Drug Rediscovery Protocol shows modest gains in off-label cancer therapy; ancient genome study reveals pervasive directional selection in Western Eurasia

Overview

The podcast examines the Drug Rediscovery Protocol (DRP), a decade-long Dutch trial investigating off-label cancer drugs matched to patients’ tumor genetics, and pairs it with coverage of a large-scale ancient DNA study on natural selection in western Eurasia.

• Off-label cancer therapies show modest overall benefit but reveal meaningful opportunities for subgroups.
• Data sharing and infrastructure are highlighted as key takeaways for improving evidence and reimbursement decisions.
• A parallel ancient-genomics study suggests directional selection has been more widespread in recent human history than previously thought, prompting careful interpretation.

Introduction: DRP and the search for off-label opportunities

In the podcast, two Nature hosts introduce the Drug Rediscovery Protocol (DRP), a large Dutch prospective trial designed to gather high-quality data on drugs used off label to treat cancers with matching genetic profiles. Emil Voost of the Netherlands Cancer Institute explains the moral imperative to assemble and share data for patients with advanced, treatment-refractory disease, including a young patient with a metastatic cancer who catalyzed the data-collection approach.

"we have an obligation to really share that data" - Emil Voost, Netherlands Cancer Institute

Trial design: cohorts, stages, and many cohorts

The DRP trial is described as a mosaic of many small cohorts, each eight patients to start, defined by a specific tumor type and a genetic aberration, followed by expansion to 16, 24, and a larger confirmatory stage. At its peak, the study included hundreds of cohorts and up to 37 targeted drugs, highlighting the logistical and analytical challenges of aggregating data across diverse cancers to draw evidence about off-label activity.

"we had at the peak, 37 different targeted drugs that we could use" - Emil Voost, Netherlands Cancer Institute

Outcomes: a sobering but nuanced picture

Reported outcomes cover 1,610 treated patients. About one third showed tumor regression or remained stable for more than four months, and roughly 7% had very long responses spanning years, with some possibly cured. Yet the vast majority did not benefit substantially, underscoring the limits of off-label strategies and the essential need to weigh benefits against substantial side effects. The presenters frame these results as both humbling and instructive for future practice and policy.

"Of the 1610 patients that we treated, roughly one third of the patients had either tumors that were partially regressing or they stayed stable for more than four months" - Emil Voost

Policy and data-sharing implications: a model for the world

The podcast emphasizes that when clinical data are trial-grade, they can influence regulators and reimbursement bodies, helping to broaden access to drugs that were previously off-label or unavailable for certain cancers. Voost argues for sharing DRP’s data to enable global collaboration, particularly for rare cancers that struggle to attract industry investment. The conversation also acknowledges the ongoing need to balance patient safety with opportunities to extend life through accessible, evidence-based treatments.

"we thought, okay, we create one big resource that now scientific and clinical community can also have access to" - Emil Voost

Section: Ancient genome study: natural selection in western Eurasia

After the cancer-focused discussion, the podcast shifts to Ewan Calloway’s report on a monumental ancient-genomics study. The study aggregates thousands of ancient genomes from western Eurasia to test whether directional natural selection has shifted allele frequencies over the last 10,000 years. David Reich, a co-author, describes the method as straightforward: for each variant, assess whether its frequency trends upward or downward when accounting for ancestry and genetic drift, across space and time. The study identifies hundreds of variants with strong directional signals and thousands more with weaker evidence, painting a picture of widespread evolutionary change during a transformative period of human history.

"stupidly simple, I think" - David Reich, co-author

What the findings mean and how researchers view them

The segment discusses the kinds of traits implicated, including immunity, metabolism, and even traits linked to modern GWAS signals such as educational attainment and mental-health risk factors. The podcast cautions that these signals reflect broad patterns in modern variants and that inferring precise past phenotypes is risky. Reactions in the field were mixed: the results are seen as a rich resource, yet some researchers question the breadth of the claimed directional signals and the potential confounding effects of ancestry and population movements. The host reflects on the implications for the study of human evolution and the replicability of signals across diverse global populations.

"directional selection is a lot more pervasive than we thought" - Ewan Calloway, senior reporter

Take-home messages

Overall, the podcast frames both stories as steps toward more credible, data-driven science in medicine and in understanding human history. The DRP study provides a blueprint for evidence gathering and sharing that could improve access to potential therapies, while the ancient-genome work invites ongoing debate about how selection shapes the modern genome and how best to interpret signals in the face of complex demographic histories. The podcast closes with encouragement to consider data-sharing as a path to better patient outcomes and a deeper comprehension of our own species’ past.

"If you buy the method, directional selection is much more pervasive" - interview host