Identification of diagnostic and prognostic extracellular vesicle biomarker signatures in inflammatory breast cancer

Background: Inflammatory breast cancer (IBC) is a rare and aggressive type of BC with a very poor prognosis and accounts for 10% of BC-related deaths. Diagnosis of IBC is particularly challenging, as its symptoms resemble mammary infection and the tumor often presents with metastases at the time of diagnosis. Despite growing body of work, tools for identification and risk stratification of patients with IBC are lacking. Small extracellular vesicles (sEVs) mediate cell-to-cell communication and are released in the blood circulation, thus potentially serving as predictive and prognostic biomarkers. We have previously shown that sEVs are functional determinants of BC progression, but their role in IBC outcomes is not known. Methods: sEVs were isolated from longitudinal plasma samples collected from patients with IBC (n=20), matched non-IBC patients (n=11), and healthy controls (HC, n=15) by ultracentrifugation and liquid chromatography mass spectrometry. Statistical analyses were conducted in R: (a) Kruskal-Wallis tests, followed by pairwise Wilcoxon rank-sum tests with Benjamini-Hochberg false-discovery rate (FDR) correction (IBC vs. non-IBC and HC); (b) Spearman's correlation with FDR correction for multiple testing (association with clinical parameters); (c) receiver operating characteristic (ROC) curve analysis (diagnostic performance); (d) Kaplan-Meier survival analysis with log-rank testing and Cox proportional hazards regression (association with overall survival [OS]). Results: Out of the 2540 sEV proteins detected, we have identified a signature of 20+ proteins enriched more than 2-fold in plasma sEVs from IBC patients compared to non-IBC or HC and associated with immune response pathways. In univariate analysis, a total of 10 sEV biomarkers achieved an area under the curve (AUC) > 0.88, including MARCKSL1, which demonstrated the highest discriminatory power for IBC vs. non-IBC (AUC = 0.955, confidence interval [CI]: 0.878-1.000, sensitivity 100%, specificity 81.8%). When sEV proteins were compared with clinicopathological features, 998 significant associations were identified (FDR < 0.05 and p-value < 0.05), with the strongest associations observed for Ki67 proliferation index, tumor grading, and OS. Exploratory survival analyses identified a set of 20+ overrepresented sEV proteins in patients with IBC and worse OS. High composite scores of the top three sEV biomarkers (correlation > 0.81, p < 0.0001) showed a trend toward increased mortality with good prognostic discrimination (C-index = 0.705). Conclusions: Our comprehensive proteomic analysis revealed extensive and robust associations between circulating sEV protein biomarkers, IBC detection, and clinical parameters, suggesting potential clinical utility as “liquid biopsy” biomarkers for IBC diagnosis and prognosis. Further validation in larger independent cohorts is warranted.