Comparison & Use of Priors
- A "prior" is just an older study of the same patient. Comparing to it turns a single snapshot into a movie — and the movie is where the diagnosis usually lives.
- The question is almost never "is this abnormal?" — it's "is this new, bigger, smaller, or stable?" Priors are how you answer.
- Stability over a meaningful stretch of time is one of the most reassuring findings in all of radiology.
- Compare like with like: same modality, same technique, same window, same anatomy. Mismatched studies create fake change.
- Always say which prior you compared to and its date. "No prior available" is itself a clinically important sentence.
Here's a confession that took me embarrassingly long to internalize: most of the time, a radiologist's superpower isn't spotting the lump. It's knowing whether the lump was there last year. A 9 mm lung nodule is a shrug or a five-alarm fire depending entirely on whether it's brand new or has sat there, unchanged and bored, since the Obama administration. Same pixel. Completely different story.
That story comes from the prior — radiology slang for an older imaging study on the same patient. Comparing the current study to priors is so routine that we barely mention it, which is exactly why beginners skip it and then wonder why the attending keeps asking, "Did you look at the old one?"
Why a single image is a liar
A single study is a photograph. It tells you what something looks like right now and nothing about where it came from or where it's going. Priors turn that photograph into a flipbook.
Think of finding a half-eaten donut on the counter. By itself: mildly suspicious, but you can't convict anyone. Now add a photo from five minutes ago showing a whole donut and your roommate holding it. That's the entire job. The finding didn't change; your certainty did, because now you have a before.
This is why the most useful word in a radiology report is often stable. A messy-looking but unchanged finding over years is the radiology equivalent of "this house has always made that noise" — usually benign, because aggressive disease rarely holds still.
The clinical question is rarely "abnormal yes/no." It's "new, growing, shrinking, or stable?" You cannot answer that from one study. You need the prior.
Compare like with like
Here's the trap nobody warns you about: comparing two studies that aren't actually comparable, and inventing change that was never there.
For a fair comparison, you want the studies to match on:
| Match this | Why it matters |
|---|---|
| Modality | A CT "mass" and an ultrasound "mass" measure different things; don't treat them as one ruler. |
| Technique / phase | Contrast vs non-contrast, or a different contrast phase, changes how things look and measure. |
| Window & plane | Comparing a lung-window slice to a soft-tissue one is comparing apples to a different apple. (More on this in windowing & reconstruction planes.) |
| Anatomic level | You must measure the same lesion on the same slice, not a neighbor that wandered into frame. |
Measurement is where mismatches bite hardest. If you eyeball a nodule's diameter on a thick slice today and a thin slice last year, you can manufacture "growth" out of pure technique. Real growth should survive an honest, like-for-like remeasurement.
The oldest prior isn't always the best comparison — but it's often the most powerful. To prove stability, reach for the most remote study that still shows the finding. "Unchanged since 2019" beats "unchanged since last Tuesday." For tracking active change, the most recent prior tells you the current trajectory. Smart readers cite both.
How to actually use them
Slot the comparison into your routine so it can't be forgotten. After your own systematic approach to any radiograph — where you do your independent search pattern first — then you bring up the prior and ask the only question that matters: what's different?
A sane sequence:
- Read the current study cold. Form your own impression before the prior contaminates it.
- Find the most relevant prior(s). Same body part, ideally same modality and technique. Note the dates.
- Put them side by side and walk the same anatomy in the same order on both.
- For any tracked finding, remeasure on matched slices rather than trusting the old number.
- State it explicitly: what changed, what's stable, which prior you used, and its date.
Beware satisfaction of comparison — the cousin of satisfaction of search. Once the prior says "stable nodule," it's tempting to coast and miss the new finding three slices over. The prior answers "did the old thing change?" It does not excuse you from hunting for new things.
When there is no prior
Sometimes the cupboard is bare — first-ever scan, outside studies you can't pull, a patient new to your system. That's not a footnote; it's a finding. Saying "no prior available for comparison" tells the clinician your confidence is capped and often nudges the plan toward short-interval follow-up imaging so the next study finally has a before. Many incidental-finding pathways lean entirely on this idea of watching over time — see managing incidental findings.
The one thing to remember
Radiology lives in the fourth dimension: time. A finding's danger is written less in how it looks than in how it behaves. Priors are your only window into that behavior — so find them, compare like with like, and always tell the reader exactly what you compared and when.