At the 2025 San Antonio Breast Cancer Symposium (SABCS 2025), a study titled “Gene Expression-based Subtyping of Early Triple-Negative Breast Cancer for Prediction of Response to Neoadjuvant Immune-chemotherapy in the NSABP B-59/GBG-96-GeparDouze Trial” (Abstract No. RF3-02) was presented. The study provides an in-depth analysis of the heterogeneity of benefit from immunotherapy in early triple-negative breast cancer (TNBC), revealing the predictive value of tumor-infiltrating lymphocytes (TILs) and molecular TNBC subtypes for treatment efficacy.Oncology Frontier invited Professor Jin Feng and his colleagues Professor Yu Xinmiao and Professor Zheng Ang from the First Affiliated Hospital of China Medical University to present and comment on this study, offering critical data and clinical insights to support precision immunotherapy in TNBC.
Study Overview
Study Title
Chinese: 基于基因表达的早期三阴性乳腺癌分型用于预测NSABP B-59/GBG-96-GeparDouze新辅助免疫化疗的疗效
English: Gene Expression-based Subtyping of Early Triple-Negative Breast Cancer (TNBC) for Prediction of Response to Neoadjuvant Immune-chemotherapy in the NSABP B-59/GBG-96-GeparDouze Trial
Background
In recent years, immunotherapy—particularly PD-1/PD-L1 inhibitors—has substantially reshaped the treatment landscape of TNBC. The KEYNOTE-522 trial demonstrated that neoadjuvant immunotherapy combined with chemotherapy significantly improves the pathological complete response (pCR) rate in patients with early-stage, high-risk TNBC.
However, not all TNBC patients derive benefit from immunotherapy. Against this background, the NSABP B-59/GBG-96-GeparDouze trial was designed to evaluate whether adding the PD-L1 inhibitor atezolizumab to standard neoadjuvant chemotherapy—and continuing it into the adjuvant phase—could improve outcomes. The trial showed that, compared with chemotherapy alone, the addition of atezolizumab did not significantly improve event-free survival (EFS) in the overall population.
This “overall negative” result reignited a key clinical question: Could immunotherapy still be beneficial for specific molecular subgroups, despite the lack of survival benefit in the unselected population?
Identifying predictive biomarkers capable of accurately selecting patients who will benefit from immunotherapy is therefore critical. Tumor-infiltrating lymphocytes (TILs) play a central role in the tumor immune microenvironment and are widely regarded as important predictors of immunotherapy response.
This study addressed a fundamental question: Can TIL levels and refined molecular subtyping predict response to neoadjuvant immunochemotherapy in early TNBC? Using pretreatment biopsy gene expression profiles, the investigators systematically evaluated the roles of TILs and TNBC molecular subtypes in determining treatment response, aiming to generate new evidence for individualized TNBC therapy.
Methods
A total of 1,520 patients with TNBC (tumor size >2 cm and/or clinically node-positive disease) were enrolled. Patients were randomized to receive neoadjuvant chemotherapy plus placebo or neoadjuvant chemotherapy plus atezolizumab.
To assess immunotherapy efficacy, TIL levels were evaluated in all patients, and gene expression profiling was performed in 482 patients.
Key Methodological Components
- TIL Stratification Patients were categorized based on stromal TIL levels into: Low (<10%) Intermediate (10–29%) High (≥30%)
- TNBC Molecular Subtyping (Immune-Response–Based) Basal-like immune-activated (BLIA) Basal-like immune-suppressed (BLIS) Luminal androgen receptor (LAR) Mesenchymal (MES)
- Primary Endpoints Pathological complete response (pCR) Event-free survival (EFS)
Results
TILs and Treatment Response
Higher TIL levels were strongly associated with improved pCR rates.
- High-TIL group: pCR rate of 68.6%
- Low-TIL group: pCR rate of 39.7%
Patients with high TILs also demonstrated significantly superior EFS, underscoring the critical role of immune activation in antitumor response.
Predictive Value of TILs for Immunotherapy Benefit
In patients with high TILs, the addition of atezolizumab increased the pCR rate from 65.4% to 71.9% and resulted in a statistically significant improvement in EFS. In contrast, patients with low or intermediate TILs derived minimal additional benefit from atezolizumab.
Transcriptomic Subtyping Refines Benefit Prediction
The heterogeneity observed in the intermediate-TIL group prompted further transcriptomic analysis. Using BLIA and BLIS subtyping:
- BLIA tumors exhibited high immune activation, with a pCR rate of 68.9%
- BLIS tumors showed immune suppression, with a significantly lower pCR rate of 38.6%, and poorer survival outcomes
A key finding was that the EFS benefit from atezolizumab was almost entirely confined to patients with BLIA tumors.
Within the intermediate-TIL group:
- ~51% were BLIA
- ~39% were BLIS
Atezolizumab significantly improved EFS only in the BLIA subgroup, while outcomes in BLIS patients were similar to placebo.
Conclusions
Atezolizumab significantly improves EFS in patients with immune-activated tumors, including those with TILs ≥30% or the BLIA molecular subtype. Tumors with intermediate TIL levels are biologically heterogeneous, encompassing both immune-activated (BLIA) and immune-exhausted (BLIS) states.
These findings strongly support the incorporation of TNBC molecular subtyping into clinical trials and diagnostic practice. Biomarker selection for immunotherapy continues to evolve, and future efforts should focus on more precise identification of high-TIL tumors and BLIA subtypes within intermediate-TIL populations. Validation of differential efficacy between PD-L1 and PD-1 inhibitors in independent clinical cohorts is also warranted.
Expert Commentary
This study represents a rigorous and insightful post-hoc analysis of a phase III trial with a negative primary endpoint, successfully uncovering meaningful heterogeneity in immunotherapy benefit and providing a strong foundation for precision immunotherapy in TNBC.
First, the study confirms that atezolizumab significantly improves EFS in patients with immune-activated tumors, particularly those with high TILs or BLIA subtype. In clinical practice, TIL assessment can serve as a practical tool to predict immunotherapy responsiveness.
Importantly, the study highlights pronounced heterogeneity among tumors with intermediate TIL levels, which include both immune-activated and immune-suppressed states. This explains the variable and often disappointing immunotherapy responses observed in this subgroup. Clinically, more refined molecular and immune microenvironment profiling is required to guide treatment decisions.
BLIA tumors demonstrate robust immune activation and favorable responses to immunotherapy, whereas BLIS tumors exhibit immune suppression and limited benefit. For patients with intermediate TILs, future strategies should integrate molecular subtyping and immune biomarkers to enable individualized treatment selection, avoiding a “one-size-fits-all” approach.
In summary, this work represents a high-quality “post-race analysis” in the field of immunotherapy. Different TNBC subtypes require different “tracks and pacing strategies.” Translating multigene expression profiling into clinically accessible neoadjuvant treatment algorithms—and shifting decision-making from whether to use immunotherapy to which patients should receive which immunotherapy combinations—will be a critical area of future research. Advances in artificial intelligence, multi-omics integration, and predictive modeling are expected to further optimize personalized treatment strategies for TNBC.
Professor Jin Feng
Professor Yu Xinmiao
Professor Zheng Ang
