Ozbilgic, R.; Dinc, B.; Vipparthi, K.; Seachrist, D.; Nicolas, M.; Keri, R. A.; Liu, X.; Yildirim, M.; Karaayvaz, M.

Purpose: Triple-negative breast cancer (TNBC) exhibits substantial clinical heterogeneity, with some patients experiencing early recurrence and poor survival despite similar clinicopathologic features. We sought to determine whether quantitative measures of intratumoral collagen architecture and composition derived from standard histopathologic specimens can identify patients at risk of recurrence and adverse survival outcomes. Experimental Design: We analyzed a retrospective cohort of 79 TNBC tumors assembled into a tissue microarray using a multimodal computational pathology framework integrating Massons Trichrome staining with COL1 and COL3 immunohistochemistry. Collagen architecture was quantified using fiber-based image analysis and unsupervised clustering, while collagen composition was assessed using a normalized COL3:COL1 ratio. Associations with recurrence-free interval (RFI) and overall survival (OS) were evaluated using Kaplan-Meier analysis, restricted mean survival time (RMST), and Cox proportional hazards modeling. Results: Unsupervised analysis identified four distinct collagen architectural states, which were consolidated into low-risk and high-risk groups based on recurrence patterns. High-risk collagen architecture was associated with significantly worse long-term RFI (log-rank p=0.025; RMST difference 10.1 months). Independently, a higher COL3:COL1 ratio was associated with improved OS (log-rank p=0.042; RMST difference 9.4 months). Integration of architectural and compositional biomarkers further refined risk stratification, identifying a subgroup with high-risk architecture and low COL3:COL1 ratio that exhibited the poorest survival outcomes. Notably, collagen-based stratification identified patients with divergent outcomes not readily predicted from tumor stage alone. Conclusions: Quantitative assessment of intratumoral collagen architecture and composition provides clinically meaningful prognostic information in TNBC and enables stratification of recurrence and survival risk. These findings support extracellular matrix phenotyping as a practical and scalable computational pathology approach for refining risk assessment in TNBC.