Abstract Background: Currently, only a small fraction (10%) of patients with breast cancer receiving immune checkpoint inhibitors (ICI) as standard of care (SOC) achieve clinical benefit. Additionally, immunotherapy carries serious and long-lasting immune-related adverse events, highlighting the need for effective biomarkers for patient selection. Most biomarkers evaluated thus far, including PD-L1, stromal tumor-infiltrating lymphocytes (sTILs), TMB and most gene expression profiles, are generally prognostic or predictive of chemotherapy response, and lack immunotherapy benefit specificity. Some biomarkers, like DetermaIO, tumor-specific (ts)MHC-II have been identified as potential predictive biomarkers for immunotherapy-specific response across multiple tumor types and trials while many others are generally prognostic or predictive of response in both chemotherapy- and chemo-immunotherapy-treated patients. We present the results from the early-stage cohort of TBCRC061, a non-interventional correlative trial leveraging archival tissues from patients treated with SOC ICI (neoadjuvant target accrual n=80, and metastatic n=100) to validate these biomarkers. Methods: Patients with breast cancer receiving SOC chemoimmunotherapy were identified and consented. Pre-treatment formalin-fixed and paraffin embedded archival samples were collected. Relevant patient demographics and response to treatment were extracted from the medical record. The neoadjuvant cohort completed accrual in 2025 and will be presented here. Results: Of 85 consented patients in the neoadjuvant cohort, 67 completed neoadjuvant treatment and had sufficient and suitable tissue for correlative analysis. Forty-seven patients (69%) achieved a complete or near-complete response (defined as RCB 0/I). tsMHC-II demonstrated a trend for enhanced response (ts-HLA-DR≥5%, n=40, 78% vs 59%, p=0.09, Fisher’s exact test). sTILs ≥30% defined numerically higher rates of response (83% vs 66%, p=0.14, Fisher’s exact test). Patients with tumors with immune infiltration in the tumor core (inflamed, n=29) or brisk accumulation of TILs in the tumor margin (margin predominant, n=30) achieved similar response rate (79% vs 73%), however patients with ≤5% sTILs and no accumulation on the margins (immune desert; ID) had drastically lower response rates (17%, n=6, p=0.01, Fisher’s exact test). The combination of tsMHC-I 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS2-07-15.
Balko et al. (2026) studied this question.