Beschreibung:
<jats:title>Abstract</jats:title>
<jats:p>Molecular diagnostics of cancer has undergone a revolution driven by development of high throughput molecular techniques. It is increasingly recognized that malignant tumors previously considered a single entity are in fact composed of multiple different entities that vary in their molecular genetics, underlying biology and most importantly clinical outcome. Integration of omic- and clinical data into practical classification schemes requires novel machine learning approaches to manage big data in a clinically reasonable turnaround time. Accurate molecular classification will make future clinical trials more informative and lead to development of novel therapeutic strategies. While highly informative, advanced molecular testing is expensive and not widely available. We have developed novel methods to predict mutations directly from histopathological slides using artificial intelligence (AI) and image analysis. Mutations in various genes including EGFR, STK11, and BRAF can be predicted with high accuracy by the analysis of the H&E slide image, which is a standard stain in pathology laboratories across the world. Mutational prediction by AI provides a rapid low-cost method to screen patients for potentially targetable mutations in cancer. Cloud based AI mutational predictors could also bring molecular diagnostics to remote or medically underserved areas.</jats:p>
<jats:p>Citation Format: Matija Snuderl. Machine learning and AI in molecular pathology diagnostics and clinical management of cancer [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr IA-19.</jats:p>