Proposed Staging System for Gastrointestinal Carcinoid Tumors
By Landry, Christine S McMasters, Kelly M; Scoggins, Charles R; Martin, Robert C G II
Gastrointestinal carcinoid tumors are rare neuroendocrine tumors with no staging system in existence. The goal of this study was to establish a staging system consistent with the American Joint Commission on Cancer Staging Systems using the TNM strategy. A retrospective review of our prospective database of 990 hepatopancreaticobiliary patients and our tumor registry identified 108 patients with gastrointestinal carcinoid tumors from June 1990 to September 2006. Tumors were classified into our staging system by depth of penetration, size of primary tumor, nodal status, and the presence/absence of distant metastasis. Patients were staged as Stage 1,22 per cent; Stage II, 29 per cent; Stage 3, 12 per cent; and Stage 4, 35 per cent. There were 41 men and 57 women with a median age of 58.5 years (range, 19-86 years). Primary tumors included 52 small bowel, 12 colon, 19 rectum, nine stomach, and seven of unknown primary origin. The use of our initial staging system demonstrated a trend in differences in survival across all four stages. The use of our initial staging proposal delineates the biology of the disease with accurate overall survival estimates. The addition of a dedicated American Joint Commission on Cancer staging system is needed for gastrointestinal carcinoids. Widespread use of this staging system may contribute to the future management and treatment of gastrointestinal carcinoid tumors. GASTROINTESTINAL CARCINOID TUMORS are rare neuroendocrine tumors derived from enterochromaffin cells that most commonly follow an indolent course. Often associated with hormonal symptoms, these tumors are the most common endocrine tumor of the gastrointestinal tract.1 Unfortunately, gastrointestinal carcinoid tumors can be difficult to diagnose and, as a result, frequently present with metastatic disease.
In 1963, Williams and Sandier classified carcinoid tumors by embryological decent through the foregut, midgut, and hindgut.2 Histological aspects such as mitotic activity and amount of necrosis have been used to further delineate these tumors as “typical” and “atypical.”2 Similarly, the World Health Organization attempted to further classify neuroendocrine tumors by biological behavior, depth of invasion, size of primary tumor, histological differentiation, and angioinvasion.3 However, this system covers a broad range of tumors that originate from endocrine cells, including carcinoids, pheochromocytomas, and paragangliomas. To date, there is no widely accepted formal staging system for gastrointestinal carcinoid tumors.
The overall prognosis of gastrointestinal carcinoid tumors is not currently known at various stages of the disease. As a result, the effect of treatment modalities for carcinoid tumors is difficult to define. The goal of this study was to create a global staging system for gastrointestinal carcinoid tumors so that the progression of disease can be closely monitored. Likewise, we would be able to determine which treatments provide the best overall prognosis based on stage of disease. By incorporating the size of primary tumor, depth of invasion, nodal status, and metastatic disease into a staging system consistent with the American Joint Commission on Cancer, we hope to improve prognostic knowledge of gastrointestinal carcinoid tumors.
Methods
We began by performing a literature review of factors affecting overall survival of gastrointestinal carcinoid tumors. Using this information, a staging system was created incorporating the TNM strategy consistent with American Joint Commission on Cancer staging systems (Tables 1 and 2).
A retrospective review of our prospective database of 990 hepatopancreaticobiliary patients and our tumor registry identified 108 patients diagnosed with any gastrointestinal carcinoid tumor from June 1990 to September 2006. We analyzed our data regarding survival according to size of primary tumor, lymph node status as well as depth of invasion for comparison to current literature. Clinicopathological data was compared between stages according to age, sex, race, location of primary tumor, family history of cancer, and resection. Multiple Kaplan-Meier curves were created to determine the impact of size of primary tumor, depth of invasion, lymph node involvement, and metastatic disease on survival. Similarly, a survival curve was constructed to analyze groups by stage of disease.
TABLE 1. Staging System for Gastrointestinal Carcinoid Tumors
Fisher’s exact test, Student’s t test, and Mann-Whitney U test for nominal, continuous, and ordinal variables were used to evaluate the association of independent variables to surgical complications. Proportional hazards analysis was performed on all variables found significant by bivariate analysis. Relative risk with 95 per cent confidence intervals was calculated as a measure of association. Differences of P < 0.05 were considered significant. Survival curves using the Kaplan-Meier method were created for various characteristics and treatment groups. Groups were then compared using the log rank test. Statistically significant results from univariate analysis were further examined using the Cox proportional hazard model. Statistical analysis was performed using JMP software (JMP; SAS Institute Inc., Cary, NC).
TABLE 2. Staging System for Gastrointestinal Carcinoid Tumors
Results
Clinicopathological Features
Ninety-eight patients had adequate pathological data to be classified according to the newly developed staging system. Among this group, 41 patients were male and 57 patients were female. Average age at diagnosis was 58.5 years with a median age of 58.5 years (range, 19-86 years). Eighty-two patients were of Caucasian descent, whereas 16 patients were of African American descent. Location of primary tumor included nine stomach, 52 small bowel, 12 colon, and 19 rectum. Seventy-three patients underwent surgical resection of the primary tumor, 22 patients were treated endoscopically, and three patients had no surgical intervention. Twenty-one patients died; 12 died of disease, six died of unknown causes, and two died of other causes. Thirty-three patients are currently alive with disease, and 44 patients currently have no evidence of disease with 17 patients having a history of additional malignancies at the time of carcinoid diagnosis or during their follow-up period. For instance, six patients had skin cancer, one prostate cancer, one laryngeal cancer, three colon cancer, two Mullerian mixed tumor, one brain cancer, one pancreatic cancer, and two patients with MEN I.
Patients were classified into stages according to the following: Stage I, 22 per cent; Stage II, 29 per cent; Stage III, 12 per cent; and Stage IV, 35 per cent. Groups were compared by stage according to sex, age, race, surgical treatment, family history of cancer, and location of primary tumor (Table 3). Characteristics varied by location of primary tumor (P < 0.01). More patients with rectal tumors presented in Stage I, whereas small bowel tumors presented with more advanced disease at diagnosis. Also, the patients who did not undergo resection of primary tumor had a great incidence of metastatic disease with more patients with Stage I and Stage 2 disease being treated endoscopically.
Survival Analysis
To evaluate the effect of primary tumor size on overall survival, patients were divided into three groups: tumor size less than 1 cm, tumor size 1 cm or greater but not greater than 2 cm, and tumor size 2 cm or greater. A Kaplan-Meier curve was then constructed showing a trend toward decreased survival with increasing size of primary tumor (P < 0.74) (Fig. 1). Similar curves created for lymph node involvement, metastatic disease, and depth of invasion are displayed in Figures 2 through 4. There was a trend for decreased survival for lymph node involvement (P < 0.7), metastatic disease (P < 0.059), and increased depth of invasion (P < .76).
TABLE 3. Comparison of Characteristics Between Stages
FIG. 1. Length of survival according to size of primary tumor.
Survival curve was created for patients according to stage of disease (Fig. 5), which demonstrated a trend for decreased length of survival for patients with advanced stages of disease (P < 0.3). Overall 5-year survival for all tumors was 67.4 per cent, 5 year survival was 83.3 per cent, 77.8 per cent, 66.7 per cent, and 50 per cent for Stages I through IV, respectively.
Discussion
Our data are consistent with current staging literature of other malignancies by showing a trend for decreased survival according to various pathological characteristics. For instance, increased size of primary tumor has been associated with an increased likelihood of nodal and distant metastasis in multiple studies.1, 4, 5 Likewise, depth of invasion into the muscularis propria, angio and lymphatic invasion, areas of cellular atypia, focal necrosis, and high mitotic index have been linked to decreased overall survival.3, 4 Moreover, the presence of metastatic disease and lymph node involvement confers a poorer prognosis.6 Our carcinoid patient population yielded similar findings.
FIG. 2. The impact of metastatic disease on survival.
FIG. 3. Impact of lymph node metastasis on survival.
FIG. 4. Survival by depth of invasion. FIG. 5. Survival by stage.
The rarity of gastrointestinal carcinoid tumors has contributed to a small population size in our study, which represents a limitation of our study but does not diminish from the fact that even more rare cancers (e.g., gallbladder) have established staging systems that are integral to the timing and type of treatment offered. Very few large-scale studies have been performed to evaluate the overall survival of patients with gastrointestinal carcinoid tumors. In 1975, Godwin published an epidemiological study of 2837 patients with carcinoid tumors in the bronchopulmonary system and gastrointestinal tract.7 Until 1997, Godwin’s publication regarding the incidence, distribution, and survival rates of carcinoid tumors was held as the gold standard.8 Godwin’s work reported 5-year survival rates ranging from 33 per cent for sigmoid colon carcinoid tumors to 99 per cent for appendiceal carcinoids.7
In 1997, Modlin published the largest epidemiological study of carcinoid tumors known to date.8 Modlin analyzed 13,715 patients with carcinoid tumors. This study also included the 2,837 patients in Godwin’s publication from the National Cancer Institute’s SEER database. The overall 5-year survival rate for all carcinoid tumors in Modlin’s study was 67.2 per cent, which is comparable to our data (67.4 per cent overall survival). Modlin’s work demonstrated a trend toward increased incidence of carcinoid tumors overall as well as a change in the distribution of tumors in the gastrointestinal tract. Moreover, Modlin also suggested a genetic predisposition of rectal carcinoid with black and Asian populations.8
A small percentage (12.9%) of the patients in Modlin’s study were diagnosed with distant metastasis.8 The 5-year survival rate for patients in Modlin’s population with local and distant metastases was 71.7 per cent and 38.5 per cent, respectively. Our data demonstrated 35 per cent of the patients being diagnosed with Stage IV disease, a much larger proportion than Modlin’s study. Given the fact that we are a tertiary referral center, our patients may have presented with more advanced disease. Also, our patients with distant metastatic disease had a better overall 5-year survival rate than Modlin’s population (50 per cent 5-year survival rate in our study). However, the majority of our patients with metastatic disease to the liver were treated with liver resection, radiofrequency ablation, Yttrium-90 embolization, and/or chemoembolization. Treatment modalities in Modlin’s publication were not reported for patients with liver metastasis. As a result, our overall 5-year survival rate may be better than other studies.
Should Carcinoid Tumors Be Classified by Embryological Origin?
We considered creating a staging system that separates tumors by embryological origin. Separating tumors by foregut, midgut, and hindgut has been an accepted practice for tumor classification since it was first introduced by Williams and Sandier in 1963.2 Midgut carcinoid tumors are thought to have a worse overall prognosis than other gastrointestinal carcinoid tumors because they frequently present with more advanced disease as a result of difficulty in diagnosis.5 Many sources state that the best overall survival is associated with appendiceal carcinoids, whereas patients who have midgut or hindgut tumors have the worst overall survival.5
Although many sources in the literature commonly describe carcinoid tumors by embryological origin, there appears a significant difference between these tumors based on organ of origin rather than embryological descent. For example, survival varies when separating foregut tumors by organ of origin. In Modlin’s publication, overall 5-year overall survival rate was 63 per cent and 37.5 per cent for stomach and pancreas, respectively.8 Likewise, overall 5-year survival rate for descending colon tumors and rectal tumors was 68.4 per cent and 88.3 per cent, respectively.8 If both distal colon and rectal tumors were grouped together in the hindgut group, prognostic data could be misleading. As a result of the differences in survival based on organ of origin, carcinoid tumors should be separated by organ of origin rather than by embryological descent. Unfortunately, our patient population was too small to analyze patients by location of primary tumor using our staging system. Because of this finding, we are currently working with a larger database to further determine the significance of our staging proposal as well as differences in survival by location.
Incidence of Gastrointestinal Carcinoid Tumors
Currently, the anatomic distribution of gastrointestinal carcinoid tumors is uncertain.9 One of the first largely recognized population-based reviews of gastrointestinal carcinoid tumors was published in 1975 by Godwin.7 This study claimed that the most common location for primary tumor was the appendix, rectum, and ileum. Sabiston states that carcinoid tumors are most commonly located in the appendix (45%) followed by the ileum (28%) and rectum (16%).10 Similarly, Greenfield states that carcinoids are most frequently found in the appendix followed by the small bowel.11 The M.D. Anderson Handbook describes appendiceal carcinoids (40%) to be the most common followed by small intestine (27%) and rectum (15%).1 On the contrary, more recent studies have identified the small bowel to be the most common location for gastrointestinal carcinoid tumors. For instance, Modlin from Yale analyzed 13,715 carcinoid tumors from 1950 to 1999.8 He concluded that gastrointestinal carcinoid tumors are most commonly located in the small bowel (41.8%) followed by the rectum (27.4%) and stomach (8.7%). He also noted that the incidence of appendiceal carcinoid tumors has decreased over the last several years. Other studies have also identified the small bowel as the most common location for gastrointestinal carcinoid tumors.9 Similarly, our population group identified the small bowel as the most common gastrointestinal carcinoid tumor followed by the rectum, colon, and stomach. Given that the current literature is conflicting regarding the anatomical distribution of gastrointestinal carcinoid tumors, the true incidence by location is vague.
Additional Malignancies
Gastrointestinal carcinoid tumors are often associated with other malignancies.2, 6 Seventeen per cent of patients from our database had a history of a different malignancy. Four of these patients had synchronous disease. Because carcinoid tumors are typically slowgrowing, patients often die from other malignancies. As a result, if a patient with a malignancy in addition to gastrointestinal carcinoid tumors died, they were removed from our staging analysis because survival data would be skewed. Patients with gastrointestinal carcinoid tumors should be carefully evaluated for other malignancies.2
The Importance of a Staging System for Carcinoid Tumors
To date, no official staging system exists for gastrointestinal carcinoid tumors. Our study demonstrated an overall 5-year survival of 67.4 per cent suggesting that carcinoid tumors have significant malignant potential. By classifying tumors by degree of advanced disease, we will have more information on overall prognosis for patients. Moreover, patients can more accurately be compared across treatment centers.
The effect of various treatment modalities on overall prognosis is unclear. The stage of disease when treatment is administered could have significant impact on overall survival. For example, a large portion of our patients in Stage IV received aggressive treatment and had an overall 5-year survival rate of 50 per cent. This survival characteristic is much higher than Modlin’s patients with metastatic disease (5-year survival rate 38.5%).8 As more of these patients are studied and classified into a staging system, patients with more advanced disease could be identified and thus treated more aggressively. The effectiveness of treatment modalities would be measured more accurately with clinical incorporation of a staging system.
Conclusion
Gastrointestinal carcinoid tumors have been analyzed by multiple disciplines, including endocrinologists, oncologists, surgeons, and pathologists. Our intention is to create a common method of classifying gastrointestinal carcinoid tumors across treatment centers. By working together, a widely accepted staging system could be incorporated into clinical practice. We would be better able to understand the behavior of this disease and more accurately predict overall survival. Our plans to analyze our staging proposal with a larger database will help advance our knowledge of this malignancy and thus the impact of different treatments.
REFERENCES
1. Hwang R, Robinson E, Cusack J, Tyler D. Small bowel malignancies and carcinoid tumors. In: Feig B, Berger D, Fuhrman G, eds. The M.D. Anderson Surgical Oncology Handbook. Philadelphia: Lippincott Williams & Wilkins, 2003:193-214.
2. Modlin IM, Kidd M, Latich I, et al. Current status of gastrointestinal carcinoids. Gastroenterology 2005;128:1717-51.
3. Kloppel G. Tumour biology and histopathology of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007;21:15-31.
4. Rorstad O. Prognostic indicators for carcinoid neuroendocrine tumors of the gastrointestinal tract. J Surg Oncol 2005;89:151-60.
5. Onaitis MW, Kirshbom PM, Hayward TZ, et al. Gastrointestinal carcinoids: Characterization by site of origin and hormone production. Ann Surg 2000;232:549-56.
6. Shebani KO, Souba WW, Finkelstein DM, et al. Prognosis and survival in patients with gastrointestinal tract carcinoid tumors. Ann Surg 1999;229:815-21.
7. Godwin JD. Carcinoid tumors. An analysis of 2,837 cases gastrointestinal carcinoids: Characterization by site of origin and hormone production. Cancer 1975;36:560-9.
8. Modlin IM, Lye KD, Kidd M. A 5-decade analysis of 13,715 carcinoid tumors. Cancer 2003;97:934-59. 9. Maggard MA. Updated population-based review of carcinoid tumors. Ann Surg 2004;240:117- 22.
10. Evers MB. Small intestine. In: Townsend ed. Sabiston Textbook of Surgery. Philadelphia: W.B. Saunders Co., 2004: 1359-62.
11. Schrieber M, Bass BL. Small intestinal neoplasms. In: Greenfield LJ, ed. Surgery: Scientific Principles and Practice. Baltimore: Lippincott Williams & Wilkins, 2001:831-43.
CHRISTINE S. LANDRY, M.D., KELLY M. MCMASTERS, M.D. PH.D., CHARLES R. SCOGGINS, M.D., ROBERT C.G. MARTIN II, M.D.
From the Division of Surgical Oncology, University of Louisville, Louisville, Kentucky
Presented at Digestive Disease Week, Washington, DC, May 21, 2007.
Address correspondence and reprint requests to Robert C.G. Martin II, M.D., Assistant Professor of Surgery, Division of Surgical Oncology, 315 E. Broadway, #312, Louisville, KY 40202. E-mail: Robert.Martin@louisville.edu.
Copyright Southeastern Surgical Congress May 2008
(c) 2008 American Surgeon, The. Provided by ProQuest Information and Learning. All rights Reserved.