Abstract LB-211: PathologyMap, an end-to-end histopathology platform for cancer research

Abstract

Abstract Introduction: Traditional histopathology has many limitations. Most notably, maintaining physical slides introduces challenges for researchers to store, organize, annotate and mine data. Our mission is to fight cancer cooperatively instead of individually, using collaborative tools and annotated data in a digitized, centralized environment to transform how cancer research is done and enable breakthroughs. We have built a novel online platform integrating traditional histopathology services with a set of tools to enhance collaboration between cancer researchers and pathologists. It has already helped 1500+ users from 550+ lab of top institutions and pharmas around the world to accelerate their research and publish 45 peer-review papers. Methods: Previously, we demonstrated an approach to using a fully automated histology workflow to create a crowd-sourced database of 100,000+ Whole Slide Images with associated scientific metadata, called PathologyMapTM. We then built the first online pathologist network to provide on-demand diagnostic consultation and scoring of digital slides, with over 60 human and veterinary board-certified pathologists spanning different subspecialties including neuro, ocular and cancer. Here, we highlight recent tool developments for our platform spanning five categories: Viewing, Organization, Search, Annotation, and Collaboration. These subscription-based tools are secure, scalable, customizable, cloud-based, and is currently being used by our customers as well as our growing pathologist network. Results: Several tools we integrated into our platform have had impact for cancer researchers. Among the Viewing tools, our software enables researchers to dynamically compare and screenshot multiple slide images simultaneously. Over 60% of our customers use this feature and over 20% are taking weekly screenshots for use in data presentation. Slide Organization allows grouping and filtering of slides on the fly, across studies and based on numerous attributes (e.g. organ, species, age, marker). Similarly, the capacity to Search across all the slides a research team owns, and compare them to other images within the PathologyMapTM database is now greatly expanded due to keyword and parameter search as well. The areas where our tools have changed the most versus traditional pathology are within Annotation and Collaboration. Researchers now have the ability to make comments and spatial annotations on slides. Access can be shared securely amongst a flexible team structure within labs, and both internal and external collaborations. This functionality also enables the next generation of telepathology, including distributed pathology reporting, collaboration for large studies and instant peer review for challenging diagnostic cases. Future Directions: To further transform digital histology we are developing AI tools to perform image similarity search based on species, stain, organ and cell type, as well as assist in spatial annotation and prognosis. Conclusions: We have built an end-to-end platform for the needs of any cancer researchers conducting histopathology work. These Viewing, Organization, Search, Annotation and Collaboration tools allow researchers and pathologists to annotate and compare data at institutions across the globe, and will advance diagnosis, treatment, and discovery for cancer research as well as ongoing drug development. Citation Format: Ke Cheng, Christoper Gibson. PathologyMap, an end-to-end histopathology platform for cancer research [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr LB-211.