b.limour.top/source/_posts/诺莫图（nomogram）展示回归分析的结果.md

---

title: 诺莫图（Nomogram）展示回归分析的结果 tags: [] id: '1914' categories:

• - uncategorized date: 2022-07-17 01:44:52 ---

展现回归分析的OR、HR值可以用森林图，而如果想直观预测实例化对象的生存概率，则可以使用诺莫图

安装补充包

• conda activate ggsurvplot
• conda install -c conda-forge r-rms -y

Logistic回归

• 数据来源

  dd=datadist(df)
  options(datadist="dd") 
  lrmf <- paste0("factor(dcf_status)~", paste(colnames(df)[1:10], collapse = '+'))
  lrmf
  f <- lrm(formula(lrmf) , data =  df)
  nom <- nomogram(f, fun=plogis, lp=F, funlabel="Risk")
  options(repr.plot.width=10, repr.plot.height=6)
  plot(nom, xfrac=.2)
  

Cox回归

dd=datadist(df)
options(datadist="dd") 
coxmf <- paste0("Surv(dcf_time/365, dcf_status==0)~", paste(colnames(df)[1:10], collapse = '+'))
coxmf
f2 <- psm(formula(coxmf), data=df, dist='lognormal') # 构建COX比例风险模型
med <- Quantile(f2) # 计算中位生存时间
surv <- Survival(f2) # 构建生存概率函数
nom <- nomogram(f2, fun=function(x) med(lp=x),funlabel="Median Survival Time")
options(repr.plot.width=12, repr.plot.height=6)
plot(nom,xfrac=.2)
nom <- nomogram(f2, fun=list(function(x) surv(10, x)),
                             funlabel=c("10-year Survival Probability"))
options(repr.plot.width=10, repr.plot.height=6)
plot(nom, xfrac=.2)