Five integrated genomic subgroups track prostate-cancer relapse risk
Across 259 men (482 samples), integrating copy-number and expression data defined five subgroups on 100 genes that predicted biochemical relapse in discovery, validation, and a third long-follow-up cohort.
Source
Integration of copy number and transcriptomics provides risk stratification in prostate cancer: A discovery and validation cohort study
Study at a glance
- Design
- Cohort — Discovery/validation integrating CNA and array transcriptomics for risk subgroups
- N
- N=259 · 259 men / 482 samples; discovery 125, validation 103
- Population
- Men with primary prostate cancer with tumour/benign/germline samples
- Outcome
- Integrative molecular subgroups predicting biochemical relapse
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
Authors analysed 482 tumour/benign/germline samples from 259 men with primary prostate cancer, integrating CNA and array transcriptomics (eQTL-style) to define patient subgroups and associate them with future biochemical relapse, comparing integrative models with CNA- or expression-only approaches.
What they found
Five subgroups based on 100 discriminating genes in discovery (n=125) and validation (n=103) consistently predicted biochemical relapse (p=0.0017 and p=0.016) and validated again in a third long-follow-up cohort (p=0.027). Integrative analysis outperformed single-data-type approaches and nominated many progression-linked genes not seen with either modality alone.
The limits
What it doesn't show
Biochemical relapse prediction is not metastasis-free or overall-survival proof for every subgroup. Subgroup labels are cohort-derived classifiers — not a bedside kit by themselves. Integrative gain does not mean every nominated gene is a drug target.
Key terms
- CNA
- Copy-number alterations integrated with expression to define subgroups.
- Five patient subgroups
- Outcome-linked clusters from 100 discriminating genes.
- Biochemical relapse
- PSA-based recurrence endpoint used for subgroup prediction.
- Integrative genomics
- Combining CNA and transcriptomics rather than either alone.
- Discovery/validation/third cohort
- 125 / 103 / additional long-follow-up men used to test subgroup–relapse links.
Flashcards
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Endpoint linked to the five subgroups?
Common questions
How many men/samples?
259 men contributing 482 tumour, benign, and germline samples.
How many subgroups?
Five subgroups based on 100 discriminating genes.
What clinical endpoint did subgroups predict?
Biochemical relapse, replicated across discovery, validation, and a third cohort.
Why integrate CNA with expression?
Authors show improved power versus either data type alone for subgroup/phenotype links.
Is this a treatment algorithm?
No — it is risk stratification biology; treatment-by-subgroup trials are separate.
More on Oncology outcomes