Aims and objectives
The degree of mammographic breast density (i.e.
the proportion of fibroglandular or “dense” tissue in the breast) has a considerable impact on the sensitivity of screening mammography.
In women with very fatty breasts,
sensitivity is often reported between 75% and 98%,
whereas in dense breasts,
sensitivity has been reported as low as 48% [1-4].
A large-scale trial of digital mammography (DMIST) noted sensitivity was less than 60% in younger,
pre- and perimenopausal women with dense breasts,
compared to 86% in women with fatty breasts [5]....
Methods and materials
Image Dataset
200 full-field digital mammography cases (all from Hologic Selenia Systems),
acquired between 2003 and 2008 from the Preventicon screening centre in Utrecht (the Netherlands),
were used in this retrospective reader study.
All images were anonymized and institutional review board approval was waived.
Only cases with bilateral craniocaudal (CC) and mediolateral oblique (MLO) views were selected.
The final dataset consisted of 63 randomly selected screen-detected cancers,
17 cancers that were retrospectively judged to be visible,
20 false-positive referrals,
and 100 non-cancer cases.
Interpretation of...
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...
Conclusion
In the United States,
breast cancer screening typically comprises annual mammography and single reading with CAD,
whereas European guidelines recommend bi-ennial mammography with independent double reading.
Double interpretation of mammograms is a proven,
effective strategy to improve cancer detection rates [6].
The significant improvement in sensitivity that was observed in this study (for up to 12 readers) is in agreement with previous work [7],
and indicates that there is wide scope for improving detection rates.
The benefits of additional readers on sensitivity was more pronounced...
Personal information
Nico Karssemeijer,
PhD,
Department of Radiology,
Radboud University Nijmegen Medical Centre,
Netherlands;
[email protected]
Rianne Hupse,
PhD,
Department of Radiology,
Radboud University Nijmegen Medical Centre,
Netherlands;
[email protected]
Gerard J den Heeten,
MD,
PhD; Academic Medical Centre,
Amsterdam,
Netherlands and National Expert and Training Centre for Breast Cancer Screening,
Nijmegen,
Netherlands;
[email protected]
Marc B Lobbes,
MD,
PhD; Maastricht University Medical Center,
Maastricht,
Netherlands;
[email protected]
Ritse M Mann,
MD,
PhD; Department of Radiology,
Radboud University Nijmegen Medical Centre,
Netherlands;
[email protected]
References
[1] Carney et al.
Ann Intern Med.
2003; 138: 168-175
[2]Kolb et al.
Radiology.
2002; 225:165-175
[3] Elmore et al.
Arch Int Med.
2004 ; 164: 1140-1147
[4] Kerlikowske et al.
Ann Intern Med.
2011; 155: 493-502
[5]Pisano et al.
Radiology.
2008; 246:376-383
[6] Helvie.
Radiol Clin North Am.
2007; 801-811
[7] Karssemeijer et al.
Proc.
SPIE 5372,
Medical Imaging 2004: Image Perception,
Observer Performance,
and Technology Assessment,
82 (May 4,
2004); doi:10.1117/12.535225