The TNBC Depth Score: External Validation in GSE25066 as a Universal Patient Selector for EZH2-Targeting Strategies in Triple-Negative Breast Cancer (HR=1.509, p=0.0001, n=508)

The TNBC Depth Score is a geometry-derived continuous prognostic instrument quantifying Waddington attractor lock depth in triple-negative breast cancer (TNBC). It integrates EZH2 expression level, AR (androgen receptor) expression as a continuous epigenetic depth axis (r=−0.378, p=6.23×10⁻¹⁴⁷, n=4,312 TNBC cells, GSE176078), and FOXA1 suppression into a single composite attractor position score. The score was derived entirely from 19,542 single cancer cells (GSE176078) using Waddington attractor geometry, outcome-blind, and then applied without modification to an independent external validation cohort: GSE25066 (n=508 TNBC patients treated with neoadjuvant anthracycline/taxane chemotherapy; Hatzis et al., Lancet Oncology, 2011). Validation result: HR=1.509, p=0.0001 for distant recurrence-free survival (DRFS). Higher depth score predicted significantly worse DRFS. The score outperforms single-marker EZH2 expression in the same cohort (EZH2 alone: HR=1.363, p=0.0047, confirmed in companion document DOI: 10.5281/zenodo.18891318). The score is further characterised as a universal patient selection instrument identifying the mechanistic target population for EZH2-targeting interventions (tazemetostat). It de-selects patients for AR- targeting therapies (low AR = high depth score = AR-targeting inappropriate). It selects patients for tazemetostat maintenance post-chemotherapy (CS-LIT-17, DOI: 10.5281/zenodo.18884089) and tazemetostat → fulvestrant sequential therapy (CS-LIT-16, DOI: 10.5281/zenodo.18884003). No published composite attractor-depth prognostic score for TNBC with these properties exists as of 2026-03-06. NOVEL verdict. Part of OrganismCore BRCA Cross-Subtype Analysis (BRCA-S8h, 2026-03-05). All predictions locked before literature review.