:: Volume 18, Issue 4 (Iranian South Medical Journal 2015) ::
Iran South Med J 2015, 18(4): 711-719 Back to browse issues page
Applied the additive hazard model to predict the survival time of patient with diffuse large B- cell lymphoma and determine the effective genes, using microarray data
Arefa Jafarzadeh Kohneloo1, Ali reza Soltanian2, Jalal Poorolajal2, Hosean Mahjub 3
1- Department of Biostatistics & Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Iran
2- Research Center for Modeling of Non-communicable Diseases and Department of Biostatistics & Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Iran
3- Research Center for Health Sciences and Department of Biostatistics & Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Iran , mahjub@umsha.ac.ir
Abstract:   (2324 Views)

Background: Recent studies have shown that effective genes on survival time of cancer patients play an important role as a risk factor or preventive factor. Present study was designed to determine effective genes on survival time for diffuse large B-cell lymphoma patients and predict the survival time using these selected genes. Materials & Methods: Present study is a cohort study was conducted on 40 patients with diffuse large B-cell lymphoma. For these patients, 2042 gene expression was measured. In order to predict the survival time, the composition of the semi-parametric additive survival model with two gene selection methods elastic net and lasso were used. Two methods were evaluated by plotting area under the ROC curve over time and calculating the integral of this curve. Results: Based on our findings, the elastic net method identified 10 genes, and Lasso-Cox method identified 7 genes. GENE3325X increased the survival time (P=0.006), Whereas GENE3980X and GENE377X reduced the survival time (P=0.004). These three genes were selected as important genes in both methods. Conclusion: This study showed that the elastic net method outperformed the common Lasso method in terms of predictive power. Moreover, apply the additive model instead Cox regression and using microarray data is usable way for predict the survival time of patients.

Keywords: Lymphoma, Survival analysis, Variable selection, Elastic net, Lasso
Full-Text [PDF 678 kb]   (799 Downloads)    
Type of Study: Original | Subject: Pathology
Received: 2015/09/12 | Accepted: 2015/09/12 | Published: 2015/09/12


XML   Persian Abstract   Print



Volume 18, Issue 4 (Iranian South Medical Journal 2015) Back to browse issues page