Authors: Zeraati H, (PhD) Mahmoudi M,(PhD) Kazemnejad A.(PhD) Mohammad K,(PhD) Haddad P,(MD).

Introduction: In recent years, various reports indicate that gastric cancer is highly prevalent in our country it is the second most common cancer in men, and fourth in the general population. This study was designed and carried out to determine the five-year survival rate of gastric cancer patients who had undergone surgical treatment at one of the most important cancer treatment centers, the Iran Cancer Institute, and to assess its associated factors. During a study period, patients may often experience events that are likely to affect the final outcome as well. It has been suggested that such variables be assigned in the model as time-dependent covariates, but using models with joint distribution of time before death and time before an intermediate event (in this study, relapse), will certainly provide more accurate results. In this study we analyzed the data using a non-homogenous semi-Markovian stochastic process, regarding life span and intensity rate as the doubly stochastic processes.

Methods: One hundred and 29 gastric cancer patients with adenocarcinomatous pathology and lymph node metastasis who had been admitted to and operated on at the Iran Cancer Institute between March 1995 and March 1999 were enrolled in this study. The patients’ life expectancy after surgery and its relationship with variables of age at the time of surgery, gender, and factors related to the disease such as the cancer site, number of affected lymph nodes, type of gastrectomy, presence of liver metastases or distance metastasis, number of complementary or secondary treatments received, relapse, and time to relapce were assessed. In the analyses, methods of Kaplan-Meier, Cox proportional hazards model, non-homogenous Markovian process, and Breslow estimator were used. The software used for the analyses were S Plus 2000 and R, and an alpha level of 0.05 was considered significant.

Results: The five-year survival rate and the median life expectancy in the studied patients were 18.6% and 18.90 months, respectively. The Cox proportional hazards model was used to assess the effect of different variables simultaneously, and it showed that age, and relapse influenced the chances of survival. It was also shown (based on non-homogenous Markovian stochastic processes) that sex, type of gastrectomy and cancer site correlated with time of relapse, while age affected survival after relapse, and age correlated with survival of patients without relapse.

Conclusion: Gastric cancer patients in Iran have a low five-year survival rate. One of the most important reasons seems to be delayed consultation and diagnosis. Most patients are seen first with the disease in the late stages. At this point, most have lymph node, liver, or even distant metastases which makes treatment even more complex. Thus, it is necessary to employ mass media for extensive public education about the early warning signs of the disease and performing periodic examinations.

 �

Keywords: Survival, gastric cancer, Cox proportional hazards model, semi-Markovian process, doubly stochastic process.

  Hakim 2006 8(4) 15-20.