Numerous diagnostic tests report their results quantitatively, using continuous scales. Receiver operating characteristic curve (ROC) analysis provides a statistical method for the assessment of the diagnostic accuracy of these tests, being used for three specific purposes: determine of the cutoff value with the highest sensitivity and specificity, evaluate the discriminative capacity of the diagnostic test, in other words, its ability to differentiate healthy versus sick individuals, and compare the discriminative capacity of two or more diagnostic tests that express their results as continuous scales. Based on a real clinical investigation, this article illustrates theoretical aspects regarding the construction of ROC curves, being its objective to help readers and investigators interpret correctly their results.