Outcome After Surgical Resection of Gastrointestinal Stromal Tumors (GIST) in the Second Part of Duodenum

the Medical records of patient with second part duodenal GIST were reviewed retrospectively for symptoms, clinical examination, preoperative workup including standard imaging by multi-slice computed tomography (CT) of abdomen and pelvis with oral and intravenous contrast. Endoscopic ultrasound was routinely perfumed to assess the depth of tumor invasion (T-stage) and to obtain a tissue sample by fine needle aspiration cytology (FNAC). Definitive diagnosis was based on CT images and post-operative pathology report.

Patients were subjected to curative localized resection with free margins by frozen section examination or pancreaticoduodenectomy (PD) if the tumor was invading the major duodenal papilla.

Localized duodenal resection was performed by one of two techniques: The first one was excision of part of the wall (wedge resection). Reconstruction was performed either by primary closure without tension, provided that adequate lumen is preserved, or by side to side roux-en-Y duodenojejunostomy (DJ) if there was tension on the edges of the duodenal defect. The second technique was employed in case of larger size tumors and entailed excision of D2 (segmental duodenectomy) to be followed by anastomosis (side to side roux-en-Y DJ or end-to-end DJ). Great care to avoid tumor rupture was emphasized in all operations. Standard lymph node dissection was not performed.

Pathologic data (tumor location, size, margins, and mitoses per 50 high-power fields [HPF]) and immunohistochemical analysis were collected and tumors classified into very low, low, moderate and high risk based on Miettinen classification that also incorporates tumor size.

All patients (those who underwent localized resection with safety margin and those who underwent PD) were followed up and re-evaluated at one, three, six and twelve months then once per year for a total follow-up period of 3 years. Contrast-enhanced multi-slice CT abdomen/pelvis was performed after 6 months then yearly for detection of recurrence.

Using PASS program version 20, the minimum sample size required was 45 patients with duodenal GIST using 3% local recurrence rate and 5% error at 5% level of significance and 80% power. Data were analyzed using IBM-SPSS software package version 20. Qualitative variables were summarized using numbers and percent. Quantitative variables were summarized using mean and standard deviation (SD) as data was normally distributed by kolmogrov-smirnov test. Survival analysis was done using life tables, log-rank test and Kaplan Meier's curve. All statistical analysis was conducted at 5% level of significance.