Risk-adaptive therapy guided by dynamic ctDNA in nasopharyngeal carcinoma
Article meta
Article Date: 11 March 2026
Article URL: https://www.nature.com/articles/s41586-026-10244-w
Article Image: (not provided)
Summary
This Nature paper describes the EP-STAR trial, which evaluates a risk-adaptive therapy (RAT) strategy for nasopharyngeal carcinoma guided by dynamic circulating tumour DNA (ctDNA) measurements. The manuscript sets out the trial design, ctDNA sampling schedule, biological correlative analyses and planned comparisons with an external non-randomised no-RAT cohort drawn from the EP-SEASON study. The authors provide data availability, ethics and competing-interest statements, and supplementary extended data figures and tables covering sampling timelines, subgroup biology and cost-effectiveness assessments.
Key operational details include serial blood sampling at pretreatment (T0), after each induction chemotherapy cycle (T1–T3) and weekly during the adaptive phase for patients with detectable ctDNA (T4–T15), with ctDNA testing continued until clearance was confirmed on two consecutive tests. At completion of concurrent chemoradiotherapy, ctDNA assessment was done for all patients. The paper also reports biological correlates (single-cell RNA-seq derived signatures) across ctDNA-defined risk subgroups and presents cost-effectiveness analyses comparing RAT with conventional fixed-course treatment.
Key Points
- The EP-STAR trial tests a ctDNA-guided, risk-adaptive therapeutic approach for nasopharyngeal carcinoma, with serial ctDNA sampling to inform treatment intensity.
- Blood sampling schedule: pretreatment (T0), post-induction cycles (T1–T3) and weekly during the adaptive phase for those with detectable ctDNA (T4–T15), continuing until confirmed clearance.
- An external, non-randomised comparator cohort (no-RAT) from the EP-SEASON study was included for efficacy benchmarking against standard-of-care treatment.
- Biological correlative work (including scRNA-seq references) identifies differing tumour and immune signatures across ctDNA-defined low-, intermediate- and high-risk groups, which may underpin response differences.
- The paper includes health-economic analyses (cost-effectiveness tables) and quality-of-life comparisons for RAT versus standard fixed-course approaches.
- Data and source files are publicly deposited (Mendeley Data) and supplementary peer-review reports and extended figures/tables are available from the publisher.
- Trial funding and material support (for example, donated sintilimab) are declared; independent data monitoring and multi-centre collaborations are described.
Context and relevance
Using ctDNA dynamics to personalise cancer therapy is an expanding area across tumour types. This study applies that concept to nasopharyngeal carcinoma, addressing two clinical needs: real-time risk stratification during treatment and potential tailoring of therapy intensity to reduce overtreatment or escalate for those who remain at high risk. The inclusion of biological correlates and cost-effectiveness analyses strengthens the translational relevance for clinicians and health systems considering biomarker-guided strategies.
Why should I read this?
Short and blunt: if you care about personalised radiotherapy/chemotherapy decisions, ctDNA as a way to dial treatment up or down is exactly the sort of development you need to know about. The paper walks through how the trial ran, how often they tested blood, what biology might explain different ctDNA behaviours, and whether a ctDNA-led approach could be economically sensible. We’ve done the skimming for you — read the full article if you’re deciding whether to consider ctDNA-guided protocols in trials or practice.