Person:

Park, Jihye

Summary The development of sensitive and non-invasive “liquid biopsies” presents new opportunities for longitudinal monitoring of tumor dissemination and clonal evolution. The number of circulating tumor cells (CTCs) is prognostic in multiple myeloma (MM), but there is little information on their genetic features. Here, we have analyzed the genomic landscape of CTCs from 29 MM patients, including eight cases with matched/paired bone marrow (BM) tumor cells. Our results show that 100% of clonal mutations in patient BM were detected in CTCs and that 99% of clonal mutations in CTCs were present in BM MM. These include typical driver mutations in MM such as in KRAS, NRAS, or BRAF. These data suggest that BM and CTC samples have similar clonal structures, as discordances between the two were restricted to subclonal mutations. Accordingly, our results pave the way for potentially less invasive mutation screening of MM patients through characterization of CTCs.

Determination of molecular features that mediate clinically aggressive phenotypes in prostate cancer remains a major biological and clinical challenge [1,2]. Recent advances in machine learning interpretability as applied to biomedical problems may enable discovery and prediction in clinical cancer genomics [3–5]. Here, we developed a biologically informed deep learning model (P-NET) to stratify prostate cancer patients by treatment resistance state and evaluate molecular drivers of treatment resistance for therapeutic targeting through complete model interpretability. We demonstrate that P-NET can predict cancer state using molecular data with a performance that is superior to other modeling approaches. Moreover, the biological interpretability within P-NET revealed established and novel molecularly altered candidates, such as MDM4 and FGFR1, that were implicated in predicting advanced disease and validated in vitro. Broadly, biologically informed fully interpretable neural networks enable preclinical discovery and clinical prediction in prostate cancer and may have general applicability across cancer types.