- Design
- Single-institution retrospective immunohistochemical study of molecularly verified tumours
- Population
- 108 clear cell renal cell carcinomas with IHC, 54 primary and 54 metastatic
- Primary outcome
- Extent of CA9, KER7, PAX8 and CD10 staining and its association with architecture and driver mutations
- Effect
- KER7 positive in 45% of primary versus 13% of metastatic tumours (P = 0.001); CA9 negative only in metastases (15%)
Researchers reviewed renal cell carcinomas sequenced between 2015 and 2024 to find clear cell tumours by VHL alterations or 3p loss. Of 125 identified, 108 with immunohistochemistry were analysed, 54 primary and 54 metastatic. Architecture was graded low, intermediate or high, and CA9, KER7, PAX8 and CD10 were scored for extent of staining.
KER7 was positive in 45% of primary versus 13% of metastatic tumours (P = 0.001), in 67% of low and intermediate architectural grades versus 17% of high grade (P = 0.002), and in 84% of PBRM1 wild-type versus 16% of PBRM1-mutated tumours (P = 0.021). CA9 negativity occurred only in metastatic tumours (15%). BAP1-mutated tumours more often had higher-grade architecture and lacked CD10.
Overall KER7 was positive in 30%, strong and diffuse in 16%. This is a single-institution retrospective series, so a KER7-positive clear cell carcinoma should prompt thought rather than a change of diagnosis. Sequencing-verified cases are not typical of routine practice where sequencing is unavailable.
- KER7 was positive in 30% of clear cell tumours and strongly diffuse in 16%.
- KER7 positivity was commoner in primaries (45%) than metastases (13%).
- CA9 negativity was seen only in metastatic tumours.
- BAP1-mutated tumours tended to lack CD10 and show higher-grade architecture.
- Do not exclude clear cell carcinoma on an unexpected KER7 or CA9 result alone.
Why it matters
Pathologists rely on a classic stain profile, but this series shows that deviations are common in confirmed clear cell tumours.
Don't overread it
A single-institution retrospective series of 108 cases shows associations only and does not validate a diagnostic algorithm.
The statistics, in plain English
P values here come from many comparisons in 108 cases, so some borderline associations may be chance. The PBRM1 comparison (P = 0.021) is weaker than the primary-versus-metastatic one (P = 0.001).
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 pathology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free