Editor’s Note Spotlight on SABCS, insights delivered live. The 48th San Antonio Breast Cancer Symposium (SABCS) was held from December 9 to 12, 2025, in San Antonio, USA, bringing together cutting-edge research findings and clinical insights from across the global breast cancer community. To efficiently convey the highlights of the meeting, this edition of Insight Broadcast invited Professor Junjie Li from Fudan University Shanghai Cancer Center as an on-site expert commentator. Focusing on neoadjuvant therapy for HER2-positive breast cancer, Professor Li provides an in-depth interpretation of the latest research advances on the use of biomarkers such as tumor-infiltrating lymphocytes (TILs) and circulating tumor DNA (ctDNA) to enable more precise patient selection and response prediction, while critically examining their potential, limitations, and future directions from a clinical perspective.01
TILs as Predictors of Neoadjuvant Response in HER2-Positive Breast Cancer: Promising but Far from Mature
Professor Junjie Li: The first study I will discuss today explores the predictive value of tumor-infiltrating lymphocytes (TILs) in neoadjuvant therapy for HER2-positive breast cancer. Although neoadjuvant treatment strategies for HER2-positive disease are increasingly diverse, the core challenge remains how to accurately identify patients who are more sensitive to treatment—whether for escalation or de-escalation strategies. This study provides important clues in that regard.
The ECOG-ACRIN EA1181 / CompassHER2 pCR study incorporated TILs as a biomarker for analysis. The study design itself is quite distinctive: all patients received four cycles of neoadjuvant THP (taxane + trastuzumab + pertuzumab). Patients who achieved pathological complete response (pCR) continued with adjuvant HP, while those who did not achieve pCR were considered for treatment escalation.
The key question was which patients could still benefit from de-escalated therapy. At the 2025 ASCO Annual Meeting, the HER2Dx score was reported to be associated with pCR; the current analysis focused on TILs. Because the proportion of patients with TILs ≥60% was relatively small, the study used a cutoff of 10% to stratify patients into two groups. Among 1,300 evaluable patients, those with TILs ≥10% had significantly higher pCR rates, regardless of hormone receptor status.
However, from a clinical decision-making standpoint, the value of TILs must be interpreted cautiously. Multivariable analysis showed that the odds ratio (OR) for the association between TILs and pCR was 1.52—lower than that of estrogen receptor status (OR 3.5), HER2 IHC 3+ (OR 6.5), and even lower than the HER2Dx score (OR 2–3). This indicates that while TILs are a relevant biomarker, their predictive strength is limited. They are unlikely to independently guide major treatment decisions, and their standalone use is insufficient to change current clinical practice.
This study also highlights challenges in standardizing TIL assessment. Cutoff values, evaluation methods (manual vs AI-based), and scoring criteria vary across studies, limiting generalizability and clinical translation. An ideal biomarker should not only have prognostic value but also predictive value—namely, the ability to guide treatment strategy adjustments.
In my view, TILs are a biomarker with potential, but we are still at an early stage. A truly effective predictive biomarker should inform not only prognosis but also therapeutic decision-making. Based on this retrospective analysis alone, we are far from achieving that goal. Longer follow-up, prospective validation, and genuinely standardized assessment workflows are still needed.
Study Overview
Abstract No.: GS1-04 Title: Tumor-infiltrating lymphocytes (TILs) and pathologic complete response (pCR) following taxane, trastuzumab, and pertuzumab (THP) in stage II/III HER2-positive breast cancer: Secondary results from ECOG-ACRIN EA1181 / CompassHER2 pCR trial
Background
The association between TILs and pathological complete response (pCR) or survival outcomes in HER2-positive breast cancer remains inconsistent across prior studies. This secondary analysis of the EA1181 trial evaluated the relationship between stromal TILs (sTILs) and pCR.
Methods
EA1181 (NCT04266249) enrolled patients with anatomically staged II–III HER2-positive breast cancer who received four cycles of neoadjuvant THP followed by surgery. Stromal TIL density was assessed on full-face hematoxylin and eosin–stained tumor biopsy sections. sTILs were analyzed as both continuous variables (per 10% increase) and categorical variables using the 60% cutoff proposed by Denkert et al. (2018), as well as an exploratory 30% cutoff commonly used in triple-negative breast cancer. Cox proportional hazards models were used to evaluate associations between sTILs and pCR, adjusting for available baseline factors.
Results
Among 2,141 patients treated in EA1181, 1,328 (62%) had evaluable H&E slides. The overall pCR rate was 44.5%—64% in the HER2+/ER− subgroup and 33% in the HER2+/ER+ subgroup. sTIL distribution was <10% in 47%, 10–59% in 22%, and ≥60% in 31% of patients. Increased sTILs were significantly associated with higher pCR rates in both HER2+/ER+ and HER2+/ER− disease. When analyzed as <10% versus ≥10%, sTILs were significantly associated with pCR in the overall population and in HER2+/ER+ disease.
Conclusion
Baseline sTILs were associated with pCR following THP therapy, supporting the importance of immune mechanisms in HER2-positive breast cancer and suggesting that sTILs may serve as a predictive biomarker. Associations between baseline sTILs and recurrence-free survival will be reported in future analyses.
02
Dynamic Clearance of ctDNA: A Powerful Signal for Predicting pCR in Neoadjuvant Therapy
Professor Junjie Li: The second study examined the predictive role of circulating tumor DNA (ctDNA) in neoadjuvant therapy for HER2-positive breast cancer. ctDNA is a current hotspot in oncology research and has already demonstrated prognostic and predictive value in advanced breast cancer. The key question is how ctDNA performs in early-stage HER2-positive disease during neoadjuvant treatment.
This analysis was embedded within the landmark PHERGain trial, which used early metabolic response on PET-CT to guide de-escalation strategies. Patients with a favorable metabolic response (≥40% reduction in SUV) continued dual HER2 blockade (HP) without chemotherapy, while non-responders received added chemotherapy. Ultimately, approximately 40% of patients successfully avoided chemotherapy with favorable outcomes. The translational objective of this substudy was to determine whether ctDNA could predict pCR earlier or more accurately than imaging.
The results were highly informative:
- Baseline ctDNA status was not associated with pCR, which is biologically plausible. Baseline ctDNA positivity reflects tumor burden and prognosis but does not necessarily indicate treatment resistance.
- Dynamic ctDNA clearance was critical. Conversion from detectable to undetectable ctDNA during treatment (“clearance”) was strongly associated with pCR. This association was already significant after two treatment cycles (P = 0.003) and became even stronger before surgery (P < 0.001). Notably, none of the patients with detectable ctDNA prior to surgery achieved pCR—representing an exceptionally strong negative predictive signal.
- Technical maturity is improving. The baseline ctDNA detection rate reached 71% overall and 93% in stage III patients, indicating that current technologies are capable of reliably detecting minimal residual disease in early-stage settings.
This raises a deeper question: where does ctDNA add the greatest value? If imaging already suggests complete clinical response, ctDNA clearance may be confirmatory. However, its greatest potential likely lies in patients with equivocal imaging findings or minimal residual disease, where preoperative ctDNA status could offer a more objective and precise assessment and potentially influence decisions regarding adjuvant therapy intensity.
Overall, this was a well-designed prospective biomarker study. It not only validates the predictive value of dynamic ctDNA monitoring but also provides an important framework for designing more refined, individualized neoadjuvant treatment strategies.
Study Overview
Abstract No.: GS1-06 Title: Circulating tumor DNA (ctDNA) analysis in HER2-positive early breast cancer (EBC): A translational analysis from the PHERGain neoadjuvant personalized treatment study
Background
HER2-targeted therapy has significantly improved outcomes in HER2-positive early breast cancer, prompting exploration of de-escalation strategies. ctDNA represents an emerging tool for risk stratification and real-time monitoring. This PHERGain substudy evaluated a tumor tissue–agnostic epigenomic ctDNA assay to enhance prediction of pCR and 3-year invasive disease-free survival (iDFS).
Results
ctDNA clearance after two cycles and prior to surgery was strongly associated with pCR. Baseline ctDNA positivity was associated with poorer 3-year iDFS, while all patients with detectable ctDNA before surgery failed to achieve pCR.
Conclusion
Dynamic ctDNA clearance is a powerful predictor of pCR in HER2-positive early breast cancer undergoing neoadjuvant HER2-targeted therapy. Further prospective validation is warranted.
Summary
Together, these two studies illustrate important steps toward precision neoadjuvant therapy in HER2-positive breast cancer. TILs are associated with pCR but have limited predictive strength and face challenges in standardization. In contrast, dynamic ctDNA monitoring demonstrates stronger translational potential, with clearance during treatment serving as a robust molecular signal for response prediction. Future efforts should focus on standardizing TIL assessment, prospectively validating ctDNA-guided strategies, and integrating multi-omics and artificial intelligence to ultimately achieve truly individualized neoadjuvant treatment.
Professor Junjie Li Chief Physician, Associate Professor, MD Supervisor for Master’s Students Department of Breast Surgery Fudan University Shanghai Cancer Center
