Imaging Nerd
All Systems/Oncologic Imaging/Principles/Immunotherapy Response Patterns

Immunotherapy Response Patterns

Key Points
  • Immunotherapy doesn't kill tumors directly — it wakes up the immune system to do the killing, and that changes how "response" looks on imaging.
  • A tumor can briefly grow (or new spots appear) before it shrinks, because immune cells are flooding in. This is pseudoprogression — apparent worsening that isn't real.
  • Because of that, you can't always call failure on a single scan: immune-specific criteria ask you to confirm progression on a follow-up before declaring it real.
  • The flip side is hyperprogression — genuinely faster growth on therapy — which is real and bad, not a trick of the immune system.
  • Immunotherapy also has its own side effects (immune-related adverse events) that show up on scans as inflammation: colitis, pneumonitis, thyroiditis, and more.

Old-school chemotherapy is a blunt poison: it kills fast-dividing cells, the tumor shrinks, and on the scan smaller is better. Simple. Immunotherapy plays a completely different game — and it breaks the comfortable assumption that "bigger on the scan = losing."

Here's the mental model. Chemo is the assassin who does the job directly. Immunotherapy is the coach who convinces your own immune system to do the job. And when you suddenly send a swarm of angry immune cells into a tumor, the first thing that happens isn't shrinkage — it's a crowd. The tumor can look bigger because it's now packed with lymphocytes throwing a riot. To the assassin's manager (the standard response rules from RECIST), a bigger mass means the drug failed. But here, it might mean the drug is working.

Pseudoprogression: the fake-out

Pseudoprogression is exactly what it sounds like — apparent progression that isn't real. The tumor appears to grow, or brand-new lesions pop up, and then on a later scan everything melts away. The growth was immune cells and inflammation moving in, not cancer winning.

This is the single most important idea on this page, because the conventional reflex — "new lesion, that's progression, stop the drug" — can be exactly the wrong call.

Heads Up

Under standard RECIST, a new lesion is automatic progression, full stop. Under immune-specific criteria, a new lesion is not an automatic game-over — its size gets measured and folded into the total tumor burden, and apparent progression usually needs to be confirmed on a follow-up scan a few weeks later before you call it real.

That "confirm before you condemn" rule is the whole reason immune-specific response criteria exist. The community has rolled out a few flavors over the years (you'll see names like irRC, irRECIST, and iRECIST). The fine print differs, but the shared spirit is the same: don't declare failure on a single timepoint when the patient is otherwise doing fine — wait, rescan, and see whether the "progression" was the real thing or just the immune system making a mess on its way to a win.

Hyperprogression: the real, nasty version

Now the cruel twist. Not every acceleration is a fake-out. Hyperprogression is a genuine, sometimes dramatic speeding-up of tumor growth after immunotherapy starts. It is real disease, it is bad, and it is not something a follow-up scan will wash away.

So the radiologist sits in a genuinely hard spot: a tumor that's bigger than last time could be the immune system winning (pseudoprogression) or losing badly (hyperprogression), and a single picture often can't tell you which. That's why these calls lean on the follow-up scan and the clinical picture — is the patient feeling better or falling apart? — not on one heroic read.

PatternWhat it isHow you sort it out
ResponseTumor shrinks (may be slow or delayed)Decreasing burden over serial scans
PseudoprogressionApparent growth/new lesions, then regressionConfirm on follow-up; burden later falls; patient often stable/improving
HyperprogressionGenuinely accelerated growth on therapyFaster growth than the pre-treatment pace; patient declining
Pitfall

Treat any apparent progression in a stable, clinically well patient on immunotherapy as "innocent until proven guilty." Calling progression off one scan can pull a patient off a drug that's actually working. Confirm on follow-up before you pronounce.

The immune system also attacks the bystanders

Because you've revved up the immune system, it sometimes turns on healthy organs too. These are immune-related adverse events (irAEs), and they're inflammation showing up where you don't want it. On scans they're their own little catalog: colitis (an angry, wall-thickened, stranded colon), pneumonitis (new ground-glass or patchy opacities in the lungs), thyroiditis, hepatitis, hypophysitis (an inflamed, plump pituitary), and others.

Note

irAEs matter to the radiologist for two reasons. First, they can be the thing actually making the patient sick — colitis and pneumonitis can be serious and need prompt treatment. Second, an inflamed organ can light up and mimic new metastatic disease, so don't reflexively read every new finding as cancer spreading.

Figure · CT
Axial contrast-enhanced abdominal CT showing immune-related colitis: diffuse colonic wall thickening with pericolonic fat stranding and mucosal hyperenhancement, in a patient on checkpoint-inhibitor immunotherapy.

Don't forget metabolic imaging

Anatomy alone — pure size — is a weak tool here, because a "bigger" mass might be stuffed with immune cells rather than tumor. That's where FDG-PET gets interesting, and also tricky: inflammation is metabolically hungry too, so reactive immune tissue and irAEs can glow on PET and impersonate active cancer. Useful, but read it with the same skepticism — bright is not automatically bad.

Figure · PET/CT
Fused FDG-PET/CT showing a known metastasis that is larger but with reduced central FDG uptake on immunotherapy, consistent with treatment response rather than progression despite the size increase.

The one thing to carry out the door

Immunotherapy rewrote the rule that "bigger means worse." A tumor can swell with immune cells before it shrinks (pseudoprogression), and a single scan can't always separate that from genuine acceleration (hyperprogression) — so when the patient is otherwise well, you confirm on follow-up before you call progression. And keep one eye on the rest of the body, because the same revved-up immune system that's fighting the cancer can inflame the colon, lungs, thyroid, and beyond. The picture matters, but here it only makes sense alongside the next picture and the patient.