The study leveraged machine learning to analyze over 9,000 breast tissue samples, revealing microenvironmental patterns previously invisible to pathologists. Stromal disruption describes alterations in the connective tissue architecture surrounding breast ducts and lobules, such as increased immune infiltration, fibroblast expansion, and vascular remodeling. These changes, the authors wrote, represent a disrupted immune-rich, fibroblast-enriched, angiogenic microenvironment.

Among women with benign breast disease, those whose biopsies showed high stromal disruption had significantly higher rates of aggressive breast cancer. These tumors also developed earlier: mean time to diagnosis was 3.1 years earlier for women with high versus low stromal disruption, the researchers noted. For those with invasive breast cancer, especially estrogen receptor–positive (ER+) disease, stromal disruption was strongly associated with poorer outcomes. According to the study, increased disruption was associated with high-grade, basal-like, and TP53-mutant tumors, and shorter disease-free and overall survival.

This held true even when controlling for tumor size, stage, and lymph node involvement. The team also linked stromal disruption to demographic and lifestyle factors, including Black race, younger age, obesity, multiparity, and family history, suggesting that cancer risk factors may influence breast stromal architecture via a common pathway. Perhaps most compellingly, stromal disruption is detectable using standard histopathology, without the need for molecular testing, making it a cost-effective and scalable biomarker, especially in low-resource settings.

Stromal disruption may serve as an inexpensive, scalable biomarker for identifying women with benign breast disease at high risk for aggressive breast cancer, the authors concluded, and for stratifying patients with breast cancer by risk of recurrence or death. The researchers now propose investigating anti-inflammatory drugs, lifestyle modifications, and stromal-targeted therapies as potential strategies to prevent or reverse this tissue remodeling. Because inflammation and wound healing responses are modifiable processes, they wrote, targeting these changes may represent a promising avenue for breast cancer prevention.

Source: https://www.insideprecisionmedicine.com/topics/oncology/ai-identifies-novel-biomarker-for-aggressive-breast-cancer/