Predicting Survival of Patients with Lung Cancer Using Improved Adaptive Neuro-Fuzzy Inference System

Introduction: Lung cancer is the main cause of mortality in both genders worldwide. This disease is caused by the uncontrollable growth and development of cells in both or one of the lungs. Although the early diagnosis of this cancer is not an easy task, the earlier it is diagnosed, the higher will be the chance of treating. The objective of this study was to develop an optimized prediction model of the survival of patients with lung cancer based on patients’ characteristics through data mining approach. Method: In this applied-descriptive study, the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm and the Particle Swarm Optimization (PSO) algorithm were applied to predict the survival rate of patients with lung cancer. The Surveillance, Epidemiology and End-Results (SEER) database of Louisville University, USA was also utilized. The evaluation of this proposed model was conducted based on certain criteria including accuracy, precision, error and root-mean-square error. Results: The obtained finding indicate the outperformance of ANFIS through PSO algorithm vs. its counterparts in this context with a 99.80 accuracy for one-year survival, 99.74% for two-years and 99.66% for five-years on SEER dataset. Conclusion: Applying ANFIS through PSO in predicting the survival of patients with lung cancer is a strong measure. Compared with other models, this newly proposed model was of the highest accuracy and precision and of the lowest error rate. Therefore, it is suggested to apply this model for predicting survival of patient.