The objective of this review is to report and discuss the evidence for fully automated
diagnostic instruments for cutaneous melanoma tested in a real-world clinical setting
directly compared with human diagnosis. A systematic review was performed and articles
excluded when studies did not report sensitivity or specificity for melanoma directly
compared with humans on an independent test set. Only 3 instruments have had their
diagnostic accuracy compared with a human diagnosis in the clinical field with a meaningful
sample size that could allow some generalization with the wider clinical arena. Two
of these instruments showed a significantly inferior specificity for the diagnosis of melanoma
compared with specialists. In one of these studies, the sensitivity for diagnosis,
although superior to the specialist diagnosis, did not reach statistical significance. In
contrast, one instrument had an equivalent specificity and trended superior but not significantly
for sensitivity for the diagnosis of melanoma. Other image based nonclinic studies
and studies comparing clinical management as the endpoint rather than diagnosis are also
reviewed.
Semin Cutan Med Surg 27:32-36 © 2008 Elsevier Inc. All rights reserved.