TTP and disease-free survival (DFS), are more objective measures of therapeutic response. Even TTP and DFS, however, are not without problems, because non-uniform and infrequent re-evaluation and follow-up of patients can result in the false impression of increases in outcome.

Conclusion

Dogs with metastatic ASAGAC at the time of initial diagnosis had a poorer prognosis than those dogs without metastases. Whilst surgical extirpation of metastatic regional lymph nodes is associated with an increased risk of post-operative complications, these are typically only minor, and aggressive surgical management is warranted in such cases to improve survival time. In this study, adjuvant carboplatin chemotherapy was deemed safe and well tolerated; however, the administration of chemotherapy was not associated with any statistically significant improvement in OS or TTP. The decrease in local recurrence in dogs receiving adjuvant chemotherapy is an interesting finding; however, as noted, the TTP was not different between the two groups, nor was the development of locoregional metastatic disease. The positive impact of treating progressive disease, albeit not limited to adjuvant chemotherapy, on survival time is an important finding. The limitations of this study serve to emphasize the need for prospective randomized clinical trials to firmly establish the efficacy and benefit of the various chemotherapeutic options for canine ASAGAC and to facilitate comparison of chemotherapeutic protocols and multimodality regimens in the management of the disease.

Acknowledgements

The project described was supported by the Clinical and Translational Science Award (CTSA) program, through the NIH National Center for Advancing Translational Sciences (NCATS), grant UL1TR000427 (TJS). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

The authors would also like to acknowledge the invaluable contribution of the University of Wisconsin-Madison Veterinary Medical Teaching Hospital's clients and referring veterinarians in collating the data for this study.

This work was completed at the University of Wisconsin-Madison.

Conflicts of interest

The authors declare no conflicts of interest.

Ancillary