Extracting the values from ROC curve

I could create a ROC curve for my Statistical analysis

library(pROC)
plot(roc(aSAH$outcome, aSAH$s100b,
    levels=c("Good", "Poor")))

However I would like plot the same using ggplot for additional styling, so I'm looking for some way on extracting the x-y values so that I could use them in ggplot.

I would appreciate for any pointer.

The output of roc(...) is a list-like structure of class "roc" which has members called "specificities" and "sensitivities". You could simply extract and use these for your plot:

proc_obj <- roc(aSAH$outcome, aSAH$s100b, levels=c("Good", "Poor"))

plot_df <- data.frame(Specificity = proc_obj$specificities, 
                      Sensitivity = proc_obj$sensitivities)

However, pROC contains the function coords specifically for the purpose of getting the sensitivities and specificities into a data frame, so it saves a few keystrokes to use that instead.

library(pROC)
library(ggplot2)

roc(aSAH$outcome, aSAH$s100b, levels=c("Good", "Poor")) |>
  coords() |>
  ggplot(aes(specificity, sensitivity)) +
  geom_step() +
  geom_abline(intercept = 1, slope = -1, colour = "gray") +
  coord_trans(x = "reverse", expand = FALSE, clip = "off") +
  theme_minimal(16) +
  theme(aspect.ratio = 1)

This is assuming that for some reason you don't want to create your ggplot object with ggroc, which is another function in pROC specifically for creating ggplot objects from ROC curves:

ggroc(roc(aSAH$outcome, aSAH$s100b, levels=c("Good", "Poor"))) +
  theme_minimal(16) +
  theme(aspect.ratio = 1) +
  geom_abline(intercept = 1, slope = 1, colour = "gray")