This preprint reports an analysis of the extracellular matrix (ECM) proteome in primary ductal carcinoma in situ (DCIS) to determine whether ECM composition associates with later clinical events of DCIS or invasive breast cancer (IBC). The authors note that prior transcriptional and cell-marker studies suggested ECM decreases with later events but lacked peptide-level and post-translational detail. The present work applies targeted proteomic approaches to archival tissue to resolve ECM peptide composition and test associations with later disease.
Samples were drawn from the Resource of Archival Human Breast Tissue (RAHBT) cohort. Ten tissue microarrays were analyzed. The primary DCIS set comprised 136 specimens. Among these, later events included 40 cases of subsequent DCIS and 30 cases progressing to IBC. Mean follow-up time across the cohort was 192.1 months, with a reported 95% confidence interval of 179.1 to 205.1 months.
The analytic strategy combined ECM-targeted mass spectrometry imaging with liquid chromatography tandem mass spectrometry (LC MS/MS). These approaches enabled peptide-level detection from tissue microarrays and characterization of ECM proteomic composition, including peptides reflecting collagen domains and elastin. Statistical modeling, survival analysis, and exploratory machine learning methods were applied to identify peptide signatures associated with later events.
Comparative analyses identified distinct ECM peptide profiles associated with later clinical events. Specifically, fifteen peptides derived from fibrillar collagens—COL1A1, COL1A2, COL3A1—and elastin showed significantly reduced abundance in primary DCIS samples from patients who later developed IBC. The manuscript reports that reductions in these collagen-derived peptides were a consistent feature distinguishing samples from patients with later invasive events.
Lower expression of a set of collagen peptides was associated with reduced disease-free survival for invasive breast cancer. The authors quantify a 19.9% decreased disease-free survival (95% CI 17.92 to 21.81) linked to lower peptide expression. In survival modeling adjusted for age, the lower expression group had an age-adjusted hazard ratio of 2.45 with a 95% confidence interval of 2.33 to 2.57 (P < 0.05). These reported estimates indicate a strong statistical association between decreased detection of specific ECM peptides and later invasive events in this dataset.
The study included patient-matched analyses comparing primary DCIS samples with later DCIS and later IBC from the same patients. These matched comparisons further demonstrated reductions in ECM peptide detection in later events relative to the primary lesion. Exploratory predictive modeling using the matched samples achieved high discrimination and classification performance, with reported area under the receiver operating characteristic curve (AUROC) exceeding 0.98 and accuracy greater than 93% when distinguishing primary samples from later-event samples in the tested models.
The authors note that the observed reduction of certain collagen peptides in primary DCIS is consistent with earlier findings in the RAHBT cohort. They report that reductions in these peptide signals were also seen in primary DCIS samples from patient groups previously characterized as higher risk within that cohort, reinforcing the association between ECM remodeling and elevated risk of later events.
The investigators conclude that ECM proteomic remodeling—especially decreases in specific collagen domains—is strongly associated with later clinical events of DCIS and progression to IBC. They propose that the ECM proteome could act as a regulator of breast cancer emergence and holds potential as a prognostic marker to refine risk stratification and inform clinical management of DCIS.
It is important to note that this report is a preprint and has not been peer reviewed. The source provides summary results, cohort counts, follow-up duration, peptide-level findings, survival estimates, and predictive modeling performance, but detailed methodological parameters, full peptide lists, or underlying datasets beyond those summaries were not provided in the preprint text presented here.
The preprint discloses that several authors have advisory roles and industry relationships relevant to glycomics platforms. Funding sources listed include multiple NIH grants, United States Department of Defense and Veterans Affairs awards, and support from Susan G. Komen Breast Cancer Foundation. These declarations and funding notes are reproduced from the source and indicate potential institutional support for the reported work.
Overall, the study reports that decreased abundance of defined ECM peptides, primarily from fibrillar collagens and elastin, associates with later DCIS or invasive breast cancer events and with reduced disease-free survival in this archival cohort. The authors suggest these ECM proteomic signatures merit further validation as prognostic biomarkers for DCIS management.