Ibex Medical Analytics and Kameda Medical Center (Kameda), both feature in a newly published clinical study evaluating Ibex's Galen platform, which demonstrated excellent outcomes in detecting cancer and other pathologies in prostate and breast biopsies.
The artificial intelligence (AI)-powered platform has subsequently been deployed at Kameda, a leading Japanese healthcare institute, and is now supporting its pathologists in routine practice, integrated with the Philips IntelliSite Pathology Solution.
"We were deeply impressed by Galen's performance in the study, that based on our validation, clearly showed how Ibex can support our clinicians in improving diagnostic quality and efficiency, and as a result we have decided to implement it routinely," said Professor Junya Fukuoka MD PhD, Chair of the Department of Pathology at Kameda Medical Center, Chair of the Department of Pathology Informatics at Nagasaki University and the study's primary investigator. "Interestingly, Galen identified both malignant and non-malignant findings previously undetected by pathologists and we are proud to have become the first pathology department in Japan to use digital pathology and AI routinely for primary cancer diagnosis, significantly improving our pathologists' user experience and confidence."
The study, and subsequent publication in the peer-reviewed journal Pathology authored by Kris Lami, MD, PhD, evaluated the performance of Ibex's AI platform in detecting a broad range of pathologies in 100 breast and 100 prostate biopsies taken from a Japanese patient cohort. Ibex's platform demonstrated excellent outcomes, showcasing its ability to accurately detect both cancerous and non-cancerous features. During the study, Galen Prostate demonstrated high accuracy in detecting prostate adenocarcinoma with an Area Under the Curve (AUC) of 0.988. The AI algorithm was also highly accurate in cancer grading, specifically differentiating between low and medium/high-grade Gleason scores with an AUC of 0.994, as well as in identifying perineural invasion. In several cases, Galen Prostate's AI algorithm accurately identified higher Gleason scores than previously evaluated by a pathologist and unveiled previously undetected cancer, showcasing its potential to enhance diagnostic accuracy and reduce inter-observer variability. Similarly, Galen Breast showed high accuracy in detecting invasive breast cancer (AUC of 0.997) and in detecting ductal carcinoma in situ (DCIS) (AUC of 0.996), as well as in identifying other important pathologies, such as lymphovascular invasion.
"We are proud to announce the study outcomes and publication led by such a prestigious Japanese medical centre, which come as an addition to successful studies of Ibex's AI solutions recently presented by leading United States healthcare institutions across various tissue types," remarked Dr Manuela Vecsler, VP of Clinical and Scientific Affairs at Ibex Medical Analytics. "Galen is the most widely deployed AI technology in pathology, and we are excited to see it adopted in Japan and become the new standard in cancer diagnosis, helping clinicians improve patient care."
- Lami K, Yoon HS, Parwani AV, et al. Validation of prostate and breast cancer detection artificial intelligence algorithms for accurate histopathological diagnosis and grading: a retrospective study with a Japanese cohort. Pathology. Published online April 19, 2024. doi:10.1016/j.pathol.2024.02.009