ggsurvplot技术问询：风险人数图例颜色匹配与线条替代需求

问题1：同步风险表格与曲线的颜色 #

你当前仅修改了生存曲线（p$plot）的颜色映射，但风险表格（p$table）的颜色未同步调整，导致两者不一致。只需给p$table添加相同的颜色映射即可：

# 原代码中定义颜色和线型
colors <-  rep(c("red", "red","blue","blue"))
lines <-  rep(c("solid","dashed","solid", "dashed"))

# 修改plot的颜色和线型
p$plot <- p$plot + scale_linetype_manual(values = lines) +
  scale_colour_manual(values = colors)

# 同步风险表格的颜色
p$table <- p$table + scale_colour_manual(values = colors)

问题2：用线条替代风险表格的图例（y轴文本） #

要将风险表格左侧的分组文本替换为对应线型/颜色的线条，可通过隐藏原y轴文本、添加自定义线段的方式实现：

# 获取分组名称
strata_names <- names(fit$strata)
# 构建线条数据：对应每个分组的y位置、线型、颜色
line_data <- data.frame(
  y = seq_along(strata_names),
  x_start = -0.1,
  x_end = 0,
  linetype = lines,
  color = colors
)

# 修改风险表格：隐藏原y轴文本，添加线条
p$table <- p$table +
  theme(axis.text.y = element_blank()) +
  geom_segment(data = line_data,
               aes(x = x_start, xend = x_end, y = y, yend = y,
                   linetype = I(linetype), color = I(color))) +
  xlim(-0.1, max(p$table$data$time))  # 调整x轴范围，确保线条显示

完整可运行代码 #

n <- 200  # 补充定义n的值
subjid <- 1:n

status_os <- sample(c(0,1), n, replace = TRUE)
time_os <- ifelse(status_os==0, rnorm(n,mean=365, sd=30), rnorm(n, mean=180, sd=30))

groupe <- sample(c(1,2), n, replace=TRUE)
mutation <- sample(c(1, 0), n, replace=TRUE)

data <- data.frame(Identifiant=subjid, status_os=status_os, time_os=time_os, groupe=groupe, mutation=mutation)

fit <- survfit(Surv(time_os, status_os)~groupe+mutation, data=data)

p <- ggsurvplot(fit, data = data, break.time.by = 12, risk.table = "nrisk_cumevents",
                legend.title = "", legend = c(0.2, 0.2), linetype = "strata",
                risk.table.y.text=TRUE, break.x.by = 50)

colors <-  rep(c("red", "red","blue","blue"))
lines <-  rep(c("solid","dashed","solid", "dashed"))

# 同步曲线和风险表格的颜色、线型
p$plot <- p$plot + scale_linetype_manual(values = lines) +
  scale_colour_manual(values = colors)
p$table <- p$table + scale_colour_manual(values = colors)

# 用线条替代风险表格的y轴文本
strata_names <- names(fit$strata)
line_data <- data.frame(
  y = seq_along(strata_names),
  x_start = -0.1,
  x_end = 0,
  linetype = lines,
  color = colors
)

p$table <- p$table +
  theme(axis.text.y = element_blank()) +
  geom_segment(data = line_data,
               aes(x = x_start, xend = x_end, y = y, yend = y,
                   linetype = I(linetype), color = I(color))) +
  xlim(-0.1, max(p$table$data$time))

# 合并绘图
p$plot <- p$plot / p$table + plot_layout(ncol = 1, heights = c(2, 1))

# 查看结果
print(p$plot)

说明 #

• 颜色同步：通过给p$table添加与p$plot相同的scale_colour_manual，确保风险表格的文本颜色和曲线颜色一致。
• 线条替代图例：利用geom_segment在风险表格左侧绘制对应线型和颜色的线段，同时隐藏原y轴文本，实现用线条标识分组的效果。

内容的提问来源于stack exchange，提问作者Guillaume