A Stepwise Algorithm for Linearly Combining Biomarkers under Youden Index Maximization

Combining multiple biomarkers to provide predictive models with a greater discriminatory ability is discipline that has received attention in recent years. Choosing the probability threshold corresponds highest combined marker accuracy key disease diagnosis. The Youden index statistical metric provides an appropriate synthetic for diagnostic and good criterion choosing cut-off point dichotomize biomarker. In this study, we present new stepwise algorithm linearly combining continuous maximize index. To investigate performance of our algorithm, analyzed wide range simulated scenarios compared its five other linear combination methods literature (a approach introduced by Yin Tian, min-max approach, logistic regression, parametric under multivariate normality non-parametric kernel smoothing approach). obtained results show proposed showed similar algorithms normal outperforms all non-normal scenarios. same means different covariance matrix diseased non-diseased population, rest. were also applied on two real datasets (to discriminate Duchenne muscular dystrophy prostate cancer), whose higher cancer database.