Primary colonic hemangiosarcoma in a dog

Abstract

A 13-year-old spayed female miniature dachshund dog was presented with a history of hematochezia, melena, and weight loss. A colonic mass detected by exploratory celiotomy was resected and diagnosed as hemangiosarcoma (HSA). Liver lobectomy for hepatic HSA was performed 894 days after surgery. Adjuvant chemotherapy was declined. This is the first documentation of a canine colonic HSA.

Case description

A 13-year-old, 3.5-kg spayed female miniature dachshund dog was referred to the Aikawa Veterinary Medical Center with a 7-day history of hematochezia and melena. The patient was bright, alert, and responsive with a good appetite but had lost weight (body condition score: 2/5). The remainder of this patient’s physical examination findings including digital rectal examination was unremarkable.

A complete blood (cell) count (CBC) detected a mild regenerative anemia [packed cell volume (PCV): 33.8%; reference range (RR): 37% to 55%, Reticulocyte Production Index: > 2]. White blood cell (WBC) and platelet counts and the serum biochemical profile were within normal limits. Red blood cells were detected in the feces.

Three-view thoracic and 2-view abdominal radiography and echocardiography detected no abnormalities. Abdominal ultrasonography detected 2 hypoechoic liver lesions (0.6 × 0.4 cm, 0.8 × 0.5 cm), 1 hyperechoic splenic lesion (0.2 × 0.2 cm) and no lymphadenopathy. The remaining organs (gall bladder, kidney, urinary bladder, stomach, duodenum, pancreas, and small and large intestines) were unremarkable based on abdominal ultrasound examination.

A gastrointestinal ulcer and colitis were suspected and medical treatment with famotidine (Gaster; Astellas Pharma, Tokyo, Japan), 0.7 mg/kg body weight (BW), PO, q24h, sucralfate (Ulcerlmin; Chugai Pharmaceutical, Tokyo, Japan), 500 mg, PO, q12h, metronidazole (Flagyl; Shionogi & Co., Osaka, Japan), 17.8 mg/kg BW, PO, q12h, and salazosulfapyridine (Salazopyrin tablet; Pfizer Japan, Aichi, Japan), 35.1 mg/kg BW, PO, q12h, was started. Hematologic and serum biochemistry panels, conducted at the referring veterinarian’s office on the next day, were within normal limits [PCV: 39.8%, WBC: 9900 cells/μL, total protein (TP): 54 g/L, albumin: 28 g/L]. Fecal hemoglobin measured with a latex aggregation test was 700 ng/mL (normal value: < 20 ng/mL).

Since relapse of intestinal bleeding was subsequently noted, upper and lower gastrointestinal endoscopy was performed on day 21 but detected no gastrointestinal tract inflammation, ulcers, or neoplastic lesions. Gastrointestinal biopsies were taken from the gastric cardia, fundus, angular incisure, and pylorus. Specimens were also obtained from the proximal and distal duodenum, ileum, ileocecum, and proximal and distal colon. No abnormalities were noted on the gross morphology of gastrointestinal mucosal surface or on histopathology of the collected samples. Contrast gastrointestinal radiography using barium sulfate (Baritogen-sol; Fushimi Pharmaceutical, Kagawa, Japan), 5.7 mL/kg BW, PO, did not detect filling defects, intestinal obstruction, or any other abnormalities.

The dog was readmitted 5 d later because of decreased mobility, ataxia, and progressive worsening of hematochezia and melena. Physical examination revealed pale mucous membranes, weak femoral pulses, and no abnormalities on thoracic auscultation and abdominal palpation. Neurologic examination revealed depressed but responsive mentation.

Severe anemia (PCV: 15.0%, RR: 37% to 55%, Reticulocyte Production Index: > 2) with mild thrombocytopenia (16.1 × 10⁴/μL, RR: 20.0 to 50.0 × 10⁴/μL), hypoalbuminemia (12 g/L, RR: 26 to 40 g/L), and hypokalemia (2.9 mEq/L, RR: 3.8 to 5.0 mEq/L) was observed. Since a definitive cause of the intestinal bleeding was not identified, an exploratory laparotomy was carried out the following day. The dog was administered atropine sulfate (Atropin sulfate injection; Mitsubishi Tanabe Pharma, Osaka, Japan), 0.05 mg/kg BW, SC, morphine hydrochloride (Morphine hydrochloride; Shionogi, Osaka, Japan), 0.5 mg/kg BW, IM, and midazolam (Dormicam injection; Astellas Pharma, Tokyo, Japan), 0.3 mg/kg BW, IV, before induction with propofol (Propofol 1% injection; Mylan, Tokyo, Japan), 4 mg/kg BW, IV. Anesthesia was maintained with isoflurane (Isoflu; DS Pharma Animal Health, Osaka, Japan) in oxygen. Pain management included epidural analgesia with morphine hydrochloride (0.1 mg/kg BW) and bupivacaine (Marcain Injection 0.25%; Aspen pharma, Tokyo, Japan), 0.2 mg/kg BW, after induction.

Full surgical abdominal exploration performed after whole blood transfusion (22 mL/kg BW, IV) with continuous IV potassium supplementation (0.04 mmol/kg BW per hour, IV) revealed a round dark red-violet mass (2.0 × 2.0 × 2.0 cm) located at the proximal colon (Figure 1A).

Figure 1.

A — Full surgical abdominal exploration revealed a round dark red-violet mass located between the cecum and proximal colon. B, C — Histopathological diagnosis was consistent with an intestinal HSA with complete excision.

The mass did not adhere to surrounding tissues, and the serosal surface of the colon was intact. The remainder of the abdominal organs appeared normal; therefore, the colon was considered to be the site of the primary lesion. The mass was resected with 3-cm margins and an end-to-end colonic anastomosis with 3-0 polydioxanone suture in a simple interrupted pattern was performed. The abdominal cavity was lavaged with 1.5 L of warm sterile saline (0.9% NaCl) solution prior to closure.

The resected mass was fixed in neutral-buffered 10% formalin, embedded in paraffin, sectioned in slices 5-μm thick, and stained with hematoxylin and eosin (H&E). The histopathological diagnosis was consistent with an intestinal hemangiosarcoma (HSA) with complete excision (Figures 1B, C; 2A, B). Tumor cells located between the submucous and muscular layers did not destroy the mucosal membrane.

Figure 2.

Histopathological examination of (A, B) colonic HSA and (C, D) hepatic HSA. A, C — Hematoxylin and eosin stain revealed marked nuclear atypia (arrowhead) and moderate mitotic activity of tumor cells (arrow). B, D — Immunohistochemical staining confirmed moderate membranous-cytoplasmic expression of factor VIII in tumor cells.

Hematochezia and melena subsided 2 d after surgery, and anemia improved gradually. Antibiotics were administered for 7 d (Cefazolin injection; Fujita Pharmaceutical, Tokyo, Japan), 20 mg/kg BW, IV, q12h, and as the dog continued to eat and drink without assistance, IV fluid therapy was discontinued. No complications were noted during hospitalization and the patient was discharged 7 d after surgery. The owner declined doxorubicin-based adjuvant chemotherapy.

Bimonthly postoperative recheck abdominal ultrasonography revealed increased numbers and enlargement of hypoechoic liver lesions 612 d after resection of the colonic HSA. However, fine-needle aspiration was not performed due to concerns of bleeding risks. Contrast-enhanced computed tomography was declined by the owner.

The patient was presented with mental dullness, anorexia, lethargy, and pale mucosal membranes 894 d after surgery.

A CBC revealed WBC: 9400 cells/μL, PCV: 45.2%, platelets: 18.6 × 10⁴/μL with a normal serum biochemical profile. Severe fluid accumulation in the abdominal cavity was suspected after abdominal radiography and ultrasonography. Abdominal paracentesis revealed frank hemorrhage and the character of the effusion was as follows; WBC: 9900 cells/μL, PCV: 45.2%, platelets: 6.3 × 10⁴/μL, TP 54 g/L, specific gravity (SG) 1.038, glucose 5.6 mmol/L. Bleeding from the hypoechoic liver lesions was suspected.

Since anemia and thrombocytopenia progressed over the next day (WBC: 13 400 cells/μL, PCV: 22.0%, platelets: 2.1 × 10⁴/μL), an exploratory laparotomy was performed.

Hemoabdomen due to ruptured left-sided liver lesions was diagnosed and a total left hepatic lobectomy was performed using a TA-30 stapling device with V3 cartridge (Tyco Healthcare, Princeton, New Jersey, USA) following a whole blood transfusion (22 mL/kg BW, IV). Although grossly similar lesions were detected in the right medial liver lobe (2.8 × 3.1 × 2.7 cm), this lobe was not bleeding, and lobectomy was not performed.

Contrast-computed tomography scanning performed after the liver lobectomy detected a hypodense, non-enhancing cystic lesion in the right middle liver lobe and no leakage of contrast medium from the right liver lobe lesions. The gallbladder was compressed and displaced caudally due to the mass (Figure 3). There were no metastatic lesions noted in the brain, chest, or abdominal cavity.

Figure 3.

Contrast enhanced computed tomography scanning performed after emergency liver lobectomy detected a 3-cm hypodense, non-enhancing cystic lesion in the right middle liver lobe (asterisk). The gallbladder was compressed and displaced caudally due to mass effect (arrow). Transverse image of cranial (A) and middle (B) part of liver.

Vitamin K injection (3 mg/kg BW, SC) and Ringer’s solution with 5% D-glucose, gabexate mesilate (1 mg/kg BW per hour), glutathione (1 mg/kg BW per hour), glycyrrhizin (0.1 mg/kg BW per hour), and vitamin B complex (B₁, B₂, B₆, B₁₂) infusion therapy were administered after surgery. Based on the blood test results, disseminated intravascular coagulation (DIC) was suspected [thrombocytopenia, prolonged prothrombin time, increase of fibrin degradation products (FDP)] (Table 1). Despite continuous IV supplementation with potassium (0.04 mmol/kg BW per hour) and dalteparin sodium injection (Fragmin, Kissei Pharmaceutical, Tokyo, Japan), 100 U/kg BW, SC, q24h, anemia and thrombocytepenia progressed and the patient died 5 d after the emergency liver lobectomy. Necropsy was not performed due to the owner’s decision.

Table 1.

Results of blood tests after emergency liver lobectomy.

	Value	Reference range
WBC	13.2 × 109/L	6 to 12 × 109/L
PCV	21.4%	37% to 55%
Platelets	31 × 109/L	200 to 500 × 109/L
Total protein	37 g/L	50 to 72 g/L
Albumin	18 g/L	26 to 40 g/L
ALT	> 16.7 μkat/L	0.28 to 1.30 μkat/L
AST	15.7 μkat/L	0.28 to 0.73 μkat/L
ALP	7.97 μkat/L	0.78 to 4.24 μkat/L
GGT	0.02 μkat/L	0.08 to 0.23 μkat/L
Tbil	5.1 μmol/L	1.7 to 8.6 μmol/L
NH3	15.7 μmol/L	9.3 to 43.7 μmol/L
Tchol	3.2 mmol/L	2.9 to 8.1 mmol/L
BUN	5.5 mmol/L	3.3 to 10.4 mmol/L
Glucose	8.3 mmol/L	4.2 to 7.1 mmol/L
Partial thromboplastin time	14 s	10 to 16 s
Prothrombin time	11 s	6 to 8 s
Fibrinogen	3.44 μmol/L	2.53 to 11.03 μmol/L
FDP	43.7 mg/L	0.0 to 5.0 mg/L
Antithrombin 3	147%	102% to 156%

WBC — white blood cells; PCV — packed cell volume; ALT — alanine aminotransferase; AST — aspartate aminotransferase; ALP — alkaline phosphatase; — ammonia; GGT — gamma-glutamyl aminotransferase; Tbil — total bilirubin; NH₃ Tchol — total cholesterol; BUN — blood urea nitrogen; FDP — fibrin degradation products.

The histopathological diagnosis of left-sided liver mass was consistent with HSA with incomplete excision (Figure 2C, D).

Discussion

Hemangiosarcoma (HSA) is a highly malignant neoplasm of vascular endothelial origin that develops in any tissue with blood vessels. The frequent primary sites of canine HSA are the spleen (50% to 65%), right atrium (3% to 25%), skin and subcutaneous tissues (13% to 17%), and liver (5% to 6%) (1). Typical canine splenic HSA is characterized by early distant metastasis and poor survival rates, despite surgical resection and adjuvant chemotherapy. The mean survival times (MSTs) for dogs after splenectomy alone are 19 to 86 d, and the 12-month survival percentage is 10% or less. Primary hepatic HSA carries an equally poor prognosis (1,2).

In contrast, several reports described long-term survival of digs with HSA of other sites of origin. A dog with primary lumbar extradural HSA survived over 15 mo after incomplete surgical excision and chemotherapy (3). Three canine patients with penile HSA treated with complete resection with or without chemotherapy survived 236 d, 206 d, and 20 mo, respectively (4–6). A dog with intranasal HSA treated by incomplete surgical excision with radiation therapy survived more than 30 mo (7). Fourteen dogs with renal HSA treated with surgery with or without adjuvant chemotherapy had improved 1-year survival rates (29%) and longer MSTs (278 d) compared with previous reports of dogs with splenic and hepatic HSA (8). The MSTs of 20 dogs with surgically treated lingual HSA combined with or without adjuvant chemotherapy was 553 d (9).

Thus, prognosis of canine HSA varies, depending on the site of origin. Furthermore, the MSTs for dogs with stage 1 dermal HSA which had complete surgical resection was 780 d, stage 2 (subcutaneous) and stage 3 (intramuscular) had MSTs of 172 to 307 d (2), suggesting the prognosis is different depending on the histopathological stage and status of surgical excision. The clinical course of canine intestinal HSA has not been well-documented (10,11).

In human HSA, frequent primary sites are breast (35%), skin (20%), muscle or subcutaneous tissue (13%), heart (5%), bone (4%), spleen (4%), and liver (4%) (12–14). As with canine HSA, human HSA of the heart, spleen, and liver carry a poor prognosis (5-year survival rate: 0%). But human HSA at other sites have better 5-year survival rates (muscle and subcutaneous: 74%, breast: 51%, skin: 43%, bone: 36%) (13). The histological grade of human mammary HSA also determines the 5-year postoperative disease-free rate (low grade: 76%, intermediate grade: 70%, and high grade: 15%) (15).

Intestinal HSA is also rare in humans and only 33 cases of small intestinal HSA have been reported in the English literature over the past 42 y (12). Reported survival times after being diagnosed varied (several days to 1 y: 21 reports, 1 to 2 y: 6 reports, 2 to 5 y: 3 reports, unknown: 3 reports). Although human small intestinal HSA was histologically classified into well-differentiated, poorly differentiated, and epithelioid angiosarcoma, the prognosis was generally poor regardless of the histological grade (12). Two reports that described feline intestinal HSA (total of 20 cats) indicated a poor prognosis (MSTs was not described) (16,17).

The reason for the poor prognosis of human and feline intestinal HSA may be partly due to late diagnosis. While cutaneous and subcutaneous neoplasms are easy to identify because of their conspicuousness, the presenting symptoms associated with intestinal tumors are non-specific and often poorly defined (12). The clinical manifestations of human patients with HSA of the small intestine included anorexia, weight loss, lethargy, vomiting, weakness, diarrhea, altered intestinal function, tenesmus, nonspecific abdominal pain, melena, anemia, hypoglycemia, acute abdominal signs and/or symptoms of ileus, and even nonspecific chest pain (13,14). Clinical signs often relate to location of the tumor within the GI tract. Small intestinal lesions commonly result in weight loss and large bowel lesions result in hematochezia and tenesmus. Since clinical symptoms may vary, a combination of multiple diagnostic tests is often required.

Radiography, contrast radiography, ultrasonography, and gastrointestinal endoscopy modalities that are commonly used in the diagnosis and staging of animals with suspected intestinal neoplasms (18). When these non-invasive or minimally invasive diagnostics fail to confirm a diagnosis, an exploratory laparotomy may be indicated for humans and dogs with persistent signs of intestinal disease (12,18). In the present case, the tumor was not detected by non-invasive examinations. While computed tomography (CT) may reveal important information about the gastrointestinal tract, it was not applied in this case due to its emergent nature (19).

Cutaneous, subcutaneous, and intramuscular HSA should be excised with 3-cm lateral margins and a minimum of 1 fascial layer for deep margins. Wide surgical margins (i.e., 5 cm) are generally recommended for intestinal tumors (2,20). In this case, to preserve the ileocolic valve, colonic resection, and anastomosis was performed with 3-cm margins of normal intestine on either end of the mass.

Histological examination of the mass confirmed complete resection and revealed typical anastomosing neoplastic vascular channels lined by spindle-shaped tumor cells with marked nuclear atypia and moderate mitotic activity, leading to the diagnosis of HSA. There were multifocal areas of necrosis and tumor destroyed structures of the inner circular muscular layer, but tumor cells did not destroy mucosal membranes. In addition, immunohistochemical staining confirmed moderate membranous-cytoplasmic expression of factor VIII in tumor cells (Figure 2A, B).

There was no ulcer formation on the mucosal surface; however, several small areas of bleeding were seen at the bottom of intestinal villi. This may be the reason bleeding was difficult to observe with non-invasive diagnostic techniques (Figure 4).

Figure 4.

Histologically, several bleeding points were seen at the bottom of intestinal villi. There was no ulcer formation on the mucosal surface.

Local recurrence was not observed by gastrointestinal palpation at the emergency liver lobectomy and postoperative contrast enhanced CT scanning. Histological diagnosis of the left-sided liver mass was consistent with HSA and expression of factor VIII in tumor cells was confirmed by immunohistochemical staining. The patient died due to rupture of the hepatic HSA and associated DIC 5 d after surgery.

The authors cannot confirm whether the 2 hypo-echoic liver lesions identified at the first medical examination were the HSA that slowly enlarged. It is unclear whether the liver lesions were metastatic lesions of the colonic mass, or if they were synchronous lesions of HSA arising in many tissues simultaneously.

It is possible that survival times of canine HSA are different depending on histological grading (21,22), as reported in human breast HSA. The relation between histological grading and prognosis has not been well-described in canine HSA.

Postoperative metastasis after long-term local control as reported in human intestinal HSA is another possibility. Liver metastasis 33 mo after colonic HSA (primary lesion was breast) resection (23), lung metastasis 21 mo after duodenal HSA resection (23), and abdominal wall metastasis 21 mo after duodenal and ileal HSA resection (24) have been described in humans. It is also possible that the hepatic HSA of the case herein was a new primary lesion unrelated to the colonic HSA or a synchronous lesion.

To our knowledge, this is the first documentation of a canine HSA arising from the colon. This case suggests that canine HSA arising from the large intestine may have a better chance for long-term survival after complete surgical excision compared with splenic and hepatic HSA.