AbstractPurpose: The EMERALD trial led to the approval of elacestrant for estrogen receptor (ER)–positive, HER2-negative, estrogen receptor 1 (ESR1)–mutated advanced or metastatic breast cancer (mBC) with disease progression following at least one line of endocrine therapy (ET). Subgroup analyses provided evidence suggesting that elacestrant enables ET sequencing in the second line before other targeted combinations, which could delay chemotherapy-based regimens. Experimental Design: This study used claims data from the Komodo Research Dataset linked with Foundation Medicine Inc. clinical genomic data from patients with estrogen receptor+/HER2− mBC harboring an ESR1 mutation treated with elacestrant. The primary outcome measure was time to next treatment (TTNT). Results: Among 306 patients, 93.8% had prior ET ± cyclin-dependent kinase 4/6 inhibitor for ≥12 months, 50.0% had prior chemotherapy, and 72.2% had prior fulvestrant. Median TTNT (mTTNT) was 8.2 months [95% confidence interval (CI), 6.3–13.0] in patients with 1 to 2 prior lines and 7.5 months (95% CI, 7.1–9.9) in those with ≥3 prior lines of ET. In patients with coexisting ESR1- and PI3K-pathway–mutated tumors, mTTNT was 6.3 months (95% CI, 4.8–7.9). mTTNT was 7.9 months (95% CI, 7.1–9.8) in all patients and was also sustained in patients with no prior fulvestrant [12.9 months (95% CI, 7.2–not reached)], no prior chemotherapy [8.4 months (95% CI, 7.1–13.3)], visceral metastasis [7.9 months (95% CI, 7.0–9.9)], and liver metastases [7.2 months (95% CI, 6.3–9.0)]. Conclusions: Elacestrant demonstrated a durable benefit in real-world clinical practice, particularly in earlier lines and in patients with prolonged prior ET exposure. Despite coexisting ESR1 and PI3K pathway mutations, TTNT remained clinically meaningful, reinforcing the role of elacestrant in personalized ET sequencing strategies prior to chemotherapy, antibody–drug conjugates, or targeted combinations. See related article by Lloyd et al., p. 169

PURPOSE:The EMERALD trial led to the approval of elacestrant for estrogen receptor (ER)-positive, HER2-negative, estrogen receptor 1 (ESR1)-mutated advanced or metastatic breast cancer (mBC) with disease progression following at least one line of endocrine therapy (ET). Subgroup analyses provided evidence suggesting that elacestrant enables ET sequencing in the second line before other targeted combinations, which could delay chemotherapy-based regimens. EXPERIMENTAL DESIGN:This study used claims data from the Komodo Research Dataset linked with Foundation Medicine Inc. clinical genomic data from patients with estrogen receptor+/HER2- mBC harboring an ESR1 mutation treated with elacestrant. The primary outcome measure was time to next treatment (TTNT). RESULTS:Among 306 patients, 93.8% had prior ET ± cyclin-dependent kinase 4/6 inhibitor for ≥12 months, 50.0% had prior chemotherapy, and 72.2% had prior fulvestrant. Median TTNT (mTTNT) was 8.2 months [95% confidence interval (CI), 6.3-13.0] in patients with 1 to 2 prior lines and 7.5 months (95% CI, 7.1-9.9) in those with ≥3 prior lines of ET. In patients with coexisting ESR1- and PI3K-pathway-mutated tumors, mTTNT was 6.3 months (95% CI, 4.8-7.9). mTTNT was 7.9 months (95% CI, 7.1-9.8) in all patients and was also sustained in patients with no prior fulvestrant [12.9 months (95% CI, 7.2-not reached)], no prior chemotherapy [8.4 months (95% CI, 7.1-13.3)], visceral metastasis [7.9 months (95% CI, 7.0-9.9)], and liver metastases [7.2 months (95% CI, 6.3-9.0)]. CONCLUSIONS:Elacestrant demonstrated a durable benefit in real-world clinical practice, particularly in earlier lines and in patients with prolonged prior ET exposure. Despite coexisting ESR1 and PI3K pathway mutations, TTNT remained clinically meaningful, reinforcing the role of elacestrant in personalized ET sequencing strategies prior to chemotherapy, antibody-drug conjugates, or targeted combinations. See related article by Lloyd et al., p. 169.

Importance:Progression-free survival (PFS) is a widely accepted primary end point in metastatic breast cancer (mBC) trials; however, it remains susceptible to assessment bias. To reduce this risk, blinded independent central review (BICR) is often incorporated into trial designs, but emerging evidence has suggested discordance between investigator- and BICR-assessed PFS. Objective:To compare BICR- and investigator-assessed PFS of patients with hormone receptor-positive/ERBB2-negative mBC. Data Sources:PubMed, Embase, CENTRAL, and major conference proceedings were searched from database inception to November 27, 2025, to identify all potential eligible studies. Study Selection:Phase 2 or 3 randomized clinical trials (RCTs) including patients with hormone receptor-positive/ERBB2-negative mBC with progression following CDK4/6 inhibitor therapy. Eligible studies required or permitted prior CDK4/6 inhibitor use and reported PFS assessed by both local investigators and BICR. Data Extraction and Synthesis:Data from the eligible RCTs were extracted by 2 reviewers (L.V.R. and Z.B.). A trial-level random-effect meta-analysis was performed. Main Outcomes and Measures:Outcomes were the discrepancy index (DI)-the ratio of hazard ratios for BICR- vs investigator-assessed PFS-and the absolute difference in median PFS between the 2 assessments. Results:Of 4989 identified records, 21 RCTs with 9165 patients met inclusion criteria. Eleven RCTs used BICR as the primary assessment, 13 were open label, and 7 required prior treatment with CDK4/6 inhibitors as an inclusion criterion. No statistically significant difference was observed between investigator- and BICR-assessed PFS overall, with a DI of 1.02 (95% CI, 0.95-1.10; P = .57), which was consistent across meta-regression analyses. The pooled PFS difference was 0.26 months (95% CI, -0.16 to 0.68 months; P = .22) in interventional arms and 0.44 months (95% CI, 0.09-0.78 months; P = .01) in control arms. Conclusions and Relevance:Results of this systematic review and meta-analysis show that despite a trend toward longer median PFS with BICR in recent RCTs of patients with hormone receptor-positive/ERBB2-negative mBC, there was no statistically significant DI between BICR- and investigator-assessed PFS. Results were consistent across meta-regression analyses and in outlier trials, suggesting that both approaches yield comparable PFS estimates.

Differences in elacestrant efficacy in patients with polyclonal versus single ESR1-mutant tumors. Among this cohort of patients with a baseline ESR1 mutation detected within 6 months of elacestrant start, clinical outcomes in patients with two, three, or four or more detectable ESR1 alterations were compared with patients with one ESR1 mutation. A depicts the TTNT on elacestrant, B depicts the TTD, and C depicts the OS comparing those with two ESR1 mutations in red, three in green, and four or more in brown against patients with a single ESR1 mutation in blue. The adjusted HR and corresponding 95% CI are displayed. Vertical hash marks denote a censored patient event. ESR1 alt denotes ESR1 alteration. alt, alteration; CL, confidence limit; pts, patients.

mTTNT benefit in relevant subgroups.

Precision oncology lacks scalable tools to assess, at the patient level, systems-level tumor microenvironment (TME) programs driving therapeutic resistance. To address this gap, we trained a weakly-supervised deep learning model, using routine H&E slides as input, to derive quantitative activity for therapeutically-relevant TME phenotypes, spanning immune, metabolic, and tumor cell-intrinsic programs. Using 3111 breast cancer H&E WSIs with matched bulk transcriptomics, our model accurately infers these biological states (AUROC>0.80; PCC>0.64). Validation spanned three levels: (i) tissue-matched multiplexed immunofluorescence, showing concordance between inferred functional states and immune cell fractions (p=0.006-0.106), (ii) blinded reader assessments, confirming localization of phenotype-specific morphology (p<3×10-5), and (iii) multi-institutional patient cohorts, where model-derived phenotypes stratified for clinical response (p<0.045). Despite relying on slide-level labels for training, our model's attention mechanism identifies focal regions of tumor tissue that drive the overarching clinical phenotype or treatment response. By extracting spatially resolved TME biology from routine histology, this strategy can be applied to massive legacy biobanks to enable discovery of new morpho-molecular mediators of resistance across real-world patient populations.

Cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) are an integral first-line treatment for hormone receptor-positive metastatic breast cancer, though resistance eventually develops. In this issue of Cancer Cell, Fan et al. report results from LINUX, a randomized phase 2 trial applying an artificial intelligence digital pathology classification system to guide therapy after CDK4/6i progression.

Neoadjuvant chemotherapy is standard for stage IB-III triple-negative breast cancer (TNBC), with pathological complete response (pCR) strongly associated with survival. Although escalation with platinum and immune checkpoint inhibitors (ICI) improves pCR and long-term outcomes, patients with pCR in control arms of pivotal trials also show favorable outcomes. Whether the regimen leading to pCR impacts long-term survival is largely unknown. We conducted a systematic review and meta-analysis, searching phase II and III trials including early-stage TNBC patients with pCR. A pooled analysis of Kaplan–Meier-derived individual patient data was performed for event-free survival (EFS) and overall survival (OS), with subgroup analyses by treatment regimens. Of 2830 identified publications, 18 trials comprising 3430 patients were included. Neoadjuvant ICI with chemotherapy improved EFS (HR 0.67; 95

Intrinsic and acquired resistance to CDK4/6 inhibitors (CDK4/6i) is a critical challenge in hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2-) metastatic breast cancer (MBC). CDK2 inhibitors (CDK2i) show promise in this context, with several currently under early-phase clinical evaluation. Here, by analyzing tumor and liquid biopsies from patients treated with the selective CDK2i PF-07104091, we observed that PF-07104091 monotherapy achieved disease stabilization in a cohort of CDK4/6i-resistant HR+/HER2- MBC patients. Notably, responses were observed irrespective of RB expression and phosphorylation, whereas non-progressive disease appeared more frequently among patients whose tumors retained wild-type TP53. Studies in established and patient-derived cell lines further substantiated that the growth suppressive effects of CDK2i were independent of RB status in CDK4/6i-resistant models. CDK2i activity reduced DNA replication rate, increased DNA damage, and inhibited mitotic entry, with growth inhibition dependent on p53 expression. These findings indicate that CDK dependency shifts from CDK4/6 toward CDK2 as cells transition from CDK4/6i-sensitive to CDK4/6i-resistant state. The distinct mechanisms of CDK2i and CDK4/6i support enhanced synergistic activity in HR+/HER2- MBC with acquired resistance to CDK4/6i therapy.

PURPOSE:ESR1 mutations mediate resistance to antiestrogen therapy in hormone receptor-positive metastatic breast cancer (MBC). Elacestrant, an oral selective estrogen receptor degrader, improves progression-free survival over standard endocrine therapy in ESR1-mutant MBC. We assessed real-world elacestrant use and clinical-genomic factors associated with outcomes. EXPERIMENTAL DESIGN:This study used the GuardantINFORM database, linking >42,000 real-world breast cancer cases with sequencing and claims data. We included patients with activating ESR1 mutations detected <6 months before elacestrant initiation (January 2023-March 2024). Outcomes of time-to-treatment-discontinuation, time-to-next-treatment (TTNT), and overall survival were estimated with Kaplan-Meier and Cox regression analysis, adjusting for clinical variables. RESULTS:We identified 756 patients (76% with prior cyclin-dependent kinase-4/6 inhibitor and 38% with prior chemotherapy exposure), and 742 (98.2%) were evaluable for outcomes. The median TTNT was 6.4 months, and the time-to-treatment-discontinuation was 4.6 months. In those with ≤1 prior lines of metastatic therapy, the TTNT was 8.8 months, compared with 6.0 months in the third-line setting. Prior fulvestrant exposure trended toward shorter treatment duration (hazard ratio, 1.19; 95% confidence interval, 0.91-1.56). Higher ESR1 polyclonality (≥4 alterations; 11% of patients) correlated with a shorter TTNT of 5.2 months (hazard ratio, 1.44; 95% confidence interval, 1.01-2.06), but efficacy was consistent across ESR1 alleles (e.g., Y537S and D538G). Disease with dual ESR1 and PI3K pathway mutations (PIK3CA, AKT1, and PTEN) had a median TTNT of 5.2 months. CONCLUSIONS:In ESR1-mutant MBC, elacestrant treatment durations support the routine use of elacestrant monotherapy in appropriately selected patients. For patients with concurrent ESR1 and PI3K pathway mutations, single-agent activity was comparable with outcomes observed in phase III studies. See related article by Rugo et al., p. 179.

Elacestrant therapy duration and OS by prior fulvestrant exposure. Depicted are clinical outcomes on elacestrant therapy comparing patients with versus those without previous treatment with fulvestrant. Patients treated with fulvestrant before elacestrant are displayed in red, and those without prior fulvestrant are displayed in blue. A depicts the TTNT, B depicts the TTD, and C depicts the OS. The adjusted HR and corresponding 95% CI comparing the two groups are displayed. Vertical hash marks denote a censored patient event. CL, confidence limit.

Purpose ESR1 mutations are commonly acquired on endocrine therapy (ET) and contribute to therapeutic resistance. The impact of increasing degrees of polyclonality - including dynamics of ESR1-mutant clones, co-occurring mutations, and treatment outcomes - is not well understood. Methods We conducted a retrospective, descriptive review of patients with estrogen receptor-positive (ER+) metastatic breast cancer (MBC) and >10 concurrent ESR1 variants on circulating tumor DNA (ctDNA) using either Guardant360 or Tempus xF between August, 2020 and July, 2025, defined as extreme ESR1 polyclonality. Patient and tumor characteristics, frequency of ESR1 variants, co-occurring mutations, and clonal changes in ESR1 in response to various treatments were explored. Results Nine patients with ER+ MBC and extreme ESR1 polyclonality were identified; 5 were post-menopausal at diagnosis. Five had received adjuvant chemotherapy and 6 received adjuvant ET for a minimum of 3 years (range, 3-7). Patients received a median of 1 line of ET in the metastatic setting before their first ctDNA test, which showed 0-31 ESR1 clones on initial ctDNA (mean, 6). The most frequent ESR1 clones were Y537N, Y537S, and D538G. The most frequent co-alterations included TP53 and RB1 mutations and CCND1 amplification. Tumor mutational burden (TMB) at the time extreme ESR1 polyclonality emerged ranged from 6.7-56.5 mutations per megabase (mut/Mb); all assessable tumors were microsatellite stable (MSS). Capecitabine and olaparib were associated with a decline in ESR1 polyclonality on ctDNA after radiographic progression. Other cytotoxic agents did not have the same observed effect. Increased ESR1 polyclonality was observed after progression on the oral selective estrogen receptor degrader (SERDs) elacestrant and imlunestrant. Among the 4 patients treated with elacestrant, the median progression-free survival was 6 months (range, 4-6). At data cutoff, 3 patients had died from progressive disease, 1 had transferred care, 1 was in hospice, and 4 remained on active treatment. Conclusions This is the first case series to explore extreme ESR1 polyclonality. Characterizing the mutational dynamics associated with this phenomenon may improve our general understanding of ESR1 mutations and related impact on treatment resistance.

Clinical outcomes in patients treated with elacestrant. A depicts the TTNT since elacestrant initiation in this cohort of patients with MBC and ESR1 mutation detected within 6 months prior to therapy start. B displays the TTD in this cohort, and C displays the OS. Vertical hash marks denote a censored patient event. CL, confidence limit.

Supplemental tables 1-2, supplemental figures 1-4

Study cohort baseline characteristics.

Patient outcomes on elacestrant by prior lines of therapy. TTNT (A), TTD (B), and OS (C) on elacestrant are depicted comparing differences in outcomes by the number of prior lines of metastatic therapy. Patients who received elacestrant after one or fewer prior lines are displayed in blue, after two prior lines are displayed in red, and after three or more prior lines are displayed in green. Adjusted HR and corresponding 95% CI are displayed comparing patients with two prior lines and three or more prior lines with those with one or fewer prior lines of metastatic therapy before elacestrant initiation. Vertical hash marks denote a censored patient event. CL, confidence limit; pts, patients.

Abstract While CDK4/6 inhibitors (CDK4/6i) have improved the treatment of hormone receptor-positive (HR+) breast cancer (BC), resistance remains a major clinical challenge. Resistance to CDK4 inhibition can be driven by higher CDK2 activity, due to overexpression of cyclin E1/E2 or p53 loss of function, and other mechanisms. As such, targeting both CDK4 and CDK2 is hypothesized to achieve more durable cell cycle arrest. This study evaluated the preclinical activity of GDC-4198, a novel CDK4/2 inhibitor currently in early clinical trials, and its potential to overcome CDK2-driven resistance to CDK4 inhibition as a single-agent and in combination with giredestrant, a novel selective estrogen receptor (ER) degrader and full ER antagonist. GDC-4198 has sub-nanomolar potency against CDK4/CycD1 and is a more potent inhibitor of CDK2/CycE or CDK2/CycA than CDK6/CycD3, unlike approved CDK4/6i. Western blots and immunofluorescence data demonstrated that GDC-4198 can directly decrease levels of pNCL, a direct marker of CDK2 activity. Cells engineered with a CDK2-activity reporter further confirmed the difference in inhibition profile of GDC-4198 compared to first generation CDK4/6i. In cell viability assays, we observed that GDC-4198 can maintain more durable growth arrest than CDK4/6 or CDK4-targeted inhibitors. Moreover, the introduction of TP53 knock-out had limited effects on GDC-4198 activity, and its potency was not changed by the overexpression of CCNE1/2, in contrast to other CDK inhibitors. Notably, across these different assays, the single-agent activity of GDC-4198 was comparable to the combination of CDK4- and CDK2-targeting agents. Additional studies in patient-derived cell lines obtained in the setting of metastatic HR+ breast cancer following progression on letrozole/ribociclib, as well as in cell lines with acquired in vitro resistance to palbociclib, confirmed the enhanced anti-proliferation activity of GDC-4198 compared to CDK4/6i. When combined with giredestrant, GDC-4198 showed strong combination benefits across a range of HR+ BC cell lines. In xenograft HR+ BC models, GDC-4198 demonstrated dose-dependent tumor growth inhibition as a single agent and further improved antitumor activity in combination with giredestrant. Pharmacokinetic and pharmacodynamic analyses indicated pRb modulation in tumor samples correlated with drug exposure. Taken together, these preclinical findings demonstrate that GDC-4198 induces more durable cell cycle arrest than first generation CDK4/6i and CDK4-targeting agents by overcoming CDK2-driven adaptive and intrinsic resistance to CDK4 inhibition. These results suggest that GDC-4198 is a promising next generation CDK4/2 inhibitor with the potential to prolong clinical benefit by delaying adaptation in earlier disease settings, as well as to provide benefit to patients who have progressed on approved CDK4/6i. Citation Format: Marc Hafner, Steffan Vartanian, Luca Gerosa, Nont Kosaisawe, Michael S. Hwang, Eva Lin, Yi-Chang Wang, Jason Oeh, Tianyi Chen, Kazi N. Islam, Alice Zheng, Karen Samy, Udi Segal, John G. Moffat, DANIEL ZINGG, Annie Collier, Ioannis Sanidas, Zhi Xie, Seth A. Wander. GDC-4198, a next-generation CDK4/2 inhibitor, induces durable cell cycle arrest and shows combination benefit with giredestrant [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 1910.

BACKGROUND:PI3K/AKT pathway activation is implicated in CDK4/6 inhibitor resistance. The use of AKT inhibition with continued CDK4/6 blockade after CDK4/6 inhibitor resistance remains unexplored. We evaluated the safety of ipatasertib and an antioestrogen with or without palbociclib in patients with treatment refractory HR+/HER2- metastatic breast cancer. METHODS:This single-centre, open-label, phase 1b trial was conducted at the Massachusetts General Hospital (Boston, MA, USA). Eligible patients were women older than 18 years with biopsy proven HR+/HER2- locally advanced, unresectable, or metastatic breast cancer; an Eastern Cooperative Oncology Group performance status of 0-2; disease progression on at least one previous therapy for metastatic disease; and measurable disease or bone lesions. Patients received 400 mg oral ipatasertib with standard 500 mg intramuscular fulvestrant dosing (ipatasertib and fulvestrant group) or with an aromatase inhibitor (oral anastrozole 1 mg per day, exemestane 25 mg per day, or letrozole 2·5 mg per day; ipatasertib and aromatase inhibitor group) on days 1-28 of each cycle. The ipatasertib and fulvestrant plus palbociclib group included a dose-escalation phase with patients assigned sequentially to escalating doses of ipatasertib and palbociclib using a standard 3 + 3 design starting at the recommended dose of palbociclib (125 mg on days 1-21) and the lowest dose of ipatasertib (200 mg on days 1-21). The primary endpoint was safety and progression-free survival was a key secondary endpoint. Safety was analysed in all patients who received at least one dose of ipatasertib and progression-free survival was assessed in all enrolled participants. This study is registered with ClinicalTrials.gov, NCT03959891 (active, not recruiting). FINDINGS:Between June 5, 2019, and Feb 16, 2022, 77 patients were enrolled (19 assigned to ipatasertib and fulvestrant, 16 to ipatasertib and aromatase inhibitor, and 42 to ipatasertib and fulvestrant plus palbociclib). All patients were female (77 [100%]); 75 were White (97%) and two (3%) were Asian. The median age was 62 years (range 32-88) and 66 (86%) of 77 patients received previous CDK4/6 inhibitor (median number of previous lines was 3 [range 1-13]). The median follow-up was 12·5 months (IQR 7·6-19·7). The recommended phase 2 dose was established at 400 mg ipatasertib on days 1-21 with 100 mg palbociclib on days 8-28 and standard fulvestrant 500 mg. Median progression-free survival was 5·5 months (95% CI 3·8-7·4). Serious adverse events related to study treatment occurred in seven (17%) patients in the ipatasertib and fulvestrant plus palbociclib group and one (5%) in the ipatasertib and fulvestrant group, which were related to neutropenia, leukopenia, thrombocytopenia, and hyperglycaemia. Common grade 3-4 adverse events related to study treatment (occurring in >5% of patients) were neutropenia (30 [39%] of 77), leukopenia (15 [19%]), diarrhoea (14 [18%]), rash (seven [9%]), lymphopenia (three [4%]), and anaemia (four [5%]). Four deaths occurred during the study (one possibly treatment-related due to grade 5 hyperglycaemia in the ipatasertib and fulvestrant group and two due to infectious issues and one due to pulmonary complications in the ipatasertib and fulvestrant plus palbociclib group), deemed unrelated to study treatment. INTERPRETATION:The combination of fulvestrant, ipatasertib, and palbociclib showed preliminary signs of clinical activity and showed expected adverse events in heavily pretreated patients with HR+/HER2- metastatic breast cancer, warranting further evaluation in those with CDK4/6 inhibitor-refractory disease. FUNDING:Genentech, Howard Hughes Medical Institute, National Foundation for Cancer Research, and Breast Cancer Research Foundation.