September 2012, VOLUME 199
NUMBER 3

Recommend & Share

September 2012, Volume 199, Number 3

Gastrointestinal Imaging

Original Research

Venous Tumor Thrombus in Nonfunctional Pancreatic Neuroendocrine Tumors

+ Affiliations:
1 Department of Diagnostic Radiology, Unit 1473, University of Texas M. D. Anderson Cancer Center, PO Box 301402, Houston, TX 77030.

2 Division of Surgical Oncology, University of Texas M. D. Anderson Cancer Center, Houston, TX.

3 Department of Gastrointestinal Medical Oncology, University of Texas M. D. Anderson Cancer Center, Houston, TX.

Citation: American Journal of Roentgenology. 2012;199: 602-608. 10.2214/AJR.11.7058

ABSTRACT
Next section

OBJECTIVE. We sought to determine the incidence of venous tumor thrombus in nonfunctioning pancreatic neuroendocrine tumors.

MATERIALS AND METHODS. We reviewed CT images of patients with a diagnosis of nonfunctional pancreatic neuroendocrine tumors over a 4-year period. For patients who underwent surgery, changes to surgical plans related to the tumor thrombus were also recorded.

RESULTS. CT showed venous tumor thrombi in 29 of the 88 patients (33%; 95% CI, 23–44%). This CT finding was not accurately reported in 18 of the 29 patients (62%; 95% CI, 42–79%). Of the 39 patients who underwent surgery, venous tumor thrombi were detected in 11 patients (28%; 95% CI, 15–45%) and were confirmed by pathology. Microscopic venous tumor thrombi in 10 patients were not detected by CT. Pathologic results showed venous tumor thrombi in 21 of the 39 patients (54%; 95% CI, 37–70%) who underwent surgery. The surgical plan was significantly changed in two of the 11 patients with gross thrombi (18%; 95% CI, 2–52%) who underwent surgery. There was no change in the surgical plan for the 10 patients with microscopic tumor thrombi.

CONCLUSION. The incidence of venous tumor thrombus detected by imaging was 33% in our study. This imaging finding was not accurately reported on the radiology report in 62% of the patients. In 18% of the patients with gross venous tumor thrombi, there was a significant alteration in the surgical plan. It is critical for the radiologist to be aware of the association of venous tumor thrombi in patients with nonfunctioning pancreatic neuroendocrine tumors and to report these findings.

Keywords: neuroendocrine tumor, nonfunctional, pancreas, venous tumor thrombus

Pancreatic neuroendocrine tumors are a subgroup of the gastroenteropancreatic neuroendocrine tumors with imaging features distinct from those of the more common pancreatic ductal adenocarcinoma. In 2000, the World Health Organization classification system [1] replaced the previous name for this entity, “islet cell tumor,” with the more generic term “pancreatic neuroendocrine tumor” used today. Although pancreatic neuroendocrine tumors account for only 1–2% of all pancreatic tumors, the incidence of pancreatic neuroendocrine tumors is increasing [2]; this increase may be related to an increase in detection rather than a true rise in incidence.

Pancreatic neuroendocrine tumors can be classified as either functional or nonfunctional. The majority of pancreatic neuroendocrine tumors, ranging from 60% to 90%, are nonfunctional [25], and thus they tend to remain occult until a late stage of disease. Nonfunctional pancreatic neuroendocrine tumors are typically detected either incidentally or during the evaluation of symptoms related to the mass effect on adjacent structures.

Nonfunctional pancreatic neuroendocrine tumors have been classically described as large well-defined hypervascular masses with hypervascular metastases [6, 7] in contrast to ill-defined hypovascular and infiltrative pancreatic ductal adenocarcinomas. The goal of surgery is similar in pancreatic neuroendocrine tumor and pancreatic adenocarcinoma: to achieve complete resection of the tumor. However, in patients with pancreatic neuroendocrine tumor, surgical resection of tumor and metastases can be performed when the entire tumor and metastatic disease can be removed or surgery can be performed for palliation.

Patterns of spread differ between nonfunctional and functional pancreatic neuroendocrine tumors. One such distinctive and less common feature that has been reported sporadically in case reports [813] and in small case series [14] is the presence of venous tumor thrombus arising from the nonfunctional pancreatic neuroendocrine tumor and growing into adjacent veins. This finding is different from the more commonly seen venous occlusion that can occur in pancreatic adenocarcinomas and in nonfunctional pancreatic neuroendocrine tumors.

The incidence of venous tumor thrombus in nonfunctional pancreatic neuroendocrine tumors is unknown. We hypothesize that the association of venous tumor thrombus with nonfunctional pancreatic neuroendocrine tumors has not been established as a result of the lack of incidence data similar to the association between renal cell carcinoma and renal vein tumor thrombus, and venous tumor thrombi in this setting are consequently underreported.

In the presence of resectable disease, the primary therapy for patients with nonfunctional pancreatic neuroendocrine tumors is complete surgical resection of the entire tumor. Venous tumor thrombus can extend beyond the primary tumor into the remaining pancreas within a vein. Inaccurate reporting of this finding may lead to incomplete surgical resection. Accurate reporting of this finding may lead to a change in surgical planning.

Another distinctive and unusual pattern of spread is intraductal growth of nonfunctional pancreatic neuroendocrine tumor, which has been reported in case reports [1518]. This finding is also different from ductal occlusion caused by mass effect that can be seen in pancreatic ductal adenocarcinoma and nonfunctional pancreatic neuroendocrine tumors.

Liver metastases and metastatic adenopathy are not discussed in this study. Venous narrowing or occlusion, arterial abutment, and arterial encasement are common to pancreatic ductal adenocarcinoma and pancreatic neuroendocrine tumors. These findings are discussed briefly in this study. The presence and assessment of these patterns of spread are well known and are common to ductal adenocarcinoma and are not the focus of this study.

Our primary objective was to determine the incidence of venous tumor thrombus in nonfunctional pancreatic neuroendocrine tumor. Our secondary objectives were to assess how often this finding was accurately reported and to determine the impact of this finding on surgery. Another secondary objective was to determine the incidence of intraductal growth in nonfunctional pancreatic neuroendocrine tumor.

Materials and Methods
Previous sectionNext section
Patient Population

This study was approved by our institutional review board, which granted a waiver of the requirement for informed consent. We retrospectively identified new patients presenting to our institution with the pathologic diagnosis of nonfunctional pancreatic neuroendocrine tumor over a 4-year period from October 2001 to October 2005. To do so, we queried the institutional database using the search term “nonfunctional pancreatic neuroendocrine tumor.” The patients in this study are enrolled in an ongoing project to assess treatment changes and survival and are still being monitored. Only patients with pathologic confirmation of the diagnosis of pancreatic neuroendocrine tumor at our institution were included for the purpose of this study.

Clinical Data

Medical records were reviewed for patient demographics, tumor size, and type of surgery (if performed). Pathologic data were specifically reviewed with respect to venous and ductal invasion by tumor.

CT Technique

Patients were scanned using either a multiphasic CT protocol or an unenhanced and contrast-enhanced CT protocol. The multiphasic protocol involved initially obtaining unenhanced images from the top of the liver to the bottom of the liver with reconstruction to a 2.5-mm slice thickness. Then, 125 mL of IV contrast agent was administered at a rate of 4–5 mL/s. The bolus-tracking function was used to obtain late arterial phase images after a 10-second delay once an enhancement threshold of 100 HU was reached in the aorta at the level of the celiac trunk. These images extended from the top of the liver to just below the iliac crests. After a delay of 14 seconds after the end of the late arterial phase, portal venous phase images were obtained once again from the top of the liver to just below the iliac crests. The arterial and portal venous phase images were reconstructed to a 1.25-mm slice thickness for evaluation. Delayed images (90-second delay) through the pancreas and kidneys were reconstructed to a 2.5-mm slice thickness. Oral contrast material (2% barium sulfate or meglumine diatrizoate or water) and rectal contrast material (2% barium sulfate or meglumine diatrizoate or water) were routinely used.

Patients scanned according to the unenhanced and contrast-enhanced CT protocol were imaged only in the portal venous phase of enhancement. Unenhanced images were acquired from the top of the liver to the bottom of the liver and reconstructed to a 2.5-mm slice thickness. The portal venous phase images were obtained at a 50-second delay from the top of the liver to just below the iliac crests or through the pelvis, if imaging of the pelvis had been ordered as part of the study, and were reconstructed to a 2.5-mm slice thickness. Delayed images (90-second delay) through the pancreas and kidneys were reconstructed to a 2.5-mm slice thickness. Oral and rectal barium was routinely used for the multiphasic protocol.

Image Analysis

The CT images were reviewed by two attending radiologists with subspecialty training in gastrointestinal radiology. Both reviewers are in practice at a oncologic center. At the time of the study, one reviewer had been in practice for 7 years and the other reviewer had been in practice for 20 years. The reviewers were aware that the patients had been diagnosed with pancreatic neuroendocrine tumors, but they were blinded to all other data including the pathologic data. They reviewed the images in consensus.

Our data collection was restricted to vascular and ductal involvement by pancreatic neuroendocrine tumors. Metastases and metastatic lymphadenopathy are features common to ductal adenocarcinoma and pancreatic neuroendocrine tumors and are well known and are routinely assessed. Venous tumor thrombus and intraductal tumor growth, however, are not typically seen in ductal adenocarcinoma. These features are distinctive to pancreatic neuroendocrine tumors and the focus of this study was to emphasize this association.

Because of confusion about overlap between the terms “involvement,” “occlusion,” “invasion,” and “tumor thrombus,” we defined “venous tumor thrombus” to represent direct luminal extension of the tumor into an adjacent vein with enhancement of the thrombus. We use the term “narrowing” or “occlusion” to indicate tumor narrowing or obliterating a vessel by mass effect without invading the vessel lumen. We use the terms “encasement” and “abutment” as they are used in the surgical literature. “Encasement” refers to tumor extending to more than 180° or more of the vessel circumference, and “abutment” refers to tumor extending to 180° or less of the vessel circumference.

Similarly, we use the term “intraductal growth” to indicate tumor that is directly invading the ductal lumen and the term “ductal obstruction” to indicate tumor narrowing or obliterating the duct without invading the ductal lumen.

The images were reviewed for arterial tumor thrombus, arterial occlusion, arterial abutment or encasement, venous tumor thrombus, venous occlusion, venous abutment or encasement, intraductal growth, and ductal obstruction. The vessels evaluated included the celiac trunk, common hepatic artery, superior mesenteric artery (SMA), splenic artery, splenic vein, superior mesenteric vein (SMV), and portal vein (PV). The locations of arterial tumor thrombus, arterial occlusion, venous tumor thrombus, venous occlusion, intraductal growth, and ductal obstruction were also recorded. The reference standard used for confirmation of the CT findings was the pathologic data when available. In the absence of pathologic data, follow-up imaging, if available, was used as the reference standard.

Alteration of surgical plan—The patients who underwent surgery were patients whose primary tumor met the surgical criteria for resectability. The surgical team determined surgical resectability of the primary tumor on the basis of objective radiographic criteria similar to those used for pancreatic adenocarcinoma. When necessary, staging was confirmed by multidisciplinary review. Resectable tumors included those with the following characteristics on MDCT images: no tumor extension to the SMA or celiac axis and no occlusion of the SMV or of the SMV-PV confluence [19]. Therefore, segmental resection and reconstruction of the mesenteric vasculature were performed when required to achieve negative margins [20]. The goal of surgery in all cases was complete macroscopic and microscopic (R0) resection of all disease.

The only difference between pancreatic neuroendocrine tumor surgery and pancreatic ductal adenocarcinoma surgery was that resection of the tumor was performed in the presence of metastatic disease in pancreatic neuroendocrine tumor if the entire tumor and metastatic disease could be removed surgically [21] or, alternatively, resection was performed for palliation especially for gastric varices to prevent gastrointestinal bleeds.

The presence of focal venous tumor thrombus or intraductal growth was not a contraindication to surgery. On retrospective analysis, the preoperative reporting of venous tumor thrombus and intraductal tumor growth was considered essential for complete macroscopic resection of the entire tumor.

The imaging and surgical reports of patients with venous tumor thrombus or intraductal tumor growth whose tumor was surgically resected were reviewed with a pancreatic surgical oncologist during this study. Alterations to the surgical plan were recorded in great detail by the operating surgical oncologists and were confirmed with the reviewing pancreatic surgical oncologist.

TABLE 1: Location of Venous Tumor Thrombus

Pathologic analysis—All specimens from surgery were evaluated by a pathologist with specialty training in gastrointestinal pathology. Gross pathologic data and microscopic data were recorded in detail. The variables reported under gross pathologic data included a description of the submitted specimens such as size, location, color, tumor extension, and relationship to adjacent structures. The variables reported under microscopic data included pathologic diagnosis, extrapancreatic tumor extension, lymph node involvement, the number of nodes involved and resected, arterial or venous tumor thrombus, intraductal growth, resection margin status, and description of adjacent organ involvement if present. These data were available in the electronic medical records and were obtained by accessing the patients’ records.

Statistics

Descriptive statistics were summarized for the variables of interest. The percentage of patients with thrombus or intraductal growth was calculated and is presented with its 95% CI, which was calculated using the binomial exact method. For comparison of tumor size between two groups, either the two-sample Student t test or Wilcoxon rank sum test was used depending on the distribution of the data.

figure
View larger version (232K)
Fig. 1A63-year-old woman with abdominal pain.

A, Axial contrast-enhanced CT image obtained in pancreatic parenchymal phase of enhancement shows tumor in tail of pancreas (P). Hypervascular tumor thrombus (arrows) involves entire splenic vein.

figure
View larger version (225K)
Fig. 1B63-year-old woman with abdominal pain.

B, Axial contrast-enhanced CT image obtained in pancreatic parenchymal phase of enhancement shows hypervascular tumor thrombus (arrows) involves right and left portal veins.

Results
Previous sectionNext section
Patient Characteristics

A total of 88 patients met our inclusion criteria and formed the subject population for this study. Forty-nine patients were men and thirty-nine patients were women. The patients ranged in age from 22 to 77 years, with a mean age of 55 years.

Three of the 88 patients had multiple endocrine neoplasia type 1 syndrome and one had tuberous sclerosis. However, all patients had nonfunctional pancreatic neuroendocrine tumors based on the absence of symptoms and elevated hormone production.

The tumors ranged in size from 1.0 to 14.6 cm, with a mean size of 4.97 cm and a median size of 4.65 cm.

Arterial Tumor Thrombus

None of the 88 patients had arterial tumor thrombus.

Arterial Narrowing or Occlusion

Sixteen of the 88 patients (18%; 95% CI, 11–28%) had arterial narrowing of the splenic artery. The splenic artery was encased in 15 patients, and the splenic artery was abutted by the primary tumor in one patient.

figure
View larger version (249K)
Fig. 2A59-year-old man with occasional abdominal pain.

A, Curved reformatted image (axial plane) obtained in portal phase of enhancement shows mass in body of pancreas (P) with tumor thrombus (arrow) extending into main portal vein.

figure
View larger version (234K)
Fig. 2B59-year-old man with occasional abdominal pain.

B, Curved reformatted image (coronal plane) obtained in portal phase of enhancement shows mass in body of pancreas (P) with tumor thrombus (arrow) extending into main portal vein.

Venous Narrowing or Occlusion

In 31 of the 88 patients (35%; 95% CI, 25–46%), the primary tumor was causing venous narrowing that involved the splenic vein (21 patients), SMV (five patients), PV (four patients), and gastrocolic trunk (one patient).

Venous Tumor Thrombus

Twenty-nine of the 88 patients (33%; 95% CI, 23–44%) had venous tumor thrombus in association with the pancreatic neuroendocrine tumor. In 18 of the 29 patients (62%; 95% CI, 42–79%), this finding was not accurately reported. In six of the 18 patients, venous tumor thrombus was not reported at all. In one patient, tumor thrombus was reported as “abutting the vein.” In eight patients, the vein was reported as being “occluded” by the mass. In three patients, venous tumor thrombus was reported as “involvement of the vessels,” but the presence of venous tumor thrombus was not mentioned.

Thirty-nine of the 88 patients underwent surgery after imaging (Table 1). Of these 39 patients, 11 (28%; 95% CI, 15–45%) had evidence of a venous tumor thrombus on CT that was subsequently confirmed on pathologic analysis. This finding was not accurately reported in seven of the 11 patients. In two of these 11 patients (18%; 95% CI, 2–52%), the surgical plan was significantly altered. One patient had jejunal vein thrombus that led to jejunal and jejunal mesenteric resection—surgery that is more extensive than the traditional pancreaticoduodenectomy. Incidentally, this decision to alter the surgical plan was not reported at the time of initial imaging; the decision to expand the surgery was made in the operating room. In the second patient, surgery was aborted because the tumor thrombus within the splenic vein extended to the PV confluence. The presence of tumor thrombus within the PV is a contraindication to performing distal pancreatectomy. In the remaining nine patients, surgery was not significantly altered.

In addition, 10 patients had findings of only microscopic venous tumor thrombus on pathologic analysis with no gross pathologic evidence of this feature. In two of these 10 patients, the pancreatic neuroendocrine tumor did not abut any vein by CT. In four patients, the pancreatic neuroendocrine tumor abutted the splenic vein without narrowing. In one patient, the pancreatic neuroendocrine tumor abutted the SMV without narrowing. In three patients, the pancreatic neuroendocrine tumor abutted the splenic vein with narrowing of the splenic vein. No venous tumor thrombus was evident on CT in all 10 patients. These microscopic venous tumor thrombi would not be expected to be detected by CT.

figure
View larger version (259K)
Fig. 3A51-year-old man with elevated alkaline phosphatase value.

A, Axial contrast-enhanced CT image obtained in pancreatic parenchymal phase of enhancement shows tumor in head of pancreas (P). There is hypervascular tumor thrombus (arrow) involving gastrocolic vein.

figure
View larger version (176K)
Fig. 3B51-year-old man with elevated alkaline phosphatase value.

B, Axial contrast-enhanced CT image obtained in portal venous phase of enhancement shows tumor in head of pancreas. There is hypervascular tumor thrombus (arrow) involving gastrocolic vein.

No venous tumor thrombus was detected either on CT or by pathologic analysis in the remaining 18 patients. The sensitivity of CT in detecting any type of venous tumor thrombus, either gross or microscopic, was 52.4% and specificity, 100%.

The remaining 49 of 88 patients did not undergo surgery. CT evidence of venous tumor thrombus was identified in 18 of these patients (37%; 95% CI, 23–52%) (Figs. 1A, 1B, 2A, and 2B). Of these 18 patients, follow-up CT was not performed in six patients, showed persistent tumor thrombus in eight patients, showed a decrease in the size of the tumor thrombus and a decrease in the size of the primary tumor (likely representing response to treatment) in three patients, and showed an increase in the size of the tumor thrombus with a stable primary tumor in one patient. The remaining 31 patients had no evidence of a venous tumor thrombus on CT.

figure
View larger version (150K)
Fig. 4A45-year-old man with nausea.

A, Axial contrast-enhanced CT image obtained in portal phase of enhancement shows tumor in head of pancreas (P). There is tumor thrombus (arrow) in jejunal vein.

figure
View larger version (153K)
Fig. 4B45-year-old man with nausea.

B, Axial contrast-enhanced CT image obtained at more inferior location than A in portal phase of enhancement shows tumor thrombus (thin arrow) in jejunal vein. Contrast material (thick arrow) within uninvolved jejunal vein is also shown.

Thus, 29 of the 88 patients (33%) had evidence of tumor thrombus formation on CT images (Figs. 3A, 3B, 4A, 4B, and 5). The tumor thrombus was located in the splenic vein (15 patients [Fig. 5]), at the splenic or PV confluence (five patients), in the SMV (five patients), at the junction of the splenic vein and inferior mesenteric vein (one patient), in jejunal and pancreatic veins (one patient [Figs. 4A and 4B]), in a gastrocolic vein (one patient [Figs. 3A and 3B]), and at the junction of the PV and SMV (one patient). For these patients with tumor thrombus, the size of the tumors ranged from 3.1 to 9.3 cm with a mean size of 5.8 cm. The size of the tumor in patients without tumor thrombus ranged from 1 to 14.6 cm with a mean size of 4.5 cm. A two-sample Student t test comparing the tumor size between patients with and those without tumor thrombus yields a p value of 0.005, indicating evidence that patients with tumor thrombus had significantly larger tumors.

Ductal obstruction—Thirty-three of the 88 patients (38%; 95% CI, 27–48%) had ductal obstruction caused by the primary pancreatic neuroendocrine tumor.

Intraductal growth—In five of the 88 patients (6%; 95% CI, 2–13%), CT showed intraductal growth of the tumor (Figs. 6A, 6B, 6C, and 7). The tumor sizes in these patients ranged from 5.4 to 11.4 cm with a mean size of 7.36 cm. The Wilcoxon rank sum test yielded a p value of 0.02, indicating evidence that patients with intraductal growth had larger tumors than those who did not. Of these five patients, three did not undergo surgery and two did. Surgical confirmation of intraductal growth in these two patients was obtained. The surgical plan was altered to ensure a negative margin in one of the two patients (Fig. 7) because the intraductal tumor extended beyond the primary tumor.

Discussion
Previous sectionNext section

Surgery for pancreatic neuroendocrine tumors differs from that for pancreatic ductal adenocarcinoma. In pancreatic neuroendocrine tumors, the surgical goal is to completely resect the tumor and metastases if feasible. Palliative surgery [21] is also performed in pancreatic neuroendocrine tumors especially in the setting of splenic vein occlusion and subsequent gastric varices likely related to the difference in prognosis. Familiarity with the different patterns of spread in pancreatic neuroendocrine tumors when compared with the more common ductal adenocarcinoma is necessary. Assessments of venous narrowing or occlusion, arterial abutment or encasement, liver metastases, and metastatic adenopathy are common to pancreatic ductal adenocarcinomas and pancreatic neuroendocrine tumors and are well known. Distinctive and less well-known patterns of spread include venous tumor thrombus and intraductal growth and were the focus of this study.

figure
View larger version (230K)

Fig. 5 —37-year-old man with abdominal pain. Axial contrast-enhanced CT image obtained in portal phase of enhancement shows tumor in tail of pancreas (P). There is tumor thrombus (arrows) involving splenic vein. Multiple liver metastases are seen.

Venous involvement has been previously described in pancreatic neuroendocrine tumors. The term “venous involvement” in the past could represent either venous narrowing or occlusion or venous tumor thrombus. Because of the uncommon presentation of pancreatic neuroendocrine tumors, both functional and nonfunctional pancreatic neuroendocrine tumors have often been considered together for purposes of assessment of venous involvement.

Bok et al. [14] evaluated conventional angiograms of 76 patients with either functional or nonfunctional pancreatic neuroendocrine tumors. Only three of the 76 patients (4%) had tumor thrombus detected in the PV. In seven other patients, venous occlusion or encasement was detected.

In their study of 124 patients with functional or nonfunctional pancreatic neuroendocrine tumors, Buetow et al. [22] evaluated CT features. They used the term “vascular invasion” to describe tumor involvement of the splenic vein, SMV, and splenic artery; 28 patients (23%) were reported to have vascular invasion. However, separation of arterial versus venous involvement and a description of the “venous involvement” were not reported. Venous tumor thrombus was not recorded as a separate entity.

figure
View larger version (225K)
Fig. 6A72-year-old man with abdominal pain.

A, Axial contrast-enhanced CT image obtained in pancreatic parenchymal phase of enhancement shows tumor in tail of pancreas (P). There is tumor (arrow) growing into pancreatic duct at pancreatic neck.

figure
View larger version (204K)
Fig. 6B72-year-old man with abdominal pain.

B, Axial contrast-enhanced CT image obtained caudad to A in pancreatic parenchymal phase of enhancement shows tumor in tail of pancreas (P). There is tumor (arrow) growing into pancreatic duct at pancreatic neck. Portal venous (PV) confluence is seen posterior to tumor.

figure
View larger version (192K)
Fig. 6C72-year-old man with abdominal pain.

C, Reconstructed coronal image from contrast-enhanced CT in pancreatic parenchymal phase of enhancement shows tumor (arrow) growing into pancreatic duct and extending to neck of pancreas.

Stafford-Johnson et al. [23] reported that one of six patients (17%) with nonfunctional pancreatic neuroendocrine tumors evaluated by CT had a venous tumor thrombus. In addition, several case reports have documented tumor thrombus in patients with pancreatic neuroendocrine tumors [813].

To our knowledge, our study is the largest series to date to focus on the presence and incidence of venous tumor thrombus on CT in patients with nonfunctional pancreatic neuroendocrine tumors. In our study, 29 of 88 patients (33%) had venous tumor thrombus detected by CT. The mean size of the nonfunctioning pancreatic neuroendocrine tumors in our study was approximately 5 cm compared with an average size of 7.7 cm reported by Buetow et al. [22] and average sizes of 9 cm [14] and 10 cm [7] reported in other studies. The incidence of venous tumor thrombus in our study group is higher than what has been previously reported. This difference in incidence may be because our study population consisted of patients with solely nonfunctional pancreatic neuroendocrine tumors who tend to present with larger tumors and more advanced disease than patients with functional pancreatic neuroendocrine tumors.

The presence of venous tumor thrombus changed the surgical plan in two of the 11 patients as described in the Results section. The decision to change the surgery was made in an intraoperative setting in both these patients. Preoperative knowledge of the presence and extent of venous tumor thrombus is clearly crucial for surgical planning.

Microscopic venous tumor thrombi were present in 10 additional patients who underwent surgery. These thrombi were not expected to be seen by CT and were not seen at the time of gross pathology evaluation. The presence of microscopic venous tumor thrombi did not change the surgical plan in any of these 10 patients.

To our knowledge, intraductal tumor growth in pancreatic neuroendocrine tumors has been described only in case reports [1518]. This series is the only one to date that describes the incidence of intraductal tumor growth in nonfunctional pancreatic neuroendocrine tumors. In our study, five of 88 patients (6%) had intraductal tumor growth. This CT finding altered the surgical plan in one patient as described in the Results section.

figure
View larger version (208K)

Fig. 7 —45-year-old man with nausea (same patient as in Figs. 4A and 4B). Axial contrast-enhanced CT image obtained in portal phase of enhancement shows tumor in head of pancreas (P). There is tumor (arrow) growing along pancreatic duct toward tail of pancreas. Duct (D) is dilated in tail of pancreas with atrophic changes of tail of pancreas.

Liver metastases and metastatic lymphadenopathy were not reported in this study. The presence and evaluation of those patterns of spread are well known.

Despite the numerous strengths of this study, its limitations must also be acknowledged. First, this study was a retrospective one with patients selected from a database by searching for specific terms. Depending on variances in coding and recorded diagnoses, we may have inadvertently missed a few patients with nonfunctional pancreatic neuroendocrine tumors. The second limitation is that our institution is a tertiary care center specializing in cancer care and treatment. Therefore, the patients referred to us likely presented with complex disease, late-stage disease, or both. This fact along with the improvement in CT technology may account for the higher incidence of tumor thrombus in our study when compared with the study by Bok et al. [14]. The third potential limitation is that we did not have pathologic proof of tumor thrombus in 18 patients. This limitation is somewhat offset by the fact that follow-up imaging was performed in 12 patients, and the follow-up imaging findings reinforced the diagnosis of venous tumor thrombus. Another limitation is the inclusion of the following four patients: three patients with multiple endocrine neoplasia, type 1, and one patient with tuberous sclerosis. These pancreatic neuroendocrine tumors were included only because they were nonfunctioning based on the absence of symptoms and the absence of elevated hormone levels.

In conclusion, pancreatic neuroendocrine tumor spread differs from pancreatic ductal adenocarcinoma spread. Two of the less well-known patterns of spread are venous tumor thrombus and intraductal growth. Venous tumor thrombus formation may be more common than previously suspected in patients with nonfunctional pancreatic neuroendocrine tumors. In our series, which is the largest to date and which specifically focuses on nonfunctional pancreatic neuroendocrine tumors, one third of patients with nonfunctional pancreatic neuroendocrine tumors had venous tumor thrombi. The presence and extent of venous tumor thrombus and intraductal tumor growth can change surgical planning and should be specifically addressed in the radiology report.

References
Previous sectionNext section
1. Solcia EKG, Sobin LH. Histological typing of endocrine tumours. New York, NY: Springer-Verlag, 2001 [Google Scholar]
2. Halfdanarson TR, Rabe KG, Rubin J, Petersen GM. Pancreatic neuroendocrine tumors (PNETs): incidence, prognosis and recent trend toward improved survival. Ann Oncol 2008; 19:1727–1733 [Google Scholar]
3. Falconi M, Plockinger U, Kwekkeboom DJ, et al. Well-differentiated pancreatic nonfunctioning tumors/carcinoma. Neuroendocrinology 2006; 84:196–211 [Google Scholar]
4. Zerbi A, Falconi M, Rindi G, et al; AISP-Network Study Group. Clinicopathological features of pancreatic endocrine tumors: a prospective multicenter study in Italy of 297 sporadic cases. Am J Gastroenterol 2010; 105:1421–1429 [Google Scholar]
5. Panzuto F, Nasoni S, Falconi M, et al. Prognostic factors and survival in endocrine tumor patients: comparison between gastrointestinal and pancreatic localization. Endocr Relat Cancer 2005; 12:1083–1092 [Google Scholar]
6. Procacci C, Carbognin G, Accordini S, et al. Nonfunctioning endocrine tumors of the pancreas: possibilities of spiral CT characterization. Eur Radiol 2001; 11:1175–1183 [Google Scholar]
7. Eelkema EA, Stephens DH, Ward EM, Sheedy PF. CT features of nonfunctioning islet cell carcinoma. AJR 1984; 143:943–948 [Abstract] [Google Scholar]
8. Nguyen BD. Pancreatic neuroendocrine tumor with portal vein tumor thrombus: PET demonstration. Clin Nucl Med 2005; 30:628–629 [Google Scholar]
9. Smith TM, Semelka RC, Noone TC, Balci NC, Woosley JT. Islet cell tumor of the pancreas associated with tumor thrombus in the portal vein. Magn Reson Imaging 1999; 17:1093–1096 [Google Scholar]
10. Obuz F, Bora S, Sarioglu S. Malignant islet cell tumor of the pancreas associated with portal venous thrombus. Eur Radiol 2001; 11:1642–1644 [Google Scholar]
11. Watase M, Sakon M, Monden M, et al. A case of splenic vein occlusion caused by the intravenous tumor thrombus of nonfunctioning islet cell carcinoma. Surg Today 1992; 22:62–65 [Google Scholar]
12. Sainz-Esteban A, Prasad V, Baum RP. Interesting image: pancreatic neuroendocrine tumor with involvement of the inferior mesenteric vein diagnosed by Ga-68 DOTA-TATE PET/CT. Clin Nucl Med 2010; 35:40–41 [Google Scholar]
13. Bedirli A, Patiroglu TE, Sakrak O, Aritas Y. Portal vein resection for a portal vein thrombus caused by nonfunctioning islet cell carcinoma: report of a case. Surg Today 2004; 34:802–804 [Google Scholar]
14. Bok EJ, Cho KJ, Williams DM, Brady TM, Weiss CA, Forrest ME. Venous involvement in islet cell tumors of the pancreas. AJR 1984; 142:319–322 [Abstract] [Google Scholar]
15. Shimizu K, Shiratori K, Toki F, et al. Nonfunctioning islet cell tumor with a unique pattern of tumor growth. Dig Dis Sci 1999; 44:547–551 [Google Scholar]
16. Kitami CE, Shimizu T, Sato O, et al. Malignant islet cell tumor projecting into the main pancreatic duct. J Hepatobiliary Pancreat Surg 2000; 7:529–533 [Google Scholar]
17. Akatsu T, Wakabayashi G, Aiura K, et al. Intraductal growth of a nonfunctioning endocrine tumor of the pancreas. J Gastroenterol 2004; 39:584–588 [Google Scholar]
18. Inagaki M, Watanabe K, Yoshikawa D, et al. A malignant nonfunctioning pancreatic endocrine tumor with a unique pattern of intraductal growth. J Hepatobiliary Pancreat Surg 2007; 14:318–323 [Google Scholar]
19. Katz MH, Pisters PW, Evans DB, et al. Borderline resectable pancreatic cancer: the importance of this emerging stage of disease. J Am Coll Surg 2008; 206:833–846; discussion, 846–848 [Google Scholar]
20. Tseng JF, Raut CP, Lee JE, et al. Pancreaticoduodenectomy with vascular resection: margin status and survival duration. J Gastrointest Surg 2004; 8:935–949; discussion, 949–950 [Google Scholar]
21. Burns WR, Edil BH. Neuroendocrine pancreatic tumors: guidelines for management and update. Curr Treat Options Oncol 2012; 13:24–34 [Google Scholar]
22. Buetow PC, Parrino TV, Buck JL, et al. Islet cell tumors of the pancreas: pathologic-imaging correlation among size, necrosis and cysts, calcification, malignant behavior, and functional status. AJR 1995; 165:1175–1179 [Abstract] [Google Scholar]
23. Stafford-Johnson DB, Francis IR, Eckhauser FE, Knol JA, Chang AE. Dual-phase helical CT of nonfunctioning islet cell tumors. J Comput Assist Tomogr 1998; 22:335–339 [Google Scholar]

Address correspondence to A. Balachandran ().

Recommended Articles

Venous Tumor Thrombus in Nonfunctional Pancreatic Neuroendocrine Tumors

Full Access, , ,
American Journal of Roentgenology. 2012;199:W335-W344. 10.2214/AJR.11.8027
Abstract | Full Text | PDF (1096 KB) | PDF Plus (1090 KB) 
Full Access, , , , , ,
American Journal of Roentgenology. 2012;199:609-615. 10.2214/AJR.11.7349
Abstract | Full Text | PDF (671 KB) | PDF Plus (750 KB) 
Full Access, , , , , ,
American Journal of Roentgenology. 2012;199:W331-W334. 10.2214/AJR.11.8056
Abstract | Full Text | PDF (650 KB) | PDF Plus (664 KB) 
Full Access, , , ,
American Journal of Roentgenology. 2012;199:309-318. 10.2214/AJR.12.8627
Abstract | Full Text | PDF (1322 KB) | PDF Plus (1335 KB) 
Full Access, ,
American Journal of Roentgenology. 2014;203:1192-1204. 10.2214/AJR.13.12386
Abstract | Full Text | PDF (1122 KB) | PDF Plus (1179 KB) | Erratum 
Full Access, , ,
American Journal of Roentgenology. 2007;189:67-72. 10.2214/AJR.06.1076
Abstract | Full Text | PDF (1538 KB) | PDF Plus (1619 KB)