Introduction

Mammary gland tumors (MGTs) are the most common neoplasm diagnosed in intact female dogs.^1–6 Overall incidence in a population varies based on whether dogs are routinely spayed for preventative medicine. In the US, where canine ovariohysterectomy is commonly performed at a young age, MGTs are somewhat less prevalent, but are still of significant concern.^7,8 Previous studies have shown that malignancy rates of canine MGTs generally range from 15% to 53%, with higher malignancy and incidence rates reported in European countries^6,9 and lower rates reported in certain parts of the US.⁷

Dogs that present primarily for MGT removal can be considered nonincidental cases, while dogs presenting for a different medical condition that have MGTs diagnosed during physical examination can be considered incidental cases. The primary presenting complaint of dogs with incidentally diagnosed MGTs varies. Intact female dogs or dogs spayed later in life are at increased risk of developing MGTs relative to dogs spayed before their first or second estrus cycle,¹⁰ and pyometra is a common presenting complaint of middle-aged to older intact female dogs, affecting up to 25% of intact female dogs before the age of 10.^11–13 Therefore, due to their intact hormone status, it is not uncommon for dogs that present with pyometra to have incidental MGTs found on examination.^13,14 Mastectomy or lumpectomy may be recommended at the time of ovariohysterectomy or another surgical procedure in a stable patient with concurrent incidental MGTs. To the authors’ knowledge, there has not been a study evaluating the malignancy rate of incidentally diagnosed MGTs.

The goal of this study was to determine whether MGTs found incidentally on examination in a population of dogs presenting for various surgical diseases would differ in malignancy compared with nonincidental MGTs. We hypothesized that there would be no significant difference in the proportion of malignant tumors between those that were incidental and those that were nonincidental. We also evaluated for any criteria that could be used to predict malignancy in canine MGTs.

Methods

Results

One hundred forty-three dogs had MGTs removed during the study period. There were 96 dogs that met the inclusion criteria. Of these, 44 of 96 dogs (46%) were presented with incidental MGTs, and 52 of 96 dogs (54%) were presented with nonincidental MGTs. Of the dogs with incidental MGTs, 33 of 44 (75%) were presented for pyometra, and 11 of 44 (25%) were presented for other causes. These 11 dogs were presented for various other reasons: 7 were presented for a separate reproductive issue (vaginal mass [n = 3], persistent vaginal discharge that was not consistent with a diagnosis of pyometra [2], pregnancy with a dead fetus [1], and pseudopregnancy/mastitis [1]), 2 were presented for tumors elsewhere in the body (1 each in the maxilla and tarsus), and 1 each were presented for gastric dilatation-volvulus and cranial cruciate ligament tear.

There were a total of 195 MGTs removed. Of these, 88 of 195 MGTs (45%) were incidentally diagnosed, and 107 of 195 MGTs (55%) were considered nonincidental. Of the incidental MGTs, 66 of 88 MGTs (75%) were removed from dogs that were presented for pyometra, and 22 of 88 MGTs (25%) were removed from dogs that were presented for nonpyometra causes.

Of the 44 dogs with incidental MGTs, 31 (70%) were diagnosed with MGTs on preanesthetic examination at our hospital by the attending emergency doctor or surgery resident/intern in charge of the case, 8 (18%) were diagnosed with MGTs when they were being shaved for surgery under anesthesia, and 3 (7%) were diagnosed by their primary veterinarian on examination prior to referral. In 2 of the 44 (5%) dogs, the owners reported being aware of the MGTs prior to development of signs related to the primary presenting complaint.

Three dogs that were presented with nonincidental MGTs were presented for staged bilateral mastectomy procedures. Time between surgeries was 4 months for 2 dogs and 15 months for the third dog. All other dogs had surgery only once for MGT removal.

A total of 36 breeds were represented along with 6 dogs of unknown breed. The most common types were Yorkshire Terrier or Yorkshire Terrier mix (n = 16), mixed-breed dog (8), German Shepherd Dog (6), Shih Tzu or Shih Tzu mix (5), Maltese (4), English Bulldog (4), Chihuahua (4), American Pit Bull Terrier (4), and Dachshund (3). The remainder of the other 27 breeds had ≤ 2 dogs represented. Of the dogs with incidental MGTs without concurrent pyometra, 7 of 11 (64%) were sexually intact and 4 of 11 (36%) were spayed. Of the dogs with nonincidental MGTs, 37 of 52 (71%) were sexually intact and 15 of 52 (29%) were spayed. Mean age of all dogs was 9.32 years (range, 2 to 17 years). Mean body weight was 16.26 kg (range, 1.4 to 60.5 kg). Forty-one of 96 (43%) dogs had a single MGT removed, and 55 of 96 (57%) dogs had multiple MGTs removed.

All dogs had a preoperative CBC and serum biochemistry performed. Eighty of the 96 dogs (83%) had preoperative thoracic radiographs performed. One dog had evidence of pulmonary metastatic disease on radiographs. Twenty of the 96 dogs (21%) had a preoperative abdominal ultrasound performed. Two of these 20 dogs had regional lymphadenopathy diagnosed on ultrasound, and both dogs were ultimately confirmed to have metastatic disease based on histopathology or cytology (1 each). All 3 dogs with evidence of preoperative metastatic disease were in the nonincidental group.

All dogs survived surgery and the immediate postoperative period until hospital discharge. There were no perioperative mortalities.

The most common benign histopathologic diagnoses were complex adenoma (n = 48), simple adenoma (31), benign mixed tumor (30), and basaloid (ductal) adenoma (30). The most common malignant histopathologic diagnosis was simple carcinoma (n = 24). The most common nonneoplastic histopathologic diagnosis was mammary hyperplasia (lobular, ductal, and/or cystic) (n = 6). Nonneoplastic results were included in the analysis as benign tumors.

In dogs that were presented for pyometra with incidental MGTs, 28 of 33 (85%) had only benign MGTs, and 5 of 33 (15%) had at least 1 malignant MGT. There were 61 of 66 (92%) benign tumors and 5 of 66 (8%) malignant tumors in this group of dogs. In the dogs that were presented for causes other than MGT removal or pyometra and had incidental MGTs, 10 of 11 (91%) had only benign MGTs, and 1 of 11 (9%) had at least 1 malignant MGT. There were 21 of 22 (95%) benign tumors and 1 of 22 (5%) malignant tumors in this group of dogs. When combining all dogs with incidental MGTs, 38 of 44 (86%) had only benign MGTs, and 6 of 44 (14%) had at least 1 malignant MGT. There were 82 of 88 (93%) benign incidental MGTs and 6 of 88 (7%) malignant incidental MGTs. In the dogs that were presented with nonincidental MGTs, 25 of 52 (48%) had only benign MGTs, and 27 of 52 (52%) had at least 1 malignant MGT. There were 75 of 107 (70%) benign nonincidental MGTs and 32 of 107 (30%) malignant nonincidental MGTs. These findings are summarized on a per-dog evaluation (Table 1) and on a per-tumor evaluation (Table 2).

Dogs that were presented with nonincidental MGTs were 6.84 times as likely to have a malignant MGT removed compared with dogs with incidental MGTs (OR, 6.84; 95% CI, 2.47 to 18.94; P < .001). Nonincidental MGTs were significantly (OR, 5.83; 95% CI, 2.31 to 14.73; P = .001) more likely to be malignant compared with incidental MGTs. Sterilization status was not a significant predictor of malignancy (OR, 0.86; 95% CI, 0.28 to 2.67; P = .792). Age was not a significant predictor of malignancy (OR, 1.16; 95% CI, 0.97 to 1.39; P = .099). Body weight was found to be a significant predictor of malignancy. The likelihood of malignancy increased by 5% for each 1-kg increase in body weight (OR, 1.05; 95% CI, 1.01 to 1.09; P = .013). The total number of MGTs removed was not predictive of malignancy (OR, 1.06; 95% CI, 0.76 to 1.48; P = .731). The results of the multivariate regression analysis are summarized (Table 3).

Size of MGTs was available for 158 of 195 tumors (81%), which included at least 1 tumor from all 96 dogs (Table 4). The proportion of T3 tumors that were malignant (63% [19/30 MGTs]) was significantly greater (P = .001) than the proportion of T1 tumors that were malignant (7% [8/115 MGTs]). In dogs that were presented for pyometra with incidental MGTs, 47 of 52 (90%) MGTs with a documented size were in T1, 3 of 52 (6%) were in T2, and 2 of 52 (4%) were in T3. All (21/21) the incidental MGTs in dogs without concurrent pyometra were in T1. In dogs that were presented with nonincidental MGTs, 47 of 85 (55%) MGTs were in T1, 10 of 85 (12%) were in T2, and 28 of 85 (33%) were in T3. The proportion of incidental tumors that were in T1 (68/73 [93%]) was significantly greater (P = .001; Table 5) than the proportion of nonincidental tumors that were in T1 (47/85 [55%]).

Location of MGTs was available for 165 tumors in 88 of 96 (92%) dogs (Table 6). Eight of 165 (5%) tumors were in the first mammary gland, 24 of 165 (15%) were in the second mammary gland, 35 of 165 (21%) were in the third mammary gland, 52 of 165 (31%) were in the fourth mammary gland, and 46 of 165 (28%) were in the fifth mammary gland. There was no statistically significant association between tumor location and proportion of malignant tumors (P = .442).

Discussion

In our cohort of 96 dogs, those with incidentally diagnosed MGTs were more likely to have benign MGTs compared with dogs that were presented with nonincidental MGTs. Incidental tumors had a malignancy rate of 7%, while nonincidental tumors had a malignancy rate of 30%. When evaluating the data per dog, 14% of dogs with incidental MGTs had at least 1 malignant tumor removed, and 52% with nonincidental MGTs had at least 1 malignant tumor removed. Based on our data, predictors of malignancy including tumor size and patient body weight may be used to help guide treatment of MGTs in future canine patients.

There is evidence that some MGTs may progress on a biologic and histopathologic continuum, where tumors that are initially benign develop characteristics of malignancy over time, especially as they grow in size.^8,16–18 To try to address the concern of an increase in malignancy rate with increasing tumor size, the best opportunity to achieve a good surgical outcome may be to remove tumors when they are smaller. In this study, 58% (25/43) of MGTs ≥ 3 cm were malignant, while 7% (8/115) of MGTs < 3 cm were malignant. As MGT size increased, malignancy rates also increased. Moreover, only 11% of the incidental MGTs found at the time of pyometra were ≥ 3 cm in size, and none of the incidental MGTs found with a nonpyometra primary condition were ≥ 3 cm, while 44% of the nonincidental MGTs were ≥ 3 cm. Therefore, the majority (93%) of incidental tumors were small (< 3 cm). Removal of incidental MGTs at the time of initial diagnosis offers a greater chance of surgical cure, since most of these tumors are small and less likely to be malignant.

The most common reason dogs with incidental MGTs were presented to our hospital was for pyometra. Three out of four dogs (75%) that were presented with incidental MGTs were presented for pyometra. Additionally, the majority (88%) of the dogs with incidental MGTs were not diagnosed with MGTs prior to referral. Most of these dogs were diagnosed with incidental MGTs during their initial examination at our hospital. Eight dogs, including 6 that were presented for pyometra, did not have the incidental MGTs diagnosed until they were under anesthesia and being shaved for surgery. This highlights the importance of performing a complete physical examination on all patients. In particular, dogs considered to be higher-risk candidates for MGT development, such as intact female dogs or female dogs known to be spayed later in life, should have a thorough palpation of both mammary chains. This is important both for general practice veterinarians during routine and sick visits and for veterinarians who work at emergency and referral hospitals. Furthermore, veterinarians should encourage owners to pursue surgical excision of incidentally diagnosed MGTs when they are found on examination, even if a patient is not presenting for a surgical disease.

Some dogs that present with surgical diseases such as pyometra may not be ideal anesthetic candidates due to their systemic illness, even with preoperative stabilization. Therefore, prolonging anesthesia to remove 1 or more MGTs may not be appropriate for every patient undergoing an emergency surgery. In those patients that are not stable enough for a longer anesthesia, staged procedures to address the pyometra or other primary condition initially, and then to remove the incidental MGTs at a later date, should be considered. Due to the fact that most incidental MGTs are benign, surgery to remove these incidental tumors is not an emergency. This is important both for client education as well as for anesthetic and surgical planning.

Pyometra is characterized by bacterial infection of the uterus, which in many cases causes systemic illness and inflammation.¹⁹ The presence of SIRS/sepsis has been diagnosed in > 50% of dogs with pyometra in multiple studies.^20–22 Even though some dogs have significant disease resulting from pyometra, mortality rates are low (3% to 10%)²³ and prognosis and outcome are generally favorable for dogs treated with ovariohysterectomy.²⁴ In fact, the pyometra mortality rate in this study was 0%. This rate is likely lower than previous studies due to the fact that all our dogs also had MGTs removed concurrently and would have needed to be considered stable enough to undergo additional anesthesia. Since most (85%) of the dogs in this study with incidental MGTs and pyometra had only benign tumors removed, prognosis for dogs with pyometra and incidental MGTs may not be worse than for those dogs with pyometra without incidental MGTs. Therefore, finding incidental MGTs on examination should not deter owners or veterinarians from pursuing surgery as treatment for pyometra.

A secondary objective of this study was to evaluate criteria that could be used to predict malignancy in canine MGTs. Aside from tumor size and patient group (incidental vs nonincidental), body weight was found to be the only other significant predictor of malignancy in this study. Contrary to some previous reports,^7,16,17,25 an increased risk of malignancy associated with heavier body weight was identified. In a 2018 study by Pastor et al,^26,27 larger dogs had an increased risk of malignancy; however, the overall malignancy rate of MGTs was much higher (81%) in their cohort. Body weight and correlation with the development of malignant MGTs could be confounded by a patient’s breed or lineage and therefore might be secondary to an underlying potential genetic predisposition for malignancy.^17,26,27 Unfortunately, due to the sizable number of breeds represented in our study, no primary breed risk factor could be identified with our data. Further research with appropriate control groups would be required to evaluate the true significance of patient body weight and possibly patient breed regarding malignant MGT development, but our results suggest that more aggressive surgical planning for tumors found in heavier dogs could be considered.

The major limitations of this study are due to its retrospective nature. The composition of the population of dogs with incidental MGTs in this study was highly influenced by selection bias because only dogs that had MGTs surgically removed and biopsied at an emergency and referral hospital were included. For this reason, these results therefore may not correlate to the patient population in every small animal general practice. Even though the medical records were evaluated to determine when the incidental MGTs were initially diagnosed, we do not definitively know whether all tumors were truly incidental. The fact that the majority of larger MGTs were in the nonincidental group likely reflects some degree of selection bias. Larger tumors are more noticeable and accordingly are more likely to be a primary reason for presentation.

Certain data were not reliably available for many dogs and were therefore not included in our analysis, which may have impacted findings regarding predictors of malignancy in our MGTs. Other pertinent factors that could be considered for evaluation include body condition score and presence of obesity, as they have previously been identified as possible risk factors for MGT development.^28,29 Additionally, clinical signs such as tumor growth rate, ulceration, and fixation to skin and underlying tissues have been correlated to malignancy in MGTs.^29–31 Preoperative staging and sentinel lymph node mapping were not standardized and could have given insight into prognosis if available. Finally, long-term follow-up information for each patient regarding development of additional MGTs was beyond the scope of this study but would have been useful to ensure that the dogs with histologically benign tumors had biologically benign tumors.

The results of this study allowed us to reject our initial hypothesis. A significantly lower malignancy rate was identified in MGTs found at the time of pyometra or in other incidental settings compared with nonincidental MGTs. Veterinarians should have confidence proceeding with surgery in dogs with MGTs identified incidentally during diagnostic evaluation of an unrelated problem, as most tumors are benign and therefore should not change a patient’s overall prognosis. Increased caution should be taken in dogs with heavier body weights and larger tumor sizes due to a higher likelihood of malignancy in these populations.