Response Assessment (RECIST)
- RECIST is a rulebook for measuring whether a cancer is growing, shrinking, or holding still on serial scans — so that everyone scores the same tumor the same way.
- You pick a few "target" lesions, add up their longest diameters, and track that one number over time. (Lymph nodes are the exception — you measure their short axis.)
- The four outcomes are Complete Response, Partial Response, Stable Disease, and Progressive Disease — and any brand-new lesion means progression, full stop.
- It's a measurement convention, not a clinical verdict. It standardizes the ruler; it doesn't read the patient's mind.
Imagine three different people measuring the same suspicious blob on a scan with three different rulers, three different definitions of "bigger," and three different moods. You'd get three different answers, and a drug company would have no idea whether their medicine worked. RECIST exists to stop exactly that chaos. It's the Response Evaluation Criteria in Solid Tumors — a shared rulebook so that a tumor measured in one trial means the same thing as a tumor measured in another.
Think of it as the metric system for cancer follow-up. Boring? A little. Essential? Completely.
What RECIST actually does
RECIST takes the squishy question "is this cancer responding to treatment?" and turns it into arithmetic. You measure some tumors at the start (the baseline), measure the same tumors on every follow-up scan, and compare the numbers. The point is consistency over time — measure the same things, the same way, and let the math tell you the story.
It's most at home in oncologic staging and follow-up, usually on a contrast-enhanced CT, though MRI can be used as long as you stay on the same modality across timepoints. Switching rulers mid-race isn't allowed.
Target vs. non-target lesions
Not every spot of cancer gets the full ruler treatment. RECIST sorts lesions into two buckets.
Target lesions are the ones you formally measure — chosen because they're big enough to measure reliably and reproducibly. You crown a handful of them as your representatives, like picking a few delegates to speak for the whole tumor burden.
Non-target lesions are everything else worth watching but not formally measured. You don't put a number on them; you just note whether they're still there, gone, or unmistakably worse.
A lesion has to be big enough to measure honestly. A blob barely larger than a pixel can't be tracked reliably — the "measurement" would be mostly noise. RECIST sets minimum size thresholds so you're not chasing rounding errors.
The one weird trick: how you measure
Here's the rule that trips everyone up, so tattoo it somewhere visible:
- For most tumors, you measure the single longest diameter.
- For lymph nodes, you measure the short axis — the shortest width across the node.
Why the special treatment for nodes? A normal lymph node is naturally long and skinny, like a grain of rice. It's the short axis swelling up — the rice grain plumping toward a marble — that signals something's wrong. The long axis can be long in perfectly healthy nodes, so it's a bad witness.
Then you add up the diameters of all your target lesions into one sum, and that sum becomes the number you follow across every scan.
The four verdicts
Every follow-up boils down to one of four categories. This is the whole point of the exercise.
| Response | Plain-English meaning |
|---|---|
| Complete Response (CR) | All target lesions gone. The scan looks clean. |
| Partial Response (PR) | The sum of diameters shrank substantially from baseline. Real, meaningful retreat. |
| Stable Disease (SD) | Not enough shrinkage to call PR, not enough growth to call PD. The tumor is loitering. |
| Progressive Disease (PD) | The sum grew meaningfully from its smallest recorded value — or any new lesion appeared. |
Two details on that PD row earn their own callout, because they're where people slip up.
Progression is judged against the nadir — the smallest the tumor sum ever got — not against the baseline. A tumor that shrank beautifully and then crept back up can be "progressing" even while still smaller than where it started. And any genuinely new lesion means PD, period, no matter how nicely the old tumors are behaving. New disease overrules good news.
For lymph nodes specifically, "complete response" doesn't require the node to vanish — a treated node can settle down to a small short axis and still count as resolved, since nodes don't usually disappear entirely.
Why a number isn't the whole story
RECIST is a beautifully standardized ruler, and like any ruler it only measures what it's pointed at — size. It doesn't know whether the tissue inside is live tumor or dead scar. That's why functional tools like FDG-PET sometimes tell a different story: a mass can stay the same size on CT while going metabolically cold.
Some treatments break RECIST's core assumption that "shrinking equals good." Immunotherapy can make tumors temporarily swell with inflammatory cells before they respond — which classic RECIST would mislabel as progression. That's why a modified framework exists for those drugs; see immunotherapy response patterns.
The takeaway
RECIST is the agreed-upon way to turn "did the cancer respond?" into a reproducible number: pick your target lesions, measure longest diameters (short axis for nodes), sum them up, and track that sum across scans into one of four verdicts. Remember that progression is measured from the nadir and that any new lesion wins automatically. It's not a diagnosis or a prognosis — it's the standardized ruler that lets everyone, everywhere, mean the same thing when they say a tumor got bigger.