列线图
最后发布时间:2022-01-15 21:50:13
浏览量:
根据中位值分组
expr_gene$Overall_Survival_Status <- as.numeric(expr_gene$Overall_Survival_Status)
diff <- survdiff(Surv(days_to_last_followup,Overall_Survival_Status)~low_high,data = expr_gene)
单因素生分析
Nomogram绘制
library(survival)
library(rms)
library(tidyverse)
data <- read.csv("Liver_clinical_arrange.csv") %>%
mutate(age = ageNum, riskScore = riskScoreNum)
View(data)
dd<-datadist(data)
options(datadist=dd)
f <- cph(Surv(futime_year,fustat) ~ age + stage+T+riskScore,data=data, x=T, y=T, surv=T)
surv <- Survival(f)
nom <- nomogram(f,fun = list(function(x)surv(1,x),function(x)surv(3,x),function(x)surv(5,x)),
lp=F,
funlabel = c("1-year Survival Probability","3-year Survival Probability","5-year Survival Probability"))
png(filename = "nomogram.png",width = 9,height = 8,res = 300,units = "in")
plot(nom)
dev.off()
预测模型校准曲线
http://www.360doc.com/content/18/0714/09/46405145_770258087.shtml
用于术前预测结直肠癌淋巴结转移的放射组学列线图的开发和验证
做了预测模型校准曲线 (Calibration curve),两个图分别是建模队列和验证队列。图的横坐标是预测概率:用预测模型对事件发生的可能性(Probability)进行预测,0到1表示发生事件可能性是0到100%。纵坐标是实际概率:患者实际的事件发生率。红色的线是拟合线,表示预测值对应的实际值。
- 如果预测值=实际值,则红线与参考线(蓝线)完全重合;
- 如果预测值>实际值,即高估了风险,则红线在蓝线上面;
- 如果预测值<实际值,即低估了风险,则红线在蓝线上面;