|
In
recent years, many methods have been suggested for the inspection of
moles oriented to the early diagnosis of melanoma. The traditional
pattern analysis of dermoscopic images provides a very high accuracy
but requires a high level of experience. The conclusion that an
inexperienced dermatologist could be confused by the new semiology
pushed many researchers to study new mechanicistic models through a non
linear combination of criteria. The evaluation of pigmented skin
lesions through subjective assessments, however, is often complex and
not reproducible. Subjective quantifications based on crispy
definitions, as employed for example in the ABCD rule and 7-Point
checklist revealed some limits in the daily practice. However the
dermoscopy itself provides a
useful tool for the early diagnosis of melanoma and many expert
dermatologists are using the pattern analysis which
provides powerful inspection techniques able to reveal early
diagnostic signals. Dermoscopy provides high accuracy and
dermatologists become more familiar with this non-invasive technique.
The experts in this field think that dermoscopy needs more
standardization in the definitions. A possible solution could be the
objective analysis. In order to overcome the
problems due to the qualitative interpretation, methods based on the
mathematical analysis of pigmented skin lesions have been recently
developed. This scientific review relates to the results obtained
through the models developed by M. Burroni, L. Andreassi, G. Dell'Eva,
P. Rubegni and colleagues. Each paragraph of this site reports a link
to the published papers. Our experience goes from the enhanced visual
inspection to the aided diagnosis. Please note how many works focus on
the objective definition and semiological studies more than on aided
diagnosis: our main goal was, infact, to provide robust and
reproducible variables able to scientifically assess the pigmentary
disorders. Image quality, reliable measurements and reproducibility
represent the starting point of our research. It should be clear how the
results related to the many studies performed using the DB-Mips System
and the many different papers indicate always very similar
trends with a mean Sensitivity close to 97% and a Specificity higher
than 85%. Here below the scientific
abstracts concerning each topic and the PubMed links have been reported.
|
|