KM 生存曲线¶

用途：Kaplan-Meier 生存曲线，展示两组或多组的时间-事件数据，附风险表 (risk table)。

交互式图表¶

悬停查看各时间点的生存概率及置信区间。

生成代码¶

RPythonSAS

library(survival)
library(survminer)
library(plotly)

# 拟合 KM
fit <- survfit(Surv(avalu, event) ~ treatment, data = adtte)

# 用 survminer 优雅展示
g <- ggsurvplot(fit, data = adtte,
                pval = TRUE, conf.int = TRUE,
                risk.table = TRUE,
                palette = c("#1f77b4", "#ff7f0e"),
                xlab = "Time (months)", 
                ylab = "Survival Probability",
                legend.title = "Treatment")

import plotly.graph_objects as go
import numpy as np
from lifelines import KaplanMeierFitter

# 拟合两条 KM 曲线
kmf_treat = KaplanMeierFitter()
kmf_ctrl = KaplanMeierFitter()

treat = adtte[adtte.treatment == "Drug"]
ctrl = adtte[adtte.treatment == "Placebo"]

kmf_treat.fit(treat.avalu, treat.event, label="Drug")
kmf_ctrl.fit(ctrl.avalu, ctrl.event, label="Placebo")

# 画图
fig = go.Figure()
fig.add_trace(go.Scatter(
    x=kmf_treat.survival_function_.index,
    y=kmf_treat.survival_function_["Drug"],
    mode="lines", name="Drug",
    line=dict(color="#1f77b4", width=2)
))
# ... 添加另一组

fig.update_layout(
    title="Kaplan-Meier Survival Curve",
    xaxis_title="Time (months)",
    yaxis_title="Survival Probability",
    legend_title="Treatment"
)
fig.show()

proc lifetest data=adtte plots=survival(atrisk=0 to 24 by 6);
    time avalu * event(0);
    strata treatment;
run;

关键参数¶

置信区间: conf.int = TRUE，显示 Greenwood 区间
p 值: log-rank 检验结果
风险表: at-risk table 是必要的辅助信息

参考方法¶

生存分析