Table 2 Summary of the diagnostic accuracy of screening instruments and evidence sources.
From: The cost-effectiveness of risk-stratified breast cancer screening in the UK
Screening instrument | Sensitivity | Source and details of calculation | FP rate per screen | Source and details of calculation |
|---|---|---|---|---|
Mammogram (MAM) | Screening sensitivity estimated using a logistic function. Maximum sensitivity assumed to be 93%. | logistic function [22] that depends on tumour diameter d (in mm), percent density m, and their interaction m/d2. The formulae is: S(d,m)= [exp(β1 + β2d +β3 m +β4 m/d2)] / [1 + exp(β1 + β2d + β3 m +β4 m/d2)] β1 = − 4.38 β2 = 0.49 β3 = − 1.34 β4 = − 7.18 | Prevalent screen: 7.04% Incident screens: 2.23% | NHSBSP data [25] |
Automated ultrasound | 47.37% relative increase applied to the sensitivity of MAM | BRAID trial detection rate of cancers with automated ultrasound compared to MAM | Prevalent screen: 7.04% Incident screens: 2.23% | Assumed to be identical to MAM [45] |
Handheld ultrasound | 10.53% decrease applied to the sensitivity of AUS | Reflects the reduction in the comparative sensitivity of handheld ultrasound to automated ultrasound found in a sample of nearly 400 women [29] | Prevalent screen: 7.04% Incident screens: 2.23% | Assumed to be identical to MAM for handheld ultrasound alone [45] Hence, MAM combined with handheld ultrasound will have doubling of the false positive rate of MAM which is also found in a meta-analysis of studies [46] |
Contrast-enhanced spectral mammogram | 54.55% relative increase applied to the sensitivity of MAM | BRAID trial detection rate of cancers with contrast-enhanced spectral MAM compared to MAM | Prevalent screen: 7.14% Incident screens: 2.59% | contrast-enhanced spectral MAM has lower specificity compared to MAM in high breast density woman of 15.92% [47] |
Abbreviated MRI | 61.57% relative increase applied to the sensitivity of MAM | BRAID trial detection rate of cancers with abbreviated MRI compared to MAM | Prevalent screen: 7.32% Incident screens: 2.65% | Abbreviated MRI has lower specificity compared to MAM in high breast density woman of 18.82% [48] |
Full MRI | 2.22% relative increase applied to the sensitivity of AMRI | The pooled sensitivity for screening studies [30] was 0.90 for abbreviated MRI and 0.92 for full MRI | Prevalent screen: 7.60% Incident screens: 2.75% (for combined screening of MAM with MRI) | In the only study [49] that compared MAM to MAM with supplemental MRI the addition of MRI led specificity to increase by 23.31% |
