Worried about your mole being cancerous?
- Stanford University experts created a database of 130,000 skin disease images
- They then trained an algorithm to decipher between a range of possible diseases
- It was found to match the diagnostic opinions of 21 registered dermatologists
Stephen Matthews For Mailonline
1
View
comments
It’s scary enough making a doctor’s appointment to see if a strange mole could be cancerous.
But putting your trust on a smartphone app to give you an accurate diagnosis is a step further.
However, that’s exactly what could happen in the near future, according to scientists.
An algorithm for diagnosing skin cancer has been developed, possibly allowing for people to be diagnosed through taking a picture of their lesion.
And experts claim that the ‘life-saving’ new method is just as accurate as visiting a dermatologist – if not more.
An algorithm for diagnosing skin cancer has been developed, possibly allowing for people to be diagnosed through taking a picture of their lesion
Stanford University researchers created a database of nearly 130,000 skin disease images to test their algorithm.
And from the very first one, it performed with high levels of accuracy – matching the opinions of 21 registered dermatologists.
It proved to be just as effective in diagnosing some of the most common and deadliest cancers, according to the study published in the journal Nature.
-
Clinical trial offers hope to boy, 3, whose brother died of…
A-lister trainer secrets: Kerry Washington’s Pilates teacher…
Do YOU use organic tampons? You’re paying over-the-odds for…
DON’T bribe children to eat vegetables, says expert – THIS…
Study author Sebastian Thrun said: ‘We realized it was feasible, not just to do something well, but as well as a human dermatologist.
‘That’s when our thinking changed. That’s when we said, “Look, this is not just a class project for students, this is an opportunity to do something great for humanity”.’
Rather than building an algorithm from scratch, the researchers began with one developed by Google.
And experts claim that the ‘life-saving’ new method is just as accurate as visiting a dermatologist – if not more
It had already been rained to identify 1.28 million images from 100 possible object categories, including cats and dogs.
But the researchers taught it to know the differences between a malignant carcinoma and a benign seborrheic keratosis.
They used images that varied in terms of angle, zoom and lighting of more than 2,000 diseases from the internet.
Susan Swetter, professor of dermatology at the university, said computer-aided classification could greatly assist those responsible for diagnosing cancer.
Melanoma, the deadliest form of the disease, has a five-year survival rate of 97 per cent if detected early.
However, if it’s not spotted until later on, this figure drops to just 14 per cent.
Currently, dermatologists look at suspicious lesions with both their naked eye and a handheld microscope to give them a better view.
If these methods prove inconclusive or lead them to believe it could be cancerous, a biopsy is then performed.
Share or comment on this article
-
e-mail
-
-
Next stop, Necker Island: Barack and Michelle Obama are all…
-
Trump’s migrant crackdown: The President will start building…
-
‘¡No, no puede!’: New White House begins with English-only…
-
What happens when you wear a Trump hat in NYC? Controversial…
-
White House press secretary Sean Spicer waged a one-sided,…
-
PIERS MORGAN: Memo to anyone who voted for Trump or Brexit -…
-
Chelsea Handler is accused of immigrant-shaming Melania for…
-
EXCLUSIVE: Now it’s Caribbean sunshine on Richard Branson’s…
-
‘Nobody believes they’re twins!’: Adorable girls with…
-
Horrifying home security video shows home aide HITTING an…
-
Sportscaster Erin Andrews reveals her secret battle with…
-
Kellyanne Conway allegedly punched a tuxedo-clad man…
Comments 1
Share what you think
-
Newest -
Oldest -
Best rated -
Worst rated
The comments below have not been moderated.
The views expressed in the contents above are those of our users and do not necessarily reflect the views of MailOnline.
Close
Your comment will be posted to MailOnline as usual
Your comment will be credited to your MailOnline persona
Close
Your comment will be posted to MailOnline as usual
Your comment will be credited to your MailOnline persona