Powering oncology drug development research with digital pathology is a core focus for PathAI. Our histological algorithms enable analyses down to the individual cellular level, spanning numerous treatment areas and molecular predictions.
Improving testing accuracy and access:
Developing an accurate and reproducible PD-L1 algorithm compatible with all four PD-L1 testing stains. More generalizable algorithms provide the flexibility and scalability to improve access to more accurate PD-L1 testing.
Identifying new biomarkers and patient sub-types:
Predicting genetic signatures and alterations from digitized H&E pathology slide images, such as c-MET overexpression and KEAP1 mutations. As drug developers look to target pathways beyond PD-L1, AI may help discover new biomarkers and identify patients eligible for new therapies quickly and cost effectively.
Quantifying CD8 classification to expand the eligible treatment population for targeted therapies. Single biomarkers such as PD-L1 imperfectly predict treatment response. AI algorithms can help identify novel biomarker combinations which may improve patient selection for precision medicine strategies.
AIM PD-L1-NSCLC: Artificial intelligence-powered PD-L1 quantification for accurate prediction of tumor proportion score in diverse, multi-stain clinical tissue samples
Machine Learning Analysis of H&E Lung Adenocarcinoma Tumor Microenvironment Shows Association of Human Interpretable Histopathological Features with KEAP1 Mutations
Since the approval of HercepTest in 1998, our understanding of HER2 as a biomarker has significantly evolved. As more HER2 targeted therapies show promise in HER2 low and ultra-low patients, there is a high need for a more highly specific and sensitive HER2 classification methodology. AI pathology can enable more granular HER2 classification than a binary HER2 positivity status which oversimplifies disease biologyOur Recent Research in HER2
In highly aggressive cancers with poor prognoses, it is essential that oncologists have the tools to accurately identify the best therapy for a patient as early as possible. Our research in ovarian cancer focuses on identifying biological features which may predict patient prognosis from the same H&E slides used in diagnosis.Ovarian Nuclear Features Research
Other Cancer Types
Across oncology research and clinical trials, we continue to see the importance of AI-powered digital pathology in unlocking new insights in the tumor microenvironment, optimizing patient selection, and discovering new biomarkers. Our research has fueled exciting developments across the oncology landscape including melanoma, hepatocellular carcinoma, head & neck, gastric, colorectal, prostate, and urothelial cancers. We are continuing to expand our research and development into other cancer types.View More of our Oncology Publications
Evolving Oncology Drug Development
The next generation of biomarker-based drug development is within reach. With AI-powered digital pathology, drug developers can explore complex relationships indiscernible to the naked eye.Learn More About our BioPharma Services