Back in
Freer and Ulissey [52] using a database
containing
patients conclude that the use of CAD in the
interpretation of screening mammograms can increase the detection
of early-stage malignancies. However, results published in
using a database of
women claimed to show that CAD, in its
present form, is not effective in that there was no significant
difference observed in cancer detection rates with CAD compared
with reading mammograms without CAD [86].
These are only two examples of an still open debate (see for
instance the work of Malich et al. [110] for more
examples), with people saying that actual CADs are effective and
people not agree with such a statement [126]. People
agree that CADs are available to detect cancers clinically missed.
However the main drawback of such systems is the number of false
positives obtained, which makes the radiologist not really trust
them [126]. This is a major issue for the low number
of malignancies within the screening population, which is supposed
to be around
out of
screened cases [172].
The full benefits of computer-aided systems will be seen in the future, when both the internal algorithms for abnormality detection and diagnosis and the preparation of the experts in reading mammograms with CAD will be improved. We are sure that CAD will become the standard of care.