Primary malignant tumors of the small bowel are rare. These are mainly adenocarcinomas followed in decreasing order by carcinoid tumors, non-Hodgkin lymphomas, gastrointestinal stromal tumors, melanomas, and other rare entities. In the United States, the incidence of all types of small bowel cancer is estimated to be approximately 5,300 cases per year and approximately 1,100 patients die from small bowel cancer each year. The development of an adenocarcinoma of the small bowel has been related to the mucosal contact time with bile acid solutions. Ross et al.  showed that the frequency of tumor distribution within the small bowel correlates with the length of mucosal contact with pancreatico-biliary secretions, implicating bile as a possible carcinogen. This is supported by findings that the active and passive transport of bile acid solutions is limited to the ileum .
In concordance with other reports, adenocarcinomas located in the small bowel, as other malignant entities of the small bowel, are observed mainly between 50 and 70 years of age [5–14]. In general, an accurate preoperative diagnosis has been reported only in 30%–72% of cases [10, 11, 15–24]. The clinical signs and symptoms may vary with the tumor site, size, and existence of ulceration. The common presenting signs and symptoms in our series were nausea, vomiting, abdominal pain, melena, weight loss, anemia, and a palpable mass, none of which was pathognomonic for small bowel tumors. All duodenal adenocarcinomas were diagnosed preoperatively by a gastro-duodenoscopy. For tumors in the jejunum and ileum, computer tomography and small bowel contrast study provided clues suggestive of small bowel tumor. Upper gastrointestinal tract series with small intestinal follow through is one of the most useful diagnostic tests. It yields an accurate diagnosis in 50 to 70% of patients with the neoplasm of small intestine . Localization of intermittent-bleeding small bowel tumors through angiography and tagged-red blood cell radioisotope scan was also helpful in our study. Depending on the clinical symptoms, an emergency operation may be necessary. Seven patients (26%) in this study received emergency surgical treatment. Three of them had gastrointestinal hemorrhage and five had intestinal obstruction. Of importance is the fact that all the tumors requiring emergency surgery were located in the jejunum or the ileum. Thus, an accurate pathologic diagnosis could be achieved intraoperatively in these cases. The rate of diagnosis of small bowel tumors of all types by laparotomy varies between 40 and 80% in the literature . In our study, the rate of diagnosis during laparotomy for small bowel adenocarcinoma was 52%. As known from the literature [14, 24], adenocarcinomas are predominant in the duodenum. The more distal tumors were found more frequently in the jejunum than the ileum, which, however, is not a case in our study (30% in the ileum, and 22% in the jejunum). Brucher et al., found no patient with adenocarcinoma of the ileum in their series . Recently, Dabaja et al., reported a 13% incidence of adenocarcinoma in the ileum .
In 1990 Sellener described an adenoma-adenocarcinoma-sequence  and In 1992 Lashner reported Crohn's disease as a risk factor in developing adenocarcinomas in the small bowel . Rodriguez-Bigas et al , found an association between hereditary nonpolyposis colorectal carcinoma (HNPCC) patients and the increased risk of small bowel adenocarcinoma. In a review by Groves et al. , a total of six out of 114 patients of familial adenomatous polyposis (FAP) developed duodenal adenocarcinoma over a follow-up period of ten years. None of the patients in our study were known to have FAP, Crohn's disease or HNPCC.
The type of surgery varied according to the operating surgeons. For duodenal adenocarcinomas, 62% (eight out of 13 patients) of patients underwent pancreaticoduodenectomy and 15% (two out of 13 patients) underwent segmental duodenal resections with curative intent. Palliative bypass procedures were performed for the remaining patients with metastatic adenocarcinomas of the duodenum. For adenocarcinomas of the jejunum and ileum, Nine out of 14 patients underwent en bloc radical resection, which included three patients with metastectomy, and five out of 14 patients with localized diseases underwent segmental resections. When performing analysis, the palliative procedures had the shortest median survival (10 months) when compared to pancreaticoduodenectomies (34 months) and radical resections of the jejunoileal diseases (40 months). The demand of a higher technical expertise for resection of duodenal tumors as compared to resectable jejunoileal tumors may explain the inferior survival of patients with duodenal tumors, as demonstrated by the fact that significant morbidity in our series occurred only in patients with the tumor located in the duodenum. Importantly, every effort should be done to obtain R0 resection when dealing with small bowel adenocarcinoma because of a significant survival advantage.
Howe et al.,  reviewed 4,995 patients with small bowel adenocarcinoma from the National Cancer Data Base from 1985–1995 and found the following factors to correlate with survival: patient age, tumor site (favoring jejunum and ileum), clinical staging, and whether curative resection was performed. Bakaeen et al.  and Ryder et al.  found tumor size, histologic grade, nodal metastases, and positive surgical margin to be prognostic factors predicting survival of adenocarcinomas of the duodenum. Brucher et al.  identified the presence of the residual tumor, tumor stage, lymph node metastasis, distant metastasis, lymphangiosis carcinomatosa, and vascular invasion as prognostic factors. Dabala et al.  recently reported that only cancer-directed surgery and lymph node involvement ratio to be independent predictors of overall survival in a multivariate analysis.
In our current study, the five-year survival was 30%, which is similar to that reported in pat literature . We also found the presence of a positive node (p < 0. 0001), vascular invasion (p < 0.0001), and poor cellular differentiation (p < 0.0001) to be prognostic indicators, which is also analogous to the report of Brucher et al .