New Pilot to Leverage API, AI for Enhanced Diagnostic Accuracy

VisualDx and Janssen Biotech recently partnered to pilot a new application programming interface (API) and equitable artificial intelligence (AI) to enhance diagnostic accuracy.  

The interface will allow providers and other healthcare technology companies to bring VisualDx’s CDS into third-party applications to improve overall health conditions, especially for marginalized patient populations. 

Notably, VisualDx technology can bring expertise on people of color to providers of all specialties. 

“The medical community is finally reckoning with its history of racism and underrepresentation of people of color in clinical training,” Art Papier, MD, CEO and co-founder at VisualDx, said in the announcement. 

“We’re thrilled to have a global leader in medicine help us in our efforts to validate our API, be a strategic partner in research and to place our platform into the hands of even more clinicians for the benefit of patients around the world,” Papier continued. 

For over 20 years, VisualDx has focused on building solutions that reduce racial and other forms of bias in the clinical decision-making process. 

The company has continued to maintain thousands of professional diagnostic medical images in its collection. Currently, there are over one-third of those images in pigmented skin.

Pharmaceutical companies are looking for ways to leverage AI within the healthcare and biotech industry. The McKinsey Global Institute estimates that AI and machine learning in the pharmaceutical industry could generate nearly $100B annually across the US healthcare system.  

VisualDx’s API allows end uses to assess skin disorders from images, assisted by the company’s AI. The AI technology is proven to analyze from light to dark skin images with the same accuracy. 

According to a Journal of Investigative Dermatology study, VisualDx’s artificial intelligence add-on feature, DermExpert, accurately identified lesion morphology across all skin types and colors.

Specifically, the device was 68 percent effective in determining the single most likely morphology from the test image blank. And when the AI’s top prediction was broadened to its top three most likely predictions, accuracy improved to 80 percent.