The evaluation of diagnostics tests based on net benefit involves both the accuracy of the tests and the clinical consequences of the diagnostic errors. Also, the benefit-risk measures approach depends on the prevalence of the underlying disease. However, for some diseases or clinical conditions, the prevalence is either unknown or different from region to region or population to population, resulting in an erroneous diagnosis. This paper introduces innovative post-test diagnostic accuracy measures and a new cut-point selection criterion based on the expected relative net benefit. Our approach does not depend on the disease’s prevalence, maximizing net benefit and reducing the clinical consequences of the diagnostic errors. We demonstrate the advantages of the proposed measures to compare different diagnostic tests and/or biomarkers, on average, the abilities for rule-in, rule-out clinical condition, and as cut-point selection criterion that maximize the expected relative net benefit diagnostic accuracy. Numerical examples, simulation studies, and real data are provided to illustrate the superiority and applicability of the proposed measures.