Why do prostate cancer risk calculators give different numbers?
Put the same PSA into four respected calculators and you can get four different answers. That doesn't mean any of them is broken. Each one was built on a different group of men, and each one answers a slightly different question.
Think of a weather forecast
Two weather services can give different chances of rain for the same afternoon. They use different data and different assumptions, and neither is lying. But the forecast that uses live radar is usually the better guide, because it can see what's actually overhead.
Prostate cancer calculators work the same way. A prostate MRI is the radar. When you have one, the calculator that uses it can see more, and its number is usually the one to lean on.
One man, four answers
Take a 62-year-old man with a PSA of 5.4 ng/mL, a normal rectal exam, no family history of prostate cancer and no prior biopsy. His MRI shows a 48 cc prostate with a PI-RADS 3 area.
| Calculator | Chance of significant cancer | Uses MRI? | Built on |
|---|---|---|---|
| PCPT 2.0 | 7% | No | Men in a screening trial, biopsied at the end of the study whatever their PSA |
| PBCG | 22% | No | Men a urologist had already referred for biopsy |
| PLUM | 11% | Yes | Men who had an MRI before their biopsy |
| Rotterdam (ERSPC) | 3% | Yes | Men in a European population screening trial |
PCPT result from the official calculator at riskcalc.org. Rotterdam result from the official SWOP risk calculator 3/4 with MRI, which also estimates a 10% chance of any cancer. PBCG and PLUM results calculated on mypsalevel from the published equations, which match the official calculators.
Why the numbers differ
Each model learned from a different group
A calculator can only reflect the men it was built on. The PCPT calculator came from a prevention trial in which men were biopsied at the end of the study regardless of their PSA, so many had nothing suspicious and its numbers run low. The PBCG calculator came from men whose urologist was already concerned enough to recommend a biopsy, so for the same PSA it gives a higher number. Its own developers' site now lists PBCG as the replacement for PCPT. The Rotterdam calculator came from a European screening trial, where cancers were found earlier and men were at lower risk than in American biopsy clinics. It also uses the older version 1 of the PI-RADS scoring system, so it gives the lowest number of the four.
MRI adds what blood tests can't see
PSA can be raised by a large prostate, infection or inflammation. An MRI shows whether there's actually a suspicious area, and how big the prostate is. The PLUM calculator uses both. When the authors compared the two, PBCG tended to overestimate risk in men who'd had an MRI, while PLUM's estimates matched what biopsies actually found.
Biopsy methods have changed
Older calculators were built when biopsies sampled the prostate in a standard pattern. Newer ones were built when biopsies also targeted the areas an MRI flagged. Because the biopsy method affects what gets found, each calculator is really answering "what would a biopsy like ours have found?"
MRI can move the estimate either way
For the same man, here is how his PLUM estimate changes with the MRI score alone. A reassuring MRI lowers his risk below the no-MRI estimate of 22%. A worrying one raises it well above.
Which number mypsalevel shows, and why
- If you enter an MRI with its PI-RADS score and prostate volume, the main estimate comes from PLUM, the model built for men who've had an MRI.
- If you haven't had an MRI, it comes from PBCG, the current model for men being considered for biopsy.
- Both are always shown when both can be calculated, so you can see how much the MRI changed things.
How to use these numbers
- Treat each number as an estimate with a range around it, not an exact answer.
- When calculators disagree, the one that uses more of your information, especially an MRI, is usually the better guide.
- No calculator knows everything your urologist knows: your exam, your symptoms, your health and what matters to you. Bring your numbers to that conversation.