PAM50

PAM50 is a gene expression classifier used in breast cancer to assign tumors to intrinsic molecular subtypes. The name stands for Prediction Analysis of Microarrays 50, referring to a set of 50 genes whose expression patterns capture tumor biology beyond conventional receptor status. The PAM50 signature was developed to identify five intrinsic subtypes: Luminal A, Luminal B, HER2-enriched, Basal-like, and Normal-like. The Normal-like category is sometimes viewed as less robust and may reflect sample contamination or technical variation.

Methodologically, PAM50 employs a nearest-centroid approach. Expression data are normalized and correlated with reference centroids for each subtype; the highest correlation determines the subtype assignment. Many implementations also compute a risk of recurrence (ROR) score, which combines subtype information with tumor size and proliferation to estimate the likelihood of distant recurrence within five to ten years, aiding risk stratification.

Clinical relevance and usage have grown through research and diagnostics. PAM50-based classification provides prognostic and, in some contexts, predictive information, particularly for estrogen receptor–positive, lymph node–negative breast cancers. The assay is marketed as Prosigna and targets formalin-fixed, paraffin-embedded tissue, delivering intrinsic subtype results along with the ROR score to inform treatment planning, including decisions about chemotherapy and endocrine therapy in early-stage disease.

Limitations include platform and processing dependence, potential instability of the Normal-like category, and varying performance across breast cancer subtypes and patient populations. Despite these caveats, PAM50 remains a widely used framework for understanding breast cancer heterogeneity and guiding research and, in select clinical settings, prognosis-driven management.