Oncology outcomes
A 22-gene classifier flags early metastasis after prostatectomy
Open access · cc by · source: Europe PMC
In Mayo radical-prostatectomy men, a 22-marker genomic classifier from primary-tumour expression predicted early clinical metastasis after PSA rise better than clinical variables (validation AUC 0.75).
Study at a glance
- Design
- Case-control — Nested case-control of Mayo Clinic radical-prostatectomy patients; genomic classifier for early metastasis
- N
- N=545 · 545 unique expression profiles from a 639-patient nested case-control sample; median follow-up 16.9 years
- Population
- Men after radical prostatectomy with biochemical recurrence risk for metastasis
- Outcome
- Genomic classifier discrimination for early clinical metastasis (validation AUC)
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
Expression profiles from 545 unique samples (median follow-up 16.9 years) supported GC validation AUC 0.75 (0.67–0.83), outperforming clinical variables and prior gene signatures. GC was the only significant prognostic factor in multivariable analyses; high GC within Gleason groups tracked earlier prostate-cancer death and reduced overall survival.
Methodology
Using a nested case-control sample of 639 Mayo Clinic radical-prostatectomy patients (1987–2001), authors built a random-forest genomic classifier (GC) of 22 expression markers on high-density arrays, focused on early clinical metastasis after biochemical recurrence, and tested performance in a withheld validation set against clinical factors and prior signatures.
Limitations
Improving metastasis prediction is not the same as proving that treating by GC score improves survival in an RCT. The design enriches for PSA-rise/metastasis cases from a surgical registry — not every newly diagnosed man. AUC 0.75 is useful discrimination, not perfect triage.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
A 22-marker genomic classifier predicts early metastasis after prostatectomy better than clinical variables alone.
In Mayo radical-prostatectomy men enriched for PSA rise, a 22-marker expression classifier achieved validation AUC 0.75 for early clinical metastasis and was the only significant factor in multivariable models.
Evidence for the claim as stated.
HDAC2 immunohistochemistry independently tracks shorter PSA relapse.
Across 192 prostate carcinomas, class I HDACs were frequently strongly expressed; HDAC2 was an independent prognostic marker associated with shorter PSA relapse and with dedifferentiation/proliferation patterns.
Scope note — different endpoint — early metastasis after BCR vs PSA-relapse timing by HDAC2 IHC
Limits the claim's scope: a different population, assay, or outcome.
Five integrative CNA+expression subgroups predict biochemical relapse across cohorts.
In 259 men, five subgroups from 100 discriminating genes predicted biochemical relapse in discovery and validation and again in a third long-follow-up cohort, with integrative analysis outperforming single data types.
Scope note — both are expression-era risk tools, but GC targets early metastasis after BCR; this paper’s subgroups target biochemical relapse
Limits the claim's scope: a different population, assay, or outcome.
GC discriminates early metastasis after PSA rise; HDAC2 IHC tracks PSA-relapse timing; integrative subgroups stratify biochemical relapse. Shared “aggressive prostate cancer” language does not make the assays interchangeable.
Evidence for the claim as stated.
Open questions
Tensions this paper is part of
From concept pages' “where studies disagree.” Disagreement means the same question; scope means different assays, populations, or outcomes.
GC discriminates early metastasis after PSA rise; HDAC2 IHC tracks PSA-relapse timing; integrative subgroups stratify biochemical relapse. Shared “aggressive prostate cancer” language does not make the assays interchangeable.
Related papers in this topic
Same topic cluster — not a recommendation engine.