Diese Website verwendet Cookies, damit wir dir die bestmögliche Benutzererfahrung bieten können. Cookie-Informationen werden in deinem Browser gespeichert und führen Funktionen aus, wie das Wiedererkennen von dir, wenn du auf unsere Website zurückkehrst, und hilft unserem Team zu verstehen, welche Abschnitte der Website für dich am interessantesten und nützlichsten sind.
The Study Reports the First Deployment of an AI Solution in Routine Clinical Use for Aiding Cancer Diagnosis in Pathology
Tel Aviv, Israel – July 28, 2020 – Ibex Medical Analytics, the pioneer in artificial intelligence (AI)-powered cancer diagnostics, today announced outstanding outcomes in a clinical validation study for the first and only AI-based solution used by pathologists in routine clinical practice. The study was conducted at the University of Pittsburgh Medical Center (UPMC) and is now reported by The Lancet Digital Health, a top-tier peer-reviewed clinical journal.
The study was led by Dr. Liron Pantanowitz and Dr. Rajiv Dhir, pathologists at UPMC Shadyside Hospital, who provided blinded whole slide images of prostate core needle biopsies for analysis by Ibex’s Galen™ Prostate solution. The UPMC team then assessed the algorithm’s output against the UPMC clinical pathology reports, followed by blinded discrepancy resolution.
The accuracy levels of Galen Prostate were found to be the highest reported in the field. The sensitivity measured for prostate cancer detection was 98.46%, specificity was 97.33% and the AUC was 0.991 – all significantly higher than previously reported metrics for AI algorithms in pathology. Moreover, this is the first algorithm to extend beyond cancer detection, reporting high performance of an AI solution in tumor grading, detection of perineural invasion and tumor sizing – clinically important features required as part of the pathology report. During the study at UPMC, the AI solution detected misdiagnosis on parts belonging to six cases (including in cancer detection, grading and perineural invasion).
The Lancet publication is the first to report on the performance metrics of an AI solution that is used by pathologists in routine clinical practice. The CE-marked Galen Prostate is deployed in pathology labs worldwide and has already demonstrated success in detecting misdiagnosed and mis-graded cancers in live clinical settings. The results of the study demonstrate that the Galen Prostate solution can provide pathologists with unprecedented diagnostic insights that will help the fight against cancer. From a broader perspective, the solution’s comprehensive capabilities, combined with its demonstrated clinical utility and implementation practicality render such technology viable in promoting improved AI-aided healthcare practice.
“This publication is an important milestone in our journey toward AI-powered digital pathology, in which computer assisted solutions become an indispensable part of cancer diagnosis” said Daphna Laifenfeld, PhD, Chief Scientific Officer at Ibex Medical Analytics. “Recent empirical data shows that up to 12% of cancer cases may be missed, with potentially severe implications on patient outcomes. The impressive results of this study demonstrate the utility of Galen Prostate in providing 100% quality control and a safety net for pathologists drastically reducing misdiagnosis of cancer with a negligible impact on their workload”.
“As pathologists, we are trained to analyze cases holistically and report multiple cellular features, not just cancer,” said Dr. Rajiv Dhir, Chief of Pathology at UPMC Shadyside Hospital. “What we found in this study is that the Ibex algorithm is accurate not only at detecting cancer, but also in cancer grading, sizing and in detection of perineural invasion.”
The article in The Lancet Digital Health is available online here. An artificial intelligence algorithm for prostate cancer diagnosis in whole slide images of core needle biopsies: a blinded clinical validation and deployment study (tldigitalhealth-D-20-00431R1).
Dr. Pantanowitz, with UPMC at the time of the study, serves on the Ibex Advisory BoardBack