We propose new summary measures of biomarker accuracy which can be used as companions to existing diagnostic accuracy measures. Conceptually, our summary measures are tantamount to the so-called Hellinger affinity and we show that they can be regarded as measures of agreement constructed from similar geometrical principles as Pearson correlation. We develop a covariate-specific version of our summary index, which practitioners can use to assess the discrimination performance of a biomarker, conditionally on the value of a predictor. We devise nonparametric Bayes estimators for the proposed indexes, derive theoretical properties of the corresponding priors, and assess the performance of our methods through a simulation study. The proposed methods are illustrated using data from a prostate cancer diagnosis study.
YoudenW. Index for rating diagnostic tests. Cancer1950; 3: 32–35.
2.
Pepe M. The statistical evaluation of medical tests for classification and prediction. New York, NY: Oxford University Press, 2003.
3.
ParodiSPistoiaVMuselliM. Not proper ROC curves as new tool for the analysis of differentially expressed genes in microarray experiments. BMC Bioinform2008; 9: 410.
4.
Silva-FortesCTurkmanMAASousaL. Arrow plot: a new graphical tool for selecting up and down regulated genes and genes differentially expressed on sample subgroups. BMC Bioinform2012; 13: 147.
5.
WangDTianL. Parametric methods for confidence interval estimation of overlap coefficients. Computat Stat Data Analysis2017; 106: 12–26.
6.
LeeWCHsiaoC. Alternative summary indices for the receiver operating characteristic curve. Epidemiology1996; 7: 605–611.
7.
WilliamsD. Probability with martingales, Cambridge, UK: Cambridge University Press, 1991.
8.
Inácio de CarvalhoVJaraAHansonT, et al.Bayesian nonparametric ROC regression modeling. Bayes Analys2013; 8: 623–646.
9.
ErkanliASungMJane CostelloE, et al.Bayesian semi-parametric ROC analysis. Stat Med2006; 25: 3905–3928.
10.
GuJGhosalSRoyA. Bayesian bootstrap estimation of ROC curve. Stat Med2008; 27: 5407–5420.
11.
BranscumAJJohnsonWOBaronAT. Robust medical test evaluation using flexible Bayesian semiparametric regression models. Epidemiol Res Int2013. Article ID 131232.
Inácio de Carvalho V, Jara A and de Carvalho M. Bayesian nonparametric approaches for ROC curve inference. In: Mitra R and Müller P (eds) Nonparametric Bayesian inference in biostatistics. Cham: Springer, pp.327–344.
14.
BranscumAJohnsonWHansonT, et al.Flexible regression models for ROC and risk analysis, with or without a gold standard. Stat Med2015; 34: 3997–4015.
15.
Johnson W and de Carvalho M. Bayesian nonparametric biostatistics. In: Mitra R and Müller P (eds) Nonparametric Bayesian inference in biostatistics. Cham, Switzerland: Springer, 2015, pp.15–54.
16.
Inácio de CarvalhoVde CarvalhoMAlonzoT, et al.Functional covariate-adjusted partial area under the specificity-ROC curve with an application to metabolic syndrome diagnosis. Ann Appl Stat2016; 10: 1472–1495.
17.
Inácio de CarvalhoVBranscumAJ. Bayesian nonparametric inference for the three-class Youden index and its associated optimal cutoff points. Stat Meth Med Res2018; 27: 0962280217742538.
18.
de CarvalhoMPageGLBarneyBJ. On the geometry of Bayesian inference. Bayesian Analys2018. (in press).
19.
BhattacharyyaA. On a measure of divergence between two multinomial populations. Sankhyā: the Indian J Stat1946; 7: 401–406. .
20.
van der VaartA. Asymptotic statistics, Cambridge, UK: Cambridge University Press, 1998.
21.
HillisSLBerbaumKS. Using the mean-to-sigma ratio as a measure of the improperness of binormal ROC curves. Academic Radiol2011; 18: 143–154.
22.
ZhouXHMcClishDKObuchowskiNA. Statistical methods in diagnostic medicine, 2nd ed. New York, NY: Wiley, 2011.
23.
DorfmanDDBerbaumKSMetzCE, et al.Proper receiver operating characteristic analysis: the bigamma model. Academic Radiol1997; 4: 138–149.
24.
FergusonT. A Bayesian analysis of some nonparametric problems. Ann Stat1973; 1: 209–230.
25.
SethuramanJ. A constructive definition of Dirichlet priors. Statistica Sinica1994; 4: 639–650.
26.
MacEachern S. Dependent Dirichlet processes. Unpublished manuscript. Department of Statistics, The Ohio State University 2000.
27.
BarrientosAJaraAQuintanaF. On the support of MacEachern's dependent Dirichlet processes and extensions. Bayesian Analys2012; 7: 277–310.
28.
De IorioMMüllerPRosnerG, et al.An ANOVA model for dependent random measures. J Am Stat Assoc2004; 99: 205–215.
29.
De IorioMJohnsonWMüllerP, et al.Bayesian nonparametric nonproportional hazards survival modeling. Biometrics2009; 65: 762–771.
30.
Ghosal S and Van der Vaart A. Fundamentals of nonparametric Bayesian inference. vol. 44. Cambridge, UK: Cambridge University Press, 2017.
PanXMetzCE. The “proper” binormal model: parametric receiver operating characteristic curve estimation with degenerate data. Academic Radiol1997; 4: 380–389.
33.
WagnerRFMetzCECampbellG. Assessment of medical imaging systems and computer aids: a tutorial review. Academic Radiol2007; 14: 723–748.
34.
BandosAIGuoBGurD. Estimating the area under roc curve when the fitted binormal curves demonstrate improper shape. Academic Radiol2017; 24: 209–219.
35.
IshwaranHJamesLF. Gibbs sampling methods for stick-breaking priors. J Am Stat Assoc2001; 96: 161–173. .
36.
IshwaranHJamesLF. Approximate Dirichlet process computing in finite normal mixtures: smoothing and prior information. J Computat Graph Stat2002; 11: 508–532.
37.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2018, https://www.R-project.org/.
38.
EtzioniRPepeMLongtonG, et al.Incorporating the time dimension in receiver operating characteristic curves: a case study of prostate cancer. Med Decis Making1999; 19: 242–251.
39.
Inácio de CarvalhoVde CarvalhoMBranscumA. Nonparametric Bayesian covariate-adjusted estimation of the Youden index. Biometrics2017; 73: 1279–1288.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
