A retrospective study of 44 canine apocrine sweat gland adenocarcinomas

Abstract

Apocrine sweat gland adenocarcinomas (AACs) are relatively uncommon skin tumors in dogs. Little prognostic or behavioral information has been published for these tumors. In this retrospective study, 44 AACs from diagnostic archives were reexamined and clinical postexcisional follow-ups for 25 of the 44 cases were obtained by a survey. There were 28 out of 44 (65.9%) AACs that invaded the capsule, stroma, or both, 5 of 44 (11.4%) invaded blood vessels and stroma, and 1 out of 25 (4%) had distant metastases. The presence or absence of stromal and vascular invasion was predicted by clinical examination with more than 80% accuracy. Postexcisional median survival of dogs with AACs was 30 mo at the time of survey. Determination of a correlation between histological features and malignant behavior of AACs was compromised by the low number of cases with clinical AAC-related problems; however, it appears that intravascular invasion is an important indicator of potential systemic metastases.

Introduction

Apocrine sweat gland adenocarcinomas (AACs) account for 0.7% to 2.2% of all skin-associated tumors in dogs (1,2,3,4,5,6,7). Even though AACs are rather uncommon neoplasms, a busy small animal clinic or diagnostic laboratory encounters at least a few of these tumors every year. Most available veterinary literature on AACs focuses predominantly on the morphologic appearance, whereas little prognostic or behavioral data have been published. In the only retrospective study reporting postexcisional behavior of AACs, none of 26 dogs diagnosed with AAC had distant metastasis, even though lymphatic invasion was present in 9 out of 40 (22.5%) AACs (4). On the other hand, most general veterinary pathology and oncology textbooks (2,3,8,9,10), as well as several review papers (6,11,12) and individual case reports (13,14,15), indicate that canine AACs have distant metastatic potential, in addition to local invasive behavior. This conflicting information in veterinary literature about behavior of AACs inevitably results in inconsistent prognoses by diagnostic pathologists for dogs with excised AACs.

The purpose of this study was to determine postexcisional behavior of canine AACs and to identify relevant histological features for prediction of clinical behavior of AACs by using cases submitted to diagnostic laboratories and clinical data obtained from veterinarians by a survey.

Materials and methods

Skin biopsy cases submitted to the Histovet Surgical Pathology (Guelph, Ontario) and Yager-Best Histovet Histological and Cytological Services (Rockwood, Ontario) between 1987 and 1995, as well as cases submitted to the Department of Veterinary Pathology at the Western College of Veterinary Medicine (Saskatoon, Saskatchewan) between 1987 and 1998 were used for this retrospective study. All AACs diagnosed during these periods were reviewed. Tumors with questionable sweat gland origin and those originating from the skin overlying or adjacent to the mammary glands, footpads, or anus were excluded, so as to remove possible mammary, eccrine, or anal sac adenocarcinomas, some of which are extremely difficult to differentiate from dermal AACs (8).

Histological diagnosis of AAC was confirmed for the remaining 44 cases, based on the following criteria: (i) capsular, stromal, or vascular invasion; (ii) desmoplasia or loss of structural differentiation; (iii) cellular and nuclear pleomorphism; and (iv) number and morphology of mitotic figures. The origin of the tumor was the apocrine sweat gland, this was confirmed for all 44 tumors based on at least 2 of the following criteria: (i) decapitation “blebs” at the apical surface of the epithelial lining, (ii) eosinophilic secretion within lumina, and (iii) double layer of epithelial lining (one layer cuboidal/columnar and the other fusiform). Tumors were classified according to the World Health Organization International Histological Classification of Tumors of Domestic Animals (8) into the following types: solid, tubular, papillary (cystic), and ductal AACs. These main types were subdivided into complex AACs, mixed AACs, or carcinosarcomas, according to the amount, metaplasia, or malignancy of myoepithelium. The degree of invasiveness, desmoplasia, and tumor-associated inflammation and necrosis was graded from 0–4 (where; 0 = absent, 1 = minimal, 2 = mild, 3 = moderate, 4 = severe). The invasiveness of AACs was subdivided into the following subcategories: capsular invasion only (if capsule was present), deeper invasion into surrounding stroma, or vascular invasion.

A survey was performed in 1996 in Ontario and in 1999 in Saskatchewan. Survey forms were sent to the veterinary clinics that submitted the biopsies to request the following information: (i) location, size, and clinical appearance (solitary/multiple, well-circumscribed/ invasive) of the tumors; (ii) treatment; and (iii) duration before excision, postexcisional behavior, and eventual outcome.

Collected histological and clinical data were entered into a database program (FileMakerPro 4.0; FileMaker, Santa Clara, California, USA) and analyzed. The Kaplan-Meier product limit method was used to generate a postexcisional survival curve by using a statistical software (GraphPad Prism 3.00; GraphPad Software, San Diego, California, USA). Live dogs with no tumor-related problems at the time of the survey were considered in this study as censored data. Natural death or euthanasia due to tumor-related and tumor-unrelated problems was considered an endpoint used to plot a postexcisional survival curve.

In this paper, postexcisional survival period was defined as the time from excision of the tumor to either the time of death or the date on which the dog was last known to be alive. Tumor-free interval was defined as the time from excision of the tumor to the date on which the tumor first recurred, the date of death, or the date on which the absence of recurrence was confirmed, if the dog was still alive.

Results

Clinical findings

Twenty-three dogs diagnosed with AACs were male (12 castrated) and 21 were female (16 spayed). Thirty-four AACs were found in 26 different purebreds and 10 tumors occurred in crossbred dogs. The median age of the affected dogs was 9 y (Figure 1). The median size of the AACs was 2 cm in diameter (based on n = 26); 89% of tumors were solitary and 11% were multiple (based on n = 29). Anatomical locations of the AACs are presented in Table 1.

Figure 1. Age distribution of 44 dogs with canine apocrine adenocarcinoma (AACs).

Table 1.

Clinical invasiveness of AACs was determined by tumor adhesion to the underlying tissue (tumor not freely movable). These data were available for 18 tumors of which 11 were well-circumscribed and movable, whereas 7 were not freely movable and interpreted as invasive. Six of the 7 tumors with clinically determined invasive behavior had histological evidence of stromal or vascular invasion. Ten of the 11 tumors determined to be clinically well circumscribed and moveable did not have histological evidence of stromal or vascular invasion; however, 4 of these 11 had histological evidence of minimal to mild capsular invasion.

Postexcisional follow-up information was available for 25 (56.8%) of the 44 dogs. One of the 25 dogs was euthanized shortly after AAC excision due to recurrence and distant metastases, 7 of the 25 dogs were euthanized due to AAC-unrelated problems, and the remaining 17 dogs were alive and healthy without any AAC-related problems at the time of survey. The Kaplan-Meier curve of survival of dogs with excised AACs revealed that the median postexcisional survival time was 30 mo at the time of survey (Figure 2).

Figure 2. Kaplan-Meier curve of survival period of dogs with excised canine apocrine adenocarcinoma (AACs) (based on 25 out of 44 dogs with available postexcisional follow-up information).

Relation between clinical behavior and histological evidence of invasiveness

Prevalence of histological types of AACs and their histological features are presented in Table 2. Thirty-three (75%) of the 44 tumors had invasive behavior: 5 invaded the stroma and blood vessels, and 28 invaded the stroma, or the capsule, or both.

Table 2.

Postexcisional follow-up was available for only 2 of the 5 patients that had AACs with stromal and vascular invasion. One of these 2 dogs was an 11-year-old crossbred spayed female. Ten days after the original excision of a solitary AAC from her neck, there were numerable satellite nodules like a string of pearls radiating outwardly from the surgical site. Microscopic examination of these nodules revealed innumerable emboli of AAC lodged within blood vessels of the superficial dermal vascular plexus, causing ischemic injury to the overlying epithelium. Over the next several weeks, the tumor continued to spread within the skin as satellite nodules, extending over the shoulder and lateral thoracic wall. Due to poor prognosis, the dog was euthanized and a cosmetic necropsy confirmed multifocal pulmonary metastases of anaplastic AAC, which had identical histological features as the previously excised dermal tumor. The originally excised solid AAC from the neck of this dog was densely cellular, well-demarcated, expansile, and thinly encapsulated with multifocal capsular invasion by neoplastic cellular nests, some of which were surrounded by fusiform endothelial-like cells (interpreted as intravascular invasion). Clean surgical excisional margins were less than 1 mm wide and, in some areas, complete excision was equivocal. Neoplastic cells were arranged predominantly in densely packed solid lobules separated by mature fibrovascular stroma. Some lobules had densely packed papillary or ribbon-like structures formed by 2 or more layers of polygonal epithelial cells. Neoplastic cells had a moderate amount of eosinophilic cytoplasm (nuclear: cytoplasmic ratio of approximately 1:1) with distinct cellular borders; they had up to 3-fold anisokaryotic nuclei containing predominantly finely clumped chromatin and, sometimes, a prominent, centrally located single nucleolus. Mitoses ranged from 2 to 4 per high power field (400×). Multifocal necrosis affected approximately 30% of the entire tumor. Moderate lymphoplasmacytic inflammatory infiltrate was present in the fibrovascular stroma and capsule (Figure 3).

Figure 3. Histological appearance of the excised canine apocrine adenocarcinoma (AACs) (primary site) that metastasized to the multiple organs. The tumor was composed predominantly of solid nests (SN) affected by multifocal necrosis (N).

The other dog (16-year-old crossbred), which was diagnosed with papillary ACC and blood vessel invasion, was euthanized shortly after the surgical excision due to age related problems. Necropsy was not performed.

Postexcisional follow-up was available for 13 of the 28 dogs that had AACs that invaded the capsule, stroma, or both, but not blood vessels. Based on histological evaluation of 1 to 3 sections of these tumors, surgical excision was not complete in 3 of the above 13 dogs, and wider surgical excision was recommended by diagnostic pathologists. One 15.5-year-old dog was euthanized due to urinary problems 48 mo after excision of the AAC. Another 6-year-old dog died due to AAC-unrelated problems 16 mo after the AAC had been excised. The remaining 11 dogs were alive without AAC-related problems at the time of the survey. The median postexcisional tumor-free interval in these 13 dogs was 12 mo (range, 6 to 51 mo).

Discussion

Postexcisional behavior of canine AAC

In this retrospective study, the AAC-related mortality rate might have been higher than 1 out of 25 dogs, because necropsies were not performed on most of the dogs that died of causes thought to be unrelated to the previously excised AAC. There was no clinical evidence of any metastasis in any of these dogs.

Kalaher et al (4) reported that postexcisional survival ranged from 6 to 36 mo in 22 out of 26 (84.6%) dogs. The postexcisional tumor-free interval range in the present study was 3 to 51 mo (median 30 mo) in 25 dogs (Figure 2). However, since neither median nor mean values for postexcisional tumor-free intervals were reported in the Kalaher et al retrospective study, comparison of postexcisional AAC-free intervals between Kalaher et al's study and the present study is limited.

In addition, Kalaher et al reported local recurrence or continued local neoplastic growth in 2 out of 26 dogs, but no systemic metastases were observed. By comparison, in the present study, 1 out of 25 dogs was euthanized due to distant metastases of the previously excised AAC, whereas local recurrence was not present in any of the 25 dogs. In a few dogs, however, further wide surgical excision was performed after a histological diagnosis of incompletely excised AAC. If all dogs with available postexcisional follow-up information are considered in both this (n = 25) and Kalaher et al's (n = 26) study, less than 2% of canine AACs were confirmed to have fatal distant metastases. To the authors' knowledge, there are no other retrospective studies of postexcisional behavior of canine AACs, even though many single case reports (6,11,13,14,15,16,17) and general text books (1,2,3,8,10,18) contain sometimes conflicting statements regarding the risk of local recurrence and invasion, as well as distant metastases, by AACs. The low metastatic rate of dermal AAC differs from the relatively high metastatic rate of apocrine adenocarcinomas of the anal sac (19).

Correlation of histologic features with clinical outcome in dogs with an AAC

Correlation of histologic features with clinical outcome in dogs with an AAC has not been previously reported (3), even though such information is sought by all diagnostic pathologists. Determination of a correlation between histological features and clinical outcome of AACs was compromised in the present study by the small number of dogs with postexcisional tumor-related problems. Specifically, only 1 AAC had distant metastases, namely, an incompletely excised, solid, and markedly necrotic AAC that invaded blood vessels.

Follow-up information was available on only 13 of the 28 AACs with capsular or stromal, but not vascular invasion. The median postexcisional AAC-free interval for these tumors was 12 mo; however, since 11 of these dogs were still alive at the time of survey, this period is most likely significantly longer.

Determination by clinical examination of the presence or absence of stromal and vascular invasion of AACs has 86% or 81% accuracy, respectively. Capsular invasion, however, appeared to be more difficult to detect by clinical examination, because it was present in 36% of well circumscribed and moveable tumors with no clinical evidence of invasion. This is not surprising because, even when viewed through a microscope, such invasion is often subtle.

Clinical and histologic findings

The gender (2,4,10,11,16,20), breed (2,4,11,16), and site (2,4,6,11,16,20) predilection of canine AACs have been variably reported in the veterinary literature. In the present study, there was no apparent influence of gender on the prevalence of AACs, which were most commonly found on the legs, thorax, and head. The age of affected dogs and clinical appearance of AACs in the present study were consistent with previously published reports (2,4,11,16,20).

The prevalence of histological types of AACs in this study was similar to those previously reported (4,8,9,20,21). Kalaher et al reported that 22.5% of AACs invaded lymphatics. In the present study, unequivocal lymphatic invasion was found to be difficult to evaluate. Quite frequently, the authors found neoplastic clusters, nests, and islands that invaded the capsule and surrounding stroma. Many of these nests were surrounded by fusiform cells that could have been endothelial cells lining lymphatics, but they could also have been myoepithelial cells or stromal cells surrounding neoplastic clusters that invaded the capsule or surrounding stroma. The authors felt that immunohistochemical staining would have been required to unequivocally confirm lymphatic invasion by AACs; however, such staining was not performed in this study. In contrast to lymphatic invasion, invasion of blood vessels was unequivocal in 5 of the 44 AACs (11.4%) in the present study; 3 of these tumors were solid AACs. The prevalence of invasive behavior of AACs has not been reported previously in the veterinary literature. In the present study, 75% of AACs invaded the capsule, surrounding stroma or vasculature, or any combination of these.

In conclusion, even though several limitations affected this study, many of which are unavoidable in most retrospective investigations, it yielded useful information about canine AACs. Stromal and vascular invasion could be detected with over 80% accuracy by a clinical examination of an AAC. Dogs with completely excised AACs that did not invade vasculature appear to have a relatively good prognosis (30 mo median survival). Confirmed distant metastatic rate of AACs appears to be very low (2%) and it is associated with intravascular invasion.