Abstract IA006: Co-evolving artificial intelligence and pathology

Abstract

Abstract A fundamental question in cancer biology is how cancers evolve heterogeneity and treatment resistance. The evolution trajectory of cancer is dictated by selective pressures from treatments and the tumor ecosystem. Artificial intelligence (AI) enables us to directly study geographical patterns of the microenvironment in pathological samples, to infer cancer habitats and niches. Significant challenges and open questions remain: how to establish multidisciplinary platforms, develop reproducible AI tools, and how to leverage pathology, genetic, molecular, and clinical data to improve personalized oncology. Charles Darwin described how the intimate coexistence between flowering plants and insects leads to reciprocal evolutionary changes; this is now known as coevolution. Today, through the demolition of disciplinary barriers, AI and pathology can co-evolve to create evolutionary changes and new paradigms. I will discuss our latest progress on combining AI and experimental technologies for spatial histology and omics data analysis. We aim to understand how cancer evolves within diverse environmental conditions. Our work has revealed a high level of geospatial variation in the tumor microenvironment, with profound implications for early diagnosis, biomarker development, and cancer therapeutics. Citation Format: Yinyin Yuan. Co-evolving artificial intelligence and pathology [abstract]. In: Proceedings of the AACR Special Conference on Rethinking DCIS: An Opportunity for Prevention?; 2022 Sep 8-11; Philadelphia, PA. Philadelphia (PA): AACR; Can Prev Res 2022;15(12 Suppl_1): Abstract nr IA006.