Health IT, Diagnostics

VisualDx using Apple’s Core ML to automate image analysis for non dermatologist doctors

Physicians take an image with their iPhone or iPad of a skin discoloration and automatically receive three possible diagnoses classified as a weak match, good match, and best match.

VisualDx DermExpert app using Apple Core ML

Physicians are increasingly reliant on their mobile devices in their day to day workflow. VisualDx is one of many companies developing clinical decision support apps in the hopes of making their job easier, through image-supported diagnosis with the goal of reducing diagnostic errors.

The company is coming out with a new version of its app which harnesses Apple’s new Core ML machine learning technology —VisualDX Plus DermExpert. Although the company has been around for many years, VisualDx CEO Dr. Art Papier noted in a phone interview that the biggest twist is the speed at which it can identify possible conditions based on the image library it has amassed over the years and “taught” the machine learning algorithm at the heart of its app.

Designed for non-dermatologists, the app enables primary care, family physicians pediatricians, ER doctors and other clinicians to take an image with their iPhone or iPad of a skin discoloration and automatically receive three possible diagnoses classified as a weak match, good match, and best match.

Papier said doctors have been asking the company for this kind of technology for years.

“The app is fully compliant with patient privacy laws as no data is transmitted and the technology has huge implications for the future of healthcare delivery,” Papier said. He notes that a consumer version of the app is under development.

VisualDx’s clinical decision support tool is among the most widely used diagnostic tools in medicine and used by 1,700 large hospitals and clinics, Papier said. It is also used for training in medical schools. Now an international sales push is underway. Recently, its search tool was translated into French, Spanish, German and Mandarin. It is also being adopted by medical schools in Singapore, Hong Kong and in Helsinki in Finland.

“Sixty-five percent of patients presenting with skin concerns are seeing doctors who are not dermatologists such as pediatricians, a family physician, a primary care doctor,” Papier said. “The skin is sometimes a window into internal medicine. Although often it is just acne or warts, rashes and skin lesions can be a symptom of a more complex probem. It can also be a symptom of a chronic disease or an infectious disease, depending on where the patient has traveled.

Beyond images, the company’s platform is used to offer a synopsis of the disease, recommended tests to confirm a condition, and the appropriate codes to use in the patient’s electronic medical record. Physicians can also add their observations to a particular image like the patient’s age, gender, which can prove educational for other users.

To support its visualization program VisualDX has a partnership with Up to Date, a leading reference support system used by doctors.

Although the company has had a version of VisualDx for Apple devices for nine years, Papier said the Core ML development enabling rapid image analysis is a game changer and will significantly speed up its program and improve its usefulness.

Asked if he thought the noise and hype surrounding other smartphone apps aimed at consumers that claimed to use camera phones to help determine if a mole was cancerous or not had impacted the company, Papier clarified his comments in an email.

“VisualDx has worked purposefully over close to 17 years to build the worlds best professional medical image collection and knowledge database.  Acting like the tortoise initially, deliberately and steadily developing a rock solid platform,  we are now using advances in machine learning to rapidly accelerate our growth.  The demand in the U.S. is matched by incredible interest overseas.  Diagnostic error is a worldwide problem, and there is huge need to provide point-of-care diagnostic support overseas as well.”