WPC-package: Implement Weighted Predictiveness Curve to Visualize the...
In WPC: Weighted Predictiveness Curve

Description Details Author(s) References See Also

The package includes ten functions and one example for illustration purpose.

Based on the nature of the data, it generates predictiveness curve by utilizing either parametric or nonparametric approaches. When the unique effect of biomarker value change is indeed multiplicative with respect to hazard rate, parametric WPC (based on COX proportional hazard model) can most efficiently estimate survival rate for each biomarker value. The function cox.wpc.est will generate the parametric WPC using Cox model, and returns point estimates, confidence intervals and their biomarker values.

For most real world cases, little is known about the data structure, and that's when nonparametric WPC should be considered. The estimates are based on a series of overlapping windows (subpopulation). There are two ways to generate the series of overlapping windows: by fixing the number of patients within each window and by fixing the biomarker scale window width. The function ns.windows uses the first approach and report the detailed information of each window, while the function ww.windows utilizes the second approach. The function npr.wpc.est incorporate those two functions and their associated parameters to generate nonparametric WPC. Similar to cox.wpc.est, the function npr.wpc.est returns point estimates, confidence intervals and their each biomarker values.

The primary functions in the package are SoloWPCCurve, DuoWPCCurve and TrioWPCCurve. They generates the graphs of single, double and triple weighted predictiveness curves based on the point estimates and confidence intervals reported by cox.wpc.est and npr.wpc.est.

The packages can be used to compare biomarkers and identify the one with the highest impact. Equally important, by simultaneously depicting several treatment-specific WPC curves, it is easy to detect treatment heterogeneity as well as treatment-specific patterns, which in turn will help us with subgroup selection and biomarker cut-off Optimization.

Index: This package was not yet installed at build time.

Hui Yang <huiy@amgen.com>, Rui Tang <rui_rui_tang@vrtx.com>, Jing Huang <jinghuang0@gmail.com>

Yang H., Tang R., Hale M. and Huang J. (2016) A visualization method measuring the performance of biomarkers for guiding treatment decisions Pharmaceutical Statistics, 15(2), 1539-1612

Therneau T (2013). A Package for Survival Analysis in S. R package version 2.37-4, http://CRAN.R-project.org/package=survival.

Christopher H. Jackson (2011). Multi-State Models for Panel Data: The msm Package for R. Journal of Statistical Software, 38(8), 1-29. http://www.jstatsoft.org/v38/i08/.

Package Survival

WPC documentation built on May 2, 2019, 6:52 a.m.