First FDA-approved autonomous AI advances medical diagnostics

Digital Diagnostics, situated in Iowa, grabbed attention in 2018 as the first FDA-approved AI (artificial intelligence) system. Using AI, it can identify diabetic retinopathy in adults without any human intervention at all, thanks to FDA clearance.

IDx-DR, its artificial intelligence diagnostic system, can detect diabetic retinopathy, a leading cause of blindness in the United States and other affluent countries, and other critical eye illnesses, such as macular edoema.

Digital Diagnostics cofounder, president, and COO Seth Rainford told VentureBeat, “There’s a strong goal and purpose for us to get our technology to patients that truly need to be tested, and certainly to clinicians that may be burnt out or be getting burnt out.”

To help bring its technology to market and meet its aim of “paving the road for AI diagnosis to become a standard-of-care, democratising healthcare, and reducing care gaps,” Digital Diagnostics today announced that it has raised $75 million in a series B fundraising round.

The importance of using AI for automating eye care

Executive chairman and cofounder Dr. Michael Abramoff conceived of Digital Diagnostics, Rainford said.

Retina expert Dr. Abramoff also holds several degrees in machine intelligence. When he was a doctor, he regularly saw patients who had been waiting months to receive a routine exam to check for vision loss. Dr. Abramoff saw a need in the market and set out to satisfy it.

The goal of his company, Digital Diagnostics, is to automate the process of detecting health hazards, starting with the human eye, by merging machine learning with his knowledge of eye health. The fundamental concept is that more people will have easier and faster access to the diagnostic tests and results they need to avoid blindness if the system is automated.

According to Rainford, the Digital Diagnostic technique has the potential to be more scalable than relying solely on human doctors, and it may also be less expensive.

According to Rainford, “if you think about what is required today to offer the same thing that we’re delivering with this technology, you begin to think about a person going through med school, then more time at a residency or fellowship, and further training.”

After years of schooling, a human doctor might prefer to practise in a bustling metropolis than a remote, neglected community. According to Rainford, Digital Diagnostics’ primary objective is to make it simpler for people to receive an accurate diagnosis of eye conditions.

It is because of this that “we are enabling or unlocking access to high quality-specialist level of care, at least from a diagnosis standpoint, and then we are able to triage those patients that actually are losing their sight to the specialists that have spent all those years going to school,” he explained.

How Self-Driving AI Fuels Next-Generation Digital Health Checkups

The goal of digital diagnostics technology is not simply to aid a clinician, but rather to replace one.

Rainford explained that the company’s FDA-approved system makes use of a fundus camera, a specialist camera designed specifically for capturing images of the human eye. Digital Diagnostics client software incorporates AI to analyse camera input.

According to Rainford, the AI is not a deep learning system but rather uses a model that has been reviewed and authorised by the Food and Drug Administration.

The FDA has given its stamp of approval to Rainford’s AI, so “that algorithm is fixed until we go back to the FDA with something different,” he explained. As a result, “there is no continuous learning happening in the background; we understand the algorithm’s logic, and its actions can be tracked in detail.”

With an eye toward the future, Digital Diagnostics is working to improve upon current methods of disease diagnosis. In 2020, Digital Diagnostics bought the dermatology-focused private company 3Derm systems.

For conditions “beyond the eye,” such as skin cancer and other skin disorders, “we’re working on the very same kind of technology,” Rainford said, “where we’re attempting to democratise access to specialist-quality testing.”