Abstract PS3-13-18: Ultrasound-guided biopsy-based tils assessment at tumor margins as a predictive biomarker of neoadjuvant treatment response in early her2-positive and triple-negative breast cancer: a multicenter study

Abstract Background: Tumor-infiltrating lymphocytes (TILs) reflect host immune response and are associated with prognosis and treatment response in HER2-positive and triple-negative breast cancer (TNBC). While the International Immuno-Oncology Biomarker Working Group recommends evaluating stromal TILs in tumor areas, intratumoral heterogeneity and small sample size in core needle biopsy (CNB) specimens limit consistent assessment. It remains unclear whether central tumor (CT) or invasive margin (IM) regions should be prioritized and what cut-off values are clinically relevant. Ultrasonographic (US) features may also reflect the immune microenvironment. We previously proposed a TILs-US score based on three US features that predicts lymphocyte-predominant breast cancer (LPBC). This study evaluated the clinical utility of TILs assessment from CT and IM regions, appropriate cut-off values for predicting response in HER2-positive and TNBC using CNB, and explored optimal biopsy targeting through pathologic and US comparisons. Methods: In this multicenter retrospective study, we analyzed 239 patients (124 HER2-positive, 115 TNBC) with cT1-3/cN0-2 breast cancer who underwent surgery after neoadjuvant chemotherapy (NAC). TILs were evaluated using CNB specimens obtained before NAC and assessed separately in the central tumor (CT) and invasive margin (IM) regions. The IM was defined as a 1-mm-wide area centered on the interface between the tumor and surrounding stroma, while the CT was defined as the remaining intratumoral area within the IM boundary. TILs were analyzed using binary cut-offs of 30% and 50%, as well as a three-tier classification: low (0-10%), intermediate (11-49%), and high (≥50%). Pre-treatment ultrasound (US) findings included tumor shape (scored 0-2), internal echo (0-2), and posterior echo (0-3), which were used to calculate a composite TILs-US score ranging from 0 to 7. Associations between TIL levels, US features, and pathological complete response (pCR) were analyzed. Results: Overall, 104 patients achieved pCR (HER2-positive: 39; TNBC: 65). Using a 30% cut-off, TILs were evenly distributed both CT and IM, while ≥50% cases were relatively rare especially CT (CT: 10%; IM: 20%). CT TILs ≥30% was significantly associated with pCR (p=0.001), whereas the ≥50% cut-off was not (p=0.128). In contrast, both 30% and 50% cut-offs in the IM region were predictive of pCR (p<0.001 and p=0.009, respectively). Subtype analysis revealed that in TNBC, TILs ≥30% in both CT and IM regions were significantly correlated with pCR. In HER2-positive cases, however, only IM TILs ≥30% was predictive of pCR. Similar trends were observed with the three-tier classification. Among US features, a small lobulated shape was associated with high TILs levels in both CT and IM regions, while an enhanced posterior echo was correlated with high TILs in the IM region. Conclusion: Our findings suggest that TILs assessment should include the IM in addition to the CT, especially in CNB specimens as provide valuable information. Given the limited tissue area available in CNB specimen, using a 30% cut-off for TILs may offer a clinically meaningful threshold to predict treatment response. In cases where US findings suggest central fibrosis, targeting only the tumor center may result in sampling fibrotic areas with sparse immune and tumor cells, potentially hindering proper evaluation. Therefore, selecting the biopsy target based on US imaging is essential. Especially in HER2-positive breast cancer, as well as TNBC, it is important for operators to consider the optimal needle insertion angle and sampling site to ensure appropriate sampling of viable tumor tissue and immune microenvironment for accurate pathological and immunological evaluation. Citation Format: Y. Kimura, R. Yamaguchi, K. Fukui, A. Kanou, M. Noma, S. Akashi-Tanaka, K. Arihiro, N. Iwamoto, Y. Ohta, T. Okuno, S. Kashiwagi, Y. Kamei, Y. Tanada, S. Nakano, A. Nagata, A. Murakami, Y. Mochitomi, A. Yamakawa, M. Yamaguchi, H. Shima, A. Emi. Ultrasound-guided biopsy-based tils assessment at tumor margins as a predictive biomarker of neoadjuvant treatment response in early her2-positive and triple-negative breast cancer: a multicenter study [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 PS3-13-18.