TPS618 Background: It is unclear whether the clinical benefit of cytotoxic adjuvant chemotherapy (CT) could be replaced by ovarian-function suppression (OFS) in premenopausal, estrogen receptor (ER)+HER2- breast cancer patients who have high clinical risk score and low genomic risk assessed by multigene assays. The TAILORx trial included a subgroup of premenopausal women with high clinical risk scores and midrange RS scores and found that CT, in addition to endocrine treatment (ET), offered clear benefits in terms of invasive disease-free survival (iDFS) and distant-recurrence-free survival (DRFS) compared to ET alone. However, a majority of the patients did not receive OFS, and the use of OFS as an alternative to chemotherapy in this population is still an area of ongoing research and debate. In addition, in premenopausal women of the RxPonder trial, which enrolled node-positive disease, it is noted that the addition of OFS to ET for at least 12 months improved iDFS numerically in the ET-alone arm, although this improvement did not reach statistical significance. We hypothesized that a favorable DRFS could be achieved by OFS plus ET without CT in premenopausal, pN1, ER+HER2- breast cancer with low genomic risk identified by an NGS-based multigene assay, the OncoFREE. Methods: The INTERSTELLAR trial is a prospective, multicenter, single-arm, non-inferiority clinical study. Premenopausal women aged ≤50 years with pT1-2 ER+HER2- breast cancer and 1-3 lymph node metastasis will be enrolled. They will be tested with OncoFREE, an NGS-based breast cancer prognosis multigene assay developed and available in South Korea, where a higher portion of the patients is premenopausal. Patients with low genomic risk (Decision Index≤20) are administered OFS plus tamoxifen or an aromatase inhibitor for five years. We hypothesize that the 5-year DRFS of the single arm treated with OFS plus ET would be not inferior to 96.1%, which is observed in the chemo-ET arm from the premenopausal subgroup of the RxPonder trial. The one-sided test with a non-inferiority margin of 3% and statistical power of 80% at a significance level of 0.05 resulted in a sample size of 380 patients with low genomic risk. Considering a 70% designation to low genomic risk by OncoFREE and a 10% drop-out rate, 604 patients will be enrolled from 15 tertiary care hospitals in South Korea. The primary endpoint will be tested in the 380 patients with low genomic risk. The patients with high genomic risk will receive CT followed by ET and will be followed for survival analysis as a secondary end-point. The trial has not enrolled its first patient yet at the time of submission. Clinical trial information: NCT05333328 .

BACKGROUND:Women with a history of breast cancer face an elevated risk of developing contralateral breast cancer (CBC). Although annual mammographic screening is recommended, its sensitivity is limited, particularly in younger women and those with dense breast tissue. This study assessed the potential of artificial intelligence (AI)-based computer-aided diagnosis (AI-CAD) to improve CBC detection in a cohort of breast cancer survivors and to enhance early CBC detection. METHODS:This study included 454 women who developed CBC and 454 matched controls without recurrence. Mammograms were analyzed using AI-CAD software assigning abnormality scores from 0 to 100, with scores over 10 indicating CBC. Standalone AI predictions were compared with radiologists' initial assessments, and its ability to detect early signs of CBC in prior mammograms was evaluated. RESULTS:The median age of patients was 53 years; the majority (79.9%) had dense breast tissue. Standalone AI detected 271 CBC cases solely from mammography, achieving 6.2% higher sensitivity compared with radiologists (59.7% vs. 53.5%, p = 0.009). Notably, AI detected 66 CBC cases (14.5%) that radiologists had missed on mammographic assessment alone. With respect to earlier detection, 81 cases (29.9%) were identified on prior mammograms at a median of 13.3 (9.6-19.9) months before clinical diagnosis. Of these, 28 cases (10.3%) were detected more than 6 months before pathological confirmation, and 53 cases (19.6%) more than 1 year earlier. CONCLUSIONS:AI-CAD may enhance the detection of CBC in breast cancer survivors and facilitate earlier identification on surveillance mammography. Further studies are needed to assess its integration into surveillance mammography and clinical impact.

Importance:Clonal hematopoiesis of indeterminate potential (CHIP) is linked to an increased incidence of cardiovascular and malignant diseases. Objective:To determine whether CHIP is associated with cardiotoxic effects in patients with breast cancer receiving trastuzumab. Design, Setting, and Participants:Analyses of human cohorts with complementary animal experimentation were performed using a nationwide population-based cohort (UK Biobank), 1 tertiary referral center (Seoul National University Hospital [SNUH]), and a controlled laboratory setting. UK Biobank participants were enrolled between 2006 and 2010, and SNUH patients were enrolled from January 2004 to March 2024. Data were analyzed from March 2021 to February 2026. Main outcome measures:Incident heart failure (HF) in the UK Biobank cohort and trastuzumab-related cardiotoxic effects in the SNUH cohort were the main outcome measures. Trastuzumab-related cardiotoxic effects were defined using established clinical criteria (European Society of Cardiology [ESC], Canadian Trastuzumab Working Group, and Cardiac Review and Evaluation Committee [CREC]). Fine-Gray competing risk models were used to account for death and myocardial infarction; multivariable models were adjusted for age and cardiovascular risk factors in both cohorts, with additional adjustment for anthracycline use in the SNUH cohort. Changes in left ventricular ejection fraction (LVEF) were assessed in Tet2-deficient bone marrow chimeric mice following trastuzumab exposure. Results:Overall, 15 729 patients with breast cancer from the UK Biobank cohort (mean [SD] age, 58.8 [7.3] years; 107 [0.68%] male) and 454 female patients with breast cancer who received trastuzumab from the SNUH cohort (mean [SD] age, 52.0 [9.6] years) were included. The corresponding 2-year cumulative incidence values for trastuzumab-related cardiotoxic effects were 15.7% vs 5.0% by ESC criteria (Gray test P = .001), 19.9% vs 10.8% by Canadian criteria (P = .01), and 20.9% vs 11.3% by CREC criteria (P = .02). Using the ESC definition, CHIP positivity (variant allele frequency ≥1.0%) was associated with cardiotoxic effects in multivariable competing risk analysis (adjusted subdistribution hazard ratio, 1.91; 95% CI, 1.32-2.76). Tet2-deficient mice demonstrated a significant LVEF reduction following trastuzumab treatment (effect size, -4.2%; 95% CI, -7.91 to -0.49; P = .03); other experimental groups showed no significant change. Conclusions and Relevance:In this cohort study, the presence of CHIP was associated with increased susceptibility to trastuzumab-related cardiotoxic effects.

BACKGROUND:The NAUTILUS trial randomized cT1-2/N0 breast cancer patients to evaluate the non-inferiority of omitting SLNB. We report the clinicopathologic characteristics and axillary lymph node (ALN) status of the patients enrolled in the NAUTILUS trial and suggest expectations based on the results of this trial, which are relevant in the context of the SOUND and INSEMA trials, where the majority of participants were aged 50 years or older. METHODS:The NAUTILUS trial randomized 1734 subjects into SLNB or no-SLNB arms. Axillary ultrasonography was mandatory to determine clinical N0. Clinicopathologic variables and ALN status in the SLNB arm were analyzed to determine expectations for the NAUTILUS trial results compared to other clinical trials. RESULTS:Among 1734 patients, 1664 subjects were available for clinicopathologic analysis; 50.4% were in the SLNB arm and 49.6% were in the no-SLNB arm. Median age was 55 (range, 29-92) years, and 40.1% were premenopausal. Overall, 1.7%, 83.9%, and 14.0% subjects were pTmic, pT1, and pT2, respectively, with a median tumor size of 1.3 cm (range, 0.1-6.0). In the SLNB arm, 11.2% had ALN metastasis, comprising 1.1%, 9.4%, and 0.6% with pN1mic, pN1, and pN2-3, respectively. ALN metastasis rates according to tumor size were 7.0%, 12.8%, and 17.2% for sizes ≤1.0 cm, >1.0 cm & ≤ 2.0 cm, and >2.0 cm & ≤ 5.0 cm, respectively. CONCLUSIONS:The NAUTILUS trial completed enrollment, with included 14.0% pT2 and 40.1% premenopausal subjects and is expected to show the impact of SLNB omission in these subgroups. TRIAL REGISTRATION:ClinicalTrials.gov Identifier: NCT04303715.

The rapid and accurate detection of multiple cancers presents considerable challenges, especially for stage I disease, due to the low concentration and heterogeneous nature of circulating tumor DNA. Here we introduce an enhanced multicancer screening assay that integrates whole-genome methylation sequencing with an innovative multimodal analytical framework for cell-free DNA. The ensemble machine learning model integrates four specific cell-free DNA characteristics: average methylation fraction, copy number variation, fragment size ratio and fragment size distribution. The model underwent testing on 1415 samples, encompassing eight primary cancer types and healthy controls. The sensitivity was 93.2%, and the specificity was 95%. The test demonstrated effectiveness in detecting cancers at early stages. The sensitivity was 92.3% for stage I and 92.2% for stage II. The multimodal technique successfully combined average methylation fraction's sensitivity to early epigenetic signals with fragmentomic characteristics. This facilitated the differentiation between healthy individuals and those with early stage cancer. The model achieved an accuracy rate of 85.7% in the top 2 category for correctly identifying the tissue of origin. The results confirm that whole-genome methylation sequencing-based multimodal analysis can improve multicancer early detection technology and revolutionize cancer screening methods.

Abstract Introduction: Triple-negative breast cancer (TNBC) has higher tumor immunogenicity than other subtypes, and immune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis have been applied in its treatment. However, responses to immune checkpoint blockade (ICB) remain limited by intrinsic and acquired resistance, highlighting the need for predictive biomarkers and new combination strategies to improve anti-PD-L1 efficacy.The aryl hydrocarbon receptor (AHR) regulates immune responses and PD-L1 expression, promoting an immunosuppressive tumor microenvironment (TME). AHR signaling may also interfere with interferon-mediated immune activation and affect ICB response. This study aimed to examine how AHR activation contributes to anti-PD-L1 resistance in PD-L1 overexpressing TNBC and whether AHR blockade enhances therapeutic efficacy. Methods: PD-L1-overexpressing (PD-L1 OE) and control (Ctl) 4T1 cells were generated. The AHR antagonist (CH-223191) and anti-PD-L1 antibody were applied in vitro and in vivo. Flow cytometry, western blotting, cell proliferation assay, quantitative RT-PCR, immunohistochemistry (IHC), and co-culture of macrophages and cancer cells were performed. An orthotopic syngeneic mouse model was established by injecting 4T1 Ctl or PD-L1 OE cells into the mammary fat pad. Mice were treated with CH-223191 (5 mg/kg, i.p., four times weekly) and anti-PD-L1 antibody (150 µg, i.p., twice weekly), alone or in combination. Results: PD-L1 OE tumors showed reduced responsiveness to anti-PD-L1 therapy compared with control in vivo, confirming intrinsic resistance to ICB. AHR expression was elevated at both RNA and protein levels in PD-L1 OE tumors after anti-PD-L1 treatment. Combination therapy with CH-223191 and anti-PD-L1 antibody tended to suppress tumor growth compared with monotherapy, suggesting an ability to overcome PD-L1-driven resistance in vivo. In vitro, CH-223191 reduced 4T1 viability in a dose-dependent manner, suggesting a potential tumor-intrinsic role of AHR signaling. Co-culture of PD-L1 OE 4T1 cells with macrophages promoted M2-like polarization, supporting an immunosuppressive microenvironment. Collectively these findings suggest that AHR activation contributes to anti-PD-L1 resistance by promoting an immunosuppressive TME, and that AHR blockade could enhance anti-PD-L1 efficacy in TNBC. Conclusions: AHR activation was linked to PD-L1-driven anti-PD-L1 resistance, and AHR inhibition enhanced responsiveness in vivo. These findings suggest that AHR mediates PD-L1 axis-driven resistance and that its inhibition may serve as a promising combination strategy in TNBC. Further studies should elucidate the mechanistic link between AHR signaling and PD-L1-mediated immune regulation and validate its therapeutic potential. Citation Format: Minkyoung Choi, Ju Hee Kim, Han-Byoel Lee, Wonshik Han. Targeting AHR overcomes PD-L1-driven resistance to immune checkpoint blockade in triple-negative breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2813.

Purpose: Langer's axillary arch (LAA) is a common muscular variant that may affect the assessment of axillary lymph nodes in breast cancer. We investigated whether the presence of the LAA had a meaningful clinical impact in a large cohort of Korean patients with breast cancer. Methods: The medical records and imaging of 2,904 patients (2,953 cases) who underwent axillary surgery for breast cancer were reviewed. The presence of LAA was determined using chest computed tomography scans. The number of retrieved lymph nodes, the accuracy of axillary ultrasonography, and survival outcomes were compared between the LAA and non-LAA groups. Results: The LAA muscle was identified in 301 patients (10.2%). Preoperative ultrasound evaluation showed a lower specificity in the LAA group (77.9% vs. 83.6%, p = 0.031). Additionally, invasive disease-free survival (iDFS) was significantly worse in the 'LAA-positive patients' after adjusting for stage, histologic grade, and treatment (adjusted hazard ratio, 2.00; 95% confidence interval, 1.00-4.01; p = 0.048), with regional axillary recurrence being notably more frequent in patients with the LAA muscle (p = 0.005). Conclusion: The presence of the LAA muscle was associated with reduced specificity on preoperative ultrasound. Patients with LAA muscle-associated tumors had worse iDFS, with regional axillary recurrence occurring more frequently. Recognition of the LAA muscle during axillary staging and surgery may improve nodal assessment accuracy and outcomes, especially in the era of axillary surgery de-escalation.

INTRODUCTION:Omission of breast surgery after neoadjuvant systemic therapy (NST) may reduce treatment burden in exceptional responders, but safe implementation requires reliable identification of pathologic complete response (pCR). This study assessed the performance of OPTIMIST imaging criteria for predicting pCR and explored whether refined thresholds could improve applicability without compromising performance. MATERIALS AND METHODS:We retrospectively reviewed patients who received NST followed by surgery at a single tertiary center between 2016 and 2020. Among 1659 patients, 499 met OPTIMIST-aligned eligibility criteria: cT1-2, cN0-2, M0 disease with HER2-positive, triple-negative, or low estrogen receptor-positive subtypes. Post-NST imaging criteria included residual tumor size, lesion-to-background signal enhancement ratio (L-to-B SER), and residual calcifications. The primary endpoint was negative predictive value (NPV) for pCR. The performance of the original OPTIMIST criteria and modified tumor size thresholds was evaluated. RESULTS:Applying all three OPTIMIST imaging criteria-post-NST tumor size ≤1 cm, L-to-B SER ≤1.6, and residual calcifications ≤2 cm-yielded an NPV of 69.3% and a false-negative rate of 16.1%. Residual tumor size and SER were predictive of pCR, whereas residual calcifications were associated with residual disease. Expanding the tumor size threshold to ≤2 cm while retaining the other criteria preserved NPV (67.5%), increased the proportion of patients with pCR meeting imaging criteria, and expanded eligibility by 19%. CONCLUSIONS:Imaging criteria alone were insufficient to support omission of breast surgery after NST. Pathologic confirmation with image-guided biopsy remains necessary. Refined imaging thresholds may improve clinical applicability without compromising predictive performance.

Purpose: Identifying high-risk patients for late recurrence is crucial for optimizing extended endocrine therapy (EET). This study investigated how clinicopathologic predictors of late recurrence diverge according to menopausal status in estrogen receptor (ER)-positive breast cancer patients who completed 5-year endocrine therapy. Methods: We retrospectively analyzed 1549 patients with stage I-III ER-positive breast cancer who completed 5 years of endocrine therapy and remained recurrence-free during that period. Median follow-up was 14.7 years. Outcomes included distant metastasis-free survival (DMFS) and overall survival (OS). Analyses were stratified by menopausal status to identify status-specific risk factors. Results: During the follow-up period, 208 patients (13.4 %) experienced late recurrence, with distant metastasis (6.3 %) being the most frequent event. Kaplan-Meier analyses demonstrated that tumor size >2 cm and higher nodal stage were significantly associated with inferior DMFS and OS in both subgroups (all P < .05). However, multivariable analysis revealed distinct prognostic patterns according to menopausal status. In premenopausal women, nodal involvement was the most potent driver of late recurrence; N1 stage significantly increased the risk of poor DMFS (HR 3.14, 95 % CI 1.54-6.40; P = .002) and OS (HR 3.39, 95 % CI 1.45-7.93; P = .005). Furthermore, age at diagnosis showed a significant inverse association with distant metastasis risk, with each one-year decrease in age increasing the risk by 6 % (HR 0.94; P = .019). In contrast, for postmenopausal patients, tumor size (>2 cm) emerged as the predominant predictor for both DMFS (HR 2.70; P = .006) and OS (HR 2.26; P = .039), whereas the prognostic impact of nodal stage was notably diminished (N1: P = .374 for DMFS; P = .999 for OS). In this group, older age was independently associated with worse OS (HR 1.13; P < .001) but did not significantly impact DMFS. Conclusion: Predictors for late recurrence are different depending on menopausal status: nodal burden and young age dominate in premenopausal patients, whereas tumor size is more critical in postmenopausal patients. These findings suggest that menopausal status must be a primary consideration in risk-stratification models for EET and long-term surveillance.

Abstract Background: The ubiquitin-proteasome system (UPS) and autophagy cooperate to maintain protein homeostasis, yet how these pathways interact to regulate breast cancer stem cells (BCSCs) remains unclear. PSMD2, a 19S regulatory subunit of the proteasome, has been linked to poor outcomes in several cancers, but its role within the autophagy-proteostasis network in ER+ breast cancer has not been defined. Here, we examined how PSMD2 influences CSC maintenance and drug response through UPS-autophagy crosstalk and explored how its suppression alters autophagy regulation and CSC vulnerability in ER+ breast cancer. Methods: Stable PSMD2 knockdown (ShPSMD2) lines were generated in ER+ breast cancer cells (MCF-7, ZR-75-1). Autophagy was modulated using hydroxychloroquine (HCQ). Western blotting assessed CSC-related markers (SOX2, OCT4, NANOG), EMT markers (E-cadherin, N-cadherin, Vimentin), and autophagy proteins (LC3B-II, p62, p-mTOR). Functional assays, including MTT viability, colony formation, migration, and invasion, were performed to evaluate PSMD2-dependent phenotypes. Mammosphere and viability assays assessed CSC formation and HCQ sensitivity. Tumor growth and metastasis were tested using mammary fat-pad and tail-vein xenografts in NSG mice. Results: PSMD2 depletion reduced CSC frequency, mammosphere formation, and colony formation, with decreased stemness markers (SOX2, OCT4, NANOG) and reversal of EMT features, shown by lower N-cadherin, Vimentin, and Snail. Loss of PSMD2 resulted in LC3B-II and p62 accumulation with mTOR suppression, indicating impaired autophagy turnover and disrupted proteostasis. PSMD2-deficient cells displayed reduced migration and invasion and were more sensitive to HCQ-induced cytotoxicity. HCQ treatment further lowered PSMD2 protein levels in a dose-dependent manner, implying feedback destabilization of the proteasome under lysosomal stress. This correlated with loss of CD44+/CD24- CSC-like populations and reduced sphere-forming ability. In xenografts, PSMD2 knockdown suppressed tumor growth and lung, liver metastasis, suggesting that PSMD2 is a critical node linking proteostasis to metastatic progression. Conclusions: PSMD2 depletion disturbs UPS-autophagy coordination, leading to loss of proteostasis and CSC depletion in ER+ breast cancer. This insight reveals PSMD2 as a key regulator of proteolytic balance and a potential target for autophagy-based therapeutic strategies. Citation Format: Yejoo Lee, Ju Hee Kim, So-Youn Jung, Wonshik Han. PSMD2 Suppression Disrupts Autophagy Turnover and Creates an Autophagy-Dependent Vulnerability in ER+ Breast Cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2203.

Breast cancer remains the most common malignancy and a leading cause of cancer-related mortality among women worldwide. Conventional in vitro and in vivo drug-testing models, including patient-derived xenografts (PDXs), are limited by low establishment efficiency, high cost, and poor reflection of tumor heterogeneity. To address these limitations, we developed a high-throughput drug-screening platform using breast cancer patient-derived organoids (PDOs) cultured on a 384-hanging pillar plate, which prevents cell attachment and preserves three-dimensional architecture, achieving a 70% culture success rate compared with 50% using standard well plates. We further established a Cancer Organoid-based Diagnosis Reactivity Prediction (CODRP) model that integrates the area under the curve (AUC), PDO growth rate, and clinical stage to enhance drug-response prediction. Validation of the CODRP model using residual tissues from both neoadjuvant chemotherapy (NAC)-treated and non-NAC patients demonstrated improved predictive performance compared with the conventional AUC index, with higher sensitivity (88.89% vs. 55.56%) and specificity (71.43% vs. 57.14%). In a cohort of triple-negative breast cancer (TNBC) patients receiving adjuvant chemotherapy (n = 9), the CODRP-classified sensitive group demonstrated prolonged recurrence-free survival (RFS) over two years. These findings indicate that combining the hanging pillar PDO culture system with the CODRP model improves concordance between preclinical drug-response data and clinical outcomes, supporting its potential as a precision-medicine platform for individualized breast cancer therapy optimization.

TPS651 Background: Omission of axillary surgery has been established as a safe option in selected patients with early-stage breast cancer undergoing upfront surgery. In the neoadjuvant systemic therapy (NST) setting, however, axillary surgery remains standard practice. Ongoing trials, including EUBREAST-01, ASICS, and ASLAN, are evaluating the safety of omitting sentinel lymph node biopsy (SLNB) after NST, but these studies have largely been limited to single-arm designs and to patients with confirmed breast pathologic complete response (pCR). Retrospective studies suggest that axillary nodal clearance after NST may be predicted using tumor subtype, treatment response, and post-NST imaging, independent of breast pCR. The NeoNAUTILUS trial was designed to prospectively evaluate the safety of omitting SLNB after NST in patients predicted to have no residual axillary disease, using axillary response–based selection rather than breast pCR. Methods: NeoNAUTILUS is a prospective, multicenter, randomized, controlled, non-inferiority trial being conducted at 23 centers in Korea. Eligible participants are women aged ≥20 years with clinical T1–T3 invasive breast cancer who have completed NST and are candidates for breast-conserving surgery (BCS). Patients with clinical N0 disease of any subtype are eligible. Selected patients with clinical N1 disease are eligible if they have HER2-positive, triple-negative, or estrogen receptor–low-positive tumors and demonstrate a >30% reduction in primary tumor size on post-NST magnetic resonance imaging. All participants must have no suspicious lymph nodes on post-NST axillary ultrasound. Participants are randomized 1:1 to BCS with SLNB or BCS without axillary surgery, stratified by clinical nodal status (cN0 vs cN1) and tumor subtype. Patients remain blinded to treatment allocation until surgery. Whole-breast irradiation is planned for all patients, with protocol-defined axillary radiation fields in the no-SLNB group according to initial nodal status. The primary endpoint is 5-year invasive disease-free survival. Secondary endpoints include overall survival, distant metastasis-free survival, ipsilateral axillary recurrence, locoregional recurrence, and patient-reported quality of life. A total of 464 patients are planned for enrollment over 3 years with a minimum follow-up of 5 years. Enrollment began in May 2025 and is ongoing. Clinical trial registration: NCT06704945; KCT0010174. Clinical trial information: NCT06704945 .

The effectiveness of adjuvant endocrine therapy for hormone receptor-positive (HR+) breast cancer (BC) varies according to menopausal status, body mass index (BMI), and tumor biology. We evaluated the association between selective estrogen receptor modulators (SERMs), aromatase inhibitors (AIs), and BC-specific mortality according to menopausal status, BMI, and molecular subtype in a nationwide Korean cohort. We analyzed data from 31,030 patients with HR+ BC who were registered in the Korean Breast Cancer Society Registry, diagnosed between 2000 and 2008, and followed through 2013. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for BC-specific mortality after adjusting for demographic and clinical factors. Of the 31,030 patients, 19,634 received SERM therapy, and 3,354 received AI therapy. SERM use was associated with lower BC-specific mortality in premenopausal women (HR, 0.75; 95% CI, 0.63-0.91), whereas AI therapy was more strongly associated with lower BC-specific mortality among postmenopausal women (HR, 0.76; 95% CI, 0.61-0.94). Lower BC-specific mortality was observed among patients with a BMI ≥ 23 kg/m² who received SERM (HR, 0.84; 95% CI, 0.72-0.98) or AI therapy (HR, 0.78; 95% CI, 0.62-0.99). The strongest association with lower BC-specific mortality was observed in postmenopausal women with luminal B tumors (HR, 0.59; 95% CI, 0.42-0.83). The association between adjuvant endocrine therapy and BC-specific mortality differed according to menopausal status, BMI, and molecular subtype. These findings suggest that menopausal status, BMI, and molecular subtype are important considerations when evaluating endocrine treatment strategies.

e12764 Background: While tumor location influences lymphatic drainage patterns, its prognostic value in node-positive disease is unknown. We evaluated whether upper outer quadrant (UOQ) tumor location is associated with distinct survival outcomes in patients with N1 breast cancer. Methods: We conducted a retrospective cohort study of 1,914 breast cancer patients with N1 disease treated at Seoul National University Hospital between January 2000 and December 2020. N1 disease was defined as biopsy-confirmed clinical N1 in neoadjuvant chemotherapy recipients or pathologic N1 in upfront surgery patients. Tumor location was determined by preoperative imaging (ultrasonography and MRI). Distant metastasis-free survival (DMFS) and overall survival (OS) were analyzed using Kaplan-Meier and Cox proportional hazards models. Results: Among enrolled patients, 938 (49.0%) had UOQ tumors and 976 (51.0%) had non-UOQ tumors. While UOQ patients were older (median 49 vs 47, p-value 0.004) and received breast-conserving surgery and radiotherapy more frequently (58.6% vs 44.8, p-value < 0.001), tumor stage, molecular subtype, histologic grade, and systemic treatment were comparable between groups. In this node-positive cohort, UOQ location was associated with significantly superior 5-year DMFS (86.3% vs 75.3%, p < 0.001) and OS (93.4% vs 87.9%, p < 0.001). On multivariate analysis, UOQ location remained a significant independent favorable prognostic factor, reducing the risk of distant metastasis by 50% (HR 0.50, 95% CI 0.42–0.60, p < 0.001) and death by 47% (HR 0.53, 95% CI 0.42–0.66, p < 0.001) after adjusting for tumor stage, molecular subtype, histologic grade. Conclusions: Among patients with node-positive breast cancer, UOQ tumor location confers a substantial survival advantage independent of conventional prognostic factors. This finding suggests that nodal positivity has a less adverse prognostic impact when the primary tumor is located in the UOQ, potentially reflecting distinct patterns of lymphatic spread. Tumor location might be incorporated into risk stratification for N1 breast cancer patients.

Background: After the ACOSOG Z0011 trial, sentinel lymph node (SLN) biopsy became the standard for staging early breast cancer, demonstrating that in select patients with limited lymph node involvement, SLN biopsy alone was sufficient. However, this trial excluded patients with extracapsular extension (ECE), leaving its impact on recurrence rates unclear. This study evaluates the impact of ECE on recurrence rates in SLN-positive breast cancer patients and compares recurrence rates in patients with ECE based on the type of axillary surgery. Methods: We conducted a cohort study of 1,579 SLN-positive breast cancer patients treated at Seoul National University Hospital from 2009 and 2018. Included were patients with invasive breast cancer, T1 or T2 tumor stage, and positive pathology results in both frozen and permanent SLN biopsy. Radiation status and type of surgery (breast-conserving or mastectomy) were not considered. Those who received neoadjuvant chemotherapy were excluded. Patients were divided into two groups based on the presence (N=266) or absence (N=1,313) of ECE. Recurrence rates were analyzed by evaluating disease free survival (DFS) using Kaplan-Meier and Cox regression models. Results: The presence of ECE was associated with significant differences in patient characteristics. ECE-positive patients were older (51.8 vs 50.2 years, p=0.023) and more likely to undergo SLNB+ALND (82.0% vs 65.7%, p<0.001) compared to ECE-negative patients. They also had larger tumors and more advanced nodal disease (both p<0.001).To reduce selection bias, we performed propensity score matching (PSM) with a 1:2 matching ratio. After PSM to account for pathologic T stage, number of positive SLNs, type of axillary surgery, and molecular subtype, the comparison between ECE-positive (N=252) and ECE-negative (N=441) patients showed no significant differences in DFS (log-rank p = 0.96). Using the same PSM data for multivariate Cox regression analysis, adjusted for tumor size, nodal involvement, and molecular subtype, there was no significant difference in DFS between ECE-positive and ECE-negative groups (HR = 0.936, 95% CI: 0.609-1.437, p = 0.762). Among the 266 ECE-positive patients, we compared DFS between those who underwent SLNB alone and those who received SLNB followed by ALND. The addition of ALND did not significantly improve DFS among ECE-positive patients (log-rank p=0.33). Conclusion: Our analysis revealed that the presence of ECE in SLN-positive breast cancer patients did not significantly affect recurrence rates. Furthermore, among ECE-positive patients, additional ALND after SLNB did not significantly improve DFS. These findings suggest that while ECE may indicate more advanced disease characteristics, it does not independently predict recurrence when other factors are adjusted for. Consequently, in certain cases, it may be feasible to omit ALND in ECE-positive patients with positive SLN. These results highlight the need for prospective studies to confirm these results and investigate potential therapeutic strategies. Citation Format: Hyelim Kang, Min Jung Lee, Changhoon Lee, Jinyoung Byeon, Eunhye Kang, Ji-Jung Jung, Hong-Kyu Kim, Han-Byoel Lee, Hyeong-Gon Moon, Wonshik Han. Impact of Extracapsular Extension on Recurrence in Sentinel Lymph Node Positive Breast Cancer Patients: A Retrospective Cohort Study [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P3-09-12.

Background: Invasive lobular carcinoma (ILC) represents the second most prevalent histological subtype of breast cancer, comprising 3-15% of all diagnosed cases. Despite its unique clinical and pathological profile when compared to invasive ductal carcinoma (IDC), there remains significant debate regarding the prognosis and optimal management strategies for ILC. This study aimed to investigate the differences in clinical and pathological features, survival between ILC and IDC and their impact on prognosis. Methods: We reviewed the data of the 15,489 IDC patients and 966 ILC patients who were treated between January 2005 and December 2022 at Seoul National University Hospital. The clinical and pathological characteristics of ILC and IDC were compared, and survival outcomes were analyzed using cox proportional hazard models. We conducted a parallel analysis using the Surveillance, Epidemiology, and End Results (SEER) database. We extracted data from 247,290 breast cancer patients diagnosed between 2010 and 2015, comprising 219,404 cases of IDC and 27,886 cases of ILC. Result: ILC had a larger tumor size, higher proportion of HR+/HER2- subtype, lower tumor grade, and lower Ki-67 index compared to IDC. Kaplan-Meier survival analysis with log-rank tests showed no statistically significant differences in breast cancer-specific survival (BCSS) or overall survival (OS) between patients with ILC and IDC (p = 0.52, 0.27, respectively). After adjusting for histologic grade and subtype, we found distinct patterns in survival determinants between ILC and IDC. In IDC, both tumor size (T stage) and nodal status (N stage) showed consistent prognostic significance. Compared to T1, T2 and T3 stages had 2.5-fold (95% CI: 2.00-3.00) and 4.4-fold (95% CI: 3.40-5.72) higher breast cancer-specific death risks, respectively. N1, N2, and N3 stages showed 2.5-fold (95% CI: 2.08-3.06), 3.8-fold (95% CI: 3.06-4.74), and 6.0-fold (95% CI: 4.66-7.67) increased risks compared to N0. Interestingly, in ILC, tumor size did not significantly impact survival (p=0.38 for T2, p=0.12 for T3 vs T1). However, advanced nodal involvement in ILC showed markedly higher risk, with N2 and N3 stages having 5.2-fold (95% CI: 1.66-16.57) and 26.1-fold (95% CI: 10.54-64.69) increased risks, respectively. SEER data analysis corroborated these findings. For IDC, T2 and T3 stages showed 2.2-fold (95% CI: 2.14-2.29) and 3.4-fold (95% CI: 3.20-3.51) increased risks, while N1, N2, and N3 stages had 1.7-fold, 2.8-fold, and 4.2-fold higher risks, respectively. In ILC, the impact of tumor size was less pronounced, with T2 and T3 stages showing 1.9-fold (95% CI: 1.69-2.04) and 2.43-fold (95% CI: 2.17-2.71) increased risks. However, nodal involvement in ILC again showed higher risk, with N2 and N3 stages having 3.5-fold (95% CI: 1.66-16.57) and 6.18-fold (95% CI: 5.53-6.90) increased risks, compared to 2.83-fold and 4.17-fold in IDC for N2 and N3, respectively. Conclusion: In conclusion, our study highlights that ILC exhibits distinct characteristics and prognostic patterns compared to IDC. Although ILC is often detected at a larger size, the impact of tumor size on survival differs from that of IDC. In ILC, tumor size alone may not be as strong a predictor of survival as it is in IDC. Conversely, nodal involvement in ILC, especially at advanced stages, carries a particularly high risk for breast cancer-specific mortality. These findings underscore the need for tailored prognostic assessments and potentially different treatment strategies for ILC and IDC. Citation Format: Eunhye Kang, Ik Beom Shin, Hyerim Kang, Changhoon Lee, Min Jung Lee, Jinyoung Byeon, Ji-Jung Jung, Hong-Kyu Kim, Han-Byoel Lee, Wonshik Han, Hyeong-Gon Moon. Prognostic implications of tumor size and nodal involvement in invasive lobular carcinoma of breast [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P5-12-01.

HRAS is often overactivated in breast cancer, whereas mutations are rare in this malignancy. This study was aimed to determine how overexpressed wild-type HRAS (HRASWT) plays an oncogenic role in human breast cancer progression. The activated form of non-mutagemic HRASWT is highly overexpressed in aggressive triple-negative breast cancer (TNBC) compared with less invasive luminal breast tumor as well as normal mammary tissues. A qPCR assay showed that TNBC cells selectively overexpress HRASWT. Notably, HRAS directly interacted with and stabilized NRF2 whereas KRAS barely bound to NRF2. Further, HRASWT exerted its oncogenic capability by inducing interaction between NRF2 and STAT3. Such interaction was observed in TNBC tissues, but not other subtypes of breast cancer. Combined silencing of NRF2 and STAT3 suppressed TNBC growth to a greater extent than that achieved with single knockdown of individual genes. RNA sequencing analysis revealed expression profiles for genes related to cell migration which may account for synergistic oncogenic activity of NRF2 and STAT3. Our results suggest that NRF2 is a potential effector of HRAS that regulates downstream signaling in a KEAP1-independent manner. Non-mutated HRAS overexpression, in cooperation with NRF2, promotes the progression of breast carcinoma and might represent a novel therapeutic target against TNBC.

The addition of a cyclin-dependent kinase 4/6 (CDK4/6) inhibitor to endocrine therapy augments biological response in breast cancer. This phase III randomized, double-blind study evaluated the efficacy of adding palbociclib to neoadjuvant endocrine therapy (NET) for operable, hormone receptor-positive human epidermal growth factor receptor 2 (HER2)-negative breast cancer. Patients randomly received 16 weeks of endocrine therapy (letrozole for postmenopausal and tamoxifen plus ovarian function suppression for pre-/perimenopausal patients) plus palbociclib or placebo. The co-primary endpoints included preoperative endocrine prognostic index (PEPI) score and EndoPredict (EPclin) risk score according to the gatekeeping procedure. Of 141 randomized patients, 130 completed the treatment with surgical samples evaluable for endpoints in 126 patients. The proportion of patients with a low, moderate, and high PEPI score was 15.2, 50.0, and 34.8% in the palbociclib arm and 13.3, 55.0, and 31.7% in the placebo arm, respectively, with no statistical difference (one-sided P = 0.563). Statistical analysis was not performed on EPclin risk score. No new safety signals were reported. Permanent treatment discontinuation by adverse events was reported for seven (9.7%) and zero patients in the palbociclib and placebo arms, respectively. In conclusion, the addition of palbociclib to NET did not improve the efficacy. ClinicalTrials.gov NCT03969121.

Association between PRS38_ASN+PRS190_EB and breast cancer risks in different age groups

Estimation of absolute risk of breast cancer by seven percentiles for women at different ages estimated using data from HEXA 1st. The absolute risk of developing breast cancer predicted by the multiple PRS model (PRS38_ASN+PRS190_EB). The dotted lines represent the average risks. (A) Lifetime absolute risk (B) 5-year absolute risk