ctDNA: what it tells you, and what it does not

What is ctDNA?

All cells release fragments of DNA into the bloodstream. In someone with cancer, a proportion of that circulating DNA comes from tumour cells. Circulating tumour DNA — ctDNA — is that tumour-derived fraction, and it can be detected and sequenced from an ordinary blood sample.

The appeal is straightforward: it can be repeated without an invasive procedure, and it samples from wherever in the body the disease is releasing DNA rather than from one biopsy site.

When can it be informative?

Four situations recur.

When tissue is insufficient. A small biopsy may not yield enough material for comprehensive profiling. Plasma provides a route to molecular information without a further procedure.

When re-biopsy is difficult and the profile may have changed. At progression the question is often what has changed since baseline. Where the disease site is hard to reach, plasma may be the practical route to that answer.

When heterogeneity matters. A single biopsy samples one location. Plasma receives contributions from multiple sites — though it cannot say which site any finding came from.

For serial monitoring. Repeated sampling gives a trajectory rather than a snapshot, and change over time is frequently more informative than any single value.

What can it tell us?

• Which alterations are detectable in circulation at the time of sampling.
• How the detected level changes across serial samples.
• Whether an alteration not present at baseline has appeared — often the most useful signal.
• In some settings, whether molecular disease is detectable after treatment intended to be curative.

What can it not tell us?

Where the disease is. ctDNA carries no anatomical information. It cannot say where a tumour is, how large it is, or whether it threatens a structure. That is what imaging is for.

What the tissue looks like. Histological diagnosis, grade and architecture require tissue. Histological transformation is the clearest illustration: a plasma assay may show no new alteration while the tumour has changed its histology entirely.

That disease is absent. Shedding varies enormously between tumour types, sites and disease burdens. A negative result in a low-shedding tumour is weak evidence of absence.

Whether a variant is tumour-derived. Clonal haematopoiesis produces variants in blood cell precursors — particularly in genes such as DNMT3A, TET2 and ASXL1 — which appear in plasma and are not from the tumour. Distinguishing them requires either awareness of the pattern or paired sequencing of white cells.

Longitudinal monitoring

Serial ctDNA is where the modality is most distinctive, because it is the one thing tissue biopsy cannot practically provide. A rising level during treatment, or the emergence of an alteration not previously present, may indicate molecular change before it becomes clinically or radiologically apparent.

Interpretation still depends on the assay. Comparing values across different assays, or across laboratories, is frequently not valid.

Residual disease questions

Detecting molecular residual disease after curative-intent treatment is an area of substantial research activity, with the most developed data in colorectal cancer. The prognostic association is well described: patients with detectable ctDNA after surgery do worse as a group.

Whether acting on that result improves outcomes — rather than identifying earlier who will do badly — is a different question, and one that trials are still addressing. The distinction between a prognostic marker and a predictive one is doing a lot of work here.

Resistance and evolution

Because it can be repeated, ctDNA is well suited to observing how a tumour changes under treatment pressure. Resistance alterations emerging on targeted therapy are among the most established applications, particularly in lung cancer.

Limitations, stated plainly

• Sensitivity depends on tumour shedding, which varies by disease, site and burden.
• Detection at very low allele fraction carries less confidence than the same call at high fraction.
• Clonal haematopoiesis can produce findings that are not tumour-derived.
• Assays differ; results are frequently not comparable across platforms.
• It complements imaging and pathology. It does not replace either.