This AI tool can help detect cancer

Researchers at MIT have created an artificial intelligence tool that can help physicians detect melanoma- a deadly type of skin tumor responsible for more than 70 per cent of all skin cancer-related deaths worldwide.

Physicians usually have to visually examine suspicious pigmented lesions (SPLs) to determine whether it is dangerous and an indication of skin cancer. This can often be challenging and time-consuming given that there are a high amount of pigmented lesions that need to be checked.

If melanoma is diagnosed early not only is it much easier to treat but can also reduce treatment costs.

Scientists develop message display fabric

Researchers are artificial using deep convolutional neural networks (DCNNs) and applying them to analyzing SPLs through the use of wide-field photography, similar to the ones we capture from our smartphones and personal cameras.

The wide-field image is captured by the smartphone camera which shows a clear picture of the large skin sections. The pigmented skin lesions are then analyzed by DCNNs to identify and screen early-stage melanoma.

The system examines the pigmented lesions and marks them (yellow = consider further inspection, red = requires further inspection or referral to dermatologist).

US researchers find way to read sealed historic letters

Extracted features are used to further assess pigmented lesions and to display results in a heatmap format.

Researchers have trained the system using 20,388 wide-field images from 133 patients at the Hospital Gregorio Marañón in Madrid along will publicly available images. Dermatologists worked with the researchers to visually classified the lesions in the images for comparison.

They found that the system achieved more than 90.3 per cent accuracy in distinguishing SPLs from nonsuspicious lesions.

“Early detection of SPLs can save lives; however, the current capacity of medical systems to provide comprehensive skin screenings at scale are still lacking,” says Luis R. Soenksen, a postdoc and a medical device expert.

“Our research suggests that systems leveraging computer vision and deep neural networks, quantifying such common signs, can achieve comparable accuracy to expert dermatologists,” Soenksen explains.

“We hope our research revitalizes the desire to deliver more efficient dermatological screenings in primary care settings to drive adequate referrals.”

The screening can be used by doctors or even by the patients themselves.