Volume 15, Issue 4 (Vol.15, No.4 2020)
irje 2020, 15(4): 343-351 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Safari M, Abbasi M, Gohari Ensaf F, Berangi Z, Roshanaei G. Identification of Factors Affecting Metastatic Gastric Cancer Patients’ Survival Using the Random Survival Forest and Comparison with Cox Regression Model. irje 2020; 15 (4) :343-351
URL: http://irje.tums.ac.ir/article-1-6424-en.html
URL: http://irje.tums.ac.ir/article-1-6424-en.html
1- Candidate of PhD, Department of Biostatistics, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran
2- Assistant Professor, Department of Internal Medicine, School of Medicine, Hamadan University of Medical Sciences, Hamadan, Iran
3- MSc of Epidemiology, Students Research Center, Hamadan University of Medical Sciences, Hamadan, Iran
4- Associate Professor, Department of Biostatistics, Modeling of Noncommunicable Disease Research Center, Hamadan University of Medical Sciences, Hamadan, Iran ,roshanaei@umsha.ac.ir.Gh
2- Assistant Professor, Department of Internal Medicine, School of Medicine, Hamadan University of Medical Sciences, Hamadan, Iran
3- MSc of Epidemiology, Students Research Center, Hamadan University of Medical Sciences, Hamadan, Iran
4- Associate Professor, Department of Biostatistics, Modeling of Noncommunicable Disease Research Center, Hamadan University of Medical Sciences, Hamadan, Iran ,
Abstract: (4387 Views)
Background and Objectives: In survival analysis, using the Cox model to determine the effective factors requires the assumptions whose failure of leads to biased results. The aim of this paper was to determine the factors affecting the survival of metastatic gastric cancer patients using the non-parametric method of Randomized Survival Forest (RSF) model and to compare its result with the Cox model.
Methods: In this retrospective cohort study, 201 patients with metastatic gastric cancer were evaluated in Hamadan Province. Patient survival was calculated from diagnosis to death or end of study. Demographic characteristics (such as gender and age) and clinical variables (including stage, tumor size, etc.) were extracted from the patient records. Factors affecting survival were determined using the Cox model and RSF. Data analysis was performed using the R3.4.3 software and RandomForestSRC and survival packages.
Results: The mean (SD) age of patients was 61.5 (12.9) years old. The Cox model showed that chemotherapy (p=0.033) was effective in survival, and the results of fitting the RSF model showed that the most important variables affecting survival were type of surgery, location of metastasis, chemotherapy, age, tumor grade, surgery, number of involved lymph nodes, sex and radiotherapy. Based on the model appropriateness, the RSF model with log-rank split rule had a better performance compared to the Cox model.
Conclusion: If the number of variables is high and there is a relationship between the variables, the RSF method identifies the important and effective variables on survival with high accuracy without requiring restrictive assumptions compared to the Cox model.
Methods: In this retrospective cohort study, 201 patients with metastatic gastric cancer were evaluated in Hamadan Province. Patient survival was calculated from diagnosis to death or end of study. Demographic characteristics (such as gender and age) and clinical variables (including stage, tumor size, etc.) were extracted from the patient records. Factors affecting survival were determined using the Cox model and RSF. Data analysis was performed using the R3.4.3 software and RandomForestSRC and survival packages.
Results: The mean (SD) age of patients was 61.5 (12.9) years old. The Cox model showed that chemotherapy (p=0.033) was effective in survival, and the results of fitting the RSF model showed that the most important variables affecting survival were type of surgery, location of metastasis, chemotherapy, age, tumor grade, surgery, number of involved lymph nodes, sex and radiotherapy. Based on the model appropriateness, the RSF model with log-rank split rule had a better performance compared to the Cox model.
Conclusion: If the number of variables is high and there is a relationship between the variables, the RSF method identifies the important and effective variables on survival with high accuracy without requiring restrictive assumptions compared to the Cox model.
Type of Study: Research |
Subject:
Epidemiology
Received: 2020/01/28 | Accepted: 2020/01/28 | Published: 2020/01/28
Received: 2020/01/28 | Accepted: 2020/01/28 | Published: 2020/01/28
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
