Canine mammary gland tumors (MGTs) are the most commonly diagnosed tumors in dogs, similar to human breast cancer, with increasing prevalence attributed to longer lifespans and better nutrition. In human breast cancer, carcinoembryonic antigen (CEA), carbohydrate antigen (CA) 15-3, and cancer antigen (CA) 27-29 originating from mammary gland have been a creative area of investigation for the early detection of the tumors. The protein CEA, which is involved in cell adhesion, can be elevated in various types of cancers (22). CA15-3 and CA27-29 are protein antigens containing carbohydrates found in the transmembrane glycoprotein MUC-1, which inhibit tumor cell lysis and reduce interactions between cells (4). In cases of primary breast cancer, several studies have shown that elevated levels of CEA at the time of diagnosis are linked to a negative prognosis (26). Similarly, increased serum CA15-3 levels at diagnosis have been associated with more advanced breast cancer stage, larger tumor size, axillary lymph node metastasis, and poorer overall survival (OS) and disease-free survival (DFS) (22). Most studies examining the two MUC-1 biomarkers have focused on CA15-3, although comparative studies have shown similar correlations with cancer stage and tumor size for CA27-29. These biomarkers clearly hold prognostic significance in early breast cancer, but they exhibit low sensitivity (less than 7%) and lack evidence from prospective randomized controlled trials to support their use in early-stage disease. Overall, the combined application of breast cancer markers improved the specificity and sensitivity of diagnosis, providing better diagnosis and treatment options (5).

In veterinary medicine, there has been limited research on the combined application of these human breast cancer markers. However, it is noteworthy that the combined detection of serum markers CA15-3, CEA, and serum ferritin has shown improved detection sensitivity for canine MGTs, providing better clinical applications (7). Hence, evaluation of CA 15-3 along with CEA would be a non-invasive technique for detecting canine mammary tumors. Moreover, the application of CA 15-3 and CEA can become noteworthy with an evaluation of deregulated circulating microRNA-21 (miR-21) (8). It can be a valuable prognostic marker for the early detection of mammary tumors in canines while miR-29b can add sensitivity to the detection of canine mammary tumors if evaluated with miR-21 (14).

Owing to various factors, there has been a rising incidence of breast cancer in humans and MGT in dogs. Hence, biomarkers play a crucial role in providing insight into a breast cancer patient’s prognosis and response to treatment. Particularly, canine MGT is notorious for malignancy, encompassing 50% of total outbreaks (18). Moreover, its aggressiveness and mortality are notable due to high rates of recurrence and metastasis (23). In this context, serum biomarkers present an appealing alternative to tissue biomarkers, particularly in the context of monitoring for disease recurrence or progression, assessing response to treatment, and identifying targetable mutations for directing therapy (10). Timely diagnosis and appropriate treatment could contribute to the disease burden to the owner and improve the OS and DFS of dogs carrying cancers.

Although CA15-3, CA27-29, and CEA have low sensitivity in early-developing mammary gland tumors like adenoma and carcinoma in situ, they have proven to be valuable in managing malignant and metastatic disease, especially in monitoring treatment response through serial measurements (25). Many clinical trials using various biomarkers, including established markers such as CEA, CA15-3, and CA27-29, as well as components of liquid biopsy such as circulating tumor cells (CTCs), cell-free DNA (cfDNA), circulating tumor DNA (ct DNA) and miRNA, have been performed, which will contribute to cancer diagnosis guided by serial biomarker assessment with traditional monitoring methods (2,9). These trials are essential in defining the roles of circulating biomarkers and have the potential to revolutionize the imaging-driven response assessment for various solid tumor malignancies (6). Moreover, the integration of liquid biopsy and machine learning, as seen in the selection of panels in ccfDNA methylation and miRNA, may facilitate the seamless integration of personalized medicine into standard-of-care practices. The field of biomarker-related research in breast cancer is continuously expanding, although only a few biomarkers ultimately become the standard of care for the clinical management of breast cancer (22).

Ethical statement

Signed informed-consents were obtained from all the owners of dogs involved in this study. All the tissue samples and blood samples were obtained under professional veterinary care. Peripheral blood samples and tissue samples were collected from dog patients. No human participants were involved in the study. Approval for clinical trials was obtained from the Institutional Animal Care and Use Committee of Northeast Agricultural University, Harbin, China (approved by the State Council on October 31st, 1988 and promulgated by Decree No. 2 of the State Science and Technology Commission on November 14th, 1988).

All dog owners involved in this study provided signed informed consent. Blood samples were collected under the supervision of professional veterinary care from dog patients.

Collection of samples

Samples were gathered from Chonnam National University (CNU) Veterinary Medical Teaching Hospital. Blood samples from the patient having either canine mammary gland, uterine, or ovarian tumors. Among the cases, 34 were identified as malignant, with ages ranging from 2 to 16 years and an average age of 10.6 ± 2.5. Additionally, 25 blood samples from normal dogs were obtained, and were used as a control group.

Histopathological examination and staging

The examination of tumor extent and the presence of distant metastases involved X-ray and ultrasound procedures. Tissue samples were dissected into small sections and stored in a freezer at -80 degrees for western blot analysis. Additionally, most samples were preserved in a 10% paraformaldehyde solution. After 24-72 hours, the samples were embedded in paraffin, cut into 3 μm sections, and stained with H&E. The tumors were classified according to the World Health Organization (WHO) TNM classification system and using histological classification systems (11). Malignant tumor tissues were graded (I, II, III) based on tubule formation, cell morphology, and mitotic rate.

ELISA for CEA, CA 15-3, and CA 27-29

The serum was obtained by centrifuging (2,500 rpm for 10 minutes) blood samples from the patients. At first, the samples were stored at 80ºC for one week at the maximum. At the end of the week, stored samples in a batch were under analysis. The serum CEA, CA 15-3, and CA27-29 levels were assessed using an electrochemiluminescence immunoassay system. The marker assays were performed using commercial ELISA kits for CEA, CA 15-3, and CA 27-29 (CUSABIO, Houston, TX, USA).

Statistical analysis

Statistical analysis was performed using Graphpad prism (version 9.4.2, GraphPad Software Inc., San Diego, CA, USA). Using the Student’s t-test, and χ² test, data were expressed as mean ± standard deviation (SD). Statistical differences were determined using one-way or two-ways ANOVA. p < 0.05 was statistically significant (*). Greater significances were expressed p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).

Evaluation of the diagnostic performance of bio-markers on malignant gland tumors in dogs

In this study, we assessed the diagnostic accuracy of biomarkers for detecting malignant gland tumors in dogs. We gathered biomarker data from 59 patients, including 25 non-cancer cases, 25 malignant gland tumors (MGT), and eight other types of cancer, such as urinary and ovarian cancer. Additionally, we collected information on patient characteristics as covariates such as weight, age, breed, and sex to account for potential influences on biomarker values. We constructed three separate logistic regression models using the patients’ biomarker values and covariates. One model coded patient as one if confirmed MGT and zero otherwise as follows,

Probability of MGT_i ~ B(π_i )

logit(π_ik ) = α + age_i + weights_i + biomaker_k

where biomarker k was examined in patient i.

Additionally, we built three new separate logistic regression models using only the patients’ cancer data to differentiate malignant gland tumors from other cancers in dogs. The statistical significance of the regression coefficient of biomarker values in the logistic regression model was examined to identify the association with MGT status. We also estimated the size of the area under the curve (AUC) of the receiver operating characteristic (ROC) with sensitivity and specificity at the optimal cut-off, given the regression model with statistical significance regression coefficient of biomarkers.

Characteristics of the population carrying MGTs and other tumors

A total of 59 dogs, including 25 healthy dogs and 34 cancer patients, was characterized as shown in Table 1. Targeting CEA, CA15-3, and CA27-29, dogs diagnosed with mammary gland tumor (MGT), ovarian tumor (OT), and uterine tumor (UT) were investigated alongside the standard normal control group. With respect to the breed distribution, Maltese (32%) was the most common breed in healthy dogs, followed by Poodle (28%), Pomeranian (12%), and the others (28%), including Bichon, Chihuahua, Coton de Tuléar, Jindo, Mixed, Shih Tzu, and Spitz. In dogs with MGT or other tumors, Maltese (32%) was the most common breed in dogs carrying tumors, followed by Poodle (26%), Pomeranian (12%), Shih Tzu (9%), and the others, including Bichon, Chihuahua, Coton de Tuléar, Jindo, Mixed, Shih Tzu, Spitz. There was no significant difference in breed distribution between healthy dogs and dogs diagnosed with tumors.

Table 1 . Clinical information of dogs tested for CEA, CA15-3, and CA27-29.

ID	Breed	Age(year)	Sex	Weight (kg)	CEA(ng/mL)	CA 15-3(U/mL)	CA 27-29(U/mL)	Results	Histopathology
1	Poodle	3	F	2.5	0.15	2.28	4.91	N
2	Shih Tzu	12	SF	4.2	0.98	2.06	5.46	N
3	Maltese	8	SF	5.1	0.67	2.45	4.67	N
4	Maltese	4	SF	4.6	0.54	2.18	5.68	N
5	Maltese	5	SF	3.9	0.30	1.96	4.99	N
6	Jindo	2	SF	14.0	0.31	1.86	5.59	N
7	Pomeranian	8	F	3.2	0.54	1.96	4.59	N
8	Spitz	4	SF	5.8	2.30	2.19	4.97	N
9	Poodle	3	SF	4.3	2.62	2.12	4.73	N
10	Chihuahua	7	SF	2.9	0.79	2.67	5.41	N
11	Maltese	5	F	3.2	2.76	2.09	4.54	N
12	Pomeranian	2	F	2.7	1.38	1.76	5.65	N
13	Poodle	5	SF	5.2	1.41	2.50	7.93	N
14	Poodle	4	SF	6.0	0.83	2.33	5.40	N
15	Mixed	5	F	8.4	1.45	2.54	4.91	N
16	Maltese	4	SF	5.2	0.60	2.01	4.46	N
17	Pomeranian	3	SF	3.4	2.05	2.57	4.68	N
18	Poodle	5	SF	3.9	2.30	2.55	7.39	N
19	Coton de Tuléar	3	SF	5.2	0.90	3.32	5.90	N
20	Maltese	6	SF	4.4	0.88	2.07	5.22	N
21	Poodle	2	SF	3.1	0.21	2.17	5.27	N
22	bichon	2	F	3.8	0.32	2.70	5.58	N
23	Maltese	1	F	2.1	0.27	2.87	4.76	N
24	Poodle	4	SF	3.8	0.49	3.03	4.79	N
25	Maltese	2	SF	3.0	0.37	2.97	5.76	N
26	Maltese	11	F	4.2	3.16	20.65	31.43	MGT	Complex carcinoma
27	Maltese	8	SF	3.7	21.29	16.98	23.22	MGT	Complex carcinoma
28	Poodle	14	F	5.0	3.78	14.32	19.84	MGT	Tubular carcinoma
29	Pomeranian	9	SF	3.3	3.35	15.98	30.78	MGT	Complex carcinoma
30	Yorkshire terrier	15	SF	2.1	3.99	16.35	21.48	MGT	Complex carcinoma
31	Chihuahua	12	F	2.2	3.64	19.38	20.28	MGT	Complex carcinoma
32	Maltese	14	F	4.3	3.98	15.78	31.95	MGT	Tubular cacinoma
33	Poodle	9	SF	3.2	10.52	11.43	21.12	MGT	Tubulopapillary carcinoma
34	Pomeranian	7	SF	4.7	5.17	11.46	17.06	MGT	Carcinoma in situ
35	French bulldog	8	SF	11.5	5.22	14.24	21.05	MGT	Tubular cacinoma
36	Maltese	12	SF	5.2	3.15	16.05	41.41	MGT	Complex carcinoma
37	Poodle	13	SF	4.1	4.32	17.94	24.02	MGT	Tubulopapillary carcinoma
38	Jindo	9	F	16.0	20.32	16.23	15.78	MGT	Solid carcinoma
39	Poodle	11	F	4.5	3.83	12.39	25.03	MGT	Tubulopapillary carcinoma
40	Pomeranian	14	F	3.8	23.03	25.40	21.24	MGT	Inflammatory cacinoma
41	Shih Tzu	16	F	6.5	9.47	14.39	17.97	MGT	Tubular cacinoma
42	Maltese	10	SF	4.1	4.20	15.29	30.34	MGT	Complex carcinoma
43	Schnauzer	14	F	5.7	16.43	19.89	19.88	MGT	Complex carcinoma
44	Poodle	12	SF	4.3	4.37	13.15	21.13	MGT	Tubulopapillary carcinoma
45	Golden Retriever	8	SF	32.0	3.95	15.54	26.64	MGT	Tubular cacinoma
46	Poodle	14	F	5.1	8.10	25.78	25.52	MGT	Tubulopapillary carcinoma
47	Shih Tzu	12	SF	5.7	4.00	29.75	61.54	MGT	Inflammatory carcinoma
48	Maltese	13	SF	4.4	18.46	15.49	42.49	MGT	Complex carcinoma
49	Maltese	8	SF	3.8	9.48	16.78	52.77	MGT	Solid carcinoma
50	Poodle	10	SF	5.0	4.32	18.78	25.45	MGT	Tubulopapillary carcinoma
51	Spitz	6	F	7.1	4.79	17.46	23.49	MGT	Complex carcinoma
52	Maltese	12	F	3.3	8.41	4.42	6.33	UT	Cervical carcinoma
53	Maltese	10	F	2.9	2.72	2.81	5.48	UT	Leiomyosarcoma
54	Poodle	6	F	4.6	6.48	2.95	10.36	UT	Cervical carcinoma
55	Pomeranian	14	F	2.8	3.98	3.59	4.36	UT	Leiomyosarcoma
56	Maltese	11	F	4.1	42.89	1.52	5.84	OT	Epithelial carcinoma
57	Shih Tzu	14	F	6.9	1.58	1.99	20.52	OT	Epithelial carcinoma
58	Poodle	15	F	4.8	5.84	5.15	11.44	OT	Epithelial carcinoma
59	Maltese	12	F	3.8	3.60	3.15	13.35	OT	Epithelial carcinoma

All samples collected at the first diagnosis were analyzed by ELISA based upon a Sandwich assay principle..

N, normal dog; MGT, mammary gland tumor; UT, uterine tumor; OT, ovarian tumor; CEA, carcinoembryonic antigen; CA, carbohydrate antigen..

Additionally, the age and weight were analyzed between normal dogs and dogs with MGT, OT, and UT (Table 2). Aging was found to have significant relevance to incidence rates of MGT, OT, and UT, respectively. However, weight was found to have no correlation with the outbreaks of the tumors.

Table 2 . Comparison of CEA, CA 15-3, and CA 27-29 in serum between healthy dogs and dogs with MGT, OT, and UT.

	Diagnosis	Age (year)	Weight (kg)	CEA(ng/mL)	CA 15-3(U/mL)	CA 27-29(U/mL)
Number	Normal	25	25	25	25	25
	MGT	26	26	26	26	26
	OT	4	4	4	4	4
	UT	4	4	4	4	4
Mean	Normal	4.36	4.56	1.02	2.37	5.33
	MGT	11.10	6.21	9.09	17.20	27.40
	OT	13.00	4.90	13.50	2.95	12.80
	UT	10.50	3.40	5.40	3.44	6.63
Median	Normal	4.00	3.90	0.79	2.28	5.22
	MGT	11.50	4.45	4.34	16.10	23.80
	OT	13.00	4.45	4.72	2.57	12.40
	UT	11.00	3.10	5.23	3.27	5.90
Standard deviation	Normal	2.45	2.39	0.80	0.40	0.82
	MGT	2.75	5.98	10.60	4.35	11.00
	OT	1.83	1.40	19.70	1.62	6.06
	UT	3.42	0.83	2.55	0.73	2.61

MGT, mammary gland tumor; OT, ovarian tumor; UT, uterine tumor; CEA, carcinoembryonic antigen; CA, carbohydrate antigen..

Comparisons of serum CEA, CA 15-3, and CA 27-29 between healthy dogs and cancer patients

The alterations in the serum protein between healthy dogs and cancer patients were evaluated by independent ELISA detecting CEA, CA 15-3, and CA 27-29 (Table 2). In healthy dogs, the serum levels (mean ± SD) of CEA, CA 15-3 and CA 27-29 were 1.02 ± 0.79 ng/mL (the median = 0.79 ng/mL), 2.37 ± 2.28 U/mL (the median = 2.28 U/mL), and 5.33 ± 5.22 U/mL (the median = 5.22 U/mL), respectively. In MGT patients, the average concentration of CEA was 9.09 ± 10.60 ng/mL (the median = 4.34 ng/mL), which was significantly higher than those in normal dogs (Fig. 1A). Furthermore, CA 15-3 and CA 27-29 were significantly elevated, showing 17.20 ± 4. 35 U/mL (the median = 16.10 U/mL) and 27.40 ± 11.00 U/mL (the median = 23.80 U/mL), respectively (Fig. 1B, C). In the case of ovarian tumors, compared to normal dogs, the patients showed a significant increase in CEA levels, showing 13.50 ± 19.7 ng/mL (the median = 4.72), while CA 15-3 and CA 27-29 showed no significant overexpression in the serum. In addition, dogs carrying UTs did not exhibit a significant increase compared to normal controls for all tumor markers.

Figure 1. The serum level of CEA (ng/mL), CA15-3 (U/mL), and CA27-29 (U/mL) before treatment in patients diagnosed with either MGT, UT, or OT. Significant differences in (A); p = 0.002 for NC vs MGT, and p = 0.003 for NC vs OT. Significant differences in (B); p < 0.0001 for NC vs MGT, p < 0.0001 for MGT vs UT, and p < 0.0001 for MGT vs OT. Significant differences in (C); p < 0.0001 for NC vs MGT, p < 0.0001 for MGT vs UT, and p = 0.004 for MGT vs OT. NC, normal control; MGT, mammary gland tumor; OT, ovarian tumor; UT, uterine tumor; CEA, carcinoembryonic antigen; CA, carbohydrate antigen.

Alteration of serum CEA, CA 15-3, and CA 27-29 after tumor resection surgery in dogs with MGT and OT

For the general investigation, only 34 dogs among total hospitalized patients that received the same surgical treatment for MGTs, OT, and UT were chosen for the direct comparisons of CEA, CA 15-3, and CA 27-29. Dogs with MGTs were subject to partial mastectomy, and those with OT and UT were subject to ovariohysterectomy. No further treatments, such as chemotherapy or NSAIDs, were applied to the dogs for a better assessment of patient status after surgery, according to the owner’s consent. Aside from the surgical treatment, the serums were collected from the patients to monitor the levels at specific intervals: 1) one week after surgery, 2) one month after surgery, and 3) at six months to one year after surgery in case of any suspicions of recurrence or metastasis. For this, the patient was visited for recheck examinations every two months, during which physical examinations, radiographic imaging, and, if necessary, ultrasound evaluations were performed.

In most cases, during the recheck, the owner reported the development of new masses in the remaining mammary tissue. These reports led to subsequent consultations and diagnostic tests confirming the recurrences. All recurrences were identified within the residual mammary tissue and were confirmed as new mass formations. In brief, 25 dogs among 34 diagnosed with MGTs were under monitoring after surgery, and 6 experienced a recurrence of MGT within a year. 4 dogs, either diagnosed with UT or OT, did not show a recurrence or metastasis during the same period.

In MGT patients, CEA, CA15-3, and CA 27-29 levels dropped significantly starting from one week, showing a gradual decrease until one month after surgery (Fig. 2A-C). The OT patients showing a significant increase in CEA were also monitored, targeting only CEA, not CA 15-3 and CA 27-29. Since UT patients showed no significant alteration in all tumor markers, tracking and monitoring were not considered. As a result, no significant difference was found between dogs before and after surgery in the case of OT (Fig. 2D).

Figure 2. The serum level of CEA (ng/mL), CA15-3 (U/mL), and CA27-29 (U/mL) before treatment (BT), 1 week or 1 month after surgery in patients diagnosed with MGT and OT. Significant differences in (A); p < 0.0001 for BT vs 1 Wk, and p < 0.0001 for BT vs 1 M. Significant differences in (B); p < 0.0001 for BT vs 1 Wk, p < 0.0001 for BT vs 1 M, and p < 0.0001 for 1 Wk vs 1 M. Significant differences in (C); p < 0.0001 for BT vs 1 Wk, p < 0.0001 for BT vs 1 M, and p = 0.012 for 1 Wk vs 1 M.

Potential use for early detection of recurrence or metastasis targeting MGTs using CEA, CA 15-3, and CA 27-29

MGTs are notorious for high incidence rates (50%) of malignant tumors and high rates of recurrence or metastasis compared to other solid tumors (18,23). To identify patients at risk of recurrence at an early stage, the serum levels of the three biomarkers were monitored by regular checkups through the abovementioned ELISA method. Notably, in patients who experienced recurrence within one year, CEA levels decreased to a minimum at one-month post-surgery, followed by an increase at six months and a further elevation at one year (Fig. 3A). Similarly, CA 15-3 exhibited a gradual decrease until 6 months and started to increase at 1-year post-surgery (Fig. 3B). CA 27-29 dropped at the minimum at 1 month after surgery and began to increase at 6 months post-surgery (Fig. 3C). The data indicated a significant increase in all three biomarkers during the biological process of MGT recurrence, underlining the potential of these biomarkers for predicting and monitoring MGT recurrence.

Figure 3. The serum level of CEA (ng/mL), CA15-3 (U/mL), and CA27-29 (U/mL) during 1 Y in patients that experienced a recurrence of MGT. Significant differences in (A); p = 0.011 for 1 M vs 1 Y, and p = 0.024 for 6 M vs 1 Y. Significant differences in (B); p = 0.024 for 1 M vs 1 Y, and p = 0.015 for 6 M vs 1 Y. Significant differences in (C); p = 0.045 for 1 Wk vs 1 M, and p = 0.001 for 1 M vs 1 Y.

Evaluation of the diagnostic performance of CEA, CA 15-3, and CA 27-29 on canine MGTs

The results of evaluating the diagnostic performance of biomarkers on malignant gland tumors in dogs are summarized in Table 3, including the AUC, sensitivity, and specificity of CEA, CA 15-3 and CA 27-29. Among the three biomarkers, CA 27-29 was found to have a significant association with MGT positive (p = 0.016). Moreover, CA 27-29 had an AUC of 0.9918 (95% confidence interval: 0.975-1) with a sensitivity of 0.970 and specificity of 1, as illustrated in Fig. 4.

Table 3 . Summary of diagnostic performance of three biomarkers on malignant gland tumors using data from 59 patients in dogs.

Biomarker	Regression coefficient (p-value)	AUC (95% confidence interval)	Sensitivity	Specificity
CEA	0.04 (0.413)	0.863 (0.75-1)	0.728	1
CA 15-3	6.23 (0.999)	-	-	-
CA 27-29	0.31 (0.0164)	0.992 (0.975-1)	0.970	1

Figure 4. 1 Area under the curve (AUC) of the receiver operating characteristic (ROC) for CA 27-29 as the diagnostic accuracy to detecting malignant gland tumors in dogs.

Furthermore, CA 27-29 showed diagnostic resolution between MGT and other cancers. For differentiating the MGT from other cancers, it had an AUC of 0.9663 (95% confidence interval: 0.900-1) with a sensitivity of 0.875 and specificity of 1, as illustrated in Fig. 5.

Figure 5. Area under the curve (AUC) of the receiver operating characteristic (ROC) for CA 27-29 as the diagnostic accuracy to differentiating malignant gland tumors from other cancers in dogs.

The increasing incidence of MGT in dogs, attributed to environmental factors and extended life expectancy, is a significant concern. More than 50% of these cases are diagnosed as malignant, leading to economic burdens and treatment limitations for both owners and veterinary clinicians (18). MGT is also known for its high rates of recurrence and metastasis (23). Timely diagnosis and prevention of MGT are therefore crucial for the overall well-being of veterinary clinics and companion animals. This study focused on the application of a combination of biomarkers, CEA, CA 15-3, and CA 27-29 in dogs, and included a comparative analysis of malignant MGT with malignant types of UT and OT using the same serum. To better evaluate the diagnostic potential of the combined markers, 34 dogs with cancers that received the same surgical treatment were exclusively chosen.

Notably, in MGT patients, the serum levels of all used biomarkers were found to be significantly elevated compared to normal dogs, CEA (p = 0.002), CA 15-3 (p < 0.0001), and CA 27-29 (p < 0.0001), which indicates outstanding diagnostic performances. According to various studies, other tumors, such as UT and OT, have been found to promote the synthesis of these markers, suggesting a comparative analysis between canine MGTs and different types of tumors (1,13). In this study, the OT patient exhibited a significant increase in CEA (p = 0.003) compared to normal dogs, while the other markers did not show significant differences. The lack of significant increases in UT and OT may be attributed to the limited number of patient samples available for these tumor types, suggesting more sample collection for appropriate statistical analysis. Nevertheless, it is noted that the combined utilization of CEA, CA 15-3, and CA 27-29 shows promising diagnostic performance in detecting MGTs by patient serum, excluding other types of tumors. Further investigation with a comparative analysis between various types of tumors could ensure the ability to differentiate MGT from different cancers.

Additionally, it is noteworthy that patients who developed recurrent MGT displayed a substantial increase in CEA, CA 15-3, and CA 27-29 levels at six months and one-year post-surgery. This underscores the efficacy of these biomarkers in the proactive monitoring and early detection of MGT recurrence. However, due to the time limitations of the one-year follow-up period, the recurrence of MGT in the other patient could not be confirmed. This highlights the need for further investigations involving an expanded patient cohort and prolonged observation periods to conclusively establish the utility of these biomarkers in identifying patients susceptible to recurrent MGT.

Concerning the sensitivity and specificity of ELISA diagnostic performance, the three biomarkers revealed limited significance in detecting canine MGTs. The ROC curve analysis of CEA and CA 15-3 did not demonstrate a significant difference between healthy dogs and those with MGTs. However, CA 27-29 exhibited significant sensitivity and specificity in discriminating MGT patients from normal dogs. Furthermore, CA 27-29 displayed significant sensitivity and specificity in distinguishing canine MGT from other cancers, such as OT and UT. Notably, the use of CA 27-29 as a diagnostic marker for canine MGTs showed a promising performance, as shown by the study of human breast cancer (21). It is noteworthy that CA 27-29 is not organ-specific, as it can also be found in patients with other malignancies or benign disorders of the breast, liver, kidney, and ovarian cysts (24). Therefore, the limitation highly requires the combined application of other biomarkers and specific tumor markers.

Despite extensive research in molecular and genetic markers for tumor diagnosis and prognosis, significant challenges still need to be solved in their practical application. In the context of human breast cancer, canine MGT presents a diverse and complex array of molecular profiles, necessitating careful interpretation in conjunction with clinical diagnostic techniques. Elevated levels of CEA in breast cancer patients have been associated with lymph node involvement and tumor size, serving as indicators of disease progression (12). However, it is important to recognize that increased CEA levels can also occur in non-malignant conditions, such as chronic obstructive pulmonary disease, obesity, atherosclerosis, and metabolic syndrome (16). Similarly, CA15-3 is primarily linked to breast cancer and is utilized for monitoring treatment response as well as detecting recurrence (15). However, significant elevations in CA15-3 levels have also been documented in lung, pancreatic, ovarian, and gastric cancers. Moreover, moderate increases in CA15-3 concentrations may occur in conditions such as liver and pancreatic cancers, cirrhosis, and benign breast disorders, thereby complicating its specificity (3). CA27-29 serves a comparable role in monitoring recurrence in breast cancer, though it is not exclusively specific to mammary tumors. Elevated levels of CA27-29 can also be observed in benign conditions and various malignancies affecting the breast, liver, kidney, and ovarian cysts (24). As such, while these markers may correlate more closely with breast cancer than with other malignancies, their limited specificity for canine MGT remains a challenge.

The utility of CEA, CA15-3, and CA27-29 in the diagnosis and prognosis of canine MGT is complicated due to variability in expression levels among patients, which may obscure the interpretation. Consequently, the diagnosis and prognosis must not rely on a singular positive marker. Instead, a comprehensive analysis using combined biomarkers is recommended to enhance accuracy. Additionally, incorporating complementary clinical techniques is essential for evaluating invasion and metastasis.

Canine MGT exhibits significantly different clinical behaviors and prognoses depending on histological classification. Malignant mammary primary tumors can be subdivided mainly into carcinoma originating from epithelial cells and sarcoma from mesenchymal cells, comprising distinct immunophenotypes and molecular biology (19,20). In particular, inflammatory mammary carcinoma is characterized by heightened inflammatory responses linked to the upregulation of NF-κB, apoptosis-related genes, immune response genes, and tumor-promoting genes (17). In other words, historical features could determine molecular mechanisms affecting pathogenesis. However, depending on histological classification, this study found no significant changes in CEA, CA15-3, and CA27-29. In addition, due to realistic conditions and the owner’s unwillingness, CT scans for determination of TNM staging were not adequately carried out. So, we could not perform a comparative analysis against the biomarkers across the TNM staging.

In conclusion, ELISA measurements of CEA, CA 15-3, and CA 27-29 demonstrated significant diagnostic performance in the detection of canine MGTs. CEA levels were significantly elevated in MGT and OT patients, while CA 15-3 and CA 27-29 levels were significantly increased only in MGT patients. This suggests that the combined application of these biomarkers by ELISA can effectively differentiate MGT from other tumors and may be utilized to monitor metastasis and recurrence of canine MGT. As CA 27-29 exhibited significant diagnostic performance in detecting and differentiating MGT from other cancers, further studies are needed to evaluate the diagnostic performance of CEA and CA 15-3 using a larger sample size. Given the potential overlap of biomarkers associated with specific tumors in different organs, the combined application of tumor markers is essential to improve the sensitivity and specificity of cancer diagnosis (25). Therefore, this study highlights the potential use of combined biomarkers in veterinary medicine, aligning with recent advancements in the investigation of blood biomarkers in human oncology.

This research was supported by the Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, and Forestry (IPET) through the High-Risk Animal Infectious Disease Control Technology Development Program, funded by the Ministry of Agriculture, Food and Rural Affairs (MAFRA) (grant number: RS-2024-00399814). Additionally, this research was supported by the National Institute of Health (NIH) research project (project no. 2024-ER2101-00).

The authors have no conflicting interests.