Carcinoids of the pancreas

An analysis of 156 cases


  • Jun Soga M.D.

    Corresponding author
    1. Professor Emeritus, Niigata University, Niigata City, Japan
    • Masago 1-21-26, Niigata City, 950-2074 Niigata, Japan
    Search for more papers by this author
    • Tel/Fax: (011) 81-25-266-9230

    • Jun Soga was formerly a professor of surgery, Niigata Seiryo University, Japan, and is now a professor emeritus, Niigata University, Niigata City, Japan.



The aim of the current study was to clarify the actual clinicopathologic status of extremely rare pancreatic carcinoids. To date, statistical evaluation of such carcinoids has been hampered because an insufficient number of cases has prevented any reliable statistical analyses.


The Niigata Registry for Gut–Pancreatic Endocrinomas contains a total number of 156 cases of pancreatic carcinoids among 165 endocrinocarcinomas registered worldwide. This figure of 156 cases comprises 144 typical and 12 atypical carcinoids, which were compared statistically with carcinoids in other representative sites, according to various clinicopathologic criteria.


Pancreatic carcinoids made up 1.4% of the total number of registered cases (n = 11,343) and were characteristic in the following five ways. 1) They exhibited a high metastatic rate (66.7%), somewhat lower than that for the ileocecum (76.1%), identical to that for the ileum (66.9%), and far higher than that for the total average of 35.7%. 2) They displayed the largest average tumor size (68.6 mm), followed by that for the ovary (68.2 mm), and ileocecum (46.5 mm) against a total average of 30.2 mm. 3) They revealed a relatively high incidence of the carcinoid syndrome (23.3%), almost equal to that for the ileocecum (24.1%), and exceeded by that for the small intestine (35.8%), when compared with that for the total average of 11.0%. 4) They showed a high rate of immunohistochemical detection for serotonin (92.9%), lower than that for the testicle (100.0%), but higher than the total average of 54.9%. 5) Five-year survival rate was extremely low (28.9% ± 16.7%) compared with those for the appendix (89.7% ± 2.0%) and the small intestine (82.1% ± 3.3%). It was noteworthy that silver impregnations in the pancreatic carcinoid series indicated a result identical to that for the total average: Grimelius argyrophilia, 84.8% versus 85.4%; argyrophil cell type, 59.1% versus 58.5%; and argentaffin cell type, 22.7% versus 22.3%.


It may be said that in the treatment of patients with pancreatic carcinoids, appropriate procedures should be carried out with these distinguishing characteristics always kept in mind. Cancer 2005. © 2005 American Cancer Society.

In 1931, Arnett and Long1 described the first case of a pancreatic carcinoid as a carcinoma. Subsequent research carried out by the Swedish group, lead by Thorson et al.2 in 1954 identified it as a probable malignant carcinoid associated with a new syndrome that they were attempting to define. The so-called carcinoid syndrome was applied to cases of carcinoids in the small intestine with hepatic metastases and common characteristic symptoms.

Very few reported cases of pancreatic carcinoids figure in the available literature. Of the 12,540 cases of endocrinocarcinomas3–5 stored in the Niigata Registry for Gut–Pancreatic Endocrinomas,4 11,343 belong to the carcinoid group, excluding the 1197 that belong to the variant group. Pancreatic carcinoids, conversely, compose a mere 1.4% of the entire carcinoid group (156/11,343) and 2.2% of the gastrointestinal (GI) carcinoid group (156/7240). Similarly, in 2003 Modlin et al.6 published findings on a large series of 13,715 carcinoids registered with the National Cancer Institute (NCI) in the United States supplied by the End Result Group (ERG) and the Third National Cancer Survey (TNCS). In their study of these cases and the earlier and later findings contained in the Surveillance, Epidemiology, and End Results (SEER), Modlin et al. reported that pancreatic carcinoids accounted for only 0.58% of the entire series (79/13,715).

In the current study, the relatively small series of pancreatic carcinoids under investigation was evaluated statistically according to certain clinicopathologic criteria such as age and gender distribution, metastatic rates, the responses to silver impregnations, immunohistochemical reactions, five-year survival rates, and so on, with a view to discovering the possible link between these criteria and carcinoid cases of high or low incidence found in other representative sites.

The term “carcinoid” in this article is consistent with its use in an earlier series of articles.3–5, 7–9


Carcinoid groupings and their respective endocrinomas (both the carcinoid group and the variant group, including neuroendocrine cell carcinoma [NEC], endocrine cell carcinoma [ECC], small cell carcinoma [SCC], etc.) as well as diagnostic criteria with histologic growth patterns3, 5, 7, 8 are indicated in Figure 1. The term “endocrinocarcinoma” is proposed tentatively to cover both the carcinoid group and the variant group (Fig. 1).3, 5, 7–12

Figure 1.

Pancreatic endocrinocarcinomas: Number of cases in relation to each category. It should be noted that only small (oat) cell carcinomas with clear evidence of endocrine features are included. SCC: small cell carcinoma; OCC: oat cell carcinoma; NEC: neuroendocrine carcinoma without referring to “carcinoid.”; ECC: endocrine cell carcinoma. *With a definite evidence for endocrine features; small cell neuroendocrine carcinoma.

Of the 12,540 reported cases of endocrinocarcinomas, registered in the Niigata Registry,4 165 were recorded as primary endocrinocarcinomas of the pancreas (Table 1). Of these, 156 cases were confirmed as belonging to the carcinoid group, consisting of 144 typical and 12 atypical carcinoids.

Table 1. Distribution of Endocrinocarcinomas (A)a versus Carcinoid Group (B)
Sites of originNo. cases A/B% Digestive A/B% Total A/B
  • Data from the Niigata Registry, October 1, 2004.

  • a

    Consisting of the carcinoid group and their variants.

Small intestine1377/136717.5/18.911.0/12.1
Not specified61/570.8/0.80.5/0.5
Digestive system7859/7240100.0/100.062.7/63.8
Respiratory tract2398/226319.1/20.0
Uterine cervix222/861.8/0.8
Extradigestive system4681/410337.3/36.2

All 156 cases were evaluated statistically based on analyses of the male–female ratio, average age and gender distribution, signs and symptoms, metastatic rates and tumor sizes, the responses to silver impregnations, immunohistochemistry, and five-year survival rates. Comparisons with other representative sites, exhibiting contrasting data, were attempted in the majority of cases.

The Kaplan–Meier method was used in the statistical calculation of cumulative five-year survival rates (5YSRs), and the chi-square test for routine analysis was used to determine statistical significance, in conjunction with the Student t test or Welch t test for age and tumor-size distribution and a generalized version of the Wilcoxon test for postoperative survival rates as previously mentioned.3, 5


Distribution of Endocrinocarcinomas (A) and Carcinoid Group Excluding Endocrine Variant Types (B) (Fig. 1; Table 1)

Of the total 165 pancreatic endocrinocarcinomas, 156 were neoplasms of carcinoid group and 9 were variant endocrine types, the former consisting of 144 typical and 12 atypical carcinoids. Detailed comparative data of incidence is shown in Table 1, indicating that pancreatic carcinoids were 2.2% (156/7240) of the total in the digestive system and 1.4% (156/11,343) of the entire carcinoid system.

Male–Female Ratio in Comparison with Other Sites (Table 2)

Among patients with pancreatic carcinoids, the male–female ratio was 0.9 (72/84), exhibiting a significant statistical difference (P < 0.05) from the total average of 1.1. The higher ratio group included carcinoids of the esophagus (3.4), the larynx (3.2), and the small intestine (1.5), indicating a significant statistical difference when compared with the pancreatic carcinoid series (P < 0.01). Conversely, the lower ratio group included carcinoids of the appendix and gallbladder (0.6 each) as well as the ileocecum and liver (0.7 each), exhibiting no significant statistical difference when compared with the pancreatic carcinoid series.

Table 2. Male to Female Ratio
SitesNo. CasesMFM–F ratio
  • a versus b: P < 0.01.

  • a versus c: not significant.

  • d

    Excludes sites of origin with gender preponderance.

  • a versus e: P < 0.05.

Small intestine13678205471.5b

Average Age and Gender Distribution Compared with Other Sites (Table 3)

The average age of pancreatic carcinoid patients was calculated as 48.9 years, exhibiting a significant statistical difference when compared with that for the total average of 51.6 years (P < 0.05). The higher age group included patients with carcinoids of the gallbladder (62.1 yrs), of the larynx (61.3 yrs), and of the ileocecum (60.4 yrs), whereas the lower age group included those with carcinoids of the appendix (37.2 yrs) and the middle ear (39.7 yrs). Both higher and lower age groups revealed a significant statistical difference (P < 0.01) in average ages compared with patients with pancreatic carcinoids.

Table 3. Age and Gender: Average, Distribution, and Range
  • a versus b: P < 0.01.

  • c

    Total number of cases in the carcinoid group excluding sites of origin with sex preponderance.

  • a versus d: P < 0.05.

Middle ear39.7b7537.34342.93214–72

The Carcinoid Syndrome and Other Signs and Symptoms (Table 4)

In the pancreatic carcinoid series, the carcinoid syndrome was calculated as 23.3% (34/146) compared with 35.8% (406/1,134; P < 0.01) in the small intestine and 24.1% (20/83: without significant statistical difference) in the ileocecal region. Overall incidence of the carcinoid syndrome in the digestive system was calculated as 11.0% (636/5778; P < 0.01) whereas that of the entire system was calculated as 10.5% (953/9044; P < 0.01). Included in the low-incidence group of the carcinoid syndrome were the rectum (1.0%;11/1118), the choledochus (1.5%; 1/68), the appendix (2.1%; 17/803), and the duodenum (3.1%;26/852), all showing a significant statistical difference (P < 0.01) when compared with the incidence of the carcinoid syndrome in the pancreatic series.

Table 4. Signs and Symptoms
Sites (no.)Carcinoid syndrome no. (%)Asymptomatic no. (%)Abdominal pain no. (%)Diarrhea no. (%)Abdominal mass no. (%)Flush no. (%)Hepatomegaly no. (%)Jaundice no. (%)Anemia no. (%)
  • a versus b: P < 0.01.

  • a versus c: not significant.

  • d

    Digestive system only. In the extradigestive system, the ovarian carcinoid series indicated the highest incidence of carcinoid syndrome (22.5%; 62/275). The overall incidence of carcinoid syndrome, including both digestive and extradigestive system, was 10.5% (953/9044) versus 23.3% in the pancreatic carcinoid series (P < 0.01).

Pancreas (146)34 (23.3)a10 (6.8)59 (40.4)45 (30.8)32 (21.9)29 (19.9)22 (15.1)18 (12.3)10 (6.8)
Small intestine (1134)406 (35.8)b105 (9.3)540 (47.6)586 (42.9)170 (15.0)338 (29.8)171 (5.10)20 (1.8)127 (11.2)
Ileocecum (83)20 (24.1)c0 (0.0)44 (53.0)36 (43.4)31 (37.3)17 (20.5)13 (15.7)0 (0.0)12 (14.5)
Duodenum 852)26 (3.1)b158 (18.5)342 (40.1)63 (7.4)31 (3.6)20 (2.3)41 (4.8)138 (16.2)91 (10.7)
Appendix (803)17 (2.1)b24 (3.0)685 (85.3)50 (6.2)82 (10.2)18 (2.2)16 (2.0)5 (0.6)11 (1.4)
Choledochus (68)1 (1.5)b2 (2.9)28 (41.2)4 (5.9)3 (4.4)1 (1.5)7 (10.3)47 (69.1)5 (7.4)
Rectum (118)11 (1.0)b215 (19.2)170 (15.2)106 (9.5)61 (5.5)13 (1.2)52 (4.7)11 (1.0)29 (2.6)
Totald (5778)636 (11.0)b653 (11.3)2468 (42.7)960 (16.6)665 (11.5)540 (9.3)464 (8.0)294 (5.1)627 (10.9)

Other signs and symptoms recorded in the pancreatic carcinoid series, in descending order of frequency, were abdominal pain (40.4%), diarrhea (30.8%), abdominal mass (21.9%), flush (19.9%), and hepatomegaly suggesting hepatic metastasis (15.1%).

Carcinoid Sites in the Pancreas (Table 5)

Carcinoids were predominantly located in the head of the pancreas (44.9%; 62/138), followed by those located in the tail (21.0%; 29/138), the body-tail (14.5%; 20/138), and the body (12.3%; 17/138).

Table 5. Sites of Carcinoids in the Pancreas
Head and body75.1
Body and tail2014.5
Head and tail10.7

Metastases and Tumor Size (Table 6)

The metastatic rate in the pancreatic carcinoid series was calculated as 66.7% (104/156) of the overall value, lower than that of the ileocecum (76.1%; 70/92), and the ileum (66.9%; 743/1111) without any significant statistical difference, but significantly higher (P < 0.01) than that of the total series (35.7%; 4045/11,343). In those cases where the tumor size measured ≤ 20 mm, the metastatic rate of the pancreatic carcinoid series was 31.3% (5/16) when compared with 51.2% (155/303; P < 0.01) for the ileum.

Table 6. Rates of Metastases and Average Tumor Size Recorded on Millimeter Scale
Sites (no.)Rates of metastasesAverage tumor size
Overall≤ 20 mmOverallWith metastases
  1. a versus b: P < 0.01.

  2. a versus c: P < 0.05.

  3. a versus d: not significant.

Pancreas (156)66.7a10431.3a5/1668.6a9163.9a50
Ileocecum (92)76.1d7085.7c6/746.5b4844.6c34
Ileum (1111)66.9d74351.2d155/30322.2b45325.7b270
Appendix (1063)16.0b17013.4d31/23116.4b28926.7b60
Ovary (369)13.3b496.3b4/6468.2d26487.3c30
Middle ear (75)2.7b20.0c0/1511.1b16
Total (11,343)35.7b404521.4d821/382930.2645139.9b2046

In the pancreatic carcinoid series, the overall average tumor size was 68.6 mm (n = 91), followed by that of the ovary series (68.2 mm; n = 264), which compares with 30.2 mm (n = 6451) for the entire series (P < 0.001).

Silver Impregnations (Table 7)

Diagnostic staining of carcinoid tissue by silver impregnations was frequently used both for differential diagnosis and when categorizing the cell types of these endocrine neoplasms; argyrophil, argenntaffin, and nonreactive cell types.3, 7 In recent years, the most reliable procedure has been thought to be a combination of the Grimelius method for argyrophilia and the Fontana–Masson method for argentaffinity.

Table 7. Silver Impregnations
SitesGrimelius Argyrophilia % ( no.)Silver cell types3.7
Argyrophil % ( no.)Argentaffin % ( no.)Nonreactive % ( no.)
  1. a versus b: P < 0.01.

  2. a versus c: P < 0.05.

  3. a versus d: not significant.

Pancreas84.8a (39/46)59.1 (13/22)22.7a (5/22)18.2 (4/22)
Testicle100.0c (36/36)5.0 (1/20)95.0b (19/20)0.0 (0/20)
Ileocecum100.0d (17/17)14.3 (2/14)85.7b (12/14)0.0 (0/14)
Gallbladder98.3b (58/59)80.6 (29/36)16.7d (6/36)2.8 (1/36)
Larynx97.0b (96/99)90.2 (37/41)2.4b (1/41)7.3 (3/41)
Ileum96.8b (92/95)15.0 (9/60)81.7b (49/60)3.3 (2/60)
Middle ear75.5d (37/49)55.0 (11/20)5.0d (1/20)40.0 (8/20)
Colon67.9c (38/56)32.3 (10/31)29.0d (9/31)38.7 (12/31)
Rectum66.0b (376/570)48.4 (206/426)16.0d (68/426)35.7 (152/426)
Total85.4d (2643/3094)58.5 (929/1589)22.3d (354/1589)19.3 (306/1589)

Grimelius argyrophilia was exhibited in 84.8% (39/46) of the pancreatic carcinoid series, a value almost identical to that of the total average (85.4%; 2643/3094), but significantly lower (P < 0.01) than the 96.8% for the ileum and that of several other sites, and significantly higher (P < 0.01) than the 66.0% for the rectum and several other sites. Pancreatic carcinoids exhibited a relatively low incidence of argentaffin cell type (22.7%; 5/22), significantly lower (P < 0.01) than that of the ileum (81.7%; 46/60), almost equal to that of the total average (22.3%; 354/1589) and higher than that of the rectum (16.0%;68/426), but without any significant statistical difference.

Immunohistochemical Analysis (Table 8)

Several representative substances, including serotonin, neuron-specific enolase (NSE), and chromogranin were evaluated. A high incidence of serotonin was recorded in the pancreatic carcinoid series (92.9%; 39/42), statistically far greater (P < 0.01) than that for both the stomach (47.9%; 79/165) and the total average (54.9%; 611/1113). In the pancreatic carcinoid series, all 17 cases responded positively to NSE, as did those of the uterine cervix series (13/13), and revealed little or no statistical difference when compared with cases of the stomach (92.4%; 133/144) and with the total average (87.8%; 1012/1153). Chomogranin was also recorded positive in the pancreatic carcinoid series (90.9%; 20/22) with similarly little or no significant statistical difference when set against the rectum (70.9%; 78/110), the testicle (100.0%; 26/26) and the total average (88.2%; 1319/1496).

Table 8. Immunohistochemistry
Sites selectedSerotonin % ( no.)Gastrin % ( no.)Somatostatin % ( no.)ACTH % ( no.)PP % ( no.)NSE % ( no.)Chromogranin % ( no.)
  1. ACTH: adrenocorticotrophic hormone; [elm3]PP: NSE: neuro-specific enolase..

  2. a versus b: P < 0.01.

  3. a versus c: P < 0.05.

  4. a versus d: not significant.

Pancreas92.9%a (39/42)41.7 (10/24)28.0 (7/25)22.2 (2/9)22.2 (4/18)100.0a (17/17)90.9a (20/22)
Small intestine89.1d (82/92)25.0 (11/44)15.1 (8/53)3.1 (1/33)2.9 (1/34)93.4d (57/61)98.7d (76/77)
Appendix86.6d (71/82)15.4 (4/26)16.3 (8/49)8.7 (2/23)9.1 (2/22)74.2c (46/62)94.0d (78/83)
Stomach47.9b (79/165)26.0 (53/204)16.6 (29/175)5.0 (4/80)18.2 (18/99)92.4d (133/144)72.6d (69/95)
Rectum43.8b (60/137)11.1 (8/72)38.2 (55/144)13.3 (2/15)66.7 (80/120)85.0d (8/80)70.9d (78/110)
Testicle100.0d (14/14)33.3 (4/12)0.0 (0/11)0.0 (0/11)0.0 (0/9)95.0d (19/20)100.0d (26/26)
Uterine cervix88.9d (16/18)0.0 (0/2)30.0 (3/10)42.9 (3/7)11.1 (1/9)100.0 (13/13)100.0d (16/16)
Total54.9b (611/1113)26.3 (220/838)31.7 (331/1045)29.2 (190/651)31.8 (108/591)87.8d (1012/1153)88.2d (1319/1496)

Postoperative Outcomes (Fig. 2)

Five-year survival rates (5YSRs) within the pancreatic carcinoid series (effective n = 46) were as low as 28.9 ± 16.7%, when compared with 82.1 ± 3.3% for carcinoids of the small intestine (n = 368) and 89.7 ± 2.0% for those of the appendix (n = 565). A significant statistical difference (P < 0.05) was evident in the comparable values for the pancreas and the appendix but not for the pancreas and the small intestine.

Figure 2.

Five-year survival rate for patients with pancreatic carcinoid series compared with that of appendiceal and small intestinal series. The five-year survival rate in the pancreatic carcinoid series is extremely low compared with that of appendiceal and small intestine series.


Malignancy and the Terminology of Carcinoids

The term “carcinoid”13 has recently been criticized by several authors on the grounds that it is archaic,6, 14 outmoded,15, 16 or even a misnomer.3, 6, 8, 9, 17–19 The term “carcinoid” in this article is consistent with its use in an earlier series of articles.3–5, 7–9, 20–23, 25, 29

It is clear that submucosal lesions of GI carcinoids, often larger than primary lesions in the mucosa, are a sure indication of malignancy that results from an invasion of intramucosal primary lesions into the submucosa.20, 21

Even primary intramucosal lesions of GI microcarcinoids, ≤ 1 mm, or minute carcinoids, ≤ 5 mm, may also prove to be malignant based on histologic evidence of invasion into surrounding tissues and vascular channels.22 Moreover, they provide clear evidence of malignancy in both animal models with complete serial sections20, 21 and human materials,22 whereas small submucosal GI carcinoids of ≤ 10 mm have unexpectedly high metastatic rates calculated as 10.0% (105/1045)3 and 10.6% (114/1074).22

It is noteworthy that when early GI carcinoids and early gastric carcinomas, with invasion confined to the mucosa, are compared for evidence of malignancy, calculation revealed a metastatic rate of 13.1% (49/374) for the former and 22.5% for the latter (49/218; P < 0.01). When the lesions were restricted to a tumor of ≤ 20 mm, however, the metastatic rates proved to be identical, 9.8% versus10.3%.23

Conversely, verification of the benign nature of GI carcinoids, especially in early stages, has proved extremely difficult. Hence, the term “endocrinocarcinoma”,3, 5, 22, 24, 25 which covers both the carcinoid group and variant group (Fig. 1), was proposed to obtain further understanding of the background of malignancy of GI carcinoids.


In the pancreas, which does not have a layered structure, consideration should be given to the possibility of an identical malignant nature, one aspect of which may be represented by metastatic rates. The metastatic rate in the pancreatic carcinoid series was significantly higher (66.7%;104/156) than that of the overall total average of 35.7% (4045/11,343; P < 0.01) and somewhat lower, though statistically identical, than that in the ileum (76.1%) and the ileocecum (66.9%) in the high-metastasis group. To obtain a more detailed evaluation of metastases, the calculation of metastatic rates was based on controlled use of tumors of ≤ 20 mm (Table 6). This, however, resulted in statistically insignificant P values between metastatic rates for the pancreatic carcinoid series (31.3%; 5/16), the ileum (51.2%; 155/303), and the total average (21.4%; 821/3829).

Tumor size

The average tumor size of pancreatic carcinoids (68.6 mm) was largest of all, almost equal to that of the ovary (68.2 mm), far greater than the ileocecum (46.5 mm), and even the total average (30.2 mm). This suggests that pancreatic tumors may be difficult to detect in earlier stages without aid of image diagnosis such as ultrasonography, computed tomography, and magnetic resonance imaging, which are apparently useful in early detection of tumor nodules in the pancreas.26–28

Carcinoid syndrome

In the 1961 publication of a large series of 138 cases containing typical carcinoid syndrome, no detailed description of carcinoid syndrome in patients with pancreatic carcinoids was included in the report.12 However, a recently published evaluation of 748 cases with carcinoid syndrome (8.4%; 748/8876) includes 19 cases with this syndrome in the pancreas (17.8%; 19/107), exhibiting a significant statistical difference (P < 0.01) between both series.29 To date, numerous articles dealing with this particular syndrome have appeared,2, 10, 29–31 and it has now been confirmed that the first case of a pancreatic carcinoid is associated with this syndrome.1

Several unique and distinctive cases of pancreatic carcinoids exist. One such case is that of a 54-year old female with carcinoid syndrome who survived for > 5 years after a laparotomy, performed without removal of a lesion measuring 58 mm in diameter and accompanied by both hepatic and nodal metastases. Successful treatment with somatostatin analogues resulted in satisfactory prognosis and disappearance of the syndrome.32 A further case concerns a 55-year old man with carcinoid syndrome who survived 4 years and 11 months after a left pancreatectomy caused the syndrome to disappear. A recurrence of the lesion found in his liver with carcinoid syndrome after 4 years and 1 month was successfully treated by a liver transplant, but the patient died 10 months later of hepatic and bone metastases.33

Silver impregnations

It is worth noting that all Grimelius argyrophilia, argyrophil cell type and argentaffin cell type indicated a ratio (%) that was virtually identical to that of the total average (84.8% vs. 85.4%; 59.1% vs. 58.5%; 22.7% vs. 22.3%).


Immunohistochemical analyses revealed that carcinoids in both the pancreas and other representative sites were multisecretory. In pancreatic carcinoids, however, serotonin (5-HT) was far more dominant than other active hormonal substances, and it was suggested that neither gastrin nor somatostatin should be overlooked in immunohistochemical analyses.

At 92.9% and 100.0%, respectively, serotonin and NSE exhibited a high rate of immunohistochemical positivity in the carcinoid group. Although the former was significantly higher (P < 0.01%) than the total average (54.9%), the latter was not significantly different from the total average (87.8%). Both serotonin and NSE are without doubt necessary routine examinations as is chromogranin, which showed a high positivity rate of 90.9% identical to the total average (88.2%).

Postoperative outcomes

It was pointed out that there existed a close correlation between postoperative survival rates and histologic patterns of carcinoids.3, 5, 7, 9, 10–12 In the current series of pancreatic carcinoids, however, histologic patterns were recorded in a relatively small number of only 18 cases, insufficient for the purpose of statistically evaluating the correlation with postoperative outcomes. The cumulative 5YSR calculated on the basis of the Kaplan–Meier method (Fig. 2) was extremely low for the pancreatic carcinoid series (28.9% ± 16.7%; n = 46), significantly different from that for the appendiceal series (89.7% ± 2.0%; n = 565) but not so for the small intestine series (82.1% ± 3.3%; n = 368). Despite extremely low 5YSR for the pancreatic carcinoid series, a margin of as much as 16.7% in the rate may exist, theoretically increasing the 5YSR to a possible 45.6%. In a previous study,3 the 5YSR in this particular carcinoid series was calculated as 43.2% ± 20.9% (n = 40), but one would expect the analysis of a larger and more effective number of cases to yield a much more precise and stable value for 5YSR in the pancreatic carcinoid series.

Comparative evaluation

In the current study, several attempts have been made to compare various clinicopathologic aspects of the pancreatic carcinoid series with the carcinoid series of other representative sites. Abundant information is available in a monograph, containing a collection of carcinoids and their variant neoplasms in the systemic sites of both digestive and extradigestive organs.25


Compared with other representative sites, statistical evaluation of 156 cases of pancreatic carcinoids revealed that this series of neoplasms is characterized by a high metastatic rate (66.7%), a large average tumor size (68.6 mm), a relatively high incidence of carcinoid syndrome (23.3%), a high detection rate of immunohistochemical serotonin (92.9%) and NSE (100.0%), and a very low 5-year survival rate (28.9%).


The author thanks Dr. T. Abe at the Department of Medical Informatics, Niigata University School of Medicine, for his evaluation of P values and to Ms. Y. Yakuwa for her invaluable help with the computer analyses.