Molecular portraits of human breast tumours

Why this mattered

Before this paper, breast cancer was largely stratified by histology, grade, stage, and a small number of markers such as ER and HER2. Perou, Sørlie, Eisen and colleagues showed that genome-wide expression patterns could give each tumour a reproducible “molecular portrait,” using cDNA microarrays across thousands of genes in primary human breast tumours. The key shift was conceptual: breast cancer was not one disease with variable behavior, but a set of biologically distinct transcriptional states, visible directly from tumour RNA.

That made a new kind of oncology possible. Tumours could be grouped by coordinated gene-expression programs rather than by single markers alone, leading to the intrinsic subtype framework: luminal, basal-like, HER2-enriched, and related classes. This became a foundation for molecular taxonomy, prognostic assays, and subtype-aware clinical research, especially in distinguishing ER-positive luminal disease from basal-like/triple-negative biology.

Subsequent breast-cancer genomics, including larger expression studies and TCGA’s integrated genomic “portraits,” extended rather than replaced this idea: DNA mutations, copy-number changes, methylation, microRNA, and proteomic data were interpreted against subtype structure first made visible by expression profiling. The paper’s lasting importance is that it helped move cancer classification from microscope-centered description toward data-rich molecular diagnosis, making tumour heterogeneity a measurable organizing principle for prognosis, therapy development, and translational cancer biology.

Abstract

(no abstract available)

Related

• cite → Molecular Classification of Cancer: Class Discovery and Class Prediction by Gene Expression Monitoring — Molecular portraits extends Golub et al.'s gene-expression-based cancer class discovery and prediction to breast-tumor subtypes.
• cite → Cluster analysis and display of genome-wide expression patterns — Molecular portraits uses hierarchical clustering and heat-map display methods from genome-wide expression pattern analysis.
• enables → Comprehensive molecular portraits of human breast tumours — The intrinsic breast cancer molecular subtypes defined by gene-expression portraits were expanded and refined by TCGA's multi-platform molecular profiling.
• enables → The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups — The breast-tumor expression portraits introduced intrinsic molecular subtyping that METABRIC expanded with integrated genomic and transcriptomic data.
• cite ← Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications — The breast carcinoma study validates and refines the molecular breast-tumor subtypes introduced by the molecular portraits paper.
• cite ← Comprehensive molecular portraits of human breast tumours — TCGA's breast cancer molecular portraits extend the original intrinsic molecular subtype framework for breast tumors.
• cite ← The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups — The 2,000-tumor breast-cancer study extends the molecular-subtype framework introduced by Perou's gene-expression portraits of breast tumors.

Sources

• DOI: https://doi.org/10.1038/35021093
• OpenAlex: https://openalex.org/W2097255042