Funding sources: This work was supported in part by Grants-in-Aid for Scientific Research (KAKENHI No. 25850203) from Japan Society for the Promotion of Science (K. N.).
CONTRAST-ENHANCED ULTRASONOGRAPHIC FINDINGS IN THREE DOGS WITH PANCREATIC INSULINOMA
Article first published online: 21 MAY 2014
© 2014 American College of Veterinary Radiology
Veterinary Radiology & Ultrasound
Volume 56, Issue 1, pages 55–62, January/February 2015
How to Cite
Nakamura, K., Lim, S.-Y., Ochiai, K., Yamasaki, M., Ohta, H., Morishita, K., Takagi, S. and Takiguchi, M. (2015), CONTRAST-ENHANCED ULTRASONOGRAPHIC FINDINGS IN THREE DOGS WITH PANCREATIC INSULINOMA. Veterinary Radiology & Ultrasound, 56: 55–62. doi: 10.1111/vru.12177
- Issue published online: 15 JAN 2015
- Article first published online: 21 MAY 2014
- Manuscript Accepted: 6 MAR 2014
- Manuscript Received: 9 SEP 2013
- Japan Society for the Promotion of Science. Grant Number: 25850203
- contrast-enhanced ultrasonography;
Abdominal ultrasonography is one of the most common diagnostic imaging modalities used for dogs with suspected insulinoma; however, pancreatic masses are clearly identified in fewer than half of affected dogs and benign pancreatic nodules can be difficult to differentiate from malignant ones. The purpose of this prospective study was to describe contrast-enhanced ultrasonography (CEUS) characteristics of confirmed pancreatic insulinoma in a group of dogs. Inclusion criteria were as follows: (1) repeated hypoglycemia (blood glucose levels <60 mg/dl, twice or more); (2) elevated blood insulin levels with hypoglycemia; (3) pancreatic nodules detected with conventional ultrasonography; and (4) histological confirmation of pancreatic islet cell carcinoma. Immediately following conventional ultrasonography of the entire abdomen, CEUS of the pancreatic nodule and adjacent parenchyma was performed using contrast-specific technology pulse inversion imaging and perflubutane microbubble contrast agent. Three dogs met inclusion criteria. Pancreatic nodules in all the three dogs became more clearly demarcated after injection of the contrast agent. Each nodule showed different enhancement patterns: markedly hyperechoic for 5 s, slightly hyperechoic for 1 s, and clearly hypoechoic for over 30 s. These results were not in complete agreement with previously reported CEUS findings in human patients with insulinoma. All nodules were surgically resected and histopathologically confirmed as malignant insulinomas. Findings from the current study indicated that contrast-enhanced ultrasound may help to increase conspicuity of pancreatic insulinomas in dogs and that enhancement characteristics may be more variable in dogs than in humans.
Insulinomas are insulin-secreting beta-cell tumors that are uncommon in dogs, but have well-established and distinctive clinical characteristics. The standard method for confirming the presence of insulinoma is documentation of repeated episodes of hypoglycemia along with inappropriately elevated serum insulin levels. Complete surgical removal of the tumor often results in complete remission of clinical signs for affected dogs. However, it is often difficult to detect the location of these tumors using conventional diagnostic imaging methods because of their small size.[1, 2] Abdominal ultrasonography is one of the most commonly used diagnostic imaging modalities for dogs with suspected insulinomas; however, these and other pancreatic masses are clearly identified in fewer than half of the affected dogs.[1, 3, 4] Some pancreatic nodules are benign and can be difficult to distinguish from insulinomas.
CT plays a major role in detecting pancreatic tumors in human medicine.[5-9] Triple-phase (arterial, pancreatic, and hepatic phase) CT has been the gold standard for the diagnosis of human pancreatic tumor.[7, 8] Recently, dynamic CT was demonstrated to be a reliable diagnostic tool for detecting pancreatic insulinomas in dogs.[10, 11] However, the cost and requirement for specialized equipment, as well as the need for anesthesia, have currently limited routine use of these techniques in veterinary medicine.
Contrast-enhanced ultrasonography (CEUS) has also been described as a valuable imaging method for the diagnosis of pancreatic tumors in humans.[12, 13] Some studies have reported that the majority of pancreatic neuroendocrine tumors were hyperenhancing with CEUS, whereas pancreatic carcinomas were hypoenhancing. [12, 13] Therefore, it has been suggested that pancreatic tumors could be differentiated using CEUS. In veterinary medicine, two original articles and two review papers have been published on CEUS characteristics of the pancreas in dogs.[14-17] However, at the time of this study, there were no published studies describing the contrast-enhanced ultrasonographic characteristics of pancreatic insulinomas in dogs. The purpose of this study was to describe CEUS findings for a group of dogs with confirmed pancreatic insulinomas. Based on human studies, we hypothesized that these tumors would be hyper-enhancing with CEUS.
Materials and Methods
Dogs that were patients of Hokkaido University Veterinary Teaching Hospital were prospectively recruited. Inclusion criteria were as follows: (1) repeated hypoglycemia (blood glucose levels <60 mg/dl, twice or more); (2) elevated blood insulin levels with hypoglycemia; (3) pancreatic nodules detected with conventional ultrasonography; and (4) histological confirmation of pancreatic islet cell carcinoma.
For all included dogs, both conventional ultrasonography and CEUS were performed by a single observer (K.N.) and with the same ultrasound scanner (Aplio XG, Toshiba Medical Systems Corporation, Tochigi, Japan) with a 5–11 MHz broadband linear probe (PLT-704 AT, Toshiba Medical Systems Corporation, Tochigi, Japan). Scanning was performed in the dorsal recumbency with only manual restraint. Immediately following a complete conventional abdominal ultrasound evaluation, the left and right pancreatic lobes were scanned with the probe positioned along the stomach and duodenum in each case to detect a pancreatic nodule. The right lateral intercostal scan technique was also used to scan the right lobe and body of the pancreas. When the pancreatic nodule was detected, a location allowing clear visualization of the nodule and adjacent pancreatic parenchyma in a single view was chosen for CEUS. Contrast-enhanced ultrasonography was performed using contrast-specific technology pulse inversion imaging. The mechanical index was set at 0.21, and the imaging depth was set at 3–4 cm according to the depth and size of the lesion. Ultrasound imaging was set at 30–31 frames per second and a single focal zone was placed at the deepest part of the lesion. The gain was adjusted so that few signals from the underlying pancreas were present. A bolus of perflubutane microbubble contrast agent (Sonazoid®, Daiichi Sankyo Company, Limited, Tokyo, Japan, 0.01 ml/kg) was injected by hand through a 21-gauge butterfly catheter placed in the cephalic vein, attached to a 22-gauge intravenous catheter, and flushed with 3 ml of heparinized normal saline. Scanning was performed continuously for 1 min after injection of the contrast agent with the probe maintained in the same site. Additional diagnostic imaging, including a thoracic radiograph and CEUS of the liver were then performed to survey for metastases. Seven minutes after injection of contrast agent, the liver was scanned with the same linear probe and a 3.75 MHz convex probe (PSK-375 BT, Toshiba Medical Systems Corporation, Tochigi, Japan) to evaluate postcontrast parenchymal characteristics. All ultrasonographic images were recorded continuously to a hard disk for off-line analysis. Images were obtained at a rate of one frame per second with a video frame grabber (VLC media player, France, VideoLAN Organization, Paris, France).
Qualitative and quantitative analyses of the recorded ultrasound images were performed for all three dogs by a single observer (K.N.). The enhancement patterns for the qualitative analysis were defined based on the echogenicity of lesion in comparison with the surrounding normal parenchyma: (1) hypoechoic, (2) isoechoic, and (3) hyperechoic. For quantitative analysis of tissue intensity, image analysis freeware (ImageJ, US National Institutes of Health, Bethesda, MD) was used. In this software system, the gray scale ranged from 0 to 255 mean pixel values. Tissue intensity was calculated for each region of interest located in the pancreatic insulinoma lesion and normal pancreatic parenchyma. The region of interest was circular and drawn as large as possible without including vessels and adjacent structure. The region of interest was moved manually to maintain the same region when respiratory motion was encountered. A time–intensity curve was created for each dog. The enhancement time of the lesion was calculated for each case and was defined as the time when the difference in the gray scale level between the lesion and parenchyma was >20 mean pixel values because this magnitude of difference is generally necessary for visual recognition. All dogs underwent surgical resection of the pancreatic nodules within 2 weeks of imaging and the samples were submitted for histological examination.
Clinical findings for the three dogs are summarized in Table 1. Ultrasound findings are described for each dog below.
In dog 1, an ill-defined, round, hyperechoic nodule (6.0 × 6.0 mm) was detected in the mid portion of left pancreatic lobe using conventional ultrasonography (Fig. 1A). No other lesion was detected by complete abdominal ultrasonography or thoracic radiography. After injection of the contrast agent, the nodule became hyperechoic and more clearly demarcated (Fig. 1B). Simultaneously, the circumferential vessel was also enhanced, presenting a ring enhancement pattern (Fig. 1B and C). The echogenicity of the nodule reduced to the level of the normal parenchyma at 10 s after injection (Fig. 1C). The differences between the gray scales of the pancreatic mass and normal parenchyma reached a peak at 5 s after injection, and the peak difference was 79.6 mean pixel values and the enhancement time of the lesion was from 4 to 9 s after injection. (Fig. 4A). Parenchymal imaging of the liver revealed no metastatic lesions. A firm pancreatic mass was found at the mid portion of the left pancreatic lobe during the exploratory laparotomy, as on ultrasonography. No other visible nodule was detected in any organ including the pancreas, liver, or lymph nodes. The mass was removed by partial pancreatectomy. A histopathological examination revealed that the mass was well delineated from the exocrine pancreas by dense fibrous tissue but partially lacked a fibrous capsule. The mass was composed of neoplastic acinar cells and abundant fibrous septa. The neoplastic cells were moderately polymorphic. These cells were polyhedral and had variably sized oval or cleaved nuclei and a granular eosinophilic cytoplasm. No extensive necrosis was observed in the neoplastic tissue, although pyknotic cells were observed frequently. Neoplastic cells occasionally invaded through the fibrous capsule into adjacent exocrine tissue. The mass was diagnosed as an islet cell carcinoma. The dog was clinically normal 1 year after surgery.
|Dog||Age, Sex, BW*, Breed||Glucose† (mg/dl)||Insulin‡ (μU/ml)||Size and location of nodules||Symptoms|
|1||11 y, M§, 11.9 kg, Cocker Spaniel||49||49||6 mm, Left lobe||None|
|2||14 y, FS∥, 14.2 kg, Mongrel||39||148||11 mm, Right lobe||Collapse|
|3||8 y, M, 18.0 kg, Welsh Corgi||40||23.1||11.9 mm, Left lobe||Weakness|
In dog 2, a well defined, round, slightly hyperechoic nodule (11.0 × 11.0 mm) was detected at the caudal tip of right pancreatic lobe using conventional ultrasonography (Fig. 2A). An enlarged hepatic lymph node (4.1 mm) was detected during complete abdominal ultrasonography. A thoracic radiograph revealed no metastatic lesions. After injection of the contrast agent, the nodule became slightly hyperechoic compared with the surrounding normal parenchyma, and the circumferential vessel was also hyperechoic (Fig. 2B). The enhancement of the nodule persisted for a brief second, followed by rapid loss of enhancement (Fig. 2C). The peak difference of the gray scale between the mass and normal parenchyma was 33.6 mean pixel values at 5 s after injection and the enhancement time of the lesion was from 4 to 5 s after injection (Fig. 4B). Parenchymal imaging of the liver revealed no metastatic lesions. A firm pancreatic mass at the caudal tip of the right pancreatic lobe and an enlarged hepatic lymph node were found during the exploratory laparotomy, as on ultrasonography. No other visible nodules in any organ were observed, including the pancreas, liver, or lymph nodes. The pancreatic mass and enlarged hepatic lymph node were resected. The histopathological examination revealed that the mass was well encapsulated by relatively thin fibrous tissue and composed of neoplastic acinar cells and thin fibrous septa. The neoplastic cells were moderately polymorphic. Most of the neoplastic cells were cuboidal to polyhedral and had small-sized oval nuclei with a granular eosinophilic cytoplasm but large pleomorphic cells with large hyperchromatic nuclei were scattered throughout the parenchyma. No extensive necrosis was observed in the neoplastic tissue. The neoplastic cells locally invaded the capsule, where neoplastic emboli were noted in lymph vessels. Additionally, metastasis to lymph nodes adjacent to the mass was observed. This dog was diagnosed with an islet cell carcinoma with lymph node involvement. The dog was clinically normal 1 year after surgery.
In dog 3, an ill-defined, oval, hypoechoic region (9.2 × 11.9 mm) was detected at the tip of left pancreatic lobe using conventional ultrasonography (Fig. 3A). No other lesion was detected by complete abdominal ultrasonography or thoracic radiography. After injection of the contrast agent, a vessel running through the nodule appeared hyperechoic, although the nodule remained hypoechoic to adjacent pancreatic parenchyma and became more clearly demarcated for over 30 s (Fig. 3B and C). The peak difference of the gray scale between the mass and adjacent parenchyma was −119.8 mean pixel values at 9 s after injection and the enhancement time of the lesion was from 5 to over 30 s after injection. (Fig. 4C). Parenchymal imaging of the liver revealed no metastatic lesions. A firm, white pancreatic mass was found at the tip of the left pancreatic lobe during an exploratory laparotomy, as on ultrasonography. Moreover, it was confirmed that the mass had infiltrated into the adjacent splenic vein. Although CEUS of the liver revealed no lesion, a very small, ill-defined nodule (diameter 2 mm) was found on the surface of the liver during the exploratory laparotomy. The pancreatic mass and liver mass were resected. A histopathological examination revealed that the ill-defined mass was composed of closely packed neoplastic acinar cells, which were subdivided into variably sized nests or cords by connective tissue septa with capillaries. Peripheral condensation of connective tissue was observed, but the neoplastic cells invaded frequently through the fibrous capsule into the adjacent exocrine pancreas. Vascular invasion was also noted. Most of the neoplastic cells were well-differentiated but polyhedral cells with large atypical nuclei were occasionally observed. The neoplastic cells metastasized to the liver. The dog was diagnosed with an islet cell carcinoma with vascular invasion and the metastasis to the liver. The dog developed weakness for a few days 2 months after surgery. Hypoglycemia (blood glucose 48 mg/dl) had relapsed although no lesion was detected on abdominal ultrasonography. Prednisolone (0.5 mg/kg, sid) was started, but the dog died 3 months after surgery. A necropsy examination was not performed.
To the best of our knowledge, this is the first report describing the use of CEUS in a group of dogs with insulinoma. All pancreatic nodules histologically diagnosed as insulinoma were visible using conventional and CEUS in the present study, however CEUS resulted in more clear demarcation of the margins. Each of the three nodules showed different enhancement patterns: significantly hyperechoic for 5 s, slightly hyperechoic for 1 s, and hypoechoic for over 30 s. These results are not in complete agreement with the findings in human insulinoma patients. In a recent study on CEUS in human insulinoma patients, it was demonstrated that all nodules detected using CEUS were homogeneously hyperechoic in the earlier arterial phase (from injection to 25 s) and remained hyperechoic in the late phase (2–3 m after injection). Four previous canine reports have described the use of dual-phase computed tomographic angiography for diagnosis of insulinomas.[10, 11] Among them, three of the four nodules were markedly hyperattenuating compared with the normal pancreas during the arterial phase, whereas one was hypoattenuating during the arterial phase.[10, 11] During the venous phase, two of the four nodules were still hyperattenuating compared with the normal pancreas, and the other two were isoattenuating.[10, 11] In humans, most insulinomas are hyperdense in CT images during the arterial phase but some lesions may be found only during the parenchymal phase.[21-24] Only a few cases of atypical hypoenhanced insulinomas have been reported.[25-27] This discrepancy could be due to the difference between the typical malignancy grades of insulinoma in humans (mostly benign) versus dogs (mostly malignant). Hypoenhanced masses are generally assumed to be due to hypovascularity. Large tumors may displace vessels or exhibit neovascularity, particularly if they are malignant. However, the reported cases of hypoenhanced insulinoma in human were not malignant[25-27] and there is not an exact correlation between tumor vascularity and angiographic findings in human insulinoma. In the present study, although one dog with a hypoenhanced mass had a highly invasive tumor with liver metastasis and showed the worst prognosis, no significant difference in vascularity of the nodule was detected compared to those of the other two dogs with hyperenhanced masses. Further studies are needed to investigate the characteristic enhancement patterns and pathological implications of pancreatic insulinomas in dogs.
Detection of pancreatic insulinomas using conventional ultrasonography is challenging. Depending on the study, the sensitivity of conventional ultrasonography for detecting pancreatic insulinomas is 28–75% in dogs., [29-32] In a case series of 14 dogs with insulinomas, CT showed a higher sensitivity for detecting insulinoma mass lesions than that of conventional ultrasonography. In humans, the sensitivity of transabdominal ultrasonography for localizing pancreatic insulinomas is 9–64%.[33-35] The detection rate of pancreatic insulinoma improves from 24% to 89% with CEUS. The criteria for inclusion in the current study required that a nodule be identified with conventional ultrasonography, therefore it remains unknown whether contrast-enhanced ultrasonography can improve the detection rate of pancreatic insulinoma in dogs. However, the observation that all lesions in the present study became more clearly demarcated following CEUS suggests that pancreatic insulinoma conspicuity may be improved using this technique in dogs.
Some limitations of this study should be discussed. Respiratory motion could have affected the results of the quantitative evaluation. Although manual adjustments were made to maintain the same region of interest as much as possible, general anesthesia would be needed to increase objectivity of the quantitative evaluation. The second limitation was that the value of a differential diagnosis for pancreatic nodules was not examined. Some human studies have revealed that the majority of pancreatic adenocarcinoma are hypoenhanced. Therefore, CEUS had a potential to distinguish pancreatic nodules. Other pancreatic tumors, such as adenocarcinoma and gastrinoma, should be evaluated by CEUS and compared with findings from this study. The third limitation was a lack of comparison between CEUS and CT. Further studies are needed to assess the relative diagnostic sensitivity of CEUS and CT for diagnosing pancreatic tumors in dogs.
In conclusion, findings from this small case series indicated that contrast-enhanced ultrasound may help to increase conspicuity of pancreatic insulinomas in dogs and that enhancement characteristics may be more variable in dogs than in humans.
This work was supported in part by Grants-in-Aid for Scientific Research (KAKENHI No. 25850203) from Japan Society for the Promotion of Science (K. N.). The authors would like to thank Enago for the English language review.
- 1Beta-cell neoplasia: Insulinoma. In: Feldman EC, Nelson RW (eds): Canine and feline endocrinology and reproduction, 3rd ed. Philadelphia: WB Saunders, 2004;616–644., .
- 13Contrast-enhanced ultrasound in the diagnosis of pancreatic tumors. JOP. 2006;7:584–592., , .
- 31Canine clinical peripheral neuropathy associated with pancreatic islet cell carcinoma. Prog Vet Neurol 1994;5:57–62., , , , , , et al.