SoylentNews is people
SoylentNews
Machine learning can compete with dermatologists when it comes to diagnosing skin cancer, to an extent:
A group of Stanford researchers has trained one of Google's deep neural networks on a massive database of images that show skin lesions. By the end, the neural network was competitive with dermatologists when it came to diagnosing cancers using images. While the tests done in this paper don't fully represent the challenges a specialist would face, it's still an impressive improvement in computer performance.
[...] For the medical images, the authors relied on Stanford's extensive records focusing on skin diseases. In all, they arranged more than 2,000 individual disorders into a tree-like structure based on their relatedness. So for example, all inflammatory problems ended up on one branch of the tree, all the cancers on another. These were further subdivided until the branching reached individual diseases. Inception was then given the tree and a set of nearly 130,000 images of these disorders and was trained to properly identify each. That's over 100 times the number of images as were used for training in the largest previous study of this sort.
The authors then tested the basic classification system against two dermatologists, using a new set of images where the diagnosis had been confirmed by biopsy. On the most basic level of classification—benign, malignant, or a type called "non-neoplastic"—the accuracy of the neural network was over 70 percent while the doctors were in the 60s. When asked for a more detailed classification among nine categories, the neural network had an accuracy of about 55 percent, which is similar to the numbers put up by the dermatologists.
For a further test, the team put Inception up against 21 dermatologists, asking them to determine whether an image contained a benign or malignant lesion. Here, the neural network consistently edged out most of the doctors, and it consistently performed a bit better than their average performance.
Before you conclude that doctors are obsolete, however, remember that neither they nor the algorithm did especially well when simply handed an image of any random skin disease and asked to identify it rather than being asked to provide a yes-or-no malignancy diagnosis. In the former case, the doctors have considerable advantages: they can examine the lesion from multiple angles, feel it and its surrounding tissue to get a sense of its texture and density, ask for additional tests, and evaluate their own uncertainty. Unlike Inception, they're not limited to looking at images.
Dermatologist-level classification of skin cancer with deep neural networks (DOI: 10.1038/nature21056) (DX)
(Score: 0) by Anonymous Coward on Saturday January 28 2017, @05:29PM
Everyone who tells me they got one of these removed says it grows back after awhile. Then they have to get it removed again, and it grows back again, etc. Each time you have to be damaging the surrounding tissue a bit, and a few cancerous cells are getting into the bloodstream. If that is what happens, it may be safer and cheaper to let your immune system attempt to encapsulate the strange tissue and just hope for the best.
(Score: 2) by Runaway1956 on Saturday January 28 2017, @06:04PM
"got one of these removed says it grows back after awhile"
Got one of what removed, precisely? Doctors and computers both are fallible, when it come to identifying odd growths. We're expected to believe that some layman, with some hear-say gossip spread by other laymen, know what was removed from some unidentified, anonymous coward's acquaintance?
I think the whole point of the article is, it's difficult to identify that little blog of stuff growing on your thigh. They are working on computer programs that should help to identify that lump, but they aren't there yet. Until then, you're in the hands of trained professionals, who are still fallible.
Thank God I'm not relying on AC to identify my own little lump.
(Score: 0) by Anonymous Coward on Saturday January 28 2017, @06:18PM
Sure, you shouldn't make decisions based off hearsay you read on the internet. But what has been your experience with these skin lesions? Have you ever gotten one removed and seen it grow back?
Unfortunately this is what we are reduced to though. Due to the sad state of affairs in the medical research community, we are far more likely to be mislead by what exists in that literature than enlightened.
http://biorxiv.org/content/early/2016/12/20/095570 [biorxiv.org]
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(13)62329-6/fulltext [thelancet.com]
http://www.nature.com/news/irreproducible-biology-research-costs-put-at-28-billion-per-year-1.17711 [nature.com]
http://www.slate.com/articles/health_and_science/future_tense/2016/04/biomedicine_facing_a_worse_replication_crisis_than_the_one_plaguing_psychology.html [slate.com]
http://www.nature.com/nrd/journal/v10/n9/full/nrd3439-c1.html [nature.com]
http://www.reuters.com/article/us-science-cancer-idUSBRE82R12P20120328 [reuters.com]
http://www.nature.com/nature/journal/v483/n7391/full/483531a.html [nature.com]
http://www.sciencemag.org/content/348/6242/1411 [sciencemag.org]
http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002165 [plos.org]
http://www.nature.com/news/cancer-reproducibility-project-scales-back-ambitions-1.18938 [nature.com]
(Score: 0) by Anonymous Coward on Saturday January 28 2017, @09:34PM
It's your dick. Glad my expertise could help. Free diagnostics to subscribers, don't worry about the bill :)
(Score: 3, Insightful) by JoeMerchant on Saturday January 28 2017, @06:11PM
That's why they call it "practicing" medicine, hopefully they continue to get better with time.
If you've got an aggressive melanoma, getting it cut off is better 99% of the time - at the other end of the spectrum: cutting every mole and freckle is 99% likely to do more harm than good.
What you hope you've got is an M.D. who is genuinely looking out for you, their patient, and your best interests.
In my experience what I have found most of the time are physicians who are milking the system for maximum billable procedures, occasionally looking out for the patient to make sure that the patient share of the cost isn't so much that the patient won't come back. Believe it or not, that evil (and I mean: genuinely evil on a scale not seen before this age) insurance industry's policies and procedures are actually protecting your health somewhat better than the free market physicians who are only motivated to line their own pockets.
🌻🌻 [google.com]
(Score: 3, Informative) by aristarchus on Saturday January 28 2017, @06:43PM
We're expected to believe that some layman, with some hear-say gossip spread by other laymen, know what was removed from some unidentified, anonymous coward's acquaintance?
Better than listening to some Fox-infected, no-too-BrightBart, Beacon reading poster with a pseudonym on a news aggregation site. So it could have been worse.