Premium
Estimation of smooth ROC curves for biomarkers with limits of detection
Author(s) -
Bantis Leonidas E.,
Yan Qingxiang,
Tsimikas John V.,
Feng Ziding
Publication year - 2017
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.7394
Subject(s) - receiver operating characteristic , normality , parametric statistics , computer science , sensitivity (control systems) , false positive rate , nonparametric statistics , cancer detection , statistics , pattern recognition (psychology) , mathematics , artificial intelligence , cancer , medicine , electronic engineering , engineering
Protein biomarkers found in plasma are commonly used for cancer screening and early detection. Measurements obtained by such markers are often based on different assays that may not support detection of accurate measurements due to a limit of detection. The R O C curve is the most popular statistical tool for the evaluation of a continuous biomarker. However, in situations where limits of detection exist, the empirical R O C curve fails to provide a valid estimate for the whole spectrum of the false positive rate (FPR). Hence, crucial information regarding the performance of the marker in high sensitivity and/or high specificity values is not revealed. In this paper, we address this problem and propose methods for constructing R O C curve estimates for all possible F P R values. We explore flexible parametric methods, transformations to normality, and robust kernel‐based and spline‐based approaches. We evaluate our methods though simulations and illustrate them in colorectal and pancreatic cancer data.