Abstract
BACKGROUND: Clinical stage IV gastric cancer (GC) may need palliative procedures in the presence of symptoms such as obstruction. When palliative resection is not possible, jejunostomy is one of the options. However, the limited survival of these patients raises doubts about who benefits from this procedure. AIM: To create a prognostic score based on clinical variables for 90-d mortality for GC patients after palliative jejunostomy. METHODS: We performed a retrospective analysis of Stage IV GC who underwent jejunostomy. Eleven preoperative clinical variables were selected to define the score categories, with 90-d mortality as the main outcome. After randomization, patients were divided equally into two groups: Development (J1) and validation (J2). The following variables were used: Age, sex, body mass index (BMI), American Society of Anesthesiologists classification (ASA), Charlson Comorbidity index (CCI), hemoglobin levels, albumin levels, neutrophil-lymphocyte ratio (NLR), tumor size, presence of ascites by computed tomography (CT), and the number of disease sites. The score performance metric was determined by the area under the receiver operating characteristic (ROC) curve (AUC) to define low and high-risk groups. RESULTS: Of the 363 patients with clinical stage IVCG, 80 (22%) patients underwent jejunostomy. Patients were predominantly male (62.5%) with a mean age of 62.4 years old. After randomization, the binary logistic regression analysis was performed and points were assigned to the clinical variables to build the score. The high NLR had the highest value. The ROC curve derived from these pooled parameters had an AUC of 0.712 (95%CI: 0.537-0.887, P = 0.022) to define risk groups. In the validation cohort, the diagnostic accuracy for 90-d mortality based on the score had an AUC of 0.756, (95%CI: 0.598-0.915, P = 0.006). According to the cutoff, in the validation cohort BMI less than 18.5 kg/m(2) (P < 0.001), CCI ≥ 1 (P = 0.001), ASA III/IV (P = 0.002), high NLR (P = 0.012), and the presence of ascites on CT exam (P = 0.004) were significantly associated with the high-risk group. The risk groups showed a significant association with first-line (P = 0.012), second-line chemotherapy (P = 0.009), 30-d (P = 0.013), and 90-d mortality (P < 0.001). CONCLUSION: The scoring system developed with 11 variables related to patient's performance status and medical condition was able to distinguish patients undergoing jejunostomy with high risk of 90 d mortality.