ABSTRACT: Small bowel adenocarcinoma is a relatively rare malignancy. Only limited information is available on the incidence, prognosis, and role of chemotherapy in the treatment of this disease. We present a review of currently available clinical data to assist the practicing oncologist in the treatment of these patients. Approximately 5,300 new cases and 1,100 deaths from small bowel adenocarcinoma are reported annually in the United States. Increased incidence is seen in patients with Crohn’s disease, hereditary nonpolyposis colorectal cancer, and familial adenomatous polyposis. Factors associated with poor prognosis are age > 75 years, lack of surgical resection, advanced stage, and tumor arising in the duodenum. The median survival of patients with localized, locally advanced, and metastatic disease is 50.1, 22.2, and 8.6 months, respectively. Few data exist on the use of (neo)adjuvant or palliative chemo(radio)therapy in this setting. Fluorouracil(Drug information on fluorouracil) (5-FU)-based chemotherapy, as a single agent or in combination with others, has been used in most case series. Duodenal adenocarcinoma accounts for more than 50% of all cases of small bowel adenocarcinoma. Resectability is the key prognostic factor, along with age, performance status, tumor location, and presence of distant metastasis. Although the activity of 5-FU-based regimens has been documented, the assessment of clinical benefit is hindered by the lack of prospective, randomized data. [ONCOLOGY 16:1364-1373, 2002]
Cancer of the small bowel is a relatively rare malignancy, accounting for approximately 2% of gastrointestinal (GI) tumors. The annual incidence of small bowel cancer is about 5,300 cases, and this disease is associated with 1,100 deaths each year in the United States. An estimated 40% of small bowel cancers are adenocarcinomas, 40% are carcinoids, 15% are sarcomas (GI stromal tumors), and less than 5% are lymphomas. Small bowel adenocarcinoma includes adenocarcinomas arising in the duodenum, jejunum, and ileum. While adenocarcinomas arising from the ampulla of Vater and the periampullary region are typically included in the category of small bowel adenocarcinomas, those arising from the ileocecal valve, appendix, and Meckel’s diverticulum are excluded.
The National Cancer Data Base (NCDB), a joint project of the American College of Surgeons Commission on Cancer and the American Cancer Society, collects data on cancer patients in US hospitals. This database represents an estimated 60% of all cancer cases in the United States. During the 10-year period from 1985 to 1995, 4,995 cases of nonampullary small bowel adenocarcinoma were registered. The mean age at presentation was 65 years. Some reports suggest a slight male predisposition, but in the NCDB, there was almost an equal incidence in both sexes. Approximately 55% of small bowel adenocarcinomas occurred in the duodenum, 18% in the jejunum, and 13% in the ileum. In 14% of cases, the precise location of small bowel adenocarcinoma was not specified. Upon presentation, 26% of duodenal adenocarcinomas were found to be locoregionally advanced, while 32% of cases showed evidence of metastatic spread.
An increased risk of developing small bowel adenocarcinoma has been described in patients with Crohn’s disease (40- to 100-fold increase in relative risk), familial adenomatous polyposis (FAP, 50- to 300-fold),[5,6] and hereditary nonpolyposis colorectal cancer (HNPCC, > 100-fold). In a study involving HNPCC patients who developed small bowel adenocarcinomas, mismatch repair gene mutations were noted in 36% (15 of 42 patients). In 57% of the patients, the small intestine was the first site of carcinoma, suggesting that small bowel cancer can be the presenting neoplasm in HNPCC patients.
Japan-Hawaii Cancer Study
Stemmermann et al reported the results of the Japan-Hawaii Cancer Study (JHCS), which identified four small bowel adenocarcinomas (0.05%) among a cohort of 8,006 Hawaiian-Japanese men followed for a period of 22 years. All the tumors were located in the duodenum or proximal jejunum, and all the patients also had colorectal cancer. In three of the patients, the colorectal cancer was multicentric. The fourth patient also had gastric cancer, and his brother had multicentric colorectal cancer.
Also in this study, three men had leiomyosarcoma of the small intestine and three had adenocarcinoma of the ampulla of Vater. None of these patients had colorectal cancer, and they did not have a family history of gastric or colorectal cancers. The familial clustering of small bowel adenocarcinomas with multicenteric colorectal cancer and gastric cancer led the authors to suggest that there may be a genetic trait that makes these patients susceptible to carcinogens.
In a study of 100 patients with jejunoileal small bowel adenocarcinoma, reported by Veyrieres et al, 25% had prior or secondary—mostly intra-abdominal—tumors, and another eight patients developed secondary duodenojejunal, nonmalignant polyps. This number far exceeds the tumor incidence expected in the general population or in patients with a large bowel primary malignancy. As a result, close follow-up of patients with small bowel adenocarcinoma after curative-intent resection has been strongly recommended. As these patients are at an increased risk for other GI cancers, such as gastric and colorectal carcinomas, patients should undergo endoscopic evaluation of the GI tract at the time of diagnosis of the small bowel adenocarcinoma and during follow-up.
Other Carcinogenic and Genetic Factors
Although the duodenum accounts for only 8% to 10% of the length of the small bowel, the relatively high incidence of adenocarcinomas in the duodenum has raised the possibility that yet unidentified carcinogens may be present in bile or in gastropancreatic secretions.
K-ras mutations at codon 12 are common in gastric, pancreatic, and colon carcinomas. Interestingly, Younes et al identified K-ras mutations at codon 12 in 4 of 12 duodenal, but in none of 16 jejunoileal small bowel adenocarcinomas tested, suggesting a possible pathogenetic role for duodenal small bowel adenocarcinoma, similar to that seen in pancreatic and colorectal cancer.
Park et al studied the expression of p53 in carcinoma arising from the ampulla of Vater and correlated it with survival. Overexpression of p53 in de novo carcinomas was associated with a worse prognosis. Vaidya et al identified significant differences in the expression of c-erbB-2, Ki-67, and tenascin in carcinoma of the papilla of Vater, between patients who survived less than or longer than 5 years after diagnosis. Increased expression of c-erbB-2, Ki-67, and tenascin was associated with poorer survival 
Strong cytoplasmic carcinoembryonic antigen (CEA) staining has been described in small bowel adenocarcinomas. Immunohistochemical staining for CEA, and to a lesser extent CA 19-9, is positive in the majority of ampullary and about half of nonampullary duodenal small bowel adenocarcinomas. However, no clear relationship between the expression of these tumor antigens, tumor growth, and DNA ploidy has been observed, and their independent prognostic value has not been established.
Small bowel adenocarcinomas are similar to their colonic counterparts with respect to the adenoma-carcinoma sequence and demonstrate similar pathologic features, with slight differences in frequency of histologic types and immunohistochemical markers. Most small bowel adenocarcinomas are solitary, sessile lesions, often appearing in association with adenomas. They are usually moderately well-differentiated and almost always positive for acid mucin. As mentioned, small bowel adenocarcinomas can be positive for CEA, CA 19-9, and p53. Expression of c-erbB-2, Ki-67, and tenascin has also been described.[12-14] Small bowel adenocarcinomas arising from the ileum may show staining with neuroendocrine markers.
The clinical presentation of small bowel adenocarcinoma depends on the location of the primary tumor, its growth pattern, and extent of metastatic spread. In general, symptoms are initially vague and can vary from anemia, bleeding, abdominal pain, nausea, and vomiting to obstruction and/or perforation in the case of more locally advanced lesions. Given the lack of a specific complex of symptoms diagnostic for small bowel cancer, the time between initial development of symptoms and diagnosis is relatively long (approximately 6 to 8 months).
The absence of effective screening methods and the relative paucity of presenting symptoms contribute to the higher percentage of advanced cases at the time of diagnosis, compared to patients with colorectal cancer. Advancing small bowel adenocarcinoma can spread either contiguously (submucosally) or via the lymphohematogenous route. Common sites of metastases include locoregional lymph nodes, liver, lung, and the peritoneum.
The diagnostic and staging work-up of small bowel adenocarcinoma typically includes an upper GI series with small bowel follow through (SBFT) and computerized tomography (CT) of the abdomen and pelvis. SBFT has been reported to be abnormal in 53% to 83% of patients and is able to delineate the tumor in 30% to 44%. In a series of 35 patients with small bowel tumors, Laurent et al observed CT abnormalities in 97% of cases and showed that CT was able to predict the presence of cancer in 80%. The characteristic finding on CT is a partially obstructing, concentric narrowing in the proximal small bowel, especially seen with primary tumors larger than 3 cm.
Upper GI barium series and endoscopies have varying rates of success, but remain the mainstay in identifying proximal small bowel adenocarcinoma. No clear role for CEA or other serum tumor markers has been established for diagnosis, although cases with elevation of CEA, CA 19-9, or CA 125 have been reported.
Small bowel adenocarcinomas are staged according to the tumor-node-metastasis (TNM) criteria, as presented in Table 1. Staging is based on the depth of involvement of the bowel wall by tumor, regional nodal status, and the presence or absence of distant metastasis.
By multivariate analysis, factors significantly correlated with disease-specific survival in small bowel adenocarcinoma are patient age, tumor site, disease stage, and whether surgical resection was performed. Poor prognostic factors include patient age > 75 years, duodenal location, lack of curative-intent surgery, and advanced disease stage/presence of distant metastasis.
Although duodenal tumors do not present with more advanced disease than those from the jejunum and/or ileum, they are associated with a worse overall prognosis. One possible explanation for this observation could be that a greater proportion of patients with duodenal tumors present at an elderly age (> 75 years), compared to those with jejunoileal tumors. Moreover, fewer elderly patients undergo surgical resection for their disease. Pancreaticoduodenectomy, typically required for resection of duodenal lesions, is a more complicated surgery, compared with resection of small bowel and lymph node-bearing mesentery or ileocolectomy performed for jejunal and ileal lesions.
A strong trend for improved 10-year survival was found in well-differentiated and moderately differentiated vs undifferentiated small bowel adenocarcinomas (43% vs 0%), whereas the presence or absence of serosal involvement did not seem to affect the outcome. The prognostic significance of lymph node status for survival is controversial. It was not independently correlated with survival for patients in the NCDB. In other clinical series, however, it has been found to be a significant prognostic factor for survival.
In a series of 65 patients (34 with duodenal tumor) from the University of Rochester, for example, node positivity carried a worse prognosis. The 5-year survival was 13% for patients with node-positive small bowel adenocarcinoma, as compared with 70% in patients with node-negative small bowel adenocarcinoma, treated with surgery alone.
Upon analysis of the NCDB database, the 5-year overall survival of patients with small bowel adenocarcinoma was about 30%, with a median survival of 19.7 months. In a subgroup analysis, patients with localized disease had a median survival of 50.1 months; those with regional disease, 22.2 months; and those with distant metastatic disease, 8.6 months. The respective 5-year disease-specific survival and the median survival were reported to be 28.2% and 16.9 months for duodenal primaries; 37.6% and 28.9 months for jejunal tumors; and 37.8% and 30.8 months for ileal tumors.
Rodriguez-Bigas et al reported a study of 42 patients with HNPCC who developed small bowel tumors. The median age at diagnosis was 49 years and the 5- and 10-year survival rates were 44% and 33%. This suggests that, like colorectal carcinoma, small bowel adenocarcinomas in patients with HNPCC occur at an earlier age and may have a better prognosis, compared with those occurring in the general population.