Skip to main content
. Author manuscript; available in PMC: 2018 May 11.
Published in final edited form as: Int J Biostat. 2016 Nov 1;12(2):/j/ijb.2016.12.issue-2/ijb-2016-0017/ijb-2016-0017.xml. doi: 10.1515/ijb-2016-0017

Table 3.

This table gives the ratios of statistics for digital mammography with tomosynthesis to digital mammography alone for various assumed values of prevalence πp using data from Rose et al. [12]. The ratios are specificities (rSpec, eq. [11]), areas under a power-law ROC curve (rAUCP, eq. [18]), and areas under an equal-variance bi-normal ROC curve (rAUCN). The estimated standard errors of these ratios are in the third and seventh columns. The fifth column gives the actual differences between the estimated values and the center value estimated at πp = 0.008. The fourth column gives estimates of these differences using the bias expressions (13) and (19).

Prev.
πp 		rSpec Standard
error		 Difference
estimate		 Measured
difference
0.0055 1.03728 0.0060 −0.00010 −0.00010
0.0080 1.03739 0.0060 0.0 0.0
0.0110 1.03751 0.0060 0.00012 0.00012
rAUCP rAUCN Std. Err.
0.0055 1.102 0.08 −0.0119 −0.0088 1.08 0.04
0.0080 1.111 0.07 0.0 0.0 1.11 0.05
0.0110 1.117 0.07 0.0067 0.0064 1.14 0.05