Abstract 867: Prostate cancer: AI-markers for detecting, grading, and prognosticating 5-year survival post-biopsy

Abstract

Abstract Adenocarcinoma of the prostate is a diverse disease with multiple treatment options and inconsistent outcomes, and is the second most common cancer diagnosed in men, with approximately one in nine men diagnosed in their lifetime. The recent development and regulatory approval of whole slide image (WSI) scanning systems, together with the advent of robust computational approaches such as artificial-intelligence (AI) and machine learning, enable the development of new AI-based algorithmic markers (“AI markers”) that can be used to support the pathologist in their diagnostic process, as well as to better predict outcome and thus disease management in prostate cancer patients. Through a collaboration with Maccabi Healthcare Services, we have developed AI-based diagnostic tools that can correctly identify multiple features within prostate core needle biopsies (PCNBs), including adenocarcinoma, Gleason grade, inflammation, high grade prostatic intraepithelial neoplasia, and others. Combining these features with clinical and molecular data, we have developed multi-feature AI markers that prognosticate 5-year survival post-biopsy using thousands of slides from over 500 patients. In addition, we identify explanatory features associated with pre-biopsy PSA levels that together may help guide treatment decisions. Using AI on a combination of multiple data types, including lab tests, clinical outcome, and tissue biopsy, we demonstrate a powerful approach for the development of novel prognostic algorithms. These algorithms constitute low-cost accessible AI markers that are tissue-sparing and can be easily applied in clinical settings for optimal patient stratification for prospective clinical trials, as well as for different treatment options in prostate cancer. Further, this approach can be extended to additional cancer tissues, with the potential to revolutionize the field of precision medicine in oncology. Citation Format: Daphna Laifenfeld, Cobi Reouven, Inbal Gross, Ronen Heled, Varda Shalev, Judith Sandbank, Chaim Linhart. Prostate cancer: AI-markers for detecting, grading, and prognosticating 5-year survival post-biopsy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 867.