Skip to main content
. 2013 Jul 19;26(5):989–994. doi: 10.1007/s10278-013-9616-5

Table 2.

Proportion of data collection errors flagged by the NLP algorithm

Study type NLP % (95 % confidence interval), proportion
Screening mammography Not applicable, 0/0
Diagnostic mammography 100.0 % (2.5, 100.0 %), 1/1
Breast ultrasound 96.0 % (79.7, 99.9 %), 24/25
Combined diagnostic mammography and ultrasound 100.0 % (39.8, 100.0 %), 4/4
Breast MRI 66.7 % (29.9, 92.5 %), 6/9
All breast imaging 89.7 % (75.8, 97.1 %), 35/39