This poster is published under an
open license. Please read the
disclaimer for further details.
Keywords:
Breast, Mammography, Observer performance, Cancer
Authors:
N. Karssemeijer1, R. Hupse1, G. J. den Heeten1, M. Lobbes2, R. M. Mann1; 1Nijmegen/NL, 2Maastricht, LB/NL
DOI:
10.1594/ecr2014/C-0358
Results
Examples of screening studies categorized as non-dense and dense are shown in Figures 1 and 2,
respectively.
Increasing the number of readers improved cancer detection rates.
For non-dense cases,
the mean TPF increased from 0.68 for a single reader,
to 0.74,
0.77,
and 0.78 for two,
three and four readers,
respectively (see Figure 3).
Similarly,
in dense cases,
the mean TPF increased from 0.57 for a single reader to 0.64,
0.67 and 0.69 for two,
three and four readers,
respectively.
Interpretation of dense mammograms by three readers,
increased cancer detection rates to levels that were comparable with a single reader in non-dense breasts.
Notably,
the incremental gain in sensitivity that was observed for up to 9 readers in dense breasts,
was not observed for non-dense mammograms (see Table 1).
There was very little improvement in detection performance for more than four readers in non-dense cases,
indicating a plateau in sensitivity gain.
Compared to a single reader,
double reading increased sensitivity by 9.6% and 12.2% for non-dense and dense cases,
respectively,
whereas,
the maximum increase in sensitivity was 29.7% (12 readers) and 15.5% (six readers),
respectively.