Trial Information

Mammography is currently the only documented effective imaging tool for breast cancer
screening. However, the sensitivity of mammography may be reduced in dense breasts, and
sometimes it is difficult to even perceive a very subtle cancer which presents as a small
stellate lesion, or very faint microcalcifications, missed diagnosed thus occurs. Herein,
some researchers in Western countries developed computer-aided detection (CAD) system to
help radiologists detect subtle, easily overlooked findings to facilitate early breast
cancer detection, and most of the research regarding CAD was used in screen-film mammography
(SFM) system. Ikeda, et al, worked on the retrospective CAD usage of those negative
mammograms which later developed breast cancers. CAD could correctly mark 40% of the areas
on these mammograms reported negative previously that later developed evident cancers.
However, 80% of these are only nonspecific findings, and do not warrant recall for
additional workup even at retrospective unblinded review by well-trained mammographers. The
other research concluded that CAD could improve early cancer detection rate of mammography,
with the sensitivity of 92% in detection of breast cancer size smaller than 5mm, 94% for
cancer size 11-15mm. CAD can detect more microcalcifications than masses (sensitivity for
microcalcifications 98%, masses 84%, mass with microcalcifications 92%). CAD could mark an
average of 1.3 false positive marks per mammographic exam.

Full-field digital mammography (FFDM) is a new approved technology for breast cancer
detection after SFM era since 2000. The diagnostic accuracy of FFDM versus SFM is still
under clinical trials, and it is believed the sensitivity and accuracy of FFDM for screening
population is relatively equivalent to SFM. However, there are very few reports regarding
the CAD application in FFDM, since FFDM can offer the post-acquisition processing on
high-resolution review workstation for interpretation. Nevertheless, the spatial resolution
of soft-copy reading on monitors for FFDM is slightly inferior to but the contrast
resolution is slightly superior to that of conventional SFM. Herein, the diagnostic efficacy
and role of CAD in FFDM are still unclear. Therefore, the goal of our study is to explore
the sensitivity, false-negative (FN) and false-positive (FP) rates of combination usage of
CAD in FFDM system, in comparison with the sensitivity, FN and FP rates of interpretation
based on FFDM without CAD combination. We are also about to evaluate the efficacy,
additional time spent in adjunct application CAD in FFDM interpretation, in order to assess
the feasibility of CAD in FFDM.