Neuroendocrine Tumors of the Pancreas: Diagnosis

Jun Ushio1, Kensuke Yokoyama1, Alan Kawarai Lefor2, Kiichi Tamada1

1Division of Gastroenterology, Department of Medicine, Jichi Medical University, Japan 2Department of Surgery, Jichi Medical University, Japan

*Corresponding Author:
Jun Ushio
Department of Medicine
Division of Gastroenterology
Jichi Medical University
3311-1 Yakushiji, Shimotsuke
Tochigi 329-0498, Japan
Tel +81 285 58 7348
Fax +81 285 44 8297
E-mail [email protected]

Received Date: December 1st, 2017; Accepted Date: December 22nd, 2017

 
Visit for more related articles at JOP. Journal of the Pancreas

Abstract

Pancreatic neuroendocrine tumors are rare. In patients with functional PNETs, the excess hormones produced lead to a variety of hormone-related symptoms. Non-functioning tumors do not produce symptom-inducing hormones. Therefore, they are often discovered at an advanced stage with large tumors and metastatic spread. PNETs have a variable appearance on computed tomography scans and magnetic resonance imaging. With most functioning PNETs, Dynamic computed tomography scans and magnetic resonance imaging show well defined hypervascular small tumors. Imaging of other types of PNET show purely cystic, complex cystic or solid tumors. Functional imaging is useful both to detect the primary lesion and stage the disease. It is also useful to select candidates for peptide receptor radiometabolic treatment. Somatostatin receptor scintigraphy is the most available functional imaging technique. The gallium 68-SST analogue positron emission tomography scan is more sensitive, and is expected to be the future of functional imaging for PNETs. 18FDG PET/CT scan is useful for poorly differentiated tumors. The sensitivity of Carbidopa-assisted 18F-FDOPA PET/CT is better than somatostatin receptor scintigraphy for detecting nonfunctional PNETs. Endoscopic ultrasonography is superior to dynamic computed tomography scan to identify PNETS. Endoscopic ultrasonography-fine needle aspiration offers a high accuracy for the diagnosis of PNETs. However, tumors located in the pancreatic head and tumors with rich stromal fibrosis are associated with reduced sampling adequacy of endoscopic ultrasonography-fine needle aspiration. Some patients with poorly-differentiated PNETs have invasion of the pancreatic duct. Endoscopic retrograde cholangiopancreatography is useful for evaluating these patients.

Keywords

Neuroendocrine Tumors; Pancreas

Abbreviations

CEUS contrast-enhanced ultrasonography; MEN multiple endocrine neoplasia; PNETs pancreatic neuroendocrine tumors

INTRODUCTION

Pancreatic neuroendocrine tumors (PNETs) are a rare group of heterogeneous neoplasms [1, 2, 3]. PNETS can be classified as either functional or non-functional according to the presence of biologically active hormones and characteristic symptoms. PNETS show a wide range of malignant potential which range from slow-growing and non-infiltrative tumors to locally invasive and metastasizing tumors. Significant advances in diagnostic modalities have been made over the past decade. Although many reviews describe the utility of cross-sectional imaging modalities, the utility of endoscopic modalities in this area has not been well discussed. This article provides a comprehensive review of PNETs and an update on advances in this area.

CLINICAL AND BIOLOGICAL PRESENTATION

Functional PNETs

Some PNETs are functional. The excess hormones produced lead to a variety of hormone-related symptoms (Table 1). In these patients, early diagnosis is often possible even if the tumor is small.

image

Approximately 60% of functioning PNETs are insulinomas, which are usually benign. Other types of PNETs are often malignant. Whipple’s triad is famous as the constellation of symptoms in patients with insulinomas, and includes the presence of symptomatic hypoglycemia (in about 85% of patients), low blood sugar at the time of symptoms, and relief of symptoms with glucose administration. Therefore, many patients are obese due to overeating to avoid hypoglycemia.

Gastrinomas are gastrin-secreting tumors with symptoms typical in common peptic ulcer disease. Usually, the abdominal pain is less responsive to medical treatment. Sometimes, symptoms may relate to a complication of peptic ulcer disease, such as bleeding, gastric outlet obstruction, or perforation. Over 50% of gastrinomas are malignant. They may metastasize to regional lymph nodes and the liver. Hypergastrinemia also occurs in patients taking a proton pump inhibitor, especially with chronic renal failure as well as patients with gastrinomas. Expensive diagnostic evaluations for gastrinoma should not be conducted in these patients. Twenty percent of gastrinomas are related to multiple endocrine neoplasia (MEN) type-1 and are associated with hyperparathyroidism and pituitary adenomas. Patients with MEN type-1 associated PNETs are usually diagnosed at an earlier age than sporadic tumors. Clinical symptoms associated with other functioning tumors are summarized in Table 1.

Nonfunctional PNETs

Non-functioning tumors do not produce hormones, and are generally asymptomatic. Therefore, they are often discovered in an advanced stage with large tumors which have already metastasized at the time of diagnosis. Symptoms in patients with non-functioning PNETs are non-specific, and include abdominal pain, diarrhea, a prolonged feeling of fatigue, fainting, or weight loss. More than 50% of nonfunctioning tumors (about 40% of PNETs) are likely to be malignant (metastatic potential).

CROSS-SECTIONAL IMAGING

Ultrasonography (US)

The utility of conventional ultrasonography for the diagnosis of PNETs is limited, since it has limited ability to image the tail of the pancreas, especially in obese patients. However, recent studies have reported the utility of contrast-enhanced ultrasonography (CEUS) in this area. CEUS shows slightly, moderately or well-enhanced patterns in patients with pancreatic adenocarcinoma, inflammatory pancreatic masses, and islet cell tumors, respectively [4]. The imaging pattern on CEUS correlates with tumor grade (G1 and G2) except for G3 lesions in patients with PNETs. These findings support a possible role for information from CEUS to be used as a prognostic factor [5].

Computed Tomography (CT) Scan

Primary Tumor: CT is the main imaging modality for PNETs. For evaluation of PNETs, arterial/pancreatic phase, venous phase and delayed phase are required. The late arterial (30 s) or pancreatic phase (40 s) is suitable for the detection of small functioning PNETs in particular insulinoma [6]. It is also suitable for the detection of hepatic metastases [6, 7, 8, 9]. In our experience, however, only early arterial phase shows tumor in some patients with small functioning PNETS. The delayed phase is complementary to the arterial/pancreatic and the venous phase. Delayed phase CT scan shows delayed enhancement in some fibrous tumors [10].

Functioning Tumors: Functioning PNETs are most often manifest by endocrine symptoms. Therefore, the tumors can be detected early, often when they are small.

Insulinomas are most frequent among the functioning PNETs. Most insulinomas are under 2cm in size, solitary and benign. These tumors are detected throughout the pancreas including the head, body, and tail. Typical insulinomas are hyper-vascular and CT scans show intense enhancement during the early phase (arterial/pancreatic) (Figure 1). The enhancement is usually uniform. Sometime, a rim of enhancement is seen [11].

pancreas-hyper-enhancing

Figure 1 Hyper-enhancing tumor (Insulinoma, G1). (a, b). Computed tomography scan images in the transverse plane during the arterial and the portal venous phase show a small hyper-vascularized tumor in the pancreas with sharp margins. (c). T2-weighted magnetic resonance image shows a small lesion with a well circumscribed margin and high signal intensity.

Gastrinomas are the second most common among PNETs. Gastrinomas are also small pancreatic tumors generally 1-3 cm. About 80% are found within the ‘‘gastrinoma triangle’’ defined as the confluence of the cystic and common bile duct superiorly, the second and third portions of the duodenum inferiorly, and the neck and body of the pancreas medially. Gastrinomas also are often associated with MEN1 syndrome. In gastrinomas, dynamic CT scans often show a delayed enhancement persistent in the delayed phase due to the presence of fibrosis.

CT scan imaging of other functioning PNETs includes purely cystic tumors in 10%, a common pattern of PNETs associated with MEN1, and complex solid and cystic appearance and calcified tumors in less than 5% (Figure 2) [12].

pancreas-functional

Figure 2 Cystic and hyperenhancing tumor (Non-functional tumor, G2). (a). A Dynamic computed tomography scan image obtained during the arterial phase shows a large mass with solid and cystic components. (b). A T2-weighted magnetic resonance image shows a small lesion with high signal intensity.

The differential diagnosis of hypervascularized PNETs includes pancreatic metastases from a renal cell carcinoma and an intrapancreatic accessory spleen. In addition, pancreatic arterio-venous malformations also have an appearance similar to small hypervascularized PNETs. CT scan findings in patients with pancreatic metastases from renal cell carcinoma and patients with hypervascular PNETs were compared with a focus on the relative percentage washout (RPW). The mean RPW in the renal cell carcinoma group was significantly higher than that in the PNET group [13]. Multiple hypervascular PNETs is frequently seen in patients with MEN1 as well as in those with metastases from renal cell carcinoma.

Non-Functioning Tumors: On CT and MR imaging, nonfunctioning PNETs are seen as large pancreatic masses with heterogeneous enhancement because of necrotic and hemorrhagic changes. Patients with hypo-enhancing tumors (Figure 3) generally have a worse prognosis after resection compared to patients with hyper-enhancing lesions or a cystic appearance with hyper-enhancing tumors [14].

pancreas-tomography

Figure 3 Hypo-enhancing tumor (Non-functional tumor). (a). Computed tomography images during the arterial phase show a hypovascularized tumor of the pancreas as shown in pancreatic adenocarcinoma. The lesion has a worse prognosis compared with hyperechoic or cystic and hyper-enhancing tumors. (b). Computed tomography images during the portal venous phase show a tumor thrombus in the portal vein. (c). Magnetic Resonance images: A T2-weighted image shows a heterogenous hyperintense tumor. (d). Magnetic Resonance images: A T2-weighted image shows multiple liver metastases better depicted than on computed tomography scan. (e). The pancreatic tumor and liver metastases are also well depicted on diffusion weighted images.

Staging: CT scan plays a major role in the staging of these lesions. The TNM stage and the extent of distant metastases, especially to the liver, are the most important prognostic factors. Liver metastases correlate with the prognosis [15, 16] and are important parameters in developing the optimal treatment plan. Other typical sites for metastases from PNETs include abdominal and mediastinal lymph nodes, peritoneum, and bone. The frequency of metastatic disease in each of these sites depends on the biologic behavior of the primary tumor, the stage of disease and differentiation of the primary tumor [17].

CT scan and somatostatin receptor scintigraphy (SRS) complement each other for the staging of PNETs. While CT scan is more sensitive for detecting lung, liver and brain metastases, SRS is more sensitive for detecting metastases to bones and the mediastinum [9, 18, 19].

Magnetic Resonance (MR) Imaging

Primary Tumor: MR imaging protocols for PNETs should include T1-(T1W) and T2- weighted (T2W) sequences, dynamic three-dimensional (3D) sequences before and after intravenous administration of a gadolinium chelate with multi-arterial, venous and delayed over 5 min acquisition and diffusion-weighted (DWI) sequences. Fat suppression on T1W and T2W images is useful to maximize the signal intensity differences between the pancreatic tumor and the adjacent normal pancreatic tissue. Similar to CT scan, T1W images delayed over 5 min images improve tumor detection [10]. Diffusion-weighted images increase the sensitivity to detect the primary pancreatic tumor as well as associated liver metastases [20].

Functioning Tumors: In most functioning PNETs, MRI shows low signal intensity on T1W, high signal intensity on T2W images and intense and early enhancement on dynamic T1W sequences. The T2W sequence with fat suppression is useful to detect hypervascular tumors (typically, insulinoma). However, hypovascular tumors are better detected using the T1W sequence during the arterial phase. Non-hypervascular tumors are surrounded by enhanced normal pancreatic parenchyma. Therefore, strong enhancement of the pancreas in the arterial phase is suitable to detect it. This condition may conceal hypervascular tumors [21].

DWIs are useful to show small PNETs due to their excellent image contrast. ADC values are lower than adjacent pancreatic parenchyma in all cases of solid nodules [21]. However, ADC values are higher in the presence of a cystic pattern [22].

Non-Functioning Tumors: On MR images, most PNETs are hyperintense on T2W images and hyper- or isointense during the arterial/pancreatic phase of a dynamic study in contrast to pancreatic adenocarcinoma [23]. Tumor vein thromboses (splenic, portal and superior mesenteric veins) are frequent, and vascular invasion is less common with PNETs compared to pancreatic adenocarcinoma [21, 23] (Figure 3). Dilatation of the upstream pancreatic and common bile duct is also less common than in pancreatic adenocarcinoma [24].

Staging: MRI is more sensitive than US, CT scan, or SRS for the detection of liver metastases. It is also the imaging technique with the best inter-observer agreement [7, 8] (Figure 3). Its sensitivity is similar to that of intraoperative US. However, available pre- and intraoperative imaging techniques cannot detect about half of all liver metastases [25]. The addition of DWI sequences to standard MRI revealed additional metastases and led to modifications in patient management. Adding DWI to standard liver MRI provided additional findings for 45% of patients with 1.78 times more new lesions, and resulted in a management change for 18% of patients. DWI sequences added to whole body MRI provided additional findings for 71% of patients, with 1.72 times more lesions, resulting in a management change for 19% of patients [26].

RADIOPHARMACEUTICAL IMAGING TECHNIQUES

Somatostatin Receptor Scintigraphy (SRS)

Isotope-imaging modalities have a major role in the management of patients with PNETs. Due to the expression of multiple somatostatin receptors (SSTRS) by about 70% of PNETS, functional imaging with somatostatin (SST) analogues is used to detect NETs [27]. Functional imaging with somatostatin analogues is useful first to evaluate the expression of SSTRS. It allows us to assess disease staging, recurrence, and finally to select patient candidate for peptide receptor radio-metabolic treatment (PRRT) by Y90 (Yttrium-90) or Lu177 (Lutetium-177) SST analogues. Poorly differentiated neuroendocrine carcinomas have a low expression of SST receptors and functional imaging with SST analogues has a very limited role [28].

111-pentetrotide single photon emission computed tomography (SPECT)-SRS (OCTREOSCAN; Mallinckrodt, St Louis, MO) is the most commonly available somatostatin analog tracer with high affinity for the 2 and 5 subtypes. Scintigraphic imaging require a 2-day protocol for image acquisition (4h and 24h) with whole body 2D (anteriorposterior) evaluation at 24 h. Currently, SPECT images with 3D and fused images are available. This modality has a higher sensitivity than planar images to localize small primary tumors or distant metastases (Figure 4).

pancreas-somatostatin

Figure 4 Tumor detected by somatostatin analogue scintigraphy (Insulinoma, G1). (a). Dynamic computed tomography scan shows a hypervascular mass. (b). 111In-pentetreotide single photon emission tomography (Octreoscan) shows high uptake of somatostatin analogue in pancreas lesions.

SRS scintigraphy may help detect pancreatic primary tumors when cross-sectional imaging and EUS show no lesions. SRS scintigraphy sensitivity ranges from 40% to 70%. SRS is more sensitive in detecting well-differentiated gastrinomas, glucagonomas, VIPomas and non-functioning PNETS.

FDG PET

Conventional FDG PET/CT

Conventional FDG imaging is not considered a good tracer for NETs. Differentiated G1 NETs are most likely to express the SST receptor, to disclose high SSTA uptake and be negative on FDG PET/CT scan. For staging welldifferentiated NETs with high Ki67 indices over 10% and the poorly differentiated variants, FDG PET/CT scan is most useful. In patients with well-differentiated NET and Ki67 over 10%, FDG PET/CT showed higher sensitivity compared to Octreoscan and CT scan with a higher number of detected lesions located in the lymph nodes and bone [29]. Furthermore, FDG uptake is an independent prognostic factor for patients with low-grade gastroenteropancreatic NET (GEPNETS) [30].

F-DOPA PET/CT has had excellent performance for staging midgut tumors. However, studies comparing F-DOPA to Octreoscan in non-midgut digestive NETs showed a better performance of somatostatin analogue imaging compared to F-DOPA with a sensitivity of Octreoscan of 75% compared to 25% for F-DOPA PET/CT [31].

Ga68-SSTA PET also seems to be superior to F-DOPA in a small patient series with well differentiated NETs, including PNETs with sensitivity of 96% for Ga68-SSTA PET/CT as opposed to 56% for F-DOPA PET [32, 33]. Ga 68 PET/CT is more sensitive than Octreoscan with sensitivity of 90% to 100% versus 50% to 80% for Octreoscan. It allows for the detection of micrometastases not seen on Octreoscan, especially in the liver and loco-regional lymph nodes [34, 35]. Some studies showed higher sensitivity of Ga68-SSTA PET tracers comparing to CT scan and/or MRI in detecting distant metastases, especially bone metastases in GEPNETs with sensitivity around 95% to 100% for PET and 60% to 80% for CT. However, further studies comparing Ga68 PET and high quality cross-sectional imaging are needed to define the role of each modality [35, 36, 37, 38].

Carbidopa-Assisted FDG PET/CT

The low sensitivity of 18F-FDOPA PET to detect welldifferentiated PNETS may be due to their embryologic origin and the high physiologic radiotracer uptake and retention in the mature exocrine pancreas [39, 40]. Carbidopa (CD) is an efficient inhibitor of the peripheral aromatic amino acid decarboxylase (AADC). The administration of CD improves interpretation of 18F- FDOPA PET images by lowering physiologic pancreatic uptake and increasing tumor-tobackground uptake ratios [41, 42]. The combination of CD premedication and early acquisition of PET images improves the detection of insulinomas [43, 44].

The sensitivity of CD-assisted 18F-FDOPA PET/CT (90%) is better than that of SRS (68%) for detecting nonfunctional PNETs. Moreover, 18F-FDOPA PET/CT accurately shows nodal metastatic spread than SRS. These data suggest that the utility of CD-assisted 18F-FDOPA PET/ CT for non-functioning PNETS when 68Ga-radiolabeled SSA PET is not yet defined [45].

ENDOSCOPY

Endoscopic Ultrasonography

Endoscopic ultrasonography (EUS) is superior to detect PNETs compared to CT scan, MRI and SRS [46]. EUS is particularly useful to detect small (2 to 5 mm) pancreatic lesions, such as gastrinomas and insulinomas. The detection rates are from 79% to 94% [47, 48, 49] (Figure 5). Due to the proximity of the endoscope, the sensitivity to detect the tumor in the head of the pancreas is higher than in the tail.

pancreas-ultrasound

Figure 5 Tumor detected by Endoscopic ultrasound (Non-functioning tumor, G1). (a). Dynamic computed tomography scan shows a slightly hyper-vascular mass. (b). Endoscopic ultrasound shows a small hyperechoic tumor with well circumscribed margins.

Insulinomas are all located in the pancreas with an average size of under 2 cm at the time of diagnosis in 90% of patients. A gastrinoma can be in the pancreas, and are located in the duodenum in 40% to 50% of patients. When located in the duodenum, gastrinomas are frequently small and multiple. The sensitivity of EUS is higher for a pancreatic gastrinoma than an extrapancreatic gastrinoma, probably because of their size. EUS also can detect adjacent metastatic lymph nodes within the gastrinoma triangle. Contrast-enhanced EUS (CE-EUS) increases the potential for detection of small pancreatic tumors by their ability to detect hypervascular enhancement [50, 51, 52]. The vast majority of small PNETs are predominantly hypoechogenic in B-mode (94%) and demonstrate hyperenhancement compared to the surrounding pancreatic parenchyma after contrast injection (90%).

EUS is also used to survey patients at an increased risk of developing PNETs, particularly in MEN type 1. A prospective multi-center study in 90 patients with MEN type 1 comparing MRI and pancreatic EUS showed that 48 (53%) of patients had at least one tumor over 10 mm. EUS detected 86 tumors over 10mm compared to 67 tumors for MRI. EUS failed to identify 16% of patients with pancreatic tumor over 10mm, and 19% patients for MRI. EUS and MRI should be compensated for each other. Both modalities should be performed at the initial evaluation of patients with MEN type 1.

Endoscopic Ultrasonography Guided Fine-Needle Aspiration (EUS-FNA)

Ability of EUS-FNA to Establish the Diagnosis of PNETs: A number of reports have described the excellent ability of EUS-FNA to establish the diagnosis of PNETS, with a sensitivity of 83% to 93% [53, 54, 55, 56]. EUS-FNA is essential for the preoperative diagnosis of PNETs. However, in about 10% to 15 % of patients, EUS-FNA is of little value [53, 54, 55, 56].

Size: Tumor size was not a significant predictor of adequate sampling. Tumors under 10mm were all diagnosed by EUS-FNA [57]. The reason for such a high yield may be the result of high cellularity and minimal stromal fibrosis in these tumors.

Tumor Location: Tumor location and the amount of intratumoral fibrosis were independent predictors of adequate sampling. Tumors located in the pancreatic body or tail was associated with greater sensitivity [58, 59].

Fibrosis: When tumors contain extensive stromal fibrosis (30%), EUS-FNA has a low diagnostic rate, compared to tumors with minimal fibrosis. Intratumoral fibrosis has been postulated to result from local serotonin production [60, 61], as serotonin has been implicated in fibrogenesis. In addition, serotonin has been shown to stimulate fibroblast mitosis in cell culture [62].

Most PNETs are hyper-intense on T2-WI. However, when PNETs have abundant stromal fibrosis, T2-W1 shows an isointense or hypointense pattern [63, 64]. Therefore, MRI should be performed when EUS-FNA for PNETs is planned. If an isointense or hypointense lesion is found on T2-WI, PNETs with rich fibrosis should be suspected. In such cases, tactics to obtain adequate tissue during EUSFNA are required. CE-EUS may represent an attractive option in such cases to avoid sampling rich fibrous areas. Hypervascular sites in such lesions on CE-EUS are suitable for EUS-FNA [65]. When CE-EUS is not available, using high negative-pressure suction techniques in EUS-FNA [66-68] or using a larger gauge needle is useful options [55, 69].

False-Positives: False-positive results for PNETs have been reported. They include paraganglioma and pseudopapillary neoplasms (SPN). A report by Kari et al. [70] showed that 80% of lesions misclassified as PNETS were actually SPN. Usually, FNA samples demonstrate a pseudopapillary pattern with fibrovascular stalks in SPN. However, in some cases with material crushed during aspiration or inadequate sampling, characteristic features of SPN may not be evident. Additionally, chromogranin A and/or synaptophysin staining is sometimes positive in SPN [71].

Ohara et al. reviewed 30 surgical specimens of NETs (24 cases) and SPN (6 cases). They carried out comprehensive immunohistochemical profiling using 9 markers: synaptophysin, chromogranin A, pancytokeratin, E-cadherin, progesterone receptor, vimentin, α-1-antitrypsin, CD10, and β-catenin. E-cadherin staining in NETs, and nuclear labeling of β-catenin in SPNs were the most sensitive and specific markers. Dot-like staining of chromogranin A might indicate the possibility of SPNs rather than NETs. The other six markers were not useful because their expression overlapped widely between NETs and SPNs [72].

The remaining lesions misdiagnosed as PNETs are paraganglioma [73]. In the case of paraganglioma, EUS-FNA is relatively contraindicated because it may cause a severe hypertensive crisis during EUS-FNA [74]. Therefore, when paraganglioma is suspected, meta-iodobenzylguanidine (MIBG) scintigraphy and/or 24-h urine collection for catecholamines, metanephrines, and vanillylmandelic acid should be conducted before FNA [75].

Grading: The 2010 revised World Health Organization classification grades PNETs as NET-G1 G2 and NEC, based on Ki-67 staining or mitosis rates. Concordance rates between grading of PNETs by EUS-FNA sample and postoperative histology are 77% to 89.5% [76, 77, 78, 79, 80, 81].

As for PNET grading, some previous studies have shown that the Ki-67 index by EUS-FNA sample correlates significantly and independently with the clinical outcome of patients with PNETs [55, 82, 83, 84]. However, other authors have emphasized that metastases can also appear in PNETs with a low Ki-67 index by cytological or histological samples [85, 86].

The use of EUS-FNA samples for PNET grading has several limitations. As in surgical samples, the Ki-67 index from cytological samples can be calculated either in hot spot areas or can be estimated by dividing all positive tumor cells by total tumor cells in the smear. However, PNETs are heterogeneous tumors and the identification of hotspot areas in FNA specimens is difficult. FNA sample size is limited and cannot represent all tumor zones. Therefore, hotspot areas can be overlooked [77, 78, 85]. G2 tumors are particularly noted for their heterogeneity, and these tumors account for most of the discrepant cases. Authors who have correlated Ki-67 index in cytological and tissue samples have found both under-staging and upstaging of PNETS in cytological specimens, although under-staging has been more frequent [77].

As for the size of the lesions, Unno et al. found that there was a significant difference in tumor size in cases with concordant and discordant Ki-67 indices. When the tumor is large, the Ki-67 index obtained from EUS-FNA and surgical samples show a discrepancy [80]. They propose a tumor size of 18 mm as a cutoff to improve reliability of Ki-67 estimation in cytological samples.

In summary, PNETs with higher Ki-67 index from cytology specimens have a tendency towards a worse outcome. However, some patients with tumors classified as G1 on EUS-FNA samples died due to PNETs or had tumor progression. On cytological sample, a G2/G3 result presumably suggests worse prognosis, but a G1 result does not necessarily suggest a good outcome [86].

Endoscopic Retrograde Cholangiopancreatography (ERCP)

In typical PNETs, pancreatography has normal findings or shows only displacement of the main pancreatic duct (MPD). Usually duct obstruction is common in adenocarcinoma of the pancreas, and the intraductal growth of PNETs with narrowing or occlusion of MPD is rare. Fibrosis and compression by the tumor result in narrowing, and invasion or occupation of the tumor occur with occlusion of the MPD [87] (Figure 6). Since the prognosis of patients with PNETs within the lumen of the MPD is worse compared to typical PNETs [88], endoscopic retrograde cholangiopancreatography (ERCP) is useful to predict the prognosis of patients with PNETs. In patients with PNETs invading the pancreatic duct, poorly differentiated intraductal adenocarcinoma (IDA)s may transform into NECs [89].

pancreas-occupying

Figure 6 Tumor occupying the pancreatic duct. This lesion has a worse prognosis. (Non-functional tumor, NEC). (a). Dynamic computed tomography scan shows a hypo-enhanced tumor in the pancreatic duct. (b). Endoscopic retrograde cholangiopancreatography also shows a filling defect in the pancreatic duct.

In patients with intraductal papillary mucinous neoplasms (IPMN) of the pancreas, pancreatic endocrine neoplasms may arise as well as intraductal papillary mucinous carcinoma (IPMC) and ductal adenocarcinoma [90]. When ERCP shows intraductal growth of tumor, anaplastic-type pancreatic adenocarcinoma and acinar cell carcinoma must be considered as well as NETs [91, 92, 93, 94, 95].

CONCLUSION

Neuroendocrine tumors of the pancreas are a heterogeneous group of neoplasms that are generally slow growing. However, they may become incurable if they progress to unresectable metastatic disease. A combination of US, CT scan, MRI, radiopharmaceutical imaging techniques, and endoscopic techniques are useful for the diagnosis and grading of patients with PNETs.

Conflict of Interest

All authors declare having no conflict of interests.

References

Select your language of interest to view the total content in your interested language

Viewing options

Post your comment

Share This Article

Flyer image
 

Post your comment

captcha   Reload  Can't read the image? click here to refresh