[Artificial intelligence-guided precision medicine in hematological disorders].

Precision medicine in oncology uses genomic data to provide the right intervention in the right patients at the right time. For this purpose, next-generation sequencing (NGS) is an indispensable tool. However, further innovations are necessary, including medical informatics which uses somatic mutations to inform clinical intervention. This process, called manual interpretation or curation, is currently labor-intensive, involving experienced curators who identify the relevant evidence among a growing body of knowledge and translate it into medical practice. To address this issue, since 2015, we have been organizing a clinical sequencing (CS) team, which integrates clinical and genomic information related to hematological malignancies with the aid of artificial intelligence (AI). So far, we have collated AI-assisted CS data for more than 300 patients with hematological cancers. In this paper, we provide a brief introduction for the nonspecialist to the manual interpretation process together with an overview of a representative AI platform, Watson for Genomics. Based on our own experience, we highlight why AI is needed in the manual interpretation process in CS. We also present some of the pitfalls and limitations of AI that hematologists should be aware of when interpreting AI-output.