Treatment selection based on the presence of one or more specific biomarkers has the potential to optimize treatment outcomes; nonetheless, most patients lack a specific biomarker that is predictive of benefit from one or more targeted treatment approaches. In this Review, the authors describe the potential of computational analysis approaches to enable the discovery of more complex predictive biomarkers based on comprehensive analysis of large clinical and preclinical datasets and thus address this unmet need.

Cancer screening aims to identify either pre-malignant conditions or earlier-stage disease in non-symptomatic individuals. The authors of this Review recapitulate the general principles of cancer screening, discussing the current screening practices for several cancer types as well as advances in cancer detection that might hold promise for screening an increasing number of cancers in the future.

The management of localized bladder cancers is rapidly evolving. In this Review, the authors discuss diagnostic innovations, refinements in surgical approaches and novel intravesical or systemic treatments such as immunotherapies and antibody–drug conjugates, as well as future directions including bladder-sparing approaches for muscle-invasive disease and novel biomarkers for personalized treatment and follow-up.

RNA sequencing (RNA-seq) is increasingly being integrated into molecular tumour profiling but remains underutilized for clinical decision-making. This Perspective presents a framework for assessing the therapeutic, diagnostic and prognostic actionability of RNA-seq findings as well as an ‘actionable transcriptome’ list to help guide the application of RNA-seq findings in oncology practice.

Cancer screening is an effective strategy for reducing cancer mortality, but in resource-constrained settings the key challenge is not simply whether screening works but how its implementation should be prioritized. We propose a framework for prioritizing cancer screening under resource constraints, summarized in three ‘not all’ principles: not all cancers are suitable for screening; not all individuals need screening; and not all high-performance technologies are appropriate for population screening.

This Comment describes the clinical significance of the Mammography Screening with Artificial Intelligence (MASAI) trial, which is, to the best of our knowledge, the first randomized trial of artificial intelligence-supported mammography interpretation to examine interval cancer as an outcome. This trial used artificial intelligence to assign single-reading versus double-reading of mammograms and found that this approach does not increase interval cancers in the Swedish population.

Although oncology advances stem from clinical research, <5% of global cancer trials actively recruit patients in Latin America. Herein, I discuss why expanding inclusive trial designs and fostering multisector partnerships are essential to reduce disparities and improve the role of Latin America in global oncology research. I present examples such as the recent regulatory reforms in Brazil that aim to accelerate study approvals, as well as collaborative initiatives between local and international research groups to further strengthen clinical trial capacity.