Abstract Background: Baseline plasma levels of Tumor Necrosis Factor-related Weak Inducer of Apoptosis (TWEAK), Vascular endothelial growth factor A (VEGF-A), Programmed cell death 1 ligand 2 (PD-L2), and Osteoprotegerin (OPG) were analyzed for their predictive value in women with early-stage HER2-negative breast cancer who received neoadjuvant chemotherapy or endocrine therapy, and were randomized 1:1 to additional MUC1-based immunotherapy (tecemotide, L-BLP25). Methods: Plasma levels of the biomarkers were assessed in 314 patients from the prospective, randomized, open-label, 2-arm phase-II ABCSG-34 trial before neoadjuvant treatment. Of these, 240 received chemotherapy and 74 endocrine treatment, with (n=160) or without (n=154) tecemotide, a mucin 1 (MUC1) antigen-specific peptide vaccine. Treatment response was evaluated based on residual cancer burden (RCB) classes (0 to 3) and pathologic complete response (pCR). Long-term outcome data on invasive disease-free survival (iDFS), overall survival (OS) and distant recurrence-free survival (DRFS) were available in 227 patients. Associations of log2-transformed biomarkers with response and with survival were evaluated using ordinal logistic regression and Cox models. A potential predictive role for immunotherapy benefit was tested with interaction terms. For biomarkers demonstrating predictive value, an optimal cut-off was determined using ordinal logistic regression. Results: In vaccinated patients, higher TWEAK levels at baseline correlated with worse RCB class (OR 0.66; 95% CI, 0.48-0.90). This association differed significantly from that observed in non-vaccinated patients (interaction p = 0.01), where baseline TWEAK levels were not associated with RCB classes (OR 1.13; 95% CI, 0.86-1.50).Similar, though non-significant, findings were observed for pCR. Among vaccinated patients, higher baseline TWEAK levels were associated with lower pCR-rates (OR 0.66; 95% CI, 0.41-1.06), while in non-vaccinated patients no such association was observed (OR 1.13; 95% CI, 0.78-1.63; interaction p = 0.08). Apart from TWEAK, no associations were found between treatment response and baseline plasma levels of VEGF-A, PD-L2, or OPG. Consequently, an optimal predictive cut-off for TWEAK (450 pg/ml) for RCB classes was determined based on the highest discrimination (=c-index). After a median follow-up of 7.2 years, 81 iDFS, 53 OS, and 62 DRFS events were documented. When considering TWEAK levels as a continuous variable in the COX model, results for long-term outcomes showed a similar pattern to what was observed for short-term response; however, the number of events was low, and interaction p-values did not reach statistical significance (iDFS: p = 0.63; OS: p = 0.26; DRFS: p = 0.14). Patients with low TWEAK (≤450 pg/ml) seemed to derive greater long-term benefit from immunotherapy: compared to non-vaccinated patients, those treated with tecemotide showed improved 7-year iDFS (72% vs. 58%, HR 0.61), OS (86% vs. 66%, HR 0.43), and DRFS (84% vs. 61%, HR 0.36). In contrast, among patients with high baseline TWEAK, survival differences between groups were minimal, not exceeding 5%. Conclusion: Low baseline TWEAK levels predicted significantly better pathological response measured by RCB to MUC1-based vaccination. Although long-term outcomes did not reach statistical significance, clinically meaningful benefits in iDFS, OS, and DRFS were seen in patients with low TWEAK baseline levels. These findings support TWEAK as a potential predictive biomarker for identifying patients likely to benefit from neoadjuvant immunotherapy. Further validation in larger cohorts and in different neoadjuvant settings including currently used combinations of chemo- and immune checkpoint inhibitor therapy is warranted. Citation Format: K. Wimmer, D. Hlauschek, A. Rauch, M. Sachet, C. Ramos, V. Gerakopulos, G. Pfeiler, C. Brunner, G. Pristauz-Telsnigg, G. Rinnerthaler, A. Pichler, F. Fitzal, S. P. Gampenrieder, G. Huber, M. Seifert, D. Egle, M. Filipits, R. Oehler, C. Singer, M. Gnant, Austrian Breast 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS2-07-21.
Wimmer et al. (Tue,) studied this question.