Metastatic carcinoma of unknown origin in a dog

Abstract

Although cancer of unknown primary origin (CUP) is well-described in the human literature, it is not as well-understood within veterinary medicine. This case report represents one of few focused on describing CUP in a dog.

Key clinical message:

Metastatic CUP should be considered as a differential diagnosis despite being a rare disease entity that is infrequently reported within the veterinary literature.

Cancer of unknown primary origin (CUP) is a histologically diagnosed metastatic cancer in which the primary site cannot be identified despite a thorough diagnostic workup. In humans, only 20% of primary sites are identified prior to the patient’s death (1), and the primary tumor remains unidentified following autopsy in 70% of cases (2,3). Cancer of unknown primary origin represents 3 to 5% of all human malignancies, making it 1 of the 10 most frequently diagnosed cancers worldwide (4).

Although CUP is relatively well-described in humans, it is sparsely described in the veterinary literature. To the authors’ knowledge, there are few studies available for review (5–8) with a focus on CUP in veterinary patients. The purpose of this case report is to describe the clinical presentation of, diagnostic approach to (including the use of positron emission tomography-computed tomography, PET/CT), treatment of, and outcome of CUP in a dog.

Case description

An 11-year-old intact male Labrador retriever dog was presented to the primary veterinarian for evaluation of a right ventral cervical mass. Pertinent past medical history included well-controlled diabetes mellitus and left forelimb digit 4 amputation to address a mass with unknown histological diagnosis or physical description. On physical examination, a right-sided, 5-centimeter (diameter) submandibular mass was noted. Cytology was consistent with carcinoma. The dog was returned 10 d later for surgical removal. Preoperative blood work revealed hyperglycemia [315 mg/dL; reference interval (RI): 74 to 153 mg/dL] and hyperglobulinemia (5.1 g/dL; RI: 2.5 to 4.5 g/dL). Histopathology was consistent with narrowly excised (margins < 1 mm) adenosquamous carcinoma. The mitotic count was 18 per 10 400× fields, including bizarre mitotic figures, and vascular invasion was present. As lymphoid proliferation was identified at the periphery, metastatic carcinoma to the right mandibular lymph node was suspected and referral was recommended.

Fifty-five days following initial diagnosis, the dog was presented to The Ohio State University Integrated Oncology service (Columbus, Ohio, USA). Physical examination revealed prominent peripheral lymph nodes. Oral examination revealed no primary lesions, and the palatine tonsils were grossly normal. Three-view thoracic radiographs revealed no obvious pulmonary metastatic disease. Computed tomography (CT) of the skull and abdomen was completed under sedation using a multidetector CT, 64-slice with double z-sampling capabilities totaling 128-detector (GE Revolution EVO; GE Healthcare, Waukesha, Wisconsin, USA), at 120 kVp in a transverse plane and a slice thickness of 1.25 mm. Pre- and post-contrast imaging was completed using 60 mL (240 mg/mL) of iohexol administered IV through a peripheral indwelling catheter. The study was reformatted into dorsal, sagittal, and transverse planes. Skull CT showed moderate, mildly contrast-enhancing soft tissue stranding and lobular nodules in the region of the previously resected right mandibular lymph nodes. Given the recent surgery, these were thought to represent right mandibular subcutaneous fibrosis or cellulitis, with incompletely excised mandibular lymph nodes, or recurrent or persistent neoplasia (Figure 1). The right medial retropharyngeal lymph node was enlarged and concerning for either neoplasia or reactive inflammation. No abnormalities consistent with a primary tumor were identified. Abdominal CT showed no evidence of metastatic or primary neoplasia. The left and right medial retropharyngeal lymph nodes and left mandibular lymph node were sampled. Left mandibular and left medial retropharyngeal lymph node cytology revealed a mild expansion of intermediate-to-large lymphocytes consistent with lymphoid hyperplasia. Right medial retropharyngeal lymph node cytology revealed moderate lymphoid and plasma cell hyperplasia. Given these findings, the dog was diagnosed with metastatic carcinoma to the right mandibular lymph node of unknown origin.

Figure 1.

Initial (postoperative) skull computed tomography (CT) imaging at the level of the right mandibular mass resection, showing the mildly contrast enhancing, wispy, soft tissue stranding and ovoid to lobular nodules (arrowheads). Axial plane post-contrast image (A) and coronal reconstruction (B) at the level of the normal-appearing left mandibular lymph nodes. C — Right parasagittal post-contrast reconstruction. Soft tissue window width and window level 400/40; the slice thickness was 1.25 mm; the image was acquired with 120 kVp.

FV — Facial vein; SG — Mandibular salivary glands.

To identify the primary tumor, full-body positron emission tomography-computed tomography (PET/CT) was completed under general anesthesia 77 d following initial diagnosis. Blood glucose prior to 2-deoxy-2-[18F]-fluorodeoxyglucose (18F-FDG) injection was 301 mg/dL. A dose of 3 MBq/kg (0.08 mCi/kg) of 18F-FDG was administered intravenously. The list mode time-offlight raw data was acquired continuously on the Vereos PET/CT system (Philips, Amsterdam, Netherlands). Two-millimeter isotropic voxel data sets (288 × 288 matrix size using 576-millimeter field of view) and 90 seconds/bed were reconstructed. Transverse CT imaging was acquired using the multi-slice system at 120 kVp, 150 mAs, and reconstructed with a 0.625-millimeter slice thickness. A whole-body positron emission tomography (PET) image was acquired starting at 85 min after injection. The PET/CT showed mild to moderately increased activity in the region of the previously resected subcutaneous right mandibular neoplasm. There was mild to moderately increased left mandibular lymph node, bilateral parotid, medial retropharyngeal, and superficial cervical lymph node activity. Similar increased activity was seen in the popliteal lymph nodes bilaterally. In the palatine tonsils, there was asymmetrical increased activity (moderate on the right and mild on the left). Separately, a right caudal cervical region of mildly increased activity was seen in the dorsal subcutaneous tissues, associated with soft tissue stranding and an irregular soft tissue nodule. This lesion was suspected to be cellulitis secondary to iatrogenic insulin injections. The right mandibular subcutaneous and lymph node findings were nonspecific and were considered to represent either reactive inflammation and fibrosis secondary to the recent surgery or recurrent neoplasia. Increased activity in the palatine tonsils has been described as a common indeterminate finding due to lymphoid reactivity, inflammation, or periodontal disease (9). The remaining regions of increased activity were considered non-neoplastic. No primary neoplastic lesion was identified (Figure 2).

Figure 2.

A — Dorsal plane whole-body grayscale positron emission tomography (PET) image. B — Magnified head dorsal plane grayscale PET image. C — Axial plane fused PET/CT at the level of the right mandibular mass resection and left mandibular lymph nodes. Note the mild to moderate increased activity in the region of the previously resected subcutaneous right mandibular neoplasm, with similar increased activity in the remaining cervical lymph nodes. Note the irregular region of right caudal cervical increased activity. The increased 2-deoxy-2-[18F]-fluorodeoxyglucose (18F-FDG) uptake in the brain, salivary tissue, and abdominal structures are normal findings. The increased activity along multiple joints is associated with osteoarthritis.

Given these findings, histopathological review from a second pathologist was obtained from the original biopsy. This confirmed the initial diagnosis of a well-differentiated adenosquamous carcinoma with widespread nodal effacement and vascular invasion (Figure 3). Restaging was recommended every 3 mo with skull CT and thoracic radiographs.

Figure 3.

Histopathology. A — Light photomicrograph depicting the metastatic neoplasm expanding the subcapsular sinus and compressing the preexisting lymph node cortex (arrows). Hematoxylin and eosin, scale bar = 500 μm. B — Light photomicrograph depicting the neoplastic cell population with neutrophil infiltration. Hematoxylin and eosin, scale bar = 20 μm.

One hundred and fifty-five days following initial presentation, a prominent, firm, right superficial cervical lymph node was identified. Cytology revealed large, ovoid-to-polygonal, individualized and tightly cohesive epithelial cells with moderate nuclear to cytoplasmic ratio, moderate anisocytosis and anisokaryosis, and frequent multinucleation, consistent with metastatic squamous cell carcinoma. Thoracic radiographs revealed no evidence of pulmonary metastasis. It was recommended that the dog be returned to the clinic in 1 wk for repeat skull, cervical, and thoracic CT scans.

Computed tomography of the skull, neck, and thorax was completed under general anesthesia. Pre- and post-contrast imaging was done using 80 mL (240 mg/mL) of iohexol administered IV. The CT imaging showed similar right mandibular soft tissue stranding and nodules in the region of the previous surgical site. Progressive enlargement of the right medial retropharyngeal, left mandibular, bilateral superficial cervical, and axillary lymph nodes was noted. The lymph node changes were concerning for progressive neoplastic or reactive lymphadenopathy. No pulmonary metastasis was identified on thoracic CT. The right mandibular soft tissues, right and left medial retropharyngeal lymph nodes, and right axillary lymph node were sampled using ultrasound guidance. Cytology of the right retropharyngeal lymph node revealed large, ovoid, individualized, tightly cohesive epithelial cells displaying anisocytosis and anisokaryosis consistent with metastatic carcinoma. Cytology of the right axillary lymph node revealed the lymphoid population to be heterogeneous and consisting primarily of small, well-differentiated lymphocytes with plasma cells frequently observed; these were most consistent with moderate plasma cell hyperplasia. Given these findings, the recommendation was for surgical removal of the affected lymph nodes followed by radiation therapy and/or chemotherapy versus chemotherapy alone.

The dog was presented to the Ohio State University Integrated Oncology service 175 d following initial diagnosis, for surgery. During patient preparation, the left superficial cervical lymph node was noted to be enlarged and elected for inclusion in the planned lymph node extirpation. Routine left mandibular, bilateral medial retropharyngeal, bilateral superficial cervical, and right deep cervical lymph node extirpations were completed. The excised lymph nodes were submitted for histopathology. The right mandibular lymph node could not be identified, though there was a significant amount of scar tissue appreciated at this site. No complications were encountered in the perioperative period and the dog recovered uneventfully. The dog received ropivacaine (1.0 mg/kg) infiltration as an incisional block and carprofen (2.2 mg/kg), SC, in recovery.

Postoperatively, the dog was monitored in the intensive care unit. The dog was managed on Plasma-Lyte (Baxter Healthcare, Deerfield, Illinois, USA) (49.6 mL/kg per day, IV); acepromazine (0.02 to 0.03 mg/kg, IV, q6h as needed); hydromorphone (0.05 mg/kg, IV, q6h); trazodone (3.6 mg/kg, PO, q8h as needed); carprofen (1.8 mg/kg, PO, q12h); cephalexin (24.3 mg/kg, PO, q12h); and Novolin N (Novo Nordisk, Princeton, New Jersey, USA) [0.75 U/kg (32 units), SC, q12h if eating]. The dog was discharged from the hospital 24 h postoperatively and recovered uneventfully.

Histopathologic evaluation of the right cranial deep cervical, right superficial cervical, right retropharyngeal, and left superficial cervical lymph nodes revealed that the cortex, medulla, and subcapsular spaces were expanded and infiltrated by a poorly demarcated, unencapsulated neoplasm. The neoplasm was composed of polygonal cells forming islands and cords supported by a variable amount of fibrous stroma (Figure 2). The neoplastic cells had 1 ovoid nucleus containing 1 to 3 prominent nucleoli, abundant eosinophilic cytoplasm, and distinct cell borders. Multinucleated neoplastic cells were frequently present. Anisokaryosis and anisocytosis were both marked. There were 21 mitotic figures per 10 400× fields present, with frequent atypical mitotic figures. Individualized round neoplastic cells and hypereosinophilic angular neoplastic cells with pyknotic nuclei were scattered within the neoplasm. Small numbers of lymphocytes and plasma cells infiltrated the neoplastic tissue. These findings were supportive of moderately differentiated metastatic carcinoma in several lymph nodes, consistent with the previously diagnosed adenosquamous carcinoma. Histopathology results for the left mandibular lymph node and left retropharyngeal lymph node revealed moderate, multifocal, lymphoid nodular hyperplasia with hemorrhage and sinus histiocytosis composed of scattered, hemosiderin-laden macrophages.

One hundred and eighty-nine days following initial presentation, cytotoxic chemotherapy was started using carboplatin (300 mg/m², IV via the left lateral saphenous vein, every 3 wk) with a plan for a total of 5 to 6 treatments. Maropitant (1.9 mg/kg, PO) was administered before each carboplatin treatment.

Two hundred and twenty-three days following initial presentation, the dog developed right thoracic limb edema. Physical examination revealed the presence of a firm, enlarged, and painful right axillary lymph node. The dog was hospitalized overnight and managed on lactated Ringer’s solution (68.1 mL/kg per day, IV); methadone (0.2 mg/kg, IV, q6h as needed); gabapentin (7.2 mg/kg, PO, q8h as needed); carprofen (1.8 mg/kg, PO, q12h); maropitant (1 mg/kg, IV, q24h); and Novolin N [0.75 U/kg (32 units), SC, q12h], with a plan for CT and lymphangiogram of the right thoracic limb on the following day. Warm compressing of the limb was done q8h.

A CBC and serum biochemistry were obtained prior to the procedure. The CBC revealed a thrombocytopenia (52 K/μL; RI: 145.0 to 463.0 K/μL) and leukocytosis characterized by a mature neutrophilia [WBC: 16.95 K/μL (RI: 4.8 to 13.9 K/μL), neutrophils: 14.41 K/μL (RI: 2.6 to 10.8 K/μL)]. An elevated glucose (277 mg/dL; RI: 67 to 127 mg/dL) was noted on serum biochemistry. Computed tomography imaging of the skull, neck, forelimbs, and thorax was done under general anesthesia, using the imaging protocol initially described. Pre-contrast imaging was obtained, followed by a right forelimb lymphangiogram. Twenty milliliters (240 mg/mL) of iohexol was injected into the subcutaneous right interdigital spaces, followed by massage. Following lymphangiography, 86 mL (240 mg/mL) of iohexol was administered, IV, using a peripheral indwelling catheter for post-contrast evaluation. No lymphatic obstruction or vascular thrombus was identified to explain the cause of the thoracic limb edema. The right axillary lymph node was mildly rounded and enlarged (measuring 13 mm), with heterogenous contrast enhancement. Although a clear cause of this acute change was not identified, switching chemotherapy agents from carboplatin to toceranib (2.6 mg/kg, PO, every other day) was recommended due to suspected progressive disease.

The dog returned 232 d following initial diagnosis for a recheck examination. Although swelling of the right thoracic limb had improved, a 7-centimeter × 4-centimeter open wound, caudal to the right elbow and extending into the subcutis, had developed. Sedated open wound management was completed every other day for 2 wk until primary wound closure could be done under general anesthesia, via a right thoracodorsal axial skin flap with Jackson-Pratt drain placement. Postoperatively, the dog developed a 9-centimeter area of dehiscence that was treated with open wound management until primary closure could be completed. Throughout the course of wound management, topical and oral antibiotic therapy was instituted and adjusted based on culture results.

Restaging via sedated skull and thoracic CT imaging was carried out 288 d following initial presentation, using the imaging protocol initially described. The right mandibular subcutaneous region remained unchanged. There was no evidence of neoplastic recurrence. A similar mild right axillary lymphadenomegaly was noted that was suspected to represent reactivity or neoplastic infiltrate. No pulmonary metastasis or intrathoracic disease was present. No treatment changes were made based on these results.

The dog was evaluated by the primary veterinarian 31 d following his last visit at the Ohio State University due to reported seizure activity. The dog was hypoglycemic and the insulin dose was decreased. The dog was euthanized 7 d later due to continued poorly controlled diabetes. No necropsy was performed. In total, the patient survived for 363 d following the initial diagnosis.

Discussion

Cancer of unknown primary origin is described as a metastatic tumor in which the site of primary origin is not identified after a complete diagnostic workup (10). Clinical presentation is variable, and patients are categorized by subgroup, ranging from poorly differentiated though highly chemosensitive carrying a more favorable prognosis to uncategorized with unpredictive metastatic potential and aggressive tendencies (4). Clinical detection of the primary tumor often does not precede detection of systemic metastasis, and clinical signs displayed due to metastasis may differ from those associated with the primary tumor (11).

In humans, > 50% of patients diagnosed with CUP will present with multiple sites of involvement, including liver, bone, lung, or lymph nodes (12). Most commonly, the primary site localizes to the lungs or within the biliopancreatic tract (3). Adenocarcinomas are the most frequently diagnosed (10). Cancer of unknown primary origin represents 3 to 5% of all human cancer diagnoses (10); however, it is sparsely reported in the veterinary literature and information regarding tumor biology, prognosis, and optimal treatment is lacking.

A study examining CUP in dogs reported a median age of 10 y at the time of diagnosis; and a majority of cases (85.7%) presented with clinical signs, such as dyspnea and lethargy, that had been present for a median of 2 wk (5). A smaller percentage (14.3%) were asymptomatic at the time of presentation (5). The few other case reports of metastatic carcinoma of unknown primary origin involve an 11-year-old, a 10-year-old, and a 9-year-old dog presenting with progressive ataxia and paraparesis of duration 1 mo, lameness and pain of duration 1 mo, and 10-day history of cough and dyspnea, respectively (6–8). Similar to reports within the human literature, Rossi et al (5) reported that carcinoma was the most common diagnosis (57.1%), with bone, lymph nodes, lung, and spleen as the most common metastatic sites. This case further supports the findings herein with the final diagnosis of moderately differentiated metastatic carcinoma in several lymph nodes.

Discovery of the primary tumor allows for optimization of treatment protocols, which may positively influence patient prognosis. In human medicine, cross-sectional, whole-body imaging is the gold standard, as the unknown primary mass can be located anywhere within the body, making conventional radiography and ultrasound imaging inferior (5,10). Moreover, the combination of 18F-fluorodeoxyglucose (FDG)-PET and CT has become increasingly popular, as the use of the radiotracer 2-deoxy-2-[18F]-fluorodeoxyglucose allows for detection of small lesions or pathological changes often missed on other modalities (13); its good sensitivity and specificity are most notable in diagnosing head, neck, and lung cancers (14). In humans, FDG-PET/CT detected 37% of primary lesions with sensitivity and specificity both at 84% in a previous meta-analysis (15), though detection rates of up to 53% have been reported (16,17). The use of FDG-PET/CT did not detect the primary lesion in this case, and its utility in veterinary patients remains unknown; however, an extended workup via PET/CT imaging should be considered when whole-body CT fails to detect the primary tumor.

The use of immunohistochemistry (IHC) to suggest primary origin has been described in the human literature. It has been reported that IHC has 35 to 40% accuracy in predicting a single primary tumor in early metastatic cancers (12), with panels of markers more effective than single biomarkers in identifying the primary site in patients with a diagnosis of adenocarcinoma (18). However, it is important to note that these studies were done in patients with known metastatic cancers versus those with a diagnosis of CUP. Although IHC was not used in the present case, a small number of markers have been used in dogs to determine the site of origin and differentiate metastatic carcinomas, including thyroid transcription factors-1 (TTF-1) for primary pulmonary or thyroid neoplasms and uroplakin III for urothelial neoplasms (19,20). The current recommendation is to use markers as an aid in conjunction with physical examination and imaging findings, as applicable (5).

Overall, the prognosis of human patients diagnosed with CUP is often poor, with a median survival time of 4 to 12 mo (21). Most commonly, those within the favorable subgroups are treated with locoregional treatment or systemic, platinum-based chemotherapy, whereas those within the unfavorable subgroups are treated with empirical combination chemotherapy (22,23). Rossi et al reported a median survival time of 30 d for all dogs diagnosed with CUP and a median survival time of 80 d for those that underwent treatment (5). In this study, the survival time following surgery and systemic therapy with carboplatin and toceranib was 363 d. Other case studies have reported metastatic carcinoma of unknown primary origin involving the lumbar spine (6); scapula (7); and epicardium, sternal lymph nodes, and multiple lung lobes (8). In those studies, median survival times were 133 d, 737 d, and 457 d following metronomic chemotherapy with cyclophosphamide and piroxicam, amputation and adjuvant chemotherapy with doxorubicin, and piroxicam and conservative management, respectively (6–8). At present, specific diagnostic and treatment guidelines for veterinary patients diagnosed with CUP do not exist.

A limitation of the present case report is the lack of histopathological diagnosis following digit amputation 4 y prior to presentation. One differential diagnosis for this tumor would have been a digital squamous cell carcinoma (SCC). However, given the prolonged period from digital tumor removal to detection of the mandibular mass, this may be less likely to be related. Furthermore, the mandibular lymph node would not be the sentinel lymph node for the digits. Although histological examination of the initial lesion was most suggestive of metastasis within the mandibular lymph node, it was not clear whether the lymphoid proliferation associated with the tumor was representative of an enfaced lymph node or chronic lymphoplasmacytic peritumoral inflammation. Moreover, the biological behavior reported in the current case differs from previously described cases in which more widespread metastatic disease associated with shorter survival times were commonly reported. As such, additional potential differential diagnoses must include primary SCC originating from excretory ducts of the salivary glands or from remnants of the thyroglossal duct with associated severe inflammation. Another recognized limitation of this study is the lack of a necropsy following euthanasia. Although this is not mandatory in the diagnosis of CUP, further pathological studies to include sampling of the right axillary lymph node and region of the right mandibular lymph node may have been valuable.

This case report describes the clinical characteristics and treatment outcome of a dog diagnosed with metastatic CUP of several lymph nodes. Lymph node extirpation and adjuvant carboplatin and toceranib chemotherapy were performed for locoregional control. The dog survived for 363 d following the initial diagnosis. In conclusion, metastatic CUP should not be neglected as a differential diagnosis despite being a rare disease entity that is infrequently reported within the veterinary literature. Aggressive therapy combining surgery and chemotherapy can result in a long survival time.