Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Filtered through the analytical power of artificial intelligence, the wealth of available biomedical data promises to revolutionize cancer research, diagnosis and care. In this Viewpoint, six experts discuss some of the challenges, exciting developments and future questions arising at the interface of machine learning and oncology.
Discerning and analyzing the mutational patterns that arise in the cancer genome can provide essential information on the process of tumorigenesis. An analytical framework and web-based tool now aim to aid in mutational signature assignment for improved tumor stratification.
Tumor heterogeneity remains an obstacle to effective clinical management of breast cancer. Two new studies use high-dimensional imaging of single-cell protein expression in situ in clinical samples to link genomic alterations to multi-cellular features of the tumor microenvironment and reveal breast-cancer phenotypes associated with clinical outcome.
Identifying indicators of response to immunotherapy is key for treatment decisions. Two studies now report that early changes in T cell repertoires and CD8+ memory effector cytotoxic T cells in peripheral blood correlate with response to immune-checkpoint inhibitors in metastatic melanoma and may serve as actionable biomarkers of immune activation.
Drug repurposing is an attractive strategy for extending the arsenal of oncology therapies. Screening of a large collection of existing non-oncology compounds against a panel of cancer cell lines now identifies several drugs capable of selectively inhibiting the growth of cancer cells.
Cancer research in recent years has been marked by significant developments in understanding disease biology and foundational discoveries that have changed clinical practice. Ten cancer researchers take stock of the field, the advances that excite them, key outstanding questions and breakthroughs they anticipate looking forward.
Metastasis competence can be acquired early in tumorigenesis, although its underlying molecular intricacies remain unclear. A recent study provides key information about the function of L1CAM in conferring metastasis-initiation potential and chemoresistance in colorectal cancer by hijacking epithelial regenerative mechanisms.
Immunotherapy resistance is associated with poor T cell infiltration into tumors. Tumor-cell-intrinsic oncogenic events that contribute to this defect include Wnt–β-catenin activation. PAK4 is now identified as an upstream modulator of this pathway, thus suggesting the potential of enhancing the efficacy of immunotherapy by targeting this druggable kinase.
KRAS mutations are among the most prevalent tumor drivers, but targeting them pharmacologically has been challenging. Recent landmark studies have demonstrated promising clinical results of KRASG12C inhibition by using small molecules. Bar-Sagi, and Knelson and Sequist provide their distinct perspectives on this recent tour de force in targeting KRASG12C alterations.
Identifying cancer driver mutations is essential to understand disease biology and devise effective therapies, but remains a complex endeavor. A focused analytical approach is now presented that defines driver mutations affecting ubiquitin-mediated proteolysis through machine learning and mining of cancer multi-omics data.