Seeking Alpha

Shares of MELA Sciences (MELA) are on watch Monday after a NY Times article highlights the...

Shares of MELA Sciences (MELA) are on watch Monday after a NY Times article highlights the debate around MelaFind, the company's FDA approved melanoma-finding computer vision device which uses algorithms to pick out irregularities in pigmented skin legions. "My concern with MelaFind is that it just says everything is positive," one biostatistician who sat on the FDA review panel which evaluated the device says. "It will err on the side of caution," counters the company's director of commercialization. It will be interesting to see if the system can gain traction in what is quickly becoming a tough market for medical devices (I, II).
Comments (4)
  • B_Banzai
    , contributor
    Comments (283) | Send Message
    Interesting NY Time article. I imagine this sort of debate typifies every new technology doctors adopt. Comments: 1) 2.5 mm (=1/10") is a lot deeper than a doctor can see without MelaFind; 2) As the technology (and competing technology) is adopted, it is concievable that this will put pressure on non-users who might be considered negligent for not using it. 3) Over time, I would expect MELA to update it's calibration standards to weed out false positives - would MELA need FDA approval to do deviate from setting used in their clinical trials? (BTW: I own no shares at this time although MELA has been on my watch list for almost a year.)
    21 Jul 2013, 03:40 PM Reply Like
    , contributor
    Comments (2) | Send Message
    The article reminds me of the argument over whether the world was flat or not; or if Amazon was a good idea or not in the beginning?


    The negative tone of the article (which almost the entire article is) points out that the technology is not perfect. It missed 2 of 127 melanomas. At the same time the article says Melafind identifies too many false positives.


    I think the example in the article to make the point is totally ridiculous ... "compared the accuracy of MelaFind in distinguishing non-melanomas to a hypothetical pregnancy test which, used on 100 nonpregnant women, would mistakenly conclude that 90 of them were pregnant."


    That suggests a 90% inaccuracy rate, and only 10% accuracy specifically to say they are not pregnant.


    The article did not address that dermatologists miss 20% of melanomas using their current visual methods. Missing only 2 of 127 seems better to me.


    The article points out that MelaFind does not identify basal and squamous cell carcinomas. That infers that MelaFind isn't good enough because it only points out the melanomas that can kill you. Shouldn't the dermatologist be looking for the other two since they know Melafind does not identify them?


    The one disturbing thing is the representation that only 150 devices have been sold. In March we had data from the company saying 400 devices were in operation. Were 250 devices distributed for free and non-paying for tests, or were the numbers inaccurate? Another report will be coming out from the company for the latest quarter soon.


    Mela still has two major hurdles. They have to hire a new CEO and provide enough data to insurance companies to get them to cover the tests.
    21 Jul 2013, 09:45 PM Reply Like
  • troublesacomin
    , contributor
    Comments (37) | Send Message
    Thanks for taking the time to type this.
    2 Aug 2013, 04:56 PM Reply Like
  • thomasj37
    , contributor
    Comments (6) | Send Message
    Why hasn't the FDA approved Versiante Technologies equipment yet? They are selling it all over the rest of the world. Is this technology any more or less accurate than Melafind ????
    22 Jul 2013, 01:51 AM Reply Like
DJIA (DIA) S&P 500 (SPY)
ETF Tools
Find the right ETFs for your portfolio:
Seeking Alpha's new ETF Hub
ETF Investment Guide:
Table of Contents | One Page Summary
Read about different ETF Asset Classes:
ETF Selector