Laparoscopic Resection of a Pancreatic β Cell Tumor in a Dog
ABSTRACT
Laparoscopic partial pancreatectomy has been performed in experimental canine studies and has been evaluated in human medicine but has not been reported in a clinical veterinary case. The authors present a 9 yr old field spaniel with weakness and hypoglycemia with insulin levels and Amended Insulin: Glucose Ratio results equivocal for a pancreatic insulinoma. Multiple abdominal ultrasounds did not detect the tumor, yet dual-phase computed tomographic angiography revealed the presence of a focal hypoattenuating nodule in the left lobe of the pancreas. A 3-port laparoscopic approach to the abdomen confirmed a 1.5-cm mass in the mid-left limb of the pancreas, and resection of the mass was performed with a bipolar vessel-sealing device. The surgery was performed without complication, and the dog became normoglycemic within 4 hr following surgery. Final histopathology results revealed pancreatic neuroendocrine carcinoma of the β cells. Recurrence of hypoglycemia was noted 18 mo postoperatively; however, repeat computed tomographic angiography did not reveal pancreatic abnormalities and fine needle aspirates of liver nodules did not suggest metastatic disease. Medical management was elected and the patient was euthanized 28 mo after surgery due to refractory hypoglycemic seizures.
Introduction
Insulinoma (β cell tumor) is a rare endocrine neoplasm arising from the islets of Langerhans of the pancreas.1,2 In contrast to humans, the majority of these tumors in dogs are malignant carcinomas (60%), while the remaining 40% are adenomas.3 Metastasis to the liver and local lymph nodes may be detected in up to 50% of patients.4,5 A previous study comparing medical and surgical management of these tumors suggests that the overall survival times in dogs undergoing surgical resection are significantly longer than for those treated medically.6 Clinical signs include lethargy, weakness, collapse, and seizures. A measurement of normal or increased serum insulin concentrations in relation to blood glucose during a time of documented hypoglycemia is supportive of the diagnosis of an insulinoma.2 An amended insulin to glucose ratio of >30 may be correlated with the presence an insulinoma, but the specificity of this test is poor.2,7
Multiple imaging modalities have been documented to aid in the diagnosis of insulinoma in dogs, including ultrasound, scintigraphy, and computed tomography.3 Once a tumor is documented or suspected, open laparotomy is traditionally elected for mass removal. Laparoscopic biopsy of the pancreas has been reported in a series of both experimental and clinical patients and has been shown to be a safe and effective method for evaluating pancreatic disease.8,9 Laparoscopic partial pancreatectomy has been reported in the human literature, and compared to open laparotomy, the minimally invasive approach appears to be associated with similar or shortened operative times, less blood loss, fewer complications, and shorter length of hospitalization than the open approach.10,11 We are unaware of any clinical reports in the veterinary literature; consequently, we report our findings of a laparoscopic partial pancreatectomy for the removal of an insulinoma in a dog.
Case Report
A 9 yr old male intact field spaniel weighing 19 kg was evaluated for investigation of chronic hypoglycemia of 3 yr duration. With the exception of his low blood glucose (ranging from 33–55 mg/dL), he had a history of normal complete blood counts, chemistry profiles, an unremarkable adrenocorticotropic hormone stimulation test and normal pre- and postbile acid levels. He had multiple mild episodes of weakness noted over 3 yr that responded to honey administration. At 1 mo prior to presentation to the surgery service he had a 20-s seizure with a blood glucose of 39 mg/dL (70–143 mg/dL). An insulin panel revealed a normal insulin level of 81 pmol/L (36–288 pmol/L), a low glucose level of 1.9 mmol/L (3.3–6.9 mmol/L), and an insulin to glucose ratio at the high end of normal, 43 (reference range 14–43). Physical and neurologic examination at the time of presentation to our facility was unremarkable with the exception of hind-limb weakness. He had been administered prednisone at 2.5 mg orally, twice daily for 2 mo prior to presentation and diazoxide at 100 mg/day orally for 3 wk prior to presentation.
Diagnostic Imaging
Despite focused attention to the pancreas and surrounding lymph nodes, results of an abdominal ultrasound examination by a board-certified radiologist were normal. A dual-phase computed tomographic angiography (CTA)a evaluation of the chest and abdomen was elected to further investigate the cause of chronic hypoglycemia. Noncontrast enhanced computed tomographic imaging of the thorax was performed first to evaluate for pulmonary metastasis, the results of which were normal. A dual-phase CTA of the cranial abdomen was then performed using bolus tracking softwareb, similarly to previous reports describing the technique.12,13 A 1.5-cm nodule was isoattenuating to surrounding pancreas and deformed the margins of the left pancreas. The nodule remained isoattenuating during the arterial phase and discretely hypoattenuating on the venous phase of the CTA (Figure 1). Unlike three of four dogs from two previous reports, the nodule did not have intense arterial enhancement.12,13
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Due to the owners’ schedule, 1 mo following CTA, the dog was anesthetized for surgery with a preoperative blood glucose of 46 mg/dL. Methadone at 0.3 mg/kg and glycopyrrolate at 0.01 mg/kg were administered intramuscularly as a premedication 30 min prior to the induction of anesthesia. Anesthetic induction was achieved with a combination of 3 mg/kg propofol and diazepam at 0.5 mg/kg IV and maintained with a constant rate infusion of fentanyl (0.5 µg/kg/min) and isoflurane in 100% oxygen delivered through a cuffed endotracheal tube. Lactated Ringer’s solution with 5% dextrose was administered (10 mL/kg/h IV). Antibiotic prophylaxis (cefazolin 22 mg/kg IV) was administered 30 min prior to anesthesia and continued every 90 min throughout the surgery.
The dog was positioned in dorsal recumbency and prepared aseptically for surgery. The surgeon stood at the right and the camera operator on the left side of the dog. The video monitor was positioned by the head on the left side of the patient. Following the Hasson technique, a 5-mm primary portc was placed 2 cm caudal to the umbilicus on the ventral midline: traction sutures were placed on either side of the linea alba, tied around the fixation disks, and a Storz Hopkins II 5 mm 30° rigid telescope was introduced for examination of the abdomen.14 The peritoneal cavity was insufflated via an automatic insufflatord (Stryker 40 L highflow) with carbon dioxide to no greater than 12 mm Hg. Two 5-mm instrument portals were inserted along the midline approximately 3 cm and 6 cm cranial to the umbilicus. The abdomen was evaluated, and no free fluid, lymphadenopathy, or obvious abnormalities were noted. Tilting was not employed for the exploration of the abdomen. The liver appeared normal, and three biopsy specimens were taken from the right lateral, right medial, and left lateral liver lobes using a 5-mm cup biopsy forceps as previously described.15 Laparoscopic atraumatic grasping forceps were placed in the cranial-most portal to grasp the duodenum to fully inspect the right limb of the pancreas and used to expose and isolate the left limb of the pancreas via the opening of the omental sac. One solid, dark pink nodule was identified in the mid body region of the left limb of the pancreas (Figure 2). While grasping and lifting the extremity of the left pancreas toward the ventral body wall with atraumatic grasping forceps, a 5-mm endoscopice bipolar vessel-sealing devicef was placed in the central port and used to dissect free the left limb of the pancreas from omental attachments. Branches of the splenic artery supplying the left limb of the pancreas were not individually identified; however, all feeding vessels were ligated and divided via Ligasure during dissection. Although the working space was tight with midline placement of all ports, there was sufficient room for manipulation of tissues (Figure 3). The Ligasure device was then placed across the proximal portion of the left limb of the pancreas, approximately 2 cm cranial to the nodule. Three “bites” or sealing and dividing cycles, were employed to complete the partial pancreatectomy (Figure 3). Gross evidence of thermal spread to surrounding tissues was minimal. The middle portal was enlarged by approximately 5 mm to facilitate the insertion of a 10-mm specimen retrieval deviceg (Figure 4); loss of insufflation occurred at this time but was reversed by reinsertion of the trocar. The excised portion of the pancreas was then placed within the specimen retrieval pouch and removed through the portal. Hemostasis was confirmed prior to closure, the carbon dioxide was removed from the abdomen by manual pressure and with a Poole suction tip, and all portal incisions were closed routinely. Total operative time (skin incision to skin closure) was recorded as 85 min. Samples of the liver and pancreas were submitted for histopathology examination (Figure 5).
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Postoperative Care and Outcome
Postoperative care included the administration of analgesia (methadone 0.3 mg/kg IV q 6 hr) and an acid reducer (famotidine 0.5 mg/kg IV q 12 hr). Intravenous fluids were continued at a rate of 30 mL/kg/hr with 5% dextrose supplementation. Immediate postoperative blood glucose level was 76 mg/dL, which increased to 190 mg/dL within 4 hr of surgery. At 4 hr following surgery, his glucose supplementation was lowered to 2.5%, and 18 hr following surgery, the dog was offered a bland diet and supplementation with glucose was discontinued. Forty-eight hours following surgery, the dog was normoglycemic at 113 mg/dL and was discharged from the hospital with pain medication (tramadol 4 mg/kg per os q 8 hr as needed), instructions to limit activity, monitor for any signs of weakness or seizures, and to recheck in 2 wk for follow-up evaluation. All previous medications including prednisone and diazoxide were discontinued. The patient remained asymptomatic with blood glucose levels within the normal range on visits at 1 mo, 3 mo, and 6 mo following surgery. Weakness and hypoglycemia (37 mg/dL) recurred 18 mo postoperatively. Medical management consisting of diazoxide 200 mg three times daily was initiated, and diazoxide-induced nausea was managed with ondansetron 2 mg once daily as needed. A repeat CTA did not reveal pancreatic abnormalities but did identify several hypodense nodules of the liver. Fine needle aspirates of liver nodules did not reveal metastatic disease. The patient was asymptomatic and rechecked monthly with blood glucose levels remaining stable at 45–62 mg/dL. Hind-limb weakness remained; however, the patient had concurrent severe osteoarthritis. At 28 mo following surgery, the patient presented with refractory hypoglycemic seizures and was humanely euthanatized. No necropsy examination was performed.
Histologic Findings
Within the submitted sections of the pancreas, there was a well-demarcated, 1.5-cm diameter, densely cellular mass (Figure 6) surrounded by a variable amount of compressed fibrous connective tissue (pseudocapsule). Neoplastic cells within the mass were arranged in nests and packets, separated by a fine, fibrovascular stroma (Figure 7). The cells were round to oval with moderate amounts of finely vacuolated, basophilic to amphophilic cytoplasm and round to oval, central nuclei. Nuclei exhibited vesicular chromatin and typically a single, central nucleolus. Anisocytosis and anisokaryosis were moderate, up to threefold. There were 14 mitoses in 10 high-power fields. Neoplastic cells did not extend into the surrounding pancreatic parenchyma; however, there were multifocal regions of neoplastic infiltration into the pseudocapsule. In addition, peripheral regions of the neoplasm were intermixed with small foci of remnant exocrine pancreas, suggesting previous parenchymal infiltration. Few sections contain foci with suggestive, but not definitive, vascular invasion. No thermal damage was noted.
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Immunohistochemistry
Immunohistochemical stains for chromogranin, synaptophysin, neuron-specific enolase (neuroendocrine markers), and insulin (β cell marker) were performed. Neoplastic cells exhibited strong, positive cytoplasmic immunostaining for detectable antigen with synaptophysin, neuron-specific enolase, and insulin (Figures 8–10). Neoplastic cells exhibited mild, multifocal, faint, cytoplasmic immunostaining with chromogranin. All stains showed strong, positive, cytoplasmic staining of islet cells within the sections of pancreas surrounding the mass (Figures 8–10). The histologic and immunohistochemical features in this case were interpreted to be consistent with a carcinoma arising from β cells of the endocrine pancreas.
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Citation: Journal of the American Animal Hospital Association 53, 6; 10.5326/JAAHA-MS-6417
Discussion
We report a laparoscopic method of surgical resection of canine insulinoma. Consistent with reports in the human literature, laparoscopic partial pancreatectomy is a feasible method for the removal of canine insulinoma.10,11Advantages to minimally invasive surgery include less stress and postoperative pain, which have been reported in previous studies comparing laparoscopic to traditional procedures.16 Additionally, laparoscopy may improve magnification of the working area and visualization at surgery.
The physiologic effects of laparoscopic versus open pancreatectomy have been studied previously in a series of experimental dogs. Five dogs underwent laparoscopic distal pancreatectomy and five underwent traditional open distal pancreatectomy. The authors compared gastrointestinal transit time using radiopaque markers as well as serum IL-1 and cortisol levels to assess surgical stress, and they concluded that dogs undergoing laparoscopic distal pancreatectomy demonstrated faster recovery of bowel transit and less stress than those undergoing conventional surgery.17 Multiple studies in the human literature confirm low complication rates and faster recovery when comparing laparoscopic versus conventional pancreatectomy.10,11,18,19 Potential benefits of laparoscopic resection of cancers may also include the ability to inspect the abdomen and abort the procedure if occult widespread metastasis are discovered with minimal pain and trauma to the patient. Additionally, faster recovery times may translate to shorter waiting times to adjuvant therapies such as chemotherapy or radiation treatment if indicated.10
Use of the bipolar sealing device facilitated laparoscopic removal of the insulinoma. The Ligasure device safely seals vessels up to 7 mm in diameter by denaturing the collagen and elastin within the vessel wall and has been used for laparoscopic partial pancreatectomy in human patients.20,21 Recently, a veterinary study compared the use of this device with traditional suture fracture technique for open partial pancreatectomy in a series of clinical patients.22 They found the Ligasure device to significantly shorten surgical times and cause no intraoperative complications such as bleeding or collateral damage to adjacent tissues. The formation of a tight seal at the site of resection was noted to eliminate the complications of ligature slippage, bleeding, and leakage of pancreatic secretions. Pancreatitis was not noted in any of these patients following surgery. In a study comparing the use of a vessel-sealing device to hand-sewn distal pancreatectomy in a population of porcine patients, postoperative amylase and lipase concentrations in drainage fluid from the pancreas levels were higher in the hand-sewn group, and pancreatic fistulas were noted only in the hand-sewn group.23 Furthermore, the ability to provide a tight seal of the pancreatic ducts appears to minimize pancreatic fistulas in people, which is thought to be the leading cause of postoperative morbidity.24 Other methods of laparoscopic pancreatic division could include the ultrasonically activated scalpel or endoscopic stapling devices; further investigation is necessary to compare their efficacy to that of the bipolar sealing device.
The diagnosis of insulinoma in this patient was complicated by the chronic nature of his hypoglycemia as well as failure to demonstrate unequivocally elevated insulin levels. A previous study in eight dogs has shown variable reliability in measuring insulin levels in a series of dogs with histologically confirmed insulinomas. Three of eight cases (38%) of dogs had insulin levels within the normal range using two different commercial veterinary laboratories.25 The diagnosis of insulinoma based solely on clinicopathologic findings remains difficult.
Abdominal ultrasonography has low accuracy in identification of pancreatic masses (<50% of affected dogs) as well as low accuracy for detecting metastatic lesions.6,26,27 Computed tomography is significantly more sensitive, accurately identifying 10/14 pancreatic masses in dogs with insulinoma in one study, but specificity in diagnosis of metastatic lesions is lacking.26 During our CTA exam, the insulinoma did not enhance during the arterial phase but nicely contrasted the surrounding normal pancreas on the venous phase. We did not perform any delayed scans (2–3 min after injection) of our patient, but previous reports indicate a low yield for lesion identification at these later scan times.12,13 Exploratory laparotomy is indicated when clinical signs, hypoglycemia, and elevated insulin levels are present, even if results of other diagnostic testing are equivocal.24,26,28,29
Although insulinomas usually exhibit benign histologic features, their biologic behavior is difficult to predict based on histologic criteria of malignancy, and metastasis is seen in approximately 50% of cases.4–6,28 Metastasis is most often seen in the regional lymph nodes and liver, and pulmonary involvement is rare. Results of one study show that prognosis is dependent on the stage of disease: (1) those without metastasis at the time of surgery remain euglycemic for 13 mo with a median survival time of 18 mo and (2) those with known metastasis are euglycemic for 2 mo with a median survival time of 7 to 9 mo.4,5 Compared to those undergoing solely medical management, dogs that underwent surgery in conjunction with medical therapy are more likely to be euglycemic for longer and have improved survival times, as reduction in tumor burden improves subsequent medical management.6 Similarly, Polton et al. found the median survival time in dogs undergoing surgery alone to be 785 days and found that those receiving medical therapy for the management of recurrent signs to have a prolonged survival time of 1,316 days.29 Metastatic disease was suspected in this patient at 18 mo following surgery based on the liver nodules identified on repeat imaging, despite inability to confirm this diagnosis with cytological analysis. Our patient experienced a survival time of 10 mo following the identification of suspected metastasis with continued medical therapy.
Differentiation of benign (adenoma) versus malignant tumors (carcinoma) of the pancreatic islets relies more heavily upon infiltration into the adjacent pancreatic parenchyma as opposed to histomorphologic features of the neoplastic cells.30 This neoplasm was consistent with a carcinoma due to the multifocal regions of capsular infiltration, degree of pleomorphism, and mitotic count. Strong, cytoplasmic immunohistochemical staining for synaptophysin, neuron-specific enolase, and mild cytoplasmic staining with chromogranin confirms that these cells are of neuroendocrine origin. Strong cytoplasmic immunohistochemical staining for insulin confirms that the neoplastic cells are derived from the β cells of the pancreatic islets.
Conclusion
In conclusion, we report a laparoscopic method for the excision of a pancreatic β cell tumor, previously reported in only experimental models. Future studies on a wider animal population will be needed to determine the benefit of laparoscopic pancreatectomy over traditional surgery, as well as to determine the best method of pancreatic tissue division. As clinicopathologic findings may be inconclusive in diagnosing an insulinoma, the use of advanced imaging may aid in the diagnosis and in presurgical planning of patients with suspected pancreatic masses.
Three-millimeter transverse computed tomography images at the level of the cranial pole of the right kidney (*) made before contrast administration (A) and during the arterial (B) and venous phases (C). The arterial phase began when contrast reached the descending aorta at the level of the celiac artery (determined by bolus tracking software) and scanned from the caudal-most margin of the right pancreas to the cranial margin of the liver. The venous phase began 6 s later, scanning from the cranial margin of the liver to the pubis. On the image made prior to contrast administration (A), notice the iso- to slightly hypo-attenuating ovoid deformation of the left lobe of the pancreas (arrow) just dorsal to the transverse colon (arrowhead). Small arteries surround the nodule, but the nodule does not enhance on the arterial phase (B). This nodule is well margined, contouring to the shape of the deformed pancreatic margins and hypoattenuating to the surrounding normal pancreas on the venous phase (C).
A laparoscopic view of left-sided pancreatic nodule.
A laparoscopic view of the LigaSure device used to perform a partial pancreatectomy.
An excised portion of the pancreas placed within the specimen retrieval pouch.
Pancreas, dog. A postoperative photograph of a pancreatic mass.
Pancreas, dog. A low magnification view of a pancreatic mass. The mass is 15 mm in diameter, well demarcated, and densely cellular. Haematoxylin and eosin stain. Bar = 8 mm.
Pancreas, dog. A higher magnification view of a pancreatic mass. Note that neoplastic cells are arranged in nests and packets, separated by a fine, fibrovascular stroma. Haematoxylin and eosin stain. Bar = 200 um.
Pancreas, dog. Synaptophysin immunohistochemistry for neuroendocrine cells. There is strong, diffuse cytoplasmic immunostaining for detectable antigen within the neoplasm. Throughout the surrounding pancreas, multifocal brown regions of immunostaining represent pancreatic islets, acting as an internal control (arrow). DAB+ Chromogen, CAT hematoxylin counterstain. Bar = 2 mm.
Pancreas, dog. Insulin immunohistochemistry for β cells. There is strong cytoplasmic immunostaining for detectable antigen within the neoplasm. Throughout the surrounding pancreas, multifocal brown regions of immunostaining represent pancreatic islets, acting as an internal control. DAB+ Chromogen, CAT hematoxylin counterstain. Bar = 2 mm.
Pancreas, dog. Insulin immunohistochemistry. Note strong, diffuse immunostaining for detectable antigen within the cytoplasm of the neoplastic cells, confirming the β cell origin. DAB+ Chromogen, CAT hematoxylin counterstain. Bar = 200 um.
Contributor Notes
