The cox regression analysis of prognostic factors associated with rectal cancer and its dose-response relationships

Acta Universitatis Medicinalis Anhui 2021 11 v.56 1811-1816     font:big middle small

Found programs:

Authors:Zhang Dongdong; Wang Shengyi; Li Yongxiang

Keywords:rectal cancer;cox regression;neutrophil to lymphocyte ratio;restricted cubic splines;prognosis

DOI:10.19405/j.cnki.issn1000-1492.2021.11.026

〔Abstract〕 Objective To explore the prognostic factors of rectal cancer by routine clinicopathologic features. Methods Follow-up data of radical rectal cancer patients were collected using retrospective cohort study method.Associations between related factors and the survival of colorectal cancer were analyzed by classical cox regression,stepwise variable selection cox regression,restricted cubic splines( RCS) cox regression respectively. Results 8 out of 20 variables were associated with the survival of rectal cancer in the classical single-factor cox model. NLR was the independent prognostic factor in the classical multifactorial cox model( P<0. 05). NLR and tumor stage were the independent prognostic factorsin the stepwise variable selection cox regression model( P<0. 05). By RCScox regression analysis,patients' age,tumor diameters,neutrophil,lymphocyte,monocyte,platelet to lymphocyte ratio( PLR),lymphocyte to monocyte ratio( LMR),neutrophil to monocyte ratio( NLR) were associated with prognosis of rectal cancer patients( P<0. 05). Nonlinear relationship was found between NLR and response variable,nonlinear tendencies between some continuous independent variables and response variable were reminded,but there was no statistically significant. Conclusion NLR was the independent prognostic factor of rectal cancer.Stepwise variable selection cox regression and RCS-cox regression can overcome deficiencies of classical cox regression,and find the potential prognostic factors of rectal cancer in the routine clinicopathologic features,describe its association more comprehensively.