Malignant mesothelioma (MM) is extremely rare but present aggressive malignancy in dogs, arising from the mesothelium on the surface of the body cavities (10). MM is highly resistant to traditional anticancer drugs, including cisplatin, with no standard of care established (6,11). The prognosis of MM in dogs is dismal with a one-year survival rate of 22% (11). Chemotherapy is the sole variable independently associated with the survival of dogs with MM (6). Epirubicin – an anthracycline analogous to doxorubicin – has been considered an alternative that has similar clinical benefits against MM, minimizing the risk of cardiotoxicity in tumor-bearing dogs with pre-existing cardiac disease (5). We provide a clinicopathologic course and an extensive diagnostic workup for MM and report the first attempt to treat canine MM with epirubicin following pericardiectomy.

An 8-year-old, intact male, golden retriever dog with body weight of 34 kg was presented to a local animal hospital with chronic anorexia and lethargy. Echocardiography confirmed pericardial effusion. Complete blood count (CBC) revealed mild regenerative anemia with a hemoglobin concentration of 12.6 g/dL (Reference interval [RI], 13.1-20.5 g/dL), hematocrit of 35.2% (RI, 37.3-61.7%), RBC of 556 × 10⁶/μL (RI, 565-887 × 10⁶/μL), and reticulocyte concentration of 149 × 10⁶/μL (RI, 10-110 ×10⁶/μL). C-reactive protein (CRP) concentration was increased (15 mg/L; RI, 0-7 mg/L). No abnormalities were detected in serum chemistry, blood gas analysis, and canine pancreatic lipase immunoreactivity.

The dog was referred to the Veterinary Medical Teaching Hospital at Jeju National University. Thoracic radiography revealed an enlarged and globoid cardiac silhouette with a vertebral heart score (VHS) of 13.5 v and broncho-interstitial pulmonary patterns (Fig. 1A, B). Ultrasonography found pericardial effusion as well as, to a lesser extent, pleural effusion (Fig. 1C). A pericardiocentesis was performed on the right hemithorax with the dog in a right lateral recumbent position. Under ultrasound guidance with local anesthesia (4 mL of 2% lidocaine diluted in saline, Daehan Pharmacy, Korea), approximately 200 mL of bloody pericardial fluid was removed. Biochemical analysis of the fluid revealed hemorrhagic pericardial effusion with a hemoglobin concentration of 6.2 g/dL (RI, 13.1-20.5 g/dL), hematocrit of 17.6% (RI, 37.3-61.7%), RBC concentration of 620 × 10⁶/μL (RI, 565-887 × 10⁶/μL), total nucleated cell count of 18,350 × 10³/μL (RI, 5,050-16,760 cells/μL), total protein concentration of 7.6 g/dL (RI, 5.2-8.2 mg/dL), and albumin concentration of 3.9 g/dL (RI, 2.2-3.9 mg/dL). The pericardial effusion was smeared, and cytological examination was performed. Exfoliated cells showed moderate to marked cellular and nuclear atypia, suggesting neoplastic effusion (Fig. 2A). Occasionally, finely granular eosinophilic globules were found in the cytoplasm (Fig. 2A, insert). Nuclei were round to oval with granular to stippled to coarse chromatin and one to multiple prominent nucleoli. Bizarre mitoses were also found (Fig. 2B, arrow). Many macrophages contained vacuoles, hemosiderin, or hematoidin crystals. Differential diagnoses included hemangiosarcoma, malignant mesothelioma, and ectopic thyroid carcinoma. Pleural effusion also showed similar cytologic findings.

Figure 1. (A) Thoracic radiography. Note the globoid cardiac silhouette on the ventrodosal view. There is mild volume of pleural effusion and broncho-interstitial pulmonary pattern on the right lung field. (B) VHS is 13.5 v on the right lateral view on the first admission. (C) Thoracic ultrasonography. Note the presence of the pericardial effusion indicated by asterisk.

Figure 2. Cytopathology of pericardial mesothelioma in a dog. (A, B) Smears of pericardial fluid. (A) Neoplastic cells are observed in clusters showing marked cellular atypia. Amorphous eosinophilic materials are attached to neoplastic cells. Hematoidin crystals are scattered on the neoplastic cells. A finely granular eosinophilic globule (arrow) and paranuclear dark granules are found in the cytoplasm (insert). (B) Abnormal mitosis (arrow) is present admixed with neoplastic cells and inflammatory cells. (C, D) Fine needle aspirates of pericardial lesions. (C) Mesenchymal characteristics of neoplastic cells, such as wispy indistinct cytoplasmic tails, are present. (D) A binucleated cell shows marked anisokaryosis. The asterisk indicates the nucleus. Note the compared size of neutrophils and a macrophage. H&E stain, A, B, insert ×400 , C, D: ×1,000. Scale bars = 20 μm (A, B, insert), 50 μm (C, D).

The next day of presentation, pericardial effusion recurred. Thoracic radiography revealed a VHS of 12.8 v. Abdominal radiography and ultrasonography revealed mild splenomegaly with a hypoechoic nodule, which was cytologically diagnosed with reactive lymphoid hyperplasia. Echocardiography revealed nodule-like hypoechoic pericardial lesions with a significant amount of pericardial effusion. Computed tomography (CT) found pericardial nodules near the right side of the heart (Fig. 3A, B, arrow). A subtotal pericardiectomy was performed to remove the lesion, ensure long-term health, and prevent cardiac tamponade (Fig. 3A, B, asterisk). Thoracotomy revealed multifocal, small nodules localized on the inner surface of the thickened pericardial sac near the right atrium (Fig. 3C). The affected pericardium was surgically resected and subject to cytopathologic and histopathologic examinations. Fine needle aspirates of the pericardial mass contained more pleomorphic cells than those of the pericardial fluid. Of note, the exfoliated cells had epithelial to mesenchymal cytological features, including fusiform cells with indistinct margins and wispy cytoplasmic tails (Fig. 2C). Nuclear atypia, such as coarse chromatin and nucleoli, were marked (Fig. 2D). The cytological diagnosis was malignant mesothelioma with a differential diagnosis of hemangiosarcoma. After three days, approximately 30 mL of bloody pleural fluid was collected in thoracic drainage tube for 24 hours. The drainage tube was removed on the third day of hospitalization, and the effusion was confirmed to be spontaneously resolved within a week by thoracic radiography.

Figure 3. Diagnostic imaging and gross lesions of pericardial mesothelioma in a dog. Transverse (A) and dorsal (B) post-contrast computed tomographic images reveal a nodular lesion with heterogeneous contrast enhancement (arrow) adjacent to the pericardium on the right side of the heart, accompanied by pericardial effusion (asterisk). (C) With the phrenic nerve put aside, small nodular lesions (arrows) colored by brown to apricot are seen in a scattered manner on the inner pericardium near the right auricle.

Pericardial masses were fixed in 10% neutral-buffered formalin, processed, and embedded in paraffin wax. Three-micrometer sections were prepared and stained with hematoxylin and eosin (H&E). Microscopically, there was a papillary neoplasm comprised of clefts and papillary fronds of polygonal to cuboidal cells (Fig. 4A, B). Cells had variably discrete cell borders, a moderate amount of cytoplasm and a variably prominent nucleolus. Moderate cytologic atypia was present with occasional binucleated cells (Fig. 4C). The mitotic count was 9 in 10 high-powered 2.37 mm² fields (Fig. 4D). Chronic hemorrhage and lymphoplasmacytic infiltrates were present with a core of fibrovascular proliferation. Histopathologic diagnosis was inflamed neoplasm, with a primary differential diagnosis of epithelioid-appearing hemangiosarcoma and malignant mesothelioma. Immunohistochemistry (IHC) was performed in the serial H&E-stained sections to differentiate these two malignancies. After hydration and heat-induced antigen retrieval, endogenous peroxidase was depleted by incubating slides in 0.3% hydrogen peroxide for 1 hour. Sections were incubated at 4°C overnight with anti-cytokeratin, Vimentin, CD31, and thyroglobulin antibodies (Table 1), and then incubated with horseradish peroxidase-conjugated secondary antibody for 1 hour at room temperature. All sections were visualized using a 3,3’-diaminobenzidine solution. IHC revealed that atypical cells were extensively and strongly positive for both cytokeratin (Fig. 5A) and vimentin but tested negative for CD31 (Fig. 5B). Pre-existing blood vessels and well-differentiated thyroid tissue exhibited positive staining for CD31 (Fig. 5C) and thyroglobulin (Fig. 5D), respectively. These microscopic features were not supportive of a vascular tumor or neoplasm of ectopic thyroid origin. Given the more localized and extensive lesioned patterns coupled with the cytological and histological malignant findings, malignant mesothelioma was strongly favored.

Table 1 . Antibodies used in this study.

Antibody	Clone	Vendor	Catalogue number	Dilution
Cytokeratin	AE1/AE3	Bio SB	BSB5433	1:250
Vimentin	V9	Bio SB	BSB3027	1:1,000
CD31	Not applicable	Novus Biologicals	NB100-2284	1:250
Thyroglobulin	Not applicable	Invitrogen	PA5-82034	1:3,000

Figure 4. Histological examination of pericardial mesothelioma in a dog. (A) Low magnification of proliferative lesion on the inner pericardium. (B) Numerous papillary fronds extend from the surface of the inner pericardium. Chronic hemorrhage and lymphoplasmacytic infiltrates are present in the fibrovascular core. (C) Proliferative fronds surrounding cores of fibrovascular stroma and inflammation. (D) Higher magnification of papillary fronds with polygonal cells. Cellular morphology showing atypia and prominent nucleoli. Arrow indicates central mitotic figure. H&E stain, A: ×5, Scale bars = 200 μm, B, ×10, Scale bars = 100 μm, C, D: ×40, Scale bars = 50 μm.

Figure 5. Immunohistochemistry of pericardial mesothelioma in a dog. Proliferative cells strongly express cytokeratin (A) and vimentin (B), but not CD31 (C) and thyroglobulin (D). Note the strong positive expression of vimentin throughout the supporting fibrovascular stromal cells (B), CD31 on the endothelial cells of fibrovascular cores (C), and well-differentiated ectopic thyroid tissue (internal positive controls) (D). IHC stain, A, B, C, and D: ×100, Scale bars = 100 μm.

Postoperative CRP with a peak of 24.7 mg/dL (RI, 0.1-1 mg/dL) during the two days of hospitalization returned to within normal limits afterward. Adjuvant chemotherapy was performed on the dog after a discussion with the owner about the risk of potential recurrence and prognosis. Doxorubicin was considered one of the first-line regimens to treat canine MM (6). However, canine NT-proBNP (Bionote, Gyeonggi-do, South Korea) measured by a V200 Vcheck analyzer (Bionote) remained elevated for 14 days following the surgery (1,604.8 pmol/L; RI, 0-900 pmol/L). Meanwhile, the NT-proBNP concentration was 1,340.7 pmol/L prior a day of the surgery. No potential factors underlying the elevation in the NT-proBNP concentration were found, such as occult cardiomyopathy, proteinuria, chronic bronchitis, aspiration pneumonia, and pneumothorax.

To minimize potential chemotherapy-induced cardiotoxicity (5), the dog began to receive epirubicin (30 mg/m²) weekly every 3 weeks for 4 treatments. A total cumulative dose was designed to be less than 168 mg/m² to prevent potential cardiomyopathy (7). Thoracic radiography revealed no remarkable findings with a VHS of 10v. Two weeks later, the dog showed lethargy, vomiting, and diarrhea. CBC and manual differential count confirmed leukopenia (3,550 × 10³/μL; RI, 5,050-167,600 × 10³/μL) characterized by neutropenia (810 × 10³/μL; RI, 2,950-11,640 × 10³/μL). The 2^nd cycle dose of epirubicin was reduced to 24 mg/m², which did not result in neutropenia and improved clinical signs. Meanwhile, the NT-proBNP level decreased (1,318.6 pmol/L; RI, 0-900 pmol/L). After a week of the 2^nd cycle of chemotherapy, there was an increase in the VHS (12.1 v), suggesting cardiomegaly. The 3^rd cycle dose of epirubicin was increased to 27 mg/m², which was maintained during the 4^th cycle of chemotherapy. At the beginning of the 4^th cycle of chemotherapy, the VHS continued to increase (13 v). The total cumulative dose of epirubicin was 108 mg/m². After the completion of the 4 cycles of epirubicin treatment, the level of NT-proBNP returned within normal limits. The dog was presented for the evaluation of intermittent bleeding in the anal sac gland, which was resolved by the administration of amoxicillin-clavulanic acid (12.5 mg/kg, orally twice daily) and prednisone (0.3 mg/kg, orally twice daily). CBC results were within normal limits, except for non-anemic reticulocytosis at 151.1 × 10⁶/μL (RI, 10-110 ×10⁶/μL).

On the 30 days after the completion of chemotherapy and 108 days after pericardiectomy, thoracic radiography revealed a silhouette sign at the bilateral cranial lung lobe and caudomedial region of the left caudal lung lobe. In addition, echocardiography found mass-like soft tissue opacity near the heart with a small amount of serosanguineous fluid. Physical examination was unremarkable with no clinical signs other than a weight loss exceeding 8% over the past two weeks. A smear of the body cavity fluid revealed clusters of mesothelial cells with moderate cellular atypia, which shared numerous cytological features of pericardial effusion at the time of the tumor diagnosis. The owner denied treatment of intracavity or intravenous platinum-based drugs due to the dismal prognosis expected. The dog died at the home. The overall survival time was 132 days after the first presentation. A necropsy was not permitted by the owner.

Integrating clinical, imaging, clinicopathological, pathological, and immunohistochemical data is a critical strategy for diagnosing pericardial mesothelioma (6). Golden retrievers have a breed predisposition for pericardial MM (11). Pericardial effusion is a common finding in dogs with pericardial MM (19). Advanced diagnostic imaging, such as CT, is highly recommended to detect mesothelial lesions (13). More than 5 cytological criteria of atypia need to indicate neoplastic effusion, although MM should not be ruled out solely based on a lack of neoplastic cells. In our case, coupled with histologic features, molecular identification of tumor origin leveraged by antibody-based laboratory tests plays a critical role in accurately diagnosing pericardial mesothelioma. Notably, a localized, significant atypia, and increased mitotic activity were considered key components for recognizing atypical MM. Inanaga et al (3). support our notion, showing that localized mesothelioma is more likely to be neoplastic. Misdiagnosing pericardial MM with other malignancies, such as hemangiosarcoma, could underestimate the survival chance and prognosis of the patient when treated with chemotherapy (9). For example, dogs with hemangiosarcoma, which is the most common cause of pericardial effusion, tend to have shorter survival time than those with MM (116 days (14) vs. 234 days (11). Meanwhile, florid mesothelial hyperplasia cannot be completely precluded due to overlapping histological features, but it is a diagnostic challenge and an unavoidable limitation in this field (11,19).

Potential factors that predict the prognosis of canine MM involve surgical intervention, chemotherapy (6), drug response (11), and recurrent effusion (8,19). For example, dogs who received chemotherapy are shown to live 9 times longer than those without chemotherapy (234 days vs. 29 days, p-value = 0.001) (11). Complete resolution of effusion after the first chemotherapy administration positively correlated with survival in dogs with MM (415 days vs. 160 days, p-value < 0.01) (6). Pericardiectomy tends to improve the mean survival time of dogs with MM (366 days vs. 147 days, p-value = 0.18) (6). In our case, chemotherapy after pericardiectomy based on an early and accurate diagnosis was applied to the dog, which resolved the body cavity effusion. The time to diagnose pericardial MM from the first pericardial effusion event was 31 days, which is significantly shorter than the reported median time for dogs with MM (389 days) (11). At 108 postoperative days, despite these factors favoring a favorable prognosis, the patient underwent recurrent effusion. Effusion recurrence is a factor that significantly predicts a poor prognosis of pericardial MM in dogs (8). Our case supports previous findings that a recurrence-free interval of > 120 days could be a likelihood for the long-term survival of dogs with neoplastic mesothelial effusion (4,19).

Adjuvant chemotherapy is critical in prolonged survival in dogs with MM (19). Epirubicin – an epimer of doxorubicin – has less cardiovascular toxicity than doxorubicin but is presumed to exhibit the same clinical benefits against MM (15). In our case, we suggest clinical evidence that epirubicin positively contributes to eradicating potentially remaining neoplastic cells and/or metastasis. Epirubicin is demonstrated to induce apoptosis of MM (18) and inhibit the metastatic property of neoplastic cells (17). Of note, clinical improvement was obvious during epirubicin treatment, particularly showing no effusion recurrence. However, regarding recurrent effusion, several scenarios need to be considered. First, it might be possible that the neoplastic cells were primary to the heart and independent of the clinical course of pericardial mesothelioma. Alternatively, it might be also plausible that a subset of primary mesothelial neoplastic cells colonized in the heart and acquired resistance to epirubicin. Indeed, pericardial MM cells could diffusely encase the heart (16). A dog was reported to have undergone fulminant metastatic mesothelioma after a month of pericardiectomy (10). Epirubicin could have inhibited, at least to some extent, the potential metastasis in our case. Supporting this, subclones of mesothelioma that emerged could be more locally aggressive, more likely to metastasize, and less responsive to the treatment (11). Finally, although it is unlikely, epirubicin might have been only palliatively effective for pericardial mesothelioma (1). To attest to our notion, e.g. to determine the origin of the cardiac neoplasm, additional diagnostic investigations, including necropsy, were indispensable, but the owner declined it. Cardiac mesothelioma and hemangiosarcoma are considered the top differential diagnoses for differential diagnosis of cardiac masses. Our case strongly suggests that future studies need to optimize the treatment modality of epirubicin for pericardial mesothelioma in dogs, such as increase in the total cumulative dose of epirubicin until it causes cardiotoxicity (7), or administration of platinum-based drugs that were known to be effective for mesothelioma following epirubicin treatment (12). Clinical trials using epirubicin for palliative treatment, including a group of effusion drainage alone but without surgical intervention, might elaborate the clinical benefit of epirubicin on canine pericardial mesothelioma.

A study reported potential epirubicin cardiotoxicity (a total cumulative dose of 168 mg/m² for 32 months) in a dog who had been long-term pretreated with doxorubicin (a total cumulative dose of 126 mg/m²) (7). In our case, the total cumulative dose of epirubicin was 108 mg/m². NT-proBNP level maintained increased after pericardiectomy, which preceded the development of cardiomegaly. Although cardiomegaly occurred during the epirubicin treatment, our case favors breed predisposition especially related to large-breed dogs (4). Furthermore, acute and chronic dilated cardiomyopathy was reported in 23.1% of large breeds after partial pericardiectomy (4). Causality is unknown, but spontaneous cardiomegaly following pericardiectomy might be functionally associated with reduced left ventricular ejection fraction (2). Veterinary practitioners should be aware of cardiomegaly that could spontaneously develop after pericardiectomy in dogs with pericardial effusion.

To our knowledge, this is the first attempt to provide information on the clinicopathologic findings, diagnostic investigations, and clinical course of a dog with diagnostically challenging pericardial MM following epirubicin treatment. Our case suggests the usability of epirubicin. However, future studies are strongly warranted to draw conclusion whether epirubicin alleviates cardiac burden and achieves effective preemptive management of canine pericardial MM. Finally, our results highlight the importance of considering pericardial biopsy and coupled molecular works early in the disease course of dogs presenting recurrent pericardial effusion of unknown etiology.

We express our appreciation to all the pathologists from IDEXX Laboratories involved with the pathology and immunohistochemistry reports used in this study. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE) (#RS-2023-00241779). This research was also supported by Regional Innovation Strategy (RIS) through the NRF funded by MOE (#2023RIS-009).

The authors have no conflicting interests.