Replies to post #82746 on Biotech Values

[DewDiligence]

MELA:

I got the impression that they were pushing the product just for its ability to accurately rule out a patient having a melanoma and thus spare them from a biopsy. In those cases where it could not be ruled out, a patient would have a biopsy. Inasmuch as most moles do not turn out to be a melanoma, this would presumably significantly cut down on the number of unnecessary biopsies. So, it doesn't seem useless to me from that angle.

Sensitivity and specificity are meaningless metrics, individually—they must be considered as a pair to ascertain whether a diagnostic has any medical or commercial value.

For the sake of discussion, consider a hypothetical diagnostic that always reported “no melanoma.” The specificity of this test would be 100%, but the sensitivity would be 0%. The negative predictive value — how often a negative test result is correct — would be unknown.

[rkrw]

If you had a device with a 9% accuracy in calling a mole non-melanoma would you trust it or go ahead and get a biopsy done? That's what MELA will be trying to sell. This device is calling almost everything melanoma, that's why its sensitivity is so high.