×

Evaluating prognostic accuracy of biomarkers under competing risk. (English) Zbl 1251.62054

Summary: To develop more targeted intervention strategies, an important research goal is to identify markers predictive of clinical events. A crucial step toward this goal is to characterize the clinical performance of a marker for predicting different types of events. We present statistical methods for evaluating the performance of a prognostic marker in predicting multiple competing events. To capture the potential time-varying predictive performance of the marker and incorporate competing risks, we define time- and cause-specific accuracy summaries by stratifying cases based on causes of failure. Such definition would allow one to evaluate the predictive accuracy of a marker for each type of event and compare its predictiveness across event types. Extending the nonparametric crude cause-specific receiver operating characteristics curve estimators by P. Saha and P.J. Heagerty [ibid. 66, No. 4, 999–1011 (2010; Zbl 1274.62865)], we develop inference procedures for a range of cause-specific accuracy summaries. To estimate the accuracy measures and assess how covariates may affect the accuracy of a marker under the competing risks setting, we consider two forms of semiparametric models through the cause-specific hazard framework. These approaches enable a flexible modeling of the relationships between the marker and failure times for each cause, while efficiently accommodating additional covariates. We investigate the asymptotic property of the proposed accuracy estimators and demonstrate the finite sample performance of these estimators through simulation studies. The proposed procedures are illustrated with data from a prostate cancer prognostic study.

MSC:

62P10 Applications of statistics to biology and medical sciences; meta analysis
62N02 Estimation in survival analysis and censored data
62G08 Nonparametric regression and quantile regression
62G05 Nonparametric estimation
65C60 Computational problems in statistics (MSC2010)

Citations:

Zbl 1274.62865

Software:

cmprsk
PDFBibTeX XMLCite
Full Text: DOI Link

References:

[1] Cai, The sensitivity and specificity of markers for event time, Biostatistics 7 pp 182– (2006) · Zbl 1169.62367 · doi:10.1093/biostatistics/kxi047
[2] Cheng, Prediction of cumulative incidence function under the proportional hazards model, Biometrics 54 pp 219– (1998) · Zbl 1058.62593 · doi:10.2307/2534009
[3] Dabrowska, Smoothed Cox regression, The Annals of Statistics 25 pp 1510– (1997) · Zbl 0936.62046 · doi:10.1214/aos/1031594730
[4] Fine, Regression modeling of competing crude failure probabilities, Biostatistics 2 pp 85– (2001) · Zbl 1022.62113 · doi:10.1093/biostatistics/2.1.85
[5] Fine, A proportional hazards model for the subdistribution of a competing risk, Journal of the American Statistical Association 94 pp 496– (1999) · Zbl 0999.62077 · doi:10.1080/01621459.1999.10474144
[6] Heagerty, Semiparametric estimation of regression quantiles with application to standardizing weight for height and age in US children, Applied Statistics 48 pp 533– (1999) · Zbl 1058.62520
[7] Heagerty, Survival model accuracy and roc curves, Biometrics 61 pp 92– (2005) · Zbl 1077.62077 · doi:10.1111/j.0006-341X.2005.030814.x
[8] Heagerty, Time-dependent ROC curves for censored survival data and a diagnostic marker, Biometrics 56 pp 337– (2000) · Zbl 1060.62622 · doi:10.1111/j.0006-341X.2000.00337.x
[9] Li, ROC analysis with multiple classes and multiple tests: Methodology and its application in microarray studies, Biostatistics 9 pp 566-576– (2008) · Zbl 1143.62083 · doi:10.1093/biostatistics/kxm050
[10] Moskowitz, Quantifying and comparing the accuracy of binary biomarkers when predicting a failure time outcome, Statistics in Medicine 23 pp 1555– (2004) · doi:10.1002/sim.1747
[11] Park, Estimating subject-specific survival functions under the accelerated failure time model, Biometrika 90 pp 717– (2003) · Zbl 1436.62478 · doi:10.1093/biomet/90.3.717
[12] Pencina, Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond, Statistics in Medicine 27 pp 157– (2008) · doi:10.1002/sim.2929
[13] Pepe, Kaplan-Meier, marginal or conditional probability curves in summarizing competing risks failure time data?, Statistics in Medicine 12 pp 737– (1993) · doi:10.1002/sim.4780120803
[14] Pepe, Evaluating the ROC performance of markers for future events, Lifetime Data Analysis 14 pp 86– (2008) · Zbl 1169.62385 · doi:10.1007/s10985-007-9073-x
[15] Pollard, Empirical Processes: Theory and Applications (1990)
[16] Prentice, The analysis of failure times in the presence of competing risks, Biometrics 34 pp 541– (1978) · Zbl 0392.62088 · doi:10.2307/2530374
[17] Saha, Time-dependent predictive accuracy in the presence of competing risks, Biometrics 66 pp 999– (2010) · Zbl 1274.62865 · doi:10.1111/j.1541-0420.2009.01375.x
[18] Scheike, Predicting cumulative incidence probability by direct binomial regression, Biometrika 95 pp 205-220– (2008) · Zbl 1437.62602 · doi:10.1093/biomet/asm096
[19] Tsiatis, A nonidentifiability aspect of the problem of competing risks, Proceedings of the National Academy of Sciences of the United States of America 72 pp 20-22– (1975) · Zbl 0299.62066 · doi:10.1073/pnas.72.1.20
[20] Zheng, Prospective accuracy for longitudinal markers, Biometrics 63 pp 332– (2007) · Zbl 1140.62086 · doi:10.1111/j.1541-0420.2006.00726.x
[21] Zheng, Time-dependent predictive values of prognostic biomarkers with failure time outcome, Journal of the American Statistical Association 103 pp 362– (2008) · Zbl 1469.62385 · doi:10.1198/016214507000001481
[22] Zheng, Semiparametric models of time-dependent predictive values of prognostic biomarkers, Biometrics 66 pp 50– (2010) · Zbl 1187.62185 · doi:10.1111/j.1541-0420.2009.01246.x
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.