First Digital Path AI Tool Cleared for Market by FDA

FDA’s action marks start of a new era in pathology where artificial intelligence helps with diagnosis

CEO SUMMARY: In a first for the anatomic pathology profession, the Food and Drug Administration (FDA) has cleared for market a digital pathology image analysis product intended to aid pathologists in the diagnosis of prostate cancer. Developed by New York City-based Paige, the Paige Prostate system was determined by the FDA to improve diagnostic accuracy of pathologists using the tool by 7.3% when compared to the diagnoses of pathologists not using the tool.

IT’S A MAJOR MILESTONE ON THE ROAD to the digitization of anatomic pathology! Last week, it was announced that the federal Food and Drug Administration (FDA) granted its first-ever clearance to allow the sale and use of an artificial intelligence-powered software solution to diagnose cancer.

The clearance was issued for Paige Prostate, described as “a clinical-grade AI solution for prostate cancer detection.” This product was developed by Paige, a company in New York developing artificial intelligence (AI) software to aid pathologists in diagnosing cancer. (See next story on pages 5-6.)

Obtaining the first FDA clearance for a digital pathology image analysis products is widely-recognized as an essential milestone on the path to the wider adoption of whole-slide images (WSI) and digital pathology (DP) by pathology groups throughout the United States. For the first time, pathologists can use an FDA-cleared artificial intelligence software tool to aid the diagnosis of cancer when working with digital pathology images.

Multiple companies were racing to be first to get an artificial intelligence-powered digital pathology image analysis product reviewed and cleared by the FDA. Paige was first to reach this goal. 

It was earlier this year in The Dark Report that Ajit Singh, PhD, an expert in digital pathology and artificial intelligence at Artiman Ventures, predicted that digital pathology image analysis solutions for prostate cancer would be first to market with regulatory clearance. Now Paige has made his prediction come true. (See TDR, “Is Artificial Intelligence Ready for First Use in Anatomic Pathology?” Apr. 12, 2021.)

There are a significant number of companies known to be developing digital image analysis products intended to aid in the diagnosis of cancer and other health conditions. At the upcoming conference of the Digital Pathology Association scheduled to take place in Las Vegas on Oct. 17-19, 2021, at least 14 companies developing pathology image analysis solutions will be sponsoring and exhibiting. 

AI-Powered Image Analysis 

Investors are equally enthused about the potential of digital image analysis to transform anatomic pathology. Digital pathology companies are attracting substantial capital investments, as these examples demonstrate: 

  • Proscia, Inc., of Philadelphia raised $23 million in a Series B funding round in Dec. 2020. 
  • Paige of New York City raised $100 million in a Series C funding round in Jan. 2021.
  • Ibex Medical Analytics of Tel Aviv, Israel, raised $38 million in a Series B funding round in March 2021.
  • PathAI of Boston raised $165 million in a Series B funding round in May 2021. 

Pathologists and pathology practice administrators should consider the implications of heightened investor interest. The investment rounds shown above represent $326 million of capital flowing into just four companies developing digital pathology image analysis solutions. In only six months, investors poured one-third of a billion dollars into the development of solutions to automate the analysis of pathology images. 

Digital Pathology Predictions 

Visionaries in the anatomic pathology profession have long predicted that the profession will move to a fully-digital clinical service. These visionaries noted that two components would form the foundation of a fully-digital pathology practice. But it was the third component that would be the disruptive and revolutionary force. 

Component One would be scanners to create whole-slide images from glass slides. 

Component Two would be a digital pathology system that includes a viewer and pathologist workflow solution. 

Component Three would be automated analysis of whole-slide images. In its first iteration, it will be pathologist-directed image analysis. But eventually the technology will improve so that digital pathology image analysis tools can make a primary diagnosis with accuracy comparable to a human. 

It has always been recognized that it would be difficult to achieve component three. It would also take longer than the other two components because of the complexity of developing an artificial intelligence product capable of replicating how the human mind analyzes a pathology image. 

FDA Is Ready to Act 

That is why the decision by the FDA to clear the first-ever digital pathology image analysis product for clinical use is a watershed moment for the entire profession of pathology. It means that regulators accept the study data that indicates the performance of AI-powered digital analysis tools can perform reliably and contribute to improved accuracy and better patient care. 

The Dark Report expects this action by the FDA to trigger two developments. First, other companies will study Paige’s study data and the FDA’s review of that data. Using those insights, they will prepare and submit pre-market review applications for their own pathology image analysis products. Second, and equally significant, expect Paige to build upon this first FDA review and begin submitting algorithm products trained to analyze other types of cancer. 

Collectively, this activity may mean that pathologists will have a steadily-growing number of FDA-cleared image analysis solutions they can incorporate into their daily practice. 



You are reading premium content from The Dark Report, your primary resource for running an efficient and profitable laboratory.

Get Unlimited Access to The Dark Report absolutely FREE!

You have read 0 of 1 of your complimentary articles this month

Privacy Policy: We will never share your personal information.