Abstract Background: Transcriptomic analyses using patient-matched tumour samples collected before and after treatment enable the assessment of treatment-induced molecular changes by minimizing inter-individual biological variability. However, the absence of untreated control groups often limits the ability to distinguish true treatment-induced changes from sampling-related artefacts. This issue was highlighted in the POETIC window-of-opportunity trial. To overcome this limitation, we systematically evaluated gene expression dynamics (GED) in a cohort of patients with invasive breast cancer (BC) undergoing primary surgery without prior systemic treatment, to provide a reference for treatment-independent GED. Methods: Within the SEOM-ARTEFACT Spanish study (2023-2024), we prospectively collected paired tumour samples from 30 patients with early BC at three timepoints: diagnostic core biopsy (TP1), immediate biopsy from the surgical specimen (TP2), and the routine surgical excision sample after the usual standard fixation delay of 30 minutes to 2 hours (TP3). All patients were treated at the Institut Català d'Oncologia - Hospital Germans Trias i Pujol (Badalona) and received no neoadjuvant or window-of-opportunity therapy. Time from excision to fixation was documented for adjustment. RNA sequencing was used to assess intrinsic subtypes using the PAM50 RNAseq-adapted algorithm and GED. Here, we report transcriptomic findings from the first ten patients with high-quality material available across all three timepoints. Results: Median age was 74 years (range 24-94). Most tumours were ductal (80%) and grade 2 (80%). Surrogate IHC-based subtypes included Luminal B (60%), Luminal A (30%), and triple-negative (10%). Intrinsic and surrogate subtypes matched in 60% of cases (6/10). Only 40% of patients received adjuvant chemotherapy, while 80% were treated with adjuvant endocrine therapy. Despite the absence of preoperative treatment, we observed consistent GED changes between TP1 and TP3: 28 genes were upregulated and 20 downregulated (|log2FC| = 1.5-4.8; FDR 0.05), independent of subtype. Upregulated genes included stress-response and hypoxia-related markers (FOSB, DUSP1, EGR1 involved in FOS-JUN pathways), as well as inflammatory chemokines (IL6, CCL20, CCL4). Downregulated genes were associated with mechanical stimulus response and glucagon metabolism pathways (GNGT2, SLC9A5, TRPV3). No significant differences were found between TP1 and TP2, supporting a sampling-related rather than biological origin. Due to limited material, TP2-TP3 comparisons were not feasible. Subtype classification shifted across timepoints in 3 patients, all with borderline Luminal A/B profiles, indicating only minimal shifts in proliferation and endocrine-related genes. Conclusions: This prospective analysis shows that tumour sampling methodology significantly alters gene expression, likely reflecting transient hypoxia and wound-related immune responses. These changes, driven by upregulation of stress-response genes (FOS-JUN) and local cytokines (IL6, CCL20), highlight the biological impact of biopsy-induced microenvironmental remodelling. Such artefactual alterations represent a key confounding factor in neoadjuvant and window-of-opportunity trial designs and Log2FC values from these untreated controls may serve as correction factors in future studies. Ongoing analyses in the full ARTEFACT cohort will further define the extent and implications of these artefacts. Citation Format: M. A. Bergamino Sirvén, A. Pous, S. Cabrero, A. Castillo, G. Martinelli, L. Pons, C. Perelló, X. Zhu, M. Pascual, E. Riveira, P. Rodriguez, E. Felip, M. Luna, I. Teruel, B. Cirauqui, L. Blay, A. López-Paradís, J. Antón, F. Schettini, M. González-Rodríguez, L. Boronat, A. Mariscal Martínez, M. Margeli, C. Ríos, E. Ballana, P. Fernández-Ruíz. Unmasking Sampling Artefacts in Early Breast Cancer: Molecular Shifts Between Diagnostic and Surgical Specimens in the Absence of Therapy abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS2-10-27.
Sirvén et al. (2026) studied this question.