With the upcoming iOS 11 launch, a new VisualDx app, enabled with Apple Core ML, will help doctors provide quick and accurate diagnoses of lesions, rashes, and other skin conditions.
Billed as a clinical decision support system for physicians and frontline healthcare professionals, Rochester, New York-based VisualDx developed the technology to help non-dermatology health practitioners identify and treat a wide range of skin conditions, many of which were previously difficult to diagnose without referral to a specialist.
The technology analyzes a photo taken by a clinician and automatically classifies the photo in less than a second on the iPhone itself. The clinician views and confirms the classification of the lesion type, body location and skin type and then can immediately review accurate diagnostic possibilities and treatment options.
“You can put in the type of lesion you see on the patient -- is it a blister or an ulcer? -- and then the clinician clicks on the type of lesion they see,” said Dr. Art Papier, chief executive officer and co-founder of VisualDx. “A lot of dermatologic diagnosis is done by family physicians or emergency physicians, and they can sometimes have a very different opinion of what they’re seeing. One might look at them and describe what they see, but non-dermatologists are not very good at that.”
Papier said the app isn’t 100 percent accurate, but the machine learning nature of the technology should ensure that it becomes incrementally more accurate with each subsequent use. That way, if a family or emergency physician is charged with making a dermatologic diagnosis, the delay and possible error are mitigated; the software classifies what it’s seeing in about one second.
The technology is also designed to protect patient privacy by keeping everything on the phone, said Papier.
“The image is not going up to the cloud, and this is very important from a HIPAA point of view,” he said. “Doctors and hospitals probably would not be comfortable with a physician taking a picture of a patient and sending it to a third party company like ours because it might have a picture of a patient’s face. Our company never sees the image. There’s real patient confidentiality here.”
According to a study in the Journal of the American Academy of Dermatology, More than 65 percent of skin complaints are seen by non-dermatologists, and 15 to 20 percent of all primary care physician visits include skin-related issues. Medical students who later become primary care or generalist physicians receive on average 21 hours of training in dermatology, making it difficult to accurately diagnose skin conditions and their corresponding illnesses. Moreover, there is a growing shortage of dermatologists around the world.
The VisualDx app promises to help healthcare providers make smarter, faster and more accurate diagnoses and treatment choices based on a series of questions, including current and previous global locations, and will take into account information such as local/regional outbreaks and health and medication history. Diseases that may be prevalent in a country thousands of miles away are often unfamiliar to physicians, yet they’re important to diagnose rapidly. With the app, providers can, for example, quickly input the name of a country where the patient recently traveled, select from symptoms, and see diagnoses for consideration in a matter of seconds.
“Our efforts at VisualDx are all about diagnostic accuracy, with both the patient and the practitioner in mind,” said Papier. “Our technology provides a much-needed solution for the global medical community, with the potential to be adapted for other industries.”
The app is expected to debut this fall along with the new iOS 11. VisualDx is also developing a consumer-focused machine learning app, Aysa, which will educate and assist consumers to make better decisions regarding their own health problems.