In real life
A fasting blood sugar test can be read at different cut‑offs. Set the cut‑off low, and you catch nearly every case of diabetes (high sensitivity) but label many healthy adults as abnormal. Set it high, and almost everyone labelled abnormal really has diabetes (high specificity), but you miss some cases. Choosing the cut‑off is a trade‑off, and the ROC curve shows it in one picture.