Dichotomous classification of measurements of disease markers, useful particularly for improved diagnosis of lung cancer, using a set of classifiers

摘要

Dichotomous classification of measurement values in medical analysis. Dichotomous classification of measurement values in medical analysis comprises: (1) generating a set of at least two measurement values (MV); (2) generating a basic classifier and using this to produce N classifiers (Kn); (3) identifying a classifier (K*) that allows neither positive or negative classification of MV; (4) determining the positive and negative predictive value on the basis of the sensitivity (SE*) and specificity (SP*) of K*; and (5) classifying MV based on the positive/negative predictive values. Independent claims are also included for: (1) diagnosing diseases by measuring indicator substances (IS) and application of the new method; and (2) device for quantitative determination of IS in body fluids.