- Design
- systematic review and meta-analysis with hierarchical bivariate modelling; QUADAS-2 quality assessment; histopathology reference standard
- Population
- 1,999 patients with intermediate- or high-risk prostate cancer across 16 studies (14 nuclear imaging, 6 magnetic resonance)
- Primary outcome
- sensitivity and specificity for histologically confirmed pelvic lymph node metastasis
- Effect
- PSMA PET-CT sensitivity 54.6% (95% CI 43.7-65.1), specificity 92.7%; MRI sensitivity 33.0% (18.7-51.3), specificity 94.9%
Sixteen studies covering 1,999 patients with intermediate- or high-risk prostate cancer were pooled, all using histopathology from lymph node dissection as the reference standard — the only way to measure these tests honestly, since imaging cannot be validated against imaging.
Both modalities were specific and neither was sensitive. Pooled sensitivity for prostate-specific membrane antigen positron emission tomography with computed tomography was 54.6% (95% CI 43.7-65.1) with specificity 92.7% (90.5-94.4). Multiparametric magnetic resonance imaging gave sensitivity 33.0% (18.7-51.3) and specificity 94.9% (92.2-96.8). The area under the curve was 0.909 and 0.941 respectively.
The comparison between them is the weakest part and the authors say so twice: fourteen studies contributed to one analysis and six to the other, so this is an indirect comparison of separate bodies of evidence, not a head-to-head meta-analysis. The apparent sensitivity advantage of the nuclear study should be held loosely.
The part that is not in doubt is the consequence of a sensitivity near 50%. A negative scan misses around half of histologically confirmed nodal metastases, which means it cannot be used to avoid extended pelvic lymph node dissection in a man whose risk profile otherwise warrants it. That is the opposite of how these scans are often used in practice — as reassurance that the nodes are clear — and it is worth stating in the report and in the conversation. A positive scan, by contrast, is informative: specificity above 92% means a positive result is usually real.
- Do not use a negative PSMA PET-CT or MRI to omit extended pelvic lymph node dissection where risk otherwise indicates it
- Treat a positive scan as usually real — specificity was above 92% for both modalities
- Tell patients that these scans miss roughly half of nodal metastases confirmed at surgery
- Hold the apparent advantage of the nuclear study loosely; the comparison was indirect, not head to head
- Continue to use extended pelvic lymph node dissection as the reference standard for pathological nodal staging in selected patients
Why it matters
These scans are commonly read as clearing the nodes, and they cannot do that.
Don't overread it
The comparison between the two modalities is indirect, drawn from separate sets of studies rather than patients imaged with both.
The statistics, in plain English
Specificity above 92% with sensitivity around 55% describes a test that is good at confirming and poor at excluding: a positive result usually means disease, a negative one means little. The area under the curve of 0.909 looks impressive but summarises performance across all thresholds, not the one used in practice, which is why the paired sensitivity and specificity are the numbers to quote. The confidence interval on MRI sensitivity, 18.7% to 51.3%, is so wide that its true performance is barely characterised — six studies is not many.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for urology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free