Abstract Background: ERBB2 and PIK3CA are among the most commonly mutated genes in HER2-positive breast cancer and are known mediators of resistance to HER2-targeted therapies. However, the molecular evolution of these mutant tumors under treatment remains poorly understood. In this study, we profiled the multi-omics landscape of ERBB2- and PIK3CA-mutant tumors from the randomized PREDIX HER2 trial, which compared neoadjuvant trastuzumab emtansine (T-DM1) with dual HER2 blockade plus chemotherapy in early-stage HER2-positive disease. Methods: Fresh-frozen tumor biopsies were used for RNA sequencing and whole-exome sequencing (WES), while FFPE biopsies were used for Xenium 5K spatial transcriptomics. Longitudinal differential gene expressions were analyzed using Gaussian mixed-effects models. Tumor subtypes were assigned with a five-subtype single-sample predictor (SSP.Subtype). Somatic mutations and copy number alterations were profiled using GATK4 Mutect2 and GISTIC2. Somatic mutations were filtered and curated using a 5% variant allele frequency (VAF) cut-off. Results: Of the 190 patients with available WES data, 7 (3.7%) and 49 (25.8%) patients carried non-synonymous ERBB2 mutations (median VAF = 0.28) and PIK3CA mutations (median VAF = 0.21), respectively. In those patients who did not achieve pathological complete response (pCR) (63%), we identified two ERBB2 mutations (S310Y and R143*) that were consistently detected prior to, during, and after treatment, all located in the extracellular domain. Notably, both tumors maintained the HER2-enriched subtype and ERBB2 amplification during treatment, suggesting mutations may influence trastuzumab binding and blocking HER2 signals. Similarly, we identified 10 patients with persistent PIK3CA mutations during treatment, all of which were known gain-of-function alterations. Only one out of four tumors retained the HER2-enriched subtype, suggesting a potential shift in molecular phenotype under therapeutic selective pressure. Phylogenetic analysis further revealed recurrent clonal expansion involving PIK3CA and acquired KLB mutations, the latter implicated in the regulation of cholesterol metabolism. Further in vitro validation of KLB-FGF21 signaling is ongoing. Conclusion: This longitudinal evolutionary analysis of ERBB2- or PIK3CA-mutant, HER2+ breast cancer suggests that tumors evolve heterogeneously under HER2-targeted therapies, revealing the positive selection of pathogenic ERBB2 and PIK3CA mutations. Citation Format: Kang Wang, Ioannis Zerdes, Emmanouil G. Sifakis, Dimitrios Salgkamis, Jonas Bergh, Thomas Hatschek, Alexios Matikas, Theodoros Foukakis, . Clonal evolutionary trajectory of HER2-positive breast cancer with ERBB2 or PIK3CA mutations under neoadjuvant treatment [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 3527.

Immune checkpoint inhibitors combined with chemotherapy are established as first-line therapy for patients with programmed death ligand 1 (PD-L1)-positive metastatic triple-negative breast cancer (mTNBC). We evaluated whether immuno-PET using [89Zr]Zr-atezolizumab could more accurately identify patients with mTNBC who may benefit from immune checkpoint inhibitor therapy. Methods: Three patients with mTNBC underwent [89Zr]Zr-atezolizumab PET/CT followed by a biopsy of a targeted metastatic lesion. Patients received atezolizumab plus chemotherapy if either immunohistochemistry of the metastatic lesion or PET imaging showed PD-L1 positivity. Results: All patients had tracer-avid metastatic lesions on PET. Two of 3 biopsied, tracer-avid lesions were PD-L1 negative on immunohistochemistry. Intraindividual heterogeneity in tracer uptake was noted. All patients were treated with atezolizumab plus chemotherapy, with all demonstrating an initial response. Conclusion: A work-up with [89Zr]Zr-atezolizumab immuno-PET is clinically feasible. This molecular imaging-based strategy holds potential as a noninvasive tool to more accurately identify patients with mTNBC who may benefit from immune checkpoint inhibitor therapy.

Abstract In PREDIX LumB patients with estrogen receptor positive and human epidermal growth factor receptor negative (ER + /HER2-) breast cancer > 20 mm and/or with lymph node metastasis were randomized 1:1 to receive either paclitaxel weekly for 12 weeks followed by palbociclib and endocrine therapy for 12 weeks (arm A), or the reverse sequence (arm B). Primary endpoint is objective radiologic response at 12 weeks (ORR12), and key secondary endpoints are ORR24, pathologic complete response, event-free survival, safety and correlative studies of tissue and circulating biomarkers. Whole exome sequencing and RNA sequencing were performed on baseline fresh frozen tissue samples. In total, 179 patients comprise the intention-to-treat population. There is no statistically significant difference between the two arms in ORR12 (59% vs 45%, p = 0.058). An exploratory gene expression analysis identified differentially expressed genes and gene sets between responders and non-responders at 12 weeks. A predictive signature, CDKPredX, comprising 31 genes related to proliferation, ER signaling and immune activity was developed to identify patients resistant to chemotherapy but responding to palbociclib plus endocrine therapy (pinteraction=0.03). The predictive signature was independently validated in the CORALLEEN trial (pinteraction=0.048). Clinicaltrials.gov identifier: NCT02603679

Trophoblast cell surface antigen 2 (Trop-2) is a targetable transmembrane glycoprotein commonly overexpressed in breast cancer (BC). This study combines a systematic review and meta-analysis with a retrospective analysis of an independent early BC patient cohort, aiming to investigate the expression patterns of Trop-2 and their clinical significance in BC. A systematic literature search was conducted up to October 2025 to identify studies evaluating Trop-2 protein expression levels and clinical outcomes in patients with BC receiving Trop-2–directed antibody-drug conjugates (ADCs) or chemotherapy. Random-effects meta-analyses were performed to estimate pooled Trop-2 positivity rates and to assess progression-free (PFS) and overall survival (OS) across Trop-2 expression levels. Additionally, an association analysis between Trop-2 protein and gene expression was performed in an early BC patient cohort (n = 564) using immunohistochemistry and gene expression data. Trop-2 protein expression was assessed using both conventional observer-based methods and digital quantitative image analysis. A total of 41 studies fulfilled the inclusion criteria. The overall pooled Trop-2 protein positivity rate was 72

Abstract Background Glucocorticoid prophylaxis is routinely used in oncology. Glucocorticoids potent immunosuppressive and anti-inflammatory properties raise concerns regarding potential impairment of antitumor immune responses. Preclinical and retrospective studies suggest glucocorticoids may impair antitumor immunity, though these findings are subject to significant bias and confounding factors. Understanding the immunomodulatory effects of prophylactic glucocorticoids, isolated from other therapies, is critical to bridging the knowledge gap on their influence on antitumor immunity. Methods We performed a post hoc, retrospective analysis using cryopreserved PBMCs and plasma from patients with treatment-naïve, early stage HER2+ breast cancer, prior to neoadjuvant therapy. Of 192 patients, 129 had no prior glucocorticoid exposure and 63 received 24 mg prophylactic betamethasone before sampling. High-dimensional flow cytometry assessed monocytes, dendritic cells, and T-cell subsets. Plasma proteins were evaluated using the Olink Target 96 Immuno-Oncology proximity extension assay. Results Glucocorticoid exposure was consistently associated with a distinct immune profile. Exposed patients exhibited increased CD163+ monocytes and reduced intermediate, non-classical, and HLA-DRhi classical monocytes, plasmacytoid DCs, and conventional CD1c+ DCs. T-cell alterations included enrichment of CXCR4+ subsets, particularly terminally differentiated CD8+ T-cells, with reductions in naïve, central memory, CD27+ effector memory, and CXCR3+ populations. Inflammatory plasma proteins (IFN-γ, CCL19, CCL2, IL-12, IL-6, Granzyme A/B, CXCL10, CXCL9) were diminished, whereas IL-10 and CXCL13 were elevated. Glucocorticoid-associated immune alterations were heterogeneous across patients, highlighting interindividual variability in observed post-exposure immune states. Conclusions Prophylactic glucocorticoid exposure reshapes the circulating immune landscape, altering monocytes, dendritic cells, naïve and memory T-cell subsets, as well as key inflammatory proteins.

Background Meaningful comparisons of breast cancer (BC) survival across hospitals require appropriate case-mix adjustment. While such adjustment is routine for short-term outcomes, it remains less established for long-term BC-specific mortality (BCSM). We developed case-mix models to evaluate whether robust hospital-level benchmarking of 10-year BCSM can be achieved using routinely collected registry data. Methods Using the nationwide Swedish BC registry linked to population registers, we developed full and reduced logistic regression models for 10-year BCSM. The full model included age, tumour, socioeconomic, and comorbidity variables while the reduced model used routinely collected age and tumour variables. Model performance was assessed via discrimination and calibration. The model impact on hospital benchmarking was evaluated through observed-to-expected mortality ratios at the hospital level, comparing expected mortality estimates and shifts in outlier status. Results The study included 10,888 patients. Age, stage, and grade were the strongest predictors, while comorbidity and socioeconomic variables added limited incremental predictive value. Both the full and reduced 10-year BCSM models demonstrated similar discrimination and near-perfect calibration. Furthermore, the full and reduced models yielded highly comparable expected mortality rates across 12 hospitals (mean absolute difference 0.6 percentage points), with consistent identification of hospital outliers in all but one hospital. Conclusions Routinely collected data on age and tumour characteristics appear to offer a feasible foundation for fair benchmarking of long-term BCSM in a publicly-funded healthcare system. While external validation in contemporary and international cohorts is an important next step, these findings might be relevant to other countries with strong cancer registry infrastructure.

Background: Analyses of the 2014 EBCTCG database suggested that, in early-stage breast cancer, obesity was strongly independently associated with breast cancer mortality only in pre/peri-menopausal oestrogen-receptor-positive (ER+) disease (Pan et al ASCO 2014). Based on the far larger 2024 EBCTCG database, however, we can now test that unexpected finding and better characterise any relevance of patient characteristics to the association of body mass index (BMI) with distant recurrence and mortality. Methods: We analysed patient-level data on time to distant recurrence and death from the 206,904 women with early-stage breast cancer (entered during 1978-2017 into 147 randomised trials) in the 2024 EBCTCG database who had BMI at entry (within two years of diagnosis) recorded as 15-50 kg/m2 and with complete information on age, ER status, tumour diameter, nodal status, and randomly allocated treatment. Information on menopausal status, tumour grade, and HER2 status was available for most participants. Cox regression was used to estimate the associations of BMI with rates of distant recurrence and breast cancer mortality, calculating hazard rate ratios (RRs) per 5 kg/m2 increase of BMI or comparing 3 BMI groups (obese: BMI 30-50 [mean 34.7]; overweight: BMI 25 to <30 [mean 27.3]; lean: BMI 15 to <25 [mean 22.2] kg/m2). Results: Of the 206,904 women, 60% were postmenopausal at trial entry and 77% had ER+ disease. Their mean BMI was 27.1 (SD 5.6) kg/m2 and 26.0% (53,872) were obese (BMI ≥30 kg/m2). The prevalence of obesity increased from 19% in the early 1980s to 27% in the early 2010s. The overall adjusted rate ratio (RR) of first distant recurrence (ignoring any local or contralateral recurrences) was 1.06 (95% CI 1.05-1.07, p<0.0001) per 5 kg/m2 increase in BMI. The RR for overweight versus lean women was 1.07 (CI 1.04-1.10, p<0.0001), and that for obese versus lean women was 1.17 (95% CI 1.14-1.20, p<0.0001). This approximately log-linear association between BMI and the rate of distant recurrence was seen irrespective of patient or tumour characteristics, type of adjuvant systemic therapy, year of diagnosis, or time since diagnosis. In the 82,464 pre-menopausal women the RR per 5 kg/m2 increase of BMI was 1.08 (1.07-1.10, p<0.0001), and in the 124,440 post-menopausal women it was 1.05 (1.03-1.06, p<0.0001; heterogeneity between RRs p=0.0004). There was little heterogeneity between the RRs in ER+ and ER-poor disease. In the 159,119 women with ER+ disease the RR per 5 kg/m2 increase of BMI was 1.06 (1.05-1.08, p<0.0001), and in the 47,785 with ER-poor disease it was 1.06 (1.04-1.08, p<0.0001). The associations of BMI with breast cancer mortality mirrored those with distant recurrence. Conclusion: Overweight and obesity are associated with increased distant recurrence and breast cancer mortality in all types of patients with early-stage breast cancer, but the risk associated with a substantial (e.g. 5 kg/m2) difference in BMI is only moderate. Nevertheless, randomised assessment of the effects among overweight or obese women with early breast cancer of weight-loss interventions (perhaps utilising a GLP-1 receptor agonist) could usefully be added, using a factorial design, to some current and future adjuvant treatment trials addressing unrelated questions. Reference: Pan H, Gray R, on behalf of the EBCTCG. Effect of obesity in premenopausal ER+ early breast cancer: EBCTCG data on 80,000 patients in 70 trials. J Clin Oncol 2014; 32:5s Citation Format: Hongchao Pan, Richard Gray, Richard Peto, Jeremy Braybrooke, David Dodwell, Robert Hills, Rosie Bradley, Hellen Gelband, Hui Liu, Paul McGale, Carolyn Taylor, Mike Clarke, Wolfgang Janni, Marianne Ewertz, Pamela J Goodwin, Joseph Sparano, Kathy I Pritchard, Jonas Bergh, Sandra M Swain, for the Early Breast Cancer Trialists™ Collaborative Group (EBCTCG). Overweight, obesity and prognosis in 206,904 women in the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) database [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 GS2-09.

12024 Background: Cardiotoxicity is a known side effect of trastuzumab and combination with anthracyclines increases the risk for cardiotoxicity. Adjuvant dose-dense (DD) chemotherapy has demonstrated beneficial breast cancer (BC) outcomes in patients with high risk for relapse, but data on long-term cardiac toxicity when combined with trastuzumab is scarce. We have previously reported on the safety of trastuzumab at six years follow-up, in a subset of patients treated with dose-dense chemotherapy in the Pan-European Tailored Chemotherapy (PANTHER) phase III trial. We hereby present long-term safety data from 10-year follow-up from the same subset. Methods: This is a protocol-predefined cardiac safety study, among Swedish sites included in the PANTHER trial, including patients with HER2-positive (HER2+) and HER2-negative (HER2-) BC matched for age, treatment group and institution. Enrolled patients were up to 65 years old with node-positive or high-risk, node-negative BC were randomized 1:1 to either dose tailored (according to hematologic nadirs) and biweekly DD epirubicin and cyclophosphamide followed by docetaxel or standard 5-fluorouracil, epirubicin, and cyclophosphamide plus docetaxel every 3 weeks. Patients with HER2-positive disease received 1 year of adjuvant trastuzumab. They underwent echocardiography (ECHO) or multigated acquisition scanning and electrocardiography at baseline, at 4, 6 and 10 years of follow-up. Data on cardiac medication NT-proBNP, lipid profile and ECG were also collected. Results: ECHO at 10-years follow-up was available for 94 patients; 48 HER2+ (19 DD, 29 control) and 46 HER2- (21 DD, 25 control). Overall, incidence of cardiotoxicity was low. Mean LVEF was 58 % (range 49-68%) and 60.65% (range 50-76%) in HER2+ and HER2- respectively. Only one patient had LVEF < 50% (DD HER2+) and additional 12 patients had LVEF 50-54%, equally distributed between the treatment groups. In total, 27 patients were treated with cardiac medications at this point; 15 (56%) of which had been treated with trastuzumab and 10 of them (n = 7 HER2+), not reporting cardiac medication at previous timepoints. The majority of the patients did not report any symptoms related to heart disease, per NYHA-classification (41 patients in both groups report NYHA class 0). Overall, no significant changes were seen in the biomarkers. Conclusions: Cardiotoxicity of trastuzumab in DD anthracycline chemotherapy, examined in the context of a randomized trial sub-study, was very low. Our results underline the safety of trastuzumab in this context, providing support to offer the patients best treatment options for improving breast cancer survival. Clinical trial information: NCT00798070 .

BACKGROUND AND PURPOSE:The clinical significance of individual CYP2D6 activity for the outcome of tamoxifen treatment in early breast cancer is unclear. Our previous investigation in patients diagnosed over the period 1998-2000 indicated an association between reduced CYP2D6 activity and poor outcome in premenopausal women. The aim of this study was to investigate the association between CYP2D6 genotype and clinical outcome in a larger tamoxifen treated cohort. PATIENTS/MATERIAL AND METHODS:Swedish breast cancer patients who initiated adjuvant tamoxifen treatment over the period 2006-2014 constituted the full study cohort. Clinical information was collected from medical records. Data on endocrine treatment, use of CYP2D6 inhibitors was retrieved from the Swedish Prescribed Drug Register. CYP2D6 was genotyped and translated into predicted metabolic activity. The association between CYP2D6 activity and clinical outcome was analyzed using Cox regression, controlling for potential confounding variables. Subgroup analyses were performed based on menopausal status, tamoxifen treatment for at least 1 year and as single endocrine treatment, HER2-status and tamoxifen monotherapy. RESULTS:A total of 1,103 patients were included. A total of 761 patients received tamoxifen as monotherapy. A total of 42% were premenopausal. Median follow-up was 11.4 years. No significant association was found between CYP2D6 activity and recurrence (adjusted hazard ratio [aHR] 1.18, 95% CI 0.92; 1.52) or breast cancer mortality (aHR 1.41, 95%CI 0.93; 2.13) in the full cohort, or in the subgroup with tamoxifen monotherapy (aHR 1.39, CI 0.99; 1.96 and 1.88, CI 0.98; 3.60 respectively). INTERPRETATION:No association was noted between reduced CYP2D6 activity and poorer outcome in this early breast cancer cohort, with patients generally at lower risk of recurrence, reflecting the role of adjuvant tamoxifen in current clinical practice.

Supplementary Table S1. Representativeness of Study Participants

Background: Endocrine therapy alone, with breast surgery offered only at local progression, is still sometimes considered for older women with operable early breast cancer, but the long-term risks of deferring surgery are uncertain. We evaluated long-term outcomes in the three unconfounded randomised trials that compared immediate breast surgery plus tamoxifen versus tamoxifen alone. None of these trials scheduled radiotherapy or chemotherapy. Methods: Individual patient data meta-analyses compared effects on breast cancer outcomes in 3 trials, initiated in the 1980s, among 1082 women (age ≥70, all to receive tamoxifen for at least 5 years) comparing immediate surgery versus surgery only in the event of local progression. Primary outcomes were time to locoregional failure, to distant recurrence, and to breast cancer mortality. Locoregional failure was defined as any locoregional recurrence after surgery or, if no immediate surgery, ≥25% increase in tumour diameter. Age-adjusted intent-to-treat log-rank analyses, stratified by nodal status, were used to estimate first-event-rate ratios (RRs). Results: Median age at randomisation was 76 (IQR 73-80) years, and 63% (666/1082) had clinically estimated tumour diameter >20 mm. Mean follow-up while still alive was 7.3 woman-years. Of 518 women allocated immediate surgery, 45.7% had mastectomy, 47.3% had breast-conserving surgery, and 7.0% had neither. Locoregional failure was greatly reduced by allocation to immediate surgery (RR=0.24, 95% CI 0.19-0.30, p<0.00001). This extreme RR was little affected by age, disease stage, or time period. Although the proportional reduction in the annual rate of locoregional failure was similar during years 0-1, 2-4 and 5-9, the absolute reduction in locoregional failure was mainly before year 5 (Kaplan-Meier 5-year risks 12.1% vs 45.8%). On average over the whole follow-up period the rates of distant recurrence (RR=0.72, 0.57-0.90, p=0.003), breast cancer mortality (RR=0.68, 0.54-0.86, p=0.002), and all-cause mortality (RR=0.83, 0.72-0.97, p=0.016) were also reduced, but these benefits emerged only after years 0-1. The distant recurrence rate ratio was 0.97 (0.67-1.42) during years 0-1 after randomisation, 0.73 (0.49-1.07) during years 2-4 and 0.52 (0.36-0.76) after year 5 (trend: p=0.012). Conclusion: In early breast cancer, immediate breast surgery greatly reduces locoregional progression rates during the first 5 years and approximately halves the annual rates of distant recurrence and of breast cancer death after the first 5 years, despite having had little clinically apparent effect on distant recurrence rates during the first few years. These findings could indirectly inform the planning and interpretation of trials of less extreme de-escalation of surgery or radiotherapy. Citation Format: Robert Hills, CoRosie Bradley, Jeremy Braybrooke, Lucy Davies, David Dodwell, Gurdeep Mannu, Paul McGale, Mike Clarke, Hongchao Pan, Richard Berry, Richard Peto, Carolyn Taylor, Jonas Bergh, Sandra Swain, Stewart Anderson, Allan Hackshaw, Tom Bates, Eleftherios Mamounas, Giorgio Mustacchi, John Robertson, Richard Gray. immediate breast surgery versus deferral of surgery in women aged 70+ years with operable breast cancer: patient-level meta-analysis of the three randomised trials among 1,082 women [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 LB1-01.