Testing standardisation and optimal integration for routine clinical use remain challenging
The discovery of circulating tumour DNA (ctDNA) in the plasma of patients with cancer has led to the development of a range of assays for the assessment of treatment response and disease surveillance. However, there are challenges in terms of validating liquid biopsy platforms and standardising protocols for routine clinical implementation.
One such challenge is that tumour-derived DNA fragments exist in tiny quantities within a background of non-tumour cell-free DNA (cfDNA) – the variant allele frequency of ctDNA in cfDNA typically ranges from ~0.1–10% (Nat Rev Cancer. 2017;17:223–238) – and this presents a particular issue, especially in patients with early-stage disease or those with low-shedding tumours, as ctDNA levels may fall below the limits of detection even for the most sensitive assays, potentially resulting in false negative results (Int J Mol Sci. 2025;26:7987).
Widespread clinical adoption of ctDNA is also restricted by limited standardised guidance on how to practically apply the test results, so it can be viewed so far as a passive rather than active biomarker. High-level evidence from large-scale randomised clinical studies is lacking in most cancer settings. Currently, the phase III IMvigor011 trial provides one of the most robust datasets supporting the use of ctDNA to guide therapy. This large, prospective, randomised trial utilised ctDNA-based detection of molecular residual disease (MRD) to identify patients at high risk for recurrence after cystectomy for bladder cancer, despite no radiographic evidence of disease, and to guide subsequent adjuvant therapy with atezolizumab. The strategy resulted in significantly longer disease-free survival and overall survival compared with patients who tested ctDNA-positive but received placebo (N Engl J Med. 2025;393:2395–2408).
Interventional trials are required to determine whether ctDNA-guided treatment modifications can have a positive impact on patient outcomes. This not only involves escalating treatment or the use of targeted therapies in patients in whom ctDNA is detected, but it may also entail discontinuing maintenance therapy in low-risk, ctDNA-negative cases to reduce treatment burden and toxicities.
Recently, the Japanese Society of Clinical Oncology has addressed the need for guidance on the clinical implementation of ctDNA analysis in a position paper that includes recommendations on the appropriate clinical use of MRD testing across solid tumours (Int J Clin Oncol. 2025;30:605–654). Similarly, recently published technical recommendations from a large-scale French Delphi consensus aim to standardise ctDNA analysis across Europe – where the routine use of ctDNA testing for MRD is not currently recommended after treatment with curative intent (Ann Oncol. 2022;33:750–768) – and thus increase confidence in liquid biopsy results for routine diagnostic and therapeutic applications (Eur J Cancer. 2026; 242:116791). Huge advances in technology, with ultra-sensitive platforms or tumour-informed MRD assays, are enabling trace quantities of ctDNA to be detected and hold the promise of improved and earlier residual disease detection (Nat Commun. 2025;16:1837; Clin Cancer Res. 2025;31(Suppl 12):P2-04-23).
Looking to the future, ctDNA testing is likely to revolutionise cancer management through the discovery of actionable mutations. Non-invasive molecular classification is already being investigated as a strategy to proactively guide personalised treatment and unlock exciting new clinical opportunities. In breast cancer, for instance, the emergence of a mutation in the ESR1 gene, which encodes for the oestrogen receptor and is associated with resistance to aromatase inhibitors (AI), was used to guide treatment in a phase III study of advanced breast cancer. Patients with mutated ESR1 who were switched from an AI to the selective oestrogen receptor degrader camizestrant experienced a 56% reduction in the risk of disease progression or death compared with patients who continued the AI (N Engl J Med. 2025;393:569–580). In gynaecological oncology, ctDNA holds transformative potential, enabling non-invasive molecular classification, early detection of recurrence and personalised treatment decisions. However, we are not yet at the stage of routine clinical implementation. The immediate priorities are harmonising assay standards, building prospective interventional evidence specific to our tumour types, and learning to interpret ctDNA results – both positive and negative – with appropriate clinical nuance.
