The FDA has authorized the marketing of Paige Prostate, an artificial intelligence–based software that was designed to help pathologists scan areas of the body for suspicious cancer as a supplement to the review of digitally-scanned slide images from prostate biopsies.
The FDA has authorized the marketing of Paige Prostate, an artificial intelligence (AI)–based software that was designed to help pathologists scan areas of the body for suspicious cancer as a supplement to the review of digitally-scanned slide images from prostate biopsies.1
The software was developed to detect an area of interest on a prostate biopsy image that has the highest likelihood of harboring cancer so that the area of concern can be further assessed, especially if it had not been identified upon initial review. The software works on a broad range of data, spanning preanalytical variations, staining techniques, and scanning artifacts so that prostate cancer cases can be rapidly and accurately identified.2
“Pathologists examine biopsies of tissue suspected for diseases, such as prostate cancer, every day. Identifying areas of concern on the biopsy image can help pathologists make a diagnosis that informs the appropriate treatment,” Tim Stenzel, MD, PhD, director of the Office of In Vitro Diagnostics and Radiological Health in the FDA’s Center for Devices and Radiological Health, stated in a press release. “The authorization of this AI-based software can help increase the number of identified prostate biopsy samples with cancerous tissue, which can ultimately save lives.”
The decision to authorize Paige Prostate followed review of findings from a clinical study in which 527 slide images of prostate biopsies collected from more than 150 institutions to ensure that the system was generalized to cases from different hospitals and regions.
Of the 527 biopsies reviewed and digitized, 171 were cancerous and 356 were benign. Pathologists completed 2 assessments for each slide image collected; 1 assessment leveraged assistance from the software and the other did not.
Although the study did not examine the impact on final patient diagnosis, results indicated that the software served to improve cancer detection on individual slide images by an average of 7.3% vs unassisted reads on slide images of individual biopsies. Moreover, the use of the software also resulted in a 70% reduction in false-negative diagnoses and a 24% reduction in false-positive diagnoses.3
Notably, the improvement was observed independent of diagnostic subspecialization, the years of experience of the pathologist, and whether the analysis was done remotely or on site. Non-specialist pathologists were reported to use the software as accurately as prostate specialists who did not utilize Paige Prostate.
“The approval is a landmark achievement in the field of digital pathology and demonstrates how robust our technology is when faced with the broad range of natural variations in tissue slides encountered in day-to-day clinical practice,” Leo Grady, PhD, chief executive officer at Paige, stated in a press release. “We are grateful for everyone at Paige and our clinical partners who have brought this new generation of computational pathology products to reality.”
The full data from the analysis of Paige Prostate are anticipated to be submitted for publication in a peer-reviewed journal.