Evaluation of Prognostic Variables for Classifying the Survival In lung cancer Patients using The Decision Tree

Lung cancer due to low chance of survival is the most important and the most common fatal cancer and has the highest mortality rates compared to breast and prostate cancer and other multiple cancers. The purpose of this study is to analyze survival data of patients having lung cancer in the province of West Azerbaijan. The information of this study is relevant to 355 patients refered to hospitals exiting in West Azerbaijan province between 1386-93. The situation of the patients was followed up until the end of the Khordad 1393. The data are compromized from Histopathologic and demographic information. For analyzing the data and classifying the patients، the model of decision tree was used. All analysis and synthesis were performed using the software of R. The mean and the median time of survival for 355 patients having lung cancer were estimated 13 and 4. 8 months، respectively. The probable of 1-year surviving was about 39% and only 7% of patients had experienced 3-year survival. The model of decision tree of state therapy variables، Eastern Cooperative Oncology Group (ECOG)، smoking status، age of diagnosis and tumor stage were identified as important variables. The model of classification tree can identify the important and effective risk factors on the survival of the patients and identifying these factors in the cancer patients can prevent progression of the disease and increase survival time of patients.