Abstract 2251: PathologyMap: The world's largest online pathology database for collaborative cancer research

Introduction: Currently there is no large centralized online whole slide imaging (WSI) repository for histopathology data, making collaboration difficult and bioinformatics-driven discovery impossible. We wanted to know if it was possible to build a self-sustaining, crowd-sourced digital pathology database for cancer research discovery. If such a database existed and was searchable, it would allow researchers to compare the histology of their mouse models with that of human tissues to determine which mouse model was closest to the true disease. Furthermore, researchers could get tissue corresponding to their mouse model for subsequent DNA, RNA, or protein analysis to ensure that the model has the qualities they are most interested in. Methods: HistoWiz automates histology for cancer researchers guaranteeing a 48-hour turnaround from tissue specimen to digital slides in the cloud. This system feeds digital data into an intelligent tissue platform, PathologyMapTM. which employs a novel image-tagging technology to capture metadata from users and is searchable on the annotation fields. The search becomes more accurate with more historic data, and has the potential for breakthroughs. Results: Currently there are 48,983 scanned slides and the database is growing at 220% per year. It allows researchers to find similarities in cancers extremely quickly by searching for slides from a specific species, strain, genotype, organ, lesion, experimental treatment condition, sex, age and biomarker. A search of the database reveals 58 slides that are stained for CD8. Perhaps its greatest utility for researchers not familiar with normal histology is that abundant normal controls are searchable in the database, allowing easy comparisons between the pathology studied and normal tissue. A search for liver returns at least 30 slides depicting normal liver histology from mice, monkeys, and humans. A search for KRAS yields at least 100 unique slides. A search for p53 reveals over 150 slides. A search for GFP-infected mouse embryo yields 120 slides. PathologyMapTM also reflects the current focus of cancer therapy. For instance, there are 60 slides stained for EGFR, most likely related to the focus on precision medicine in the treatment of tumors. Conclusions: PathologyMapTM is the world9s largest, most comprehensive WSI database. It not only allows for online viewing, sharing, archival, annotation, search and meta-analysis of cancer tissue images, but also access to the corresponding cancer tissue specimens. By using machine learning (ML) algorithms and allowing histology service users to contribute, annotate and compare cancer tissue data across different laboratories and hospitals around the world, PathologyMapTM will be vital for improving cancer diagnosis, discovering insights to advance cancer research, saving money by reducing repetitive research, and accelerating drug development. Citation Format: Ke Cheng, Agedi Boto, Harini Babu. PathologyMap: The world9s largest online pathology database for collaborative cancer research [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2251.