Stratipath is excited to announce the publication of a new validation study on the prognostic performance of Stratipath Breastâthe first AI-based CE-marked medical device for risk stratification of breast cancer using routine H&E histopathology images.
The study included over 2,700 patients from two independent sites to successfully validate the prognostic performance of Stratipath Breast. In the clinically relevant subgroup of ER+/HER2- patients, a Hazard Ratio (HR) of 2.76 was observed between high- and low-risk groups in a multivariable Cox PH model adjusted for clinical factors. Additionally, in the intermediate risk group of NHG2/ER+/HER2- patients, an HR of 2.20 was observed between high- and low-risk groups.
Furthermore, the study evaluated the underlying risk score as the basis for a five-level multi-group risk stratification. A marginal HR of 9.33 was observed between the reference group and the highest risk group, which constituted 20% of the population.
This study was conducted at Karolinska Institutet in Stockholm, led by PhD student Abhinav Sharma in the groups of Mattias Rantalainen at the Department of Medical Epidemiology and Johan Hartman at the Department of Oncology-Pathology. It builds on earlier academic prototype work by Yinxi Wang et al. (Annals of Oncology, https://www.annalsofoncology.org/article/S0923-7534(21)04486-0/fulltext ) and confirms similar performance in the context of a CE-marked device for clinical use in a larger study.
Together, these studies highlight an exciting advancement in AI-based risk stratification models, offering a cost-effective and rapid alternative to traditional molecular diagnostics. We hope that AI-driven clinical decision support solutions like Stratipath Breast will improve patient access to advanced diagnostics and ultimately lead to better outcomes.
