Diagnostic performance and inter-observer concordance in lesion detection with the automated breast volume scanner (ABVS)

Table 7 Model calculation

	Interpretation of the ABVS data by 2 raters
Performance of the ABVS in our study collective (n = 100)
Prevalence of disease	18.00%
SE	83.33%
SP	78.05%
Rate of second look ultrasounds in healthy women (=100%-SP)	21.95%
Estimated performance of the ABVS in a screening collective of asymptomatic women ¹
Prevalence of disease ²	0.40%
PPV ³	1.49%
NPV ³	99.91%
FNR ^3,4	0.09%

Based on our data concerning the performance with two examiners (second-reading of the ABVS data sets), the effect in a screening setting was estimated. (SE = sensitivity; SP = specificity; PPV = positive predictive value; NPV = negative predictive value; FNR = false negative rate).
¹Asymptomatic women are defined as women, who have a normal mammogram and exhibit no symptoms
²As described in the literature, the prevalence of occult carcinomas that can be detected by conventional breast ultrasound can be estimated to be about 0.4%. This value resembles the theoretical false negative rate (FNR) of clinical examination and mammography alone.
³The new performance values of the ABVS are based upon the estimated prevalence of 0.4% in a screening collective.
⁴The new false negative rate is based upon a combination of clinical examination, mammography and ABVS in a screening collective.

ISSN: 1471-2342