Dr. Leonidas Bantis, Assistant Professor in the Department of Biostatistics and Data Science, University of Kansas Medical Center
Evaluation of the overall accuracy of biomarkers might be based on average measures of the sensitivity for all possible specificities ‐and vice versa‐ or equivalently the area under the receiver operating characteristic (ROC) curve that is typically used in such settings (a.k.a. AUC). Apart from this overall accuracy measure, it is often the case that clinical questions focus on pre-defined specificity or sensitivity levels so that over or under-diagnosis is avoided. In other settings, investigators are in need of cutoff points to determine whether intervention is required after establishing the utility of a continuous biomarker or their combinations. We will discuss all these subjects and further expand in 3-class problems that refer to trichotomous settings.
Collin J. McKinney, Director,Electronics Design Core Laboratory & Nano‐Device Characterization Laboratory (NDCL), University of North Carolina at Chapel Hill, Department of Chemistry
Download the webinar slide presentation.