A correct measurement is not enough

Measurement drift between centres

The same tumour cut at a different height of the mass yields a different longest diameter. Both radiologists measure correctly — and the conclusions are opposite.

Mechanizm

Nobody measures “wrong”. Both radiologists mark the longest diameter of the lesion exactly as RECIST requires — the problem lies elsewhere. A metastatic tumour is usually a polycyclic mass — a cluster of irregularly shaped nodules. Such a mass has a different longest diameter at every height and in every plane of section. If the follow-up study was reconstructed at a different height or angle than the baseline, the longest chord falls through a different nodule — and “grows”, even though the tumour mass has actually shrunk.

• Height and plane of section — a polycyclic mass has a different longest diameter at every level; an oblique slice catches a longer chord across neighbouring nodules, even though the tumour itself is smaller.
• Lymph-node axis — RECIST measures nodes in short axis; measuring the longest diameter (long axis) inflates the sum by several millimetres.
• No shared reference point — without a defined plane and the lesions marked on the baseline study, the measurement cannot be reliably repeated.

How we prevent it

We document every target lesion with its measurement parameters and location, and we compare each subsequent study against a reference point from the same pathway — ideally reported earlier by the same team. This keeps the numbers comparable throughout the patient’s follow-up.

Pseudoprogression in immunotherapy

A lesion “grows” because an immune response is infiltrating it — which means the treatment is just starting to work.

Mechanism

Immunotherapy does not attack the tumour directly — it activates the immune system. Immune cells flow into the lesion, bringing oedema and inflammatory infiltrate. On measurement the tumour’s dimensions grow, even though the cancer cells are being destroyed. Assessed by RECIST 1.1 alone, such a lesion would be called progression — and treatment would usually be stopped.

How we prevent it

We match the assessment criteria to the type of therapy: iRECIST for immunotherapy, Choi criteria for selected targeted therapies, RECIST 1.1 otherwise. The criteria used are always stated in the report, so the conclusion can be interpreted unambiguously in the programme documentation.

Bone flare: a lesion that did not appear — it revealed itself

An osteolytic metastasis can be barely visible. When treatment works, the bone rebuilds, the lesion sclerotises and suddenly stands out. Without comparison it reads as a new lesion.

This is the most insidious of the three pitfalls, because the measurement here is neither wrong nor doubtful. The lesion simply was not visible before — and suddenly it is. The lesion did not appear during treatment. During treatment it revealed itself.

Mechanism

An osteolytic metastasis destroys the bone trabeculae, but its own density is close to the surroundings and its borders are indistinct. On CT it can therefore be hard to see — especially in bones with rich trabecular structure and with thicker reconstruction slices.

When treatment starts to work, osteoblastic remodelling begins: the body rebuilds the destroyed bone. The lesion becomes sclerotic — dense, bright, sharply demarcated. In other words: the lesion becomes visible precisely because the treatment is working.

A radiologist who looks only at the follow-up study sees a clear lesion in the bone that “wasn’t there before”. Without the baseline study — or with it, but without a report and defined lesions — the most natural conclusion is: a new lesion, i.e. progression. And progression means the end of a working treatment.

What RECIST 1.1 says

The criteria here are unambiguous, though often overlooked:

• purely sclerotic (osteoblastic) bone lesions are non-measurable — they cannot be target lesions;
• osteolytic or mixed lesions can be measurable only if they have a distinct soft-tissue component meeting measurability criteria;
• sclerosis of a previously known lesion is not progression — it is a typical picture of treatment response.

How not to fall into this trap

1. Comparison is mandatory, not optional. Without the baseline study you cannot tell a lesion that appeared from a lesion that revealed itself.
2. Go back to the location, not the image. You must review the same location on the baseline study — the lesion is often there, just faintly visible.
3. Assess the context. If the other lesions are regressing and the “new” bone lesions are sclerotic, this is most likely healing, not spread.
4. Beware scintigraphy and PET. Increased uptake during healing (the flare phenomenon) may further suggest progression — interpretation requires correlation with the CT image and the clinical course.