Using a database of nearly 130,000 skin disease images, computer scientists at Stanford develop a deep learning algorithm able to identify cases of potential skin cancer. This program matches the diagnostic ability of a dermatologist and can be used on mobile devices to relay our health information when access to a doctor’s office is difficult.

The days of covering our cough in the waiting room while counting the minutes until a nurse calls out our name may soon be in the past. Conversing with our doctor via private message, email, or via our insurance’s web portal has become increasingly popular and may even be beneficial in health care cost savings.

The internet of things has even made discussing symptoms with our doctor irrelevant, as we are constantly tethered to devices that can collect and share a wealth of our health information. It was only a matter of time until someone questioned whether these devices could help diagnose potential life threatening diseases.

Using an algorithm that was originally developed by Google to identify 1.28 million images from 1,000 object categories including differentiating cats and dogs, a team of researchers at Stanford have developed a program that “visually” identifies cases of potential skin cancer, an illness that affects nearly 5.4 million Americans each year.

Their program uses a technique called “deep learning” wherein a computer is trained to figure out a problem as opposed to being pre-programmed with possible answers. In developing the deep learning algorithm, Stanford analyzed 130,000 images of skin lesions, representing about 2,000 different skin diseases, and focused on diagnosing keratinocyte carcinoma classification and melanoma classification. Following testing, they found that the program had the ability to match the performance of 21 board-certified dermatologists.

These results, coupled with telemedicine and the mobile apps we use for constant self-health tracking, indicate that our digital devices could be immensely beneficial to patients in rural areas, those whose access to medical care is prohibitive, or individuals whose employment does not allow adequate time to seek health care.