Independent real-world application of a clinical-grade automated prostate cancer detection system
Optimal performance achieved out-of-the-box
The study set out to test the AI in an external dataset and the outcome displayed high sensitivity and specificity.
Independent real-world application of a clinical-grade automated prostate cancer detection system
Published in The Journal of Pathology
Leonard M da Silva, Emilio M Pereira, Paulo GO Salles, Ran Godrich, Rodrigo Ceballos, Jeremy D Kunz, Adam Casson, Julian Viret, Sarat Chandarlapaty, Carlos Gil Ferreira, Bruno Ferrari, Brandon Rothrock, Patricia Raciti, Victor Reuter, Belma Dogdas, George DeMuth, Jillian Sue, Christopher Kanan, Leo Grady, Thomas J Fuchs, Jorge S Reis-Filho
J. Pathol., 254: 147-158.
ABOUT THIS STUDY
Grupo Oncoclinicas designed this study to understand whether Paige Prostate could be used as quality control tool and test its effectiveness in a setting different from where it was developed.
AI Used: Paige Prostate
BY THE NUMBERS
The Results
Paige Prostate was used to classify slides into two categories: benign or suspicious. The sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) were evaluated for a local pathologist, two central pathologists, and Paige Prostate. A total of 682 transrectal ultrasound-guided prostate needle core biopsy regions from 100 consecutive patients were assessed.
Paige Prostate identified four additional patients whose diagnoses were upgraded from benign or suspicious to malignant.
99% sensitivity for cancer detection at slide level
93% specificity for cancer detection at slide
91% negative predictive value (NPV)
65.5% reduction in time to diagnosis
Given its optimal sensitivity and NPV, Paige Prostate has the potential to be employed for the automated identification of patients whose slides could forgo full histopathologic review. In addition to providing incremental improvements in diagnostic accuracy and efficiency, this AI-based system identified patients whose prostate cancers were not initially diagnosed by three experienced histopathologists.