CT Evaluation of Gastroenteric Neuroendocrine Tumors: Relationship Between CT Features and the Pathologic Classification

Transcription

1 Gastrointestinal Imaging Original Research Feng et al. Gastrointestinal Imaging Original Research Shi-Ting Feng 1 Yanji Luo 1 Tao Chan 2 Zhenpeng Peng 1 Jie Chen 3 Minhu Chen 3 Zi-Ping Li 1 Feng ST, Luo Y, Chan T, et al. Keywords: CT, gastroenteric neuroendocrine neoplasms, pathologic classification, radiography DOI: /AJR Received May 30, 2013; accepted after revision December 6, This work was supported by the National Natural Science Foundation of China (grant ), the Zhujiang Scientific and Technological New Star Foundation (grant 2012J ), and the Fundamental Research Funds for the Central Universities (grant 10ykpy11). S. T. Feng and Y. Luo contributed equally to this work. 1 Department of Radiology, The First Affiliated Hospital, Sun Yat-Sen University, 58th, The Second Zhongshan Rd, Guangzhou, Guangdong , China. Address correspondence to Z. P. Li (liziping163@163.com). 2 Department of Medical Imaging, Union Hospital, Hong Kong. 3 Department of Gastroenterology, The First Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong, China. WEB This is a web exclusive article. AJR 2014; 203:W260 W X/14/2033 W260 American Roentgen Ray Society CT Evaluation of Gastroenteric Neuroendocrine Tumors: Relationship Between CT Features and the Pathologic Classification OBJECTIVE. The objective of our study was to compare the CT features of gastroenteric neuroendocrine neoplasms (GE-NENs) with the pathologic classification and to analyze the correlation between the CT features and classification of GE-NENs. MATERIALS AND METHODS. Fifty-six cases of pathologically and immunohistochemically proven GE-NENs, including 25 cases of neuroendocrine tumors (NETs) (i.e., G1 and G2 tumors) and 31 cases of neuroendocrine carcinomas (NECs) (i.e., G3 tumors and mixed adenoneuroendocrine carcinomas) were studied. We analyzed various CT features of the primary tumor, nodal status, and metastasis and compared these features with pathologic grading. RESULTS. The CT features that favor NEC over NET include larger tumor size (> 4.0 cm), transmural invasion, circumscribed tumor with both intra- and extraluminal involvement, circumferential growth, areas of cystic change or necrosis, ulceration, mesenteric fat infiltration, and lymphadenopathy, with p values of 0.044, 0.002, 0.024, 0.008, 0.002, 0.007, 0.002, and < 0.001, respectively. The CT features that do not distinguish between the two types of GE-NENs include tumor boundary, growth pattern, degree of enhancement, adjacent organ invasion, distant organ metastasis, and peritoneal seeding, with p values of 0.277, 0.153, 0.672, 1.000, 0.159, and 0.877, respectively. CONCLUSION. CT can be useful in the classification of GE-NENs. N euroendocrine neoplasms (NENs) can originate from anywhere in the body but most are found in the gastrointestinal tract. Gastroenteric neuroendocrine neoplasms (GE-NENs) account for approximately 55 75% of neuroendocrine tumors (NETs) in the body [1]. The incidence of this gastrointestinal malignancy (i.e., GE-NENs) comes second after colorectal carcinoma and has increased fivefold over the past 30 years [2]. The World Health Organization (WHO) [3] classifies both GE-NENs and pancreatic NENs as one single group of tumors: gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs); however, because of the significant differences in imaging characteristics between GE-NENs and pancreatic NENs, the current study focused on only GE-NENs. Currently, the WHO 2010 classification system [3] is the standard for the diagnosis of GE-NENs. According to the system, GE-NENs are divided into three grades on the basis of mitotic count and Ki-67 index and into welldifferentiated type or poorly differentiated type on the basis of histopathologic evalua- tion. Well-differentiated NENs are also known as NETs, whereas NENs that are poorly differentiated are termed neuroendocrine carcinomas (NECs) [3]. To our knowledge, the correlation between the CT features and pathologic grading of GE-NENs has not been investigated. In this study, we aimed to identify CT findings that assist in the distinction between NETs and NECs and to explore the role of the characteristic CT features in the preoperative assessment for the grading of GE-NENs. Materials and Methods Patients This study was conducted in accordance with ethical guidelines for human research and was compliant with the HIPAA. As such, this study received institutional review board and ethical committee approval, and written informed consent was obtained from all patients. The study population included all patients who had pathologically confirmed GE-NENs and had undergone preoperative contrast-enhanced CT at our institute from November 2009 through March Their medical records and images were retrospectively retrieved for analysis. W260 AJR:203, September 2014

2 TABLE 1: Classification of Gastroenteropancreatic Neuroendocrine Neoplasms (GEP-NENs) According to World Health Organization (WHO) 2010 Classification [3] GEP-NEN Tumor NET G1 G2 NEC G3 MANEC Pseudotumor lesions a Pathologic Classification Standards The WHO 2010 classification system [3] has been widely adopted for the pathologic grading of GEP-NENs (Table 1). This method is not recommended for use with carcinoids because carcinoids are relatively benign, whereas long-term follow-up studies have shown that GEP-NENs are malignant tumors [3] and that the degree of differentiation of tumor cells is correlated with the probability of metastasis and prognosis [4, 5]. CT Protocols All patients underwent cleansing enemas the night before CT and fasted for at least 6 8 hours. To achieve adequate bowel distention, patients drank L of 2.5% mannitol 1 hour before and L at 45, 30, and 15 minutes before CT examinations. The rectum was also distended with mannitol. All the patients were scanned using Mitotic count < 2 per 10 HPF, 2% Ki-67 index, or both Mitotic count 2 20 per 10 HPF, 3 20% Ki-67 index, or both WHO Definition Mitotic count > 20 per 10 HPF, > 20% Ki-67 index, or both With glandular epithelial and neuroendocrine cell phenotype, which account for at least 30% of tumor cell, occasionally with squamous cell carcinoma Hyperplastic and preneoplastic lesions Note NET = neuroendocrine tumor, HPF = high-power field, NEC = neuroendocrine carcinoma, MANEC = mixed adenoneuroendocrine carcinoma. a Pseudotumor lesions are not true tumors and therefore were not included in this study. a 64-MDCT unit (Aquilion 64, Toshiba Medical Systems). The examinations were performed from the dome of the diaphragm to the level of the iliac crest before and after ml of iodinated contrast medium (iopromide [Ultravist 300, Schering]) was IV administered at a rate of 3 4 ml/s. Contrast-enhanced images were obtained during the arterial phase (33 35 seconds after initiation of injection) and portal venous phase (60 65 seconds after initiation of injection). The following imaging acquisition parameters were used: 120 kv; mas; collimation, mm; slice thickness, 0.5 mm; slice increments, 0.5 mm; and pitch, 0.9. All data were transferred to a workstation (Vitrea 2, Toshiba Medical Systems) for the generation of multiplanar reformations, maximum intensity projections, and curved planar reformations and for volume rendering. Two different window settings a 200-HU window width with 50-HU window level and a 500-HU window width with 50-HU window level were used to display the intestinal wall and surrounding tissues. Image Analysis Two abdominal radiologists with 12 and 7 years of experience in abdominal imaging independently reviewed the original and reconstructed CT images and recorded the following features: primary tumor location, lesion size, boundary, growth pattern, transmural invasion, intratumoral areas of cystic change or necrosis, ulceration, enhancement pattern, adjacent organ invasion, mesenteric fat infiltration, lymphadenopathy, peritoneal seeding, and distant organ metastasis. Reader discrepancies, if any, were resolved by consensus after reevaluation of the images together. Fig year-old man who presented with loss of appetite; diagnosis was gastric neuroendocrine tumor (G1). Contrast-enhanced axial CT image shows small round mass (arrow) toward interior lesser curve. The boundary between mass and surrounding tissue is clear. Fig year-old man who presented with stomachache; diagnosis was gastric neuroendocrine carcinoma (G3). Contrast-enhanced axial CT image shows inhomogeneous enhanced lobulated mass (arrow) in cardia that grew toward both interior and exterior of cavity. Boundary between mass and surrounding tissue is unclear. AJR:203, September 2014 W261

3 CT Features The CT features that were evaluated in this study are listed in the following sections; all were evaluated on portal venous phase images. Lesion size The maximal diameter in the axial plane was recorded for masslike tumors; the length of the involved digestive tract was recorded for tumors that grew infiltratively. Boundary The boundary between the tumor and surrounding tissue was characterized as well defined or ill defined. It was well defined if there was a clear boundary between the tumor and surrounding tissue; otherwise, the boundary was considered ill defined. Growth pattern The growth pattern was described as either infiltrative type or circumscribed type. The infiltrative type was characterized by growth along the digestive tract wall without a definable lesion boundary, whereas the circumscribed type was defined as a readily identifiable mass that showed a well-demarcated boundary with the surrounding structures. The infiltrative tumors were further divided into tumors showing circumferential growth or eccentric growth. Circumferential growth meant growth involving the whole 360 of the intestine, and eccentric growth meant growth involving less than 360 of the intestine. The circumscribed tumors were further divided into tumors showing intraluminal, extraluminal, or both intra- and extraluminal growth based on the border convexity of the mass. A mass was considered to be intraluminal or extraluminal if it was convex on either the inner or outer surface of the intestinal wall, respectively. When the lesion produced convexity on both the inner and outer surfaces of the intestinal wall, it Feng et al. TABLE 2: General Information About the Patients With Gastroenteric Neuroendocrine Neoplasms Characteristic NETs (n = 25) NECs (n = 31) Total (n = 56) Sex (no. [%] of patients) Male (73.2) Female (26.8) Age (y) Mean ± SD ± ± ± Original tumor locationa (no. [%] of patients) Esophagus (14.3) Stomach (37.5) Duodenum (8.9) Jejunum (3.6) Ileum (1.8) Appendix (1.8) Colon (1.8) Rectum (30.4) Note NETs = neuroendocrine tumors, NECs = neuroendocrine carcinomas. a Because the sample sizes of duodenum, jejunum, ileum, appendix, and colon were too small for statistical analysis (T < 5), a 2 3 chi-square test was performed for the tumors in the esophagus, stomach, and rectum (χ 2 = ; p = 0.003, which is < 0.05). For the Bonferroni correction, α = 0.05 / [3 (3 1) / 2 + 1] = The results after Bonferroni correction were as follows: esophagus and stomach (p = 0.647), esophagus and rectum (p = 0.011), stomach and rectum (p = 0.008). was considered as showing intraluminal growth and extraluminal growth. Transmural invasion For assessment of transmural invasion, tumors were divided into groups that showed partial-thickness involvement or full-thickness involvement. If the outer edge of the involved gastrointestinal wall was smooth with clear boundaries with the surrounding tissues, it was considered partialthickness involvement; otherwise, it was considered full-thickness involvement (i.e., transmural involvement). Areas of cystic change or necrosis A lesion was considered to show areas of cystic change or necrosis if it contained a nonenhancing waterdensity area. Ulceration Ulceration manifested as abnormal morphology and discontinuity of the gastrointestinal mucosa, which may be associated with a local depression. Fig year-old woman who presented with upper abdominal pain; diagnosis was gastric neuroendocrine tumor (G1). Contrast-enhanced axial CT image shows partly thickened stomach wall (arrow) and no mass. Gastric serosa is smooth and there are no enlarged lymph nodes in area. Fig year-old woman who presented with upper abdominal pain; diagnosis was gastric neuroendocrine carcinoma (G3). Contrast-enhanced axial CT image shows irregular thickening of gastric cardia with narrow and deep ulceration (arrow). W262 AJR:203, September 2014

4 Degree of enhancement The degree of enhancement, which was measured as the change in Neuroendocrine Neoplasms and Pathologic Grade TABLE 3: Relationship Between CT Features of Gastroenteric attenuation on portal venous phase images relative to that on unenhanced images, was assessed. Histopathologic Diagnosis If the increase in the CT attenuation value was between 10 and 40 HU, it was considered mildmoderate enhancement; an increase in the CT at- No. of NETs No. of NECs % CT Features (n = 25) (n = 31) of NECs χ 2 p tenuation value of more than 40 HU was considered Tumor size marked enhancement. 4.0 cm Invasion of surrounding structures Invasion > 4.0 cm of surrounding structures, including invasion of Boundary between tumor and surrounding tissue mesenteric fat and adjacent organs, was assessed. Adjacent mesenteric fat infiltration is manifested as increased mesenteric fat density around the tumor Well defined Ill defined or a strip of soft-tissue density near the tumor. If Growth pattern the boundary of a tumor and adjacent organs was ill Infiltration defined, it was defined as adjacent organ invasion. Lymphadenopathy A nodular soft-tissue lesion Circumscribed larger than 10 mm in short-axis diameter was Convexity of circumscribed tumor a defined as regional lymphadenopathy. Intra- or extraluminal growth Peritoneal seeding If the metastatic mass was Intra- and extraluminal growth located in the peritoneum or mesenteric region and was not adjacent to the main mass, we considered Transmural invasion it to be peritoneal or mesenteric seeding. Present Distant organ metastasis Distant organ metastasis was defined as metastasis to organs other Absent Circumferential growth than the site of the primary lesion (e.g., liver, lungs). Eccentric Statistical Analysis Circumferential Statistics software (SPSS, version 16.0, IBM Areas of cystic change or necrosis Software) was used for statistical analysis. The age distributions of the two groups were subjected to Present independent-samples Student t tests, and all the remaining Absent differences with regard to CT features were Ulceration compared using the chi-square test or Fisher exact Present test. The value of chi-square reflects the difference between the observed frequency and expected frequency. Absent If the null hypothesis is true, the difference Degree of enhancement between the observed frequency and expected frequency Mild-moderate is small and there would be a large probabil- Marked ity of a small chi-square value (p > α) and vice versa [6]. A difference with a p value of < 0.05 (α = 0.05) Mesenteric fat infiltration was considered a statistically significant difference. Present Results Pathologic Grading There were 59 pathologically proven GE-NENs from November 2009 through March One of the 59 patients had undergone CT at another center, but the image quality was deemed insufficient for analysis in our investigation. Two other patients underwent CT in our hospital, but CT showed only slight thickening of the gastric and rectal mucosa, respectively, that was difficult to differentiate from normal mucosa. The radiologists did not identify these two lesions; hence, these two patients were also excluded from the study. Therefore, 56 cases of GE-NENs were included in the study: 17 cases of G1 tumors, eight cases of G2 tumors, 27 cases of G3 tumors, and four cases of mixed adenoneuro- Absent Adjacent organ invasion Present Absent Lymphadenopathy < Present Absent Distant organ metastases Present Absent Peritoneal seeding Present Absent Note Dash ( ) indicates there was no chi-square value when the Fisher exact test was used.. NETs = neuroendocrine tumors, NECs = neuroendocrine carcinomas. a The sample size was < 40, so the Fisher exact test was used. AJR:203, September 2014 W263

5 Feng et al. endocrine carcinoma (MANEC). In total, 25 cases were NETs (i.e., G1 or G2) and 31 cases were NECs (i.e., G3 or MANEC). Correlation of Pathologic Grade of Gastroenteric Neuroendocrine Neoplasms With Patient Sex and Age, Tumor Location, and Clinical Manifestations General information about all of the patients with GE-NENs is shown in Table 2. The 56 patients with GE-NENs included 41 men and 15 women between 33 and 80 years old, with a median age of 61 years. There were no statistically significant differences between the two groups of patients in terms of sex (χ 2 = 0.007, p = 0.932) and age (T = 1.697, p = 0.731). The origins of the tumors included the esophagus (n = 8); stom- Fig yearold woman who presented with tarry stool; diagnosis was gastric mixed adenoneuroendocrine carcinoma with peritoneal seeding. Contrast-enhanced axial CT image shows thickening of gastric wall (single arrow) and adjacent blurry fat plane. Abundant ascites (double arrows) is also seen. C Fig year-old man who presented with changing of bowel habits; diagnosis was rectal neuroendocrine carcinoma (G3) with liver and bone metastases. A, Contrast-enhanced axial CT image shows irregular lobulated mass (arrow) in rectum. B, CT image shows multiple low-density metastases in liver. C, CT image shows left fourth rib is distroyed. Adjacent soft-tissue mass (arrow) is seen. D, Photomicrograph (H and E stain, 400) of specimen shows high cellularity, cells that are similar size, and many mitotic figures. A B D W264 AJR:203, September 2014

6 A Fig year-old man who presented with hematochezia; diagnosis was rectal neuroendocrine carcinoma (G3) with bilateral adrenal metastases. A, Contrast-enhanced axial CT image shows circumferential thickening of rectal wall (arrow). B, CT image reveals bilateral soft-tissue masses (arrows) in adrenals, which are showing heterogeneous enhancement. ach (n = 21), including lesions in the gastric cardia (n = 10) and gastric corpus (n = 11); duodenum (n = 5); jejunum (n = 2); ileum (n = 1); appendix (n = 1); colon (n = 1); and rectum (n = 17). Multiple chi-square tests and Bonferroni corrections were used for the comparison of tumor locations. The result suggested that tumor location was significantly heterogeneous between the esophagus and rectum (p = 0.011) and between the stomach and rectum (p = 0.008). The clinical symptoms included abdominal pain (n = 14), melena and bloody stool (n = 13), dysphagia or pain swallowing (n = 7), change in bowel habits (n = 5), jaundice (n = 4), retrosternal chest pain (n = 4), weight loss (n = 3), fever (n = 1), and vomiting (n = 1). Four patients had no obvious symptoms and their lesions were detected incidentally. Relationship Between the CT Features of Gastroenteric Neuroendocrine Neoplasms and Pathologic Grade The chi-square test and Fisher exact test were used for the comparison of various CT features between the two groups (Table 3). Statistical analysis revealed that there were statistically significant differences between NETs and NECs in terms of tumor size (Figs. 1 and 2), transmural invasion, circumferential growth (Fig. 3), intra- and extraluminal involvement, areas of cystic change or necrosis, ulceration (Fig. 4), mesenteric fat infiltration, and lymphadenopathy. All of these variables carry disagreement of more than 20% between NETs and NECs. The following CT features showed no statistically significant difference between the two groups: tumor boundary, growth pattern, degree of enhancement, adjacent organ invasion, distant organ metastasis, and peritoneal seeding. Six patients had peritoneal seeding, four of whom had a primary tumor in the stomach (Fig. 5), including one G2 tumor, one gastric MANEC, and two G3 gastric NECs. None of the G1 tumors revealed peritoneal seeding. Sixteen patients had metastasis restricted to the liver, one had metastasis to the liver and bone (Fig. 6), one had bilateral adrenal metastases (Fig. 7), and one had metastases to the lung. Fig year-old man with bloody stool; diagnosis was rectal neuroendocrine carcinoma (G3). Contrast-enhanced axial CT image shows eccentric lobulated heterogeneous enhanced mass in rectal cavity. Regional wall is slightly thickened at one side of lump (arrow), forming rat tail sign. Relationship Between Unusual CT Features and Pathologic Grade Of all 56 GE-NENs, we observed four cases of NECs growing along the wall and visible as a small local nodular protrusion in the thickened wall, producing a rat tail sign appearance (Fig. 8), and five cases of NECs appearing as a local convex shape of a cusp formed by the growth of the tumor toward the cavity, producing a sharp corner sign appearance (Fig. 9). Discussion GE-NENs are a heterogeneous group of tumors that originate in the neuroendocrine system of the gastrointestinal tract [7]. The incidence of GE-NENs is approximately 5 cases per 100,000 individuals per year and the prevalence is about 35 cases per 100,000 individuals [1]. Almost 80% of GE-NENs are nonfunctional without secretory clinical syndrome [8], which makes the diagnosis of GE-NENs more difficult. Thus, it is important to be able to diagnose or even to classify GE-NENs using preoperative CT because the pathologic classification is the most important prognostic factor for patients with these tumors [1, 9]. In this study, several CT findings were shown to be helpful in the classification of GE-NENs. CT findings of lesion size larger than 4.0 cm, transmural invasion, circumferential B AJR:203, September 2014 W265

7 growth, intra- and extraluminal involvement, areas of cystic change or necrosis, ulceration, mesenteric fat infiltration, and lymphadenopathy were significantly more frequent in NECs. To our knowledge, this study is the first attempt to use CT findings to distinguish between NETs and NECs. The size of tumor was significantly different between NETs and NECs (p = 0.044). This finding suggests that the larger the tumor, the more likely it is an NEC. It was surprising that none of our patients had symptoms of intestinal obstruction, which might or might not be related to the soft consistency of these tumors. In this study, 42 (75%) tumors showed ill-defined boundaries, probably related to the fact that the majority of GE-NENs (66.1%) grew infiltratively along the gastrointestinal tract. The growth pattern of GE-NENs tended to be circumferential (p = 0.008) with transmural invasion (p = 0.002) in the NEC group. Some tumors formed a mass that showed both intra- and extraluminal growth. There were eight cases showing this feature in our study, six of which were NECs. The difference between NETs and NECs was statistically significantly (p = 0.024), suggesting that NECs tended to produce a bulk-forming mass. We found that 38 (67.9%) tumors showed areas of cystic change or necrosis; these findings were more likely to occur in NECs than in NETs (p = 0.002). In this patient population, ulceration was seen on the luminal surface of the tumor in 21 (37.5%) patients and was more common in NECs than NETs (p = 0.007). Tumor invasion into surrounding tissues included mesenteric fat infiltration and adjacent organ invasion; the percentage of patients with these changes based on CT appearances was 58.9% and 35.7%, respectively. NECs of this series were more likely to have adjacent mesenteric fat infiltration (p = 0.002) than NETs. Theoretically, NECs are more prone to distant metastasis and peritoneal seeding than NETs, but our findings suggested that these two features were not significantly different between the two groups. Perhaps this result is related to the relatively small sample size. Future larger-scale prospective studies may help to clarify this issue. GE-NENs can occur in any part of the gastrointestinal tract. In our study, the stomach (37.5%) and rectum (30.4%) were the first and second most common origins, respectively, followed by the esophagus (14.3%) and duodenum (8.9%). In the past, it was believed that most GE-NENs Fig year-old man with black stool; Feng et al. diagnosis was duodenal neuroendocrine carcinoma (G3). Contrast-enhanced axial CT image shows uneven thickening of descending duodenum wall and cusplike soft tissue protruding in widened intestinal lumen (single arrow). Boundary of thickening with adjacent fat space (double arrows) is blurred. occurred in the appendix and ileum. Yao et al. [1] previously reported that most GE- NENs arise from the rectum. Recently, Niederle et al. [10] found that most GE- NENs occurred in the stomach. It is questionable if these different findings could be related to variations in study cohorts related to ethnicity or geographic region. We also found that the different types of tumor could be distributed in different locations: Tumors in the esophagus (p = 0.011) and stomach (p = 0.008) are more likely to be NECs. This finding is consistent with the findings of Maru et al. [11] who also reported that NENs in the esophagus were more likely to be malignant. Our study has several limitations. First, this study is a retrospective review and the number of patients is relatively small. Nevertheless, this limitation might be inevitable in a single-center study because the disease is quite rare. Second, NECs outnumbered NETs in our study 31 (55.4%) to 25 (44.6%), which is contrary to the expected frequencies of NECs and NETs. We suspect this difference could be related to the fact that our hospital is a tertiary referral center for patients from multiple centers in different cities. Third, most of the patients in our hospital come from southern China, which may be the reason for the different distribution of tumor locations compared with some earlier reports [1, 2]. We believe that some of these issues could be clarified in a future multicenter study involving a larger number of patients. In this study, CT was used to accurately locate GE-NENs and provide preoperative information about the relationship of the GE-NEN to adjacent structures. This study showed that some CT features including large tumor size (> 4 cm), transmural invasion, intra- and extraluminal involvement, circumferential growth, areas of cystic change or necrosis, ulceration, mesenteric fat infiltration, and lymphadenopathy favor NEC over NET. References 1. Yao JC, Hassan M, Phan A, et al. One hundred years after carcinoid : epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. J Clin Oncol 2008; 26: Modlin IM, Oberg K, Chung DC, et al. Gastroenteropancreatic neuroendocrine tumours. Lancet Oncol 2008; 9: Rindi G, Arnold R, Bosman FT, et al. Nomenclature and classification of neuroendocrine neoplasms of the digestive system. In: Bosman FT, Carneiro F, Hruban RH, Theise ND, eds. WHO classification of tumours of the digestive system, vol. 3, 4th ed. Lyon, France: IARC Press, 2010: Dolcetta-Capuzzo A, Villa V, Albarello L, et al. Gastroenteric neuroendocrine neoplasms classification: comparison of prognostic models. Cancer 2013; 119: Pape UF, Jann H, Muller-Nordhorn J, et al. Prognostic relevance of a novel TNM classification system for upper gastroenteropancreatic neuroendocrine tumors. Cancer 2008; 113: Goonatilake R, Herath A, Herath S, et al. Intrusion detection using the chi-square goodness-of-fit test for information assurance, network, forensics and software security. J Comput Sci Coll 2007; 23: Starker L, Carling T. Molecular genetics of gastroenteropancreatic neuroendocrine tumors. Curr Opin Oncol 2008; 21: Dierdorf SF. Carcinoid tumor and carcinoid syndrome. Curr Opin Anaesthesiol 2003; 16: Sundin A, Vullierme MP, Kaltsas G, et al. ENETS consensus guidelines for the standards of care in neuroendocrine tumors: radiological examinations. Neuroendocrinology 2009; 90: Niederle MB, Hackl M, Kaserer K, et al. Gastroenteropancreatic neuroendocrine tumours: the current incidence and staging based on the WHO and European Neuroendocrine Tumour Society classification an analysis based on prospectively collected parameters. Endocr Relat Cancer 2010; 17: Maru DM, Khurana H, Rashid A, et al. Retrospective study of clinico-pathologic features and prognosis of high-grade neuroendocrine carcinoma of the esophagus. Am J Surg Pathol 2008; 32: W266 AJR:203, September 2014