Viewpoint paper on circulating tumour DNA

Researchers at the Medical University of Graz recently contributed to a Viewpoint paper in the journal eBioMedicine, which highlights the hurdles and stumbling blocks in ctDNA analysis.

Circulating tumour DNA (ctDNA) is regarded as one of the most exciting biomarkers in cancer medicine: it could help to detect minimal residual disease (MRD), refine diagnoses, predict disease progression and tailor treatments more precisely. However, a recent peer-reviewed article now urges caution. The authors demonstrate that many ctDNA studies struggle with recurring methodological and statistical problems – particularly when measurements are analysed over extended periods.

The crux of the criticism is that key performance indicators such as sensitivity and specificity change over time and must not be interpreted as constant values. Unclear definitions of ‘lead time’ (how much earlier ctDNA indicates an event) further complicate comparisons between studies. Particularly problematic is the so-called ‘immortal time bias’: if patients are classified as ‘always positive’ simply because they tested positive at some point during the course of the study, this distorts the results.

Using sample data, simulations and a re-analysis of published studies, the team provides practical recommendations for study design, analysis and reporting. The conclusion is sobering: ctDNA has great potential, but before definitive standards can be formulated, more robust evidence and consistently rigorous methodology in longitudinal analyses are required.

With this Viewpoint paper, the team aims primarily to raise awareness of these interpretation pitfalls amongst experts in the field.