Explanation:

To truly understand the potential benefit of implementing protein signatures as screening tests we must consider the background probability that an individual is at high-risk of the disease prior to testing (i.e. pre-test probability). The post-test probability will then tell us an individual’s chance of being at high-risk of the disease after screening through a standard clinical or proteomic test. This Bayesian framework is important to understand how the probability of being at high risk of the disease changes in individuals with a “positive” clinical or proteomic test. We use Fagan’s nomogram to show the post-test probability of being at high risk of a disease after performing screening through standard clinical models (without or with clinical assays) or a proteomic model, and how this increases compared to the pre-test probability. We allow for the user to change the background prevalence of the disease to enable assessing the potential to perform targeted screening in populations that are already known to be at a high absolute risk of specific diseases (e.g. people with autoimmune diseases at greater risk of coeliac disease). Fagan’s nomogram further integrates and presents the likelihood ratio. This represents the likelihood that a given test result would be expected in an individual at high risk of the disease compared to the likelihood that the same result is observed in an individual that is not at high-risk of the target disease. In other words, it is ratio of the probability that a test is correct to the probability that it isn’t. We present clinical models (without or with clinical assays) and a clinical + protein signature models in Fagan’s nomograms to illustrate the potential of proteomic signatures to improve screening strategies over and above what could be achieved by standard clinical models.