Aim: Gastric cancer is the fifth most common cause of tumor-related death in Western countries. Surgery is the only effective treatment but only 50-60% of patients can receive a curative treatment because of absent or aspecific symptoms. The aim of this study was to develop a scale for gastric cancer patients that takes into account factors related to the tumor and to the patient.
Methods: Fifty-seven patients with gastric adenocarcinoma admitted to the Department of General, Geriatric Surgery and Diagnostic and Operative Endoscopy of the University ''Federico II'' in Naples, and treated by gastrectomy from January 1998 until December 2002, were included in this retrospective cohort. The prognostic score was created according to the variables identified in Cox analysis as statistically significant (P 0.1).
Results: The 5-year mortality rate was 61%. Cox analysis identified these variables with a significant effect on mortality: age ?60 (odds ratio (OR) 4.16; P=0.015), smoking or alcoholism (OR 2.66; P=0.057), pTNM I (OR 0.04; P=0.003), pTNM II (OR 0.18; P=0.029), pTNM III (OR 0.27; P=0.023), pTNM IV (OR 3.28; P=0.012), lymph node ratio (LNR) <20% (OR 0.15; P=0.01), LNR 20% (OR 3.83; P=0.002), Lauren diffuse histotype (OR 2.41; P=0.1) and location of the neoplasm at superior third (OR 6.70; P=0.003), middle third (OR 5.60; P=0.003), or inferior third (OR 0.32; P=0.008). Patients have been randomized into three groups according to their scores (3-40.5; 41-78.5; 79-115.5) and the 5-year mortality rate was 46%, 59%, 90% in group 1, 2 and 3 respectively.
Conclusion: It is necessary to consider in prognostic stratification of gastric cancer patients not only pTNM staging but also other factors such as age, smoking or alcoholism, Lauren histotype, location and linfonodal involvement. It is possible to design a more effective prognostic score predicting the individual risk and addressing the therapy and the follow-up.