Most diagnostic accuracy measures and criteria for selecting optimal cut-points are only applicable to diseases with binary or three stages. Currently, there exist two diagnostic measures for diseases with general k stages: the hypervolume under the manifold and the generalized Youden index. While hypervolume under the manifold cannot be used for cut-points selection, generalized Youden index is only defined upon correct classification rates. This paper proposes a new measure named maximum absolute determinant for diseases with k stages (). This comprehensive new measure utilizes all the available classification information and serves as a cut-points selection criterion as well. Both the geometric and probabilistic interpretations for the new measure are examined. Power and simulation studies are carried out to investigate its performance as a measure of diagnostic accuracy as well as cut-points selection criterion. A real data set from Alzheimer’s Disease Neuroimaging Initiative is analyzed using the proposed maximum absolute determinant.
ShapiroD. The interpretation of diagnostic tests. Stat Methods Med Res1999; 8: 113–134.
2.
ZhouXHObuchowskiNMcClishD. Statistical methods in diagnostic medicine, Wiley: New York, 2002.
3.
PepeMS. The statistical evaluation of medical tests for classification and prediction, New York: Oxford University Press, Oxford Statistical Science Series Vol. 28, 2003.
4.
ZouKHLiuABandosAIet al.Statistical evaluation of diagnostic performance topics in ROC analysis, Chapman & Hall/CRC Biostatistics Series, 2011.
5.
BamberD. The area above the ordinal dominance graph and the area below the receiver operating characteristic graph. J Math Psychol1975; 12: 387–415.
6.
HanleyJAMcNeilBJ. The meaning and use of the area under the receiver operating characteristic (ROC) curve. Radiology1982; 143: 29–36.
7.
DeLongEDeLongDClarke-PearsonC. Comparing the area under two or more correlated ROC curves: a non-parametric approach. Biometrics1988; 44: 837–845.
8.
WieandSGailMHJamesBRet al.A family of nonparametric statistics for comparing diagnostic markers with paired and unpaired data. Biometrika1989; 76: 585–592.
9.
FaraggiDReiserB. Estimation of the area under the ROC curve. Stat Med2002; 21: 3093–3106.
10.
ObuchowskiN. Fundamentals of clinical research for radiologists: ROC analysis. Am J Roentgenol2005; 184: 364–372.
11.
MolodianovitchKFaraggiDReiserD. Comparing the areas under two correlated ROC curves: parametric and non-parametric approaches. Biometr J2006; 48: 745–757.
12.
ZouKHO’MalleyAJMauriL. Receiver operating characteristic analysis for evaluating diagnostic tests and predictive models. Stat Primer Cardiovasc Res2007; 115: 654–657.
YoudenWJ. Index for rating diagnostic tests. Cancer1950; 3: 32–35.
15.
SchistermanEFPerkinsNJLiuAet al.Optimal cut-point and its corresponding Youden Index to discriminate individuals using pooled blood samples. Epidemiology2005; 16: 73–81.
16.
FlussRFaraggiDReiserB. Estimation of the Youden index and its associated cutoff point. Biometr J2005; 47: 458–441.
17.
SchistermanEFPerkinsNJ. Confidence intervals for the Youden Index and corresponding optimal cut-point. Commun Stat: Simulation Comput2007; 36: 549–563.
18.
LaiCTianLSchistermanEF. Exact confidence interval estimation for the Youden index and its corresponding optimal cut-point. Comput Stat Data Anal2010; 56: 1103–1114.
19.
ScintoLFMDaffnerKR. Early diagnosis of Alzheimer’s disease, Totowa, NJ: Humana Press, 2000.
XiongCJBelleGVMillerJPet al.Measuring and estimating diagnostic accuracy when there are three ordinal diagnostic groups. Stat Med2006; 25: 1251–1273.
22.
AlonzoTANakasTCYiannoutsosCTet al.A comparison of tests for restricted orderings in the three-class case. Stat Med2009; 28: 1144–1158.
23.
MossmanD. Three-way ROCs. Med Decision Making1999; 19: 78–89.
24.
DreiseitlSOhno-MachadoLBinderM. Comparing three-class diagnostic tests by three-way ROC analysis. Med Decision Making2000; 20: 323–331.
25.
HeckerlingPS. Parametric three-way receiver operating characteristic surface analysis using Mathematica. Med Decision Making2001; 21: 409–417.
26.
NakasCTYiannoutsosCT. Ordered multiple-class ROC analysis with continuous measurement. Stat Med2004; 23: 3437–3449.
27.
XiongCJBelleGVMillerJPet al.A parametric comparison of diagnostic accuracy with three ordinal diagnostic groups. Biometr J2007; 49: 682–693.
28.
TianLXiongCLaiCet al.Exact confidence interval estimation for the difference in diagnostic accuracy with three ordinal diagnostic groups. J Stat Planning Inf2010; 141: 549–558.
29.
NakasCTAlonzoTAYiannoutsosCT. Accuracy and cut-off point selection in three-class classification problems using a generalization of the Youden index. Stat Med2010; 29: 2946–2955.
30.
Maturi TC, Elkhafifi FF and Coolen FPA. Nonparametric predictive inference for three-group ROC analysis. Technical Report 2013.
31.
LuoJXiongC. Youden index and associated cut-off points for three ordinal diagnostic groups. Commun Statist: Simul Comput2013; 42: 1213–1234.
32.
ScurfieldBK. Multiple-event forced-choice tasks in the theory of signal detectability. J Math Psychol1996; 40: 253–269.
33.
ScurfieldBK. Generalization of the theory of signal detectability to n-event m-dimensional forced-choice tasks. J Math Psychol1998; 42: 5–31.
34.
LiJFineJP. ROC analysis with multiple classes and multiple tests: methodology and its application in microarray studies. Biostatistics2008; 9: 566–576.
35.
NakasCTDalrymple-AlfordJCAndersonTJet al.Generalization of Youden index for multiple-class classification problems applied to the assessment of externally validated cognition in Parkinson disease screening. Stat Med2013; 32: 995–1003.
36.
PerkinsNJSchistermanEF. The inconsistency of ‘optimal’ cut-off points using two roc based criteria. Am J Epidemiol2006; 163: 670–675.
37.
LiuX. Classification accuracy and cut point selection. Stat Med2012; 31: 2676–2686.
38.
HeXFreyEC. Three-class ROC analysis – the equal error utility assumption and the optimality of three class ROC surface using the ideal observer. IEEE Trans Med Imaging2006; 25: 979–986.
39.
AttwoodKTianLXiongC. Diagnostic thresholds with three ordinal groups. J Biopharm Sta2014; 24: 608–633.
40.
Vinberg EB. A course in algebra. Rhode Island: American Mathematical Society, 2003.
41.
LloydCJYongZ. Kernel estimators of the ROC curve are better than empirical. Stat Probab Lett1999; 44: 221–228.
42.
HenekaMTNadrignyFRegenTet al.Locus ceruleus controls Alzheimer’s disease pathology by modulating microglial functions through norepinephrine. Proc Natl Acad Sci USA2010; 107: 6058–6063.
SkaltsaKJoverLFusterDet al.Optimum threshold estimation based on cost function in a multistate diagnostic setting. Stat Med2011; 31: 1098–1109.
45.
BattertonKASchubertCM. Confidence intervals around Bayes Cost in multi-state diagnostic settings to estimate optimal performance. Stat Med2014; 33: 3080–3099.
