Trastuzumab deruxtecan in HER2-positive advanced breast cancer with or without brain metastases: a phase 3b/4 trial

HER2-positive breast cancer represents 15–20% of cases and up to 50% of patients with advanced/metastatic disease develop brain metastases (BMs), which confer poorer prognosis. Local CNS therapies (surgery, SRS, WBRT) are standard but often yield CNS progression within 6–12 months and do not address extracranial disease; WBRT can cause cognitive decline. Therefore, additional systemic options with CNS activity are needed. Trastuzumab-based regimens are backbone treatments, and newer HER2-directed agents (e.g., tucatinib) have shown CNS benefits. Trastuzumab deruxtecan (T-DXd), an anti-HER2 antibody–drug conjugate with a topoisomerase I payload, is approved after prior anti-HER2 therapy based on DESTINY-Breast03. Preliminary pooled and small prospective studies suggested intracranial activity of T-DXd in stable and active BMs. DESTINY-Breast12 was designed to prospectively evaluate efficacy and safety of T-DXd in HER2+ mBC patients with and without baseline BMs, focusing on CNS outcomes in a real-world-like setting.

Prior work established tucatinib plus trastuzumab and capecitabine as an effective CNS-active regimen in pretreated HER2+ mBC, including patients with active BMs (HER2CLIMB), with intracranial ORR around 47% in measurable BMs and median CNS PFS near 10 months. T-DXd’s systemic superiority over T-DM1 was shown in DESTINY-Breast03. Exploratory pooled analyses from DESTINY-Breast01/02/03 indicated intracranial ORR ~45% and CNS PFS 12–18.5 months in stable and active BMs. Small prospective studies (DEBBRAH, TUXEDO-1, DESTINY-Breast07 expansion) reported encouraging CNS responses in active BMs. Retrospective ROSET-BM and other series also suggested CNS activity, including in leptomeningeal disease, though leptomeningeal metastases were excluded from the present trial. Collectively, these data motivated a large prospective evaluation of T-DXd’s intracranial efficacy.

Design: DESTINY-Breast12 was a prospective, open-label, single-arm, multicenter, international phase 3b/4 study (NCT04739761). Two parallel cohorts enrolled HER2+ advanced/metastatic breast cancer patients with baseline brain metastases (BMs; stable or active) and without baseline BMs. Key timeline: enrollment June 2021 to February 2024 across 78 sites; data cutoff 8 February 2024. Eligibility: Adults ≥18 years with pathologically confirmed HER2+ disease per ASCO/CAP guidelines; progression on ≥1 prior anti-HER2-based regimen; ≤2 prior metastatic regimens; tucatinib-naïve; ECOG PS 0–1. Patients with known or suspected leptomeningeal metastases were excluded. Patients could enroll without measurable disease. Stable BMs: radiographically stable ≥4 weeks post-treatment. Active BMs: untreated (≤2 cm lesions) or previously treated and progressing without indication for immediate local retreatment. Washouts: ≥7 days after SRS/gamma knife, ≥21 days after WBRT. Treatment: T-DXd 5.4 mg/kg IV every 3 weeks until RECIST 1.1-defined progression outside CNS, unacceptable toxicity, withdrawal, or other discontinuation criteria. Patients could continue on study after local therapy for isolated CNS progression until second progression. Assessments: Tumor imaging of chest/abdomen/pelvis (CT or MRI with IV contrast) at baseline, every 6 weeks for 48 weeks, then every 9 weeks. Brain MRI (with/without contrast) or contrast-enhanced CT at baseline and same intervals for the BMs cohort. Patients with active measurable BMs had intracranial lesions designated as target lesions for RECIST 1.1 CNS assessments. Response and progression were assessed by independent central review (ICR) per RECIST 1.1. Endpoints: Primary endpoint in BMs cohort: progression-free survival (PFS: first dose to progression or death). Secondary endpoints: CNS PFS; overall survival (OS); objective response rate (ORR); time to second progression (PFS2); time to progression; CNS ORR (patients with measurable CNS disease at baseline); duration of response (DOR); and safety. Primary endpoint in non-BMs cohort: ORR. Secondary endpoints: DOR, OS, time to progression, incidence of new symptomatic CNS metastases, and safety. Additional prespecified endpoints (e.g., site of next progression, PROs, cognitive outcomes) will be reported separately. Safety: AEs coded by MedDRA v26.1 and graded by CTCAE v5.0. Potential ILD/pneumonitis triggered treatment interruption and investigation; no formal adjudication committee was used, but an ILD advisory committee reviewed cases outside the protocol parameters. Pulmonary toxicity management per published guidance. Statistics: Single-arm descriptive analyses using SAS 9.4; no formal hypothesis testing or sample size calculation. Approximately 250 patients per cohort targeted to ensure narrow 95% CI precision for primary endpoints. Efficacy in the full analysis set (all treated). Kaplan–Meier for PFS/OS/CNS PFS/DOR; Brookmeyer–Crowley for median CIs. ORR and CNS ORR based on confirmed responses up to progression or last evaluable assessment; rules for censoring and handling missed visits specified. Analyses were separate by cohort; prespecified descriptive subgroup analyses for active vs stable BMs.

Population: 504 treated patients (baseline BMs n=263; no baseline BMs n=241). In BMs cohort: 157 stable; 106 active (39 untreated; 67 previously treated/progressing). Median prior metastatic regimens 1.0 in both cohorts. Efficacy (BMs cohort): - 12-month PFS: 61.6% (95% CI 54.9–67.6); median PFS 17.3 months (95% CI 13.7–22.1; post hoc). - By subgroup: 12-month PFS stable BMs 62.9% (95% CI 54.0–70.5); active BMs 59.6% (95% CI 49.0–68.7). Within active: untreated 47.0% (95% CI 29.6–62.7); previously treated/progressing 66.7% (95% CI 53.4–76.9) (post hoc). - 12-month CNS PFS: 58.9% (95% CI 51.9–65.3); stable 57.8% (95% CI 48.2–66.1); active 60.1% (95% CI 49.2–69.4). - Confirmed ORR (all, including non-measurable): 51.7% (95% CI 45.7–57.8); complete response 4.2% (n=11), partial response 47.5% (n=125). In measurable disease subset (n=198): ORR 64.1% (95% CI 57.5–70.8); stable 67.0% (n=109; 95% CI 58.1–75.8); active 60.7% (n=89; 95% CI 50.5–70.8). Median DOR not reached at cutoff; >50% of responses ongoing. - CNS ORR among measurable CNS disease (n=138): overall 71.7% (95% CI 64.2–79.3); stable 79.2% (95% CI 70.2–88.3); active 62.3% (95% CI 50.1–74.5). Within active: untreated 82.6% (19/23; 95% CI 67.1–98.1); previously treated/progressing 50.0% (19/38; 95% CI 34.1–65.9) (post hoc). - 12-month OS: 90.3% (95% CI 85.9–93.4); data immature (16.3% maturity). Efficacy (no baseline BMs cohort): - Confirmed ORR: 62.7% (95% CI 56.5–68.8); complete response 9.5% (n=23); partial response 53.1% (n=128). In measurable disease subset (n=215): ORR 68.4% (95% CI 62.2–74.6). Median DOR not reached; majority of responses ongoing. - 12-month OS: 90.6% (95% CI 86.0–93.8) (17.0% maturity). - Time to progression at 12 months: 72.1% (95% CI 65.4–77.8) without progression; median not reached. - New symptomatic CNS metastases: 4 patients; incidence rate 0.017 (95% CI 0.00452–0.04250). Safety: - Median treatment duration: BMs 11.5 months (range 0.1–26.9); non-BMs 12.0 (0.7–28.4). - Any AE: 98.5% (BMs); 98.3% (non-BMs). Grade ≥3 AEs: 51.0% vs 49.0%. - Common AEs (all grades): nausea (65–71%), fatigue (62–73%), constipation (~40%), neutropenia (30–36%), diarrhea (28–36%), vomiting (28–34%). - Grade ≥3 common AEs: neutropenia 16.3% (BMs) and 18.3% (non-BMs); fatigue 8.7% and 10.0%; anemia 7.2% and 5.0%. - Discontinuations due to AEs: 15.2% (BMs) and 9.5% (non-BMs); most common cause ILD/pneumonitis (10.3% and 5.4%). - Investigator-reported ILD/pneumonitis: 16.0% (BMs; grade ≥3: 3.4%; six grade 5 deaths 2.3%) and 12.9% (non-BMs; grade ≥3: 2.5%; three grade 5 deaths 1.2%). Median time to first onset ~168–169 days. Some cases co-occurred with opportunistic infections (not systematically tested), notably PJP. - Left ventricular ejection fraction decrease: 11.8% (BMs) and 10.8% (non-BMs); mostly low grade; no grade ≥4 in non-BMs; rare grade ≥3 in BMs.

This large, prospective single-arm study demonstrates substantial systemic and intracranial activity of T-DXd in HER2+ mBC with baseline BMs, including active lesions, aligning with and extending prior smaller studies. PFS at 12 months (61.6%) and CNS PFS (58.9%) with high CNS ORR (71.7% overall; 79.2% stable; 62.3% active) indicate robust intracranial control despite the agent’s macromolecular nature. Responses were durable with many ongoing at cutoff. In patients without baseline BMs, ORR was 62.7% and CNS progression as a first symptomatic event was rare. Compared with historical tucatinib-based data, direct comparisons are inappropriate due to design and population differences, yet the magnitude and durability of CNS activity observed with T-DXd support its role as an important option in previously treated patients with BMs. Safety was consistent with known T-DXd risks, notably ILD/pneumonitis; prompt recognition and management remain critical, and consideration of PJP prophylaxis in patients on chronic corticosteroids is highlighted. Overall, findings support T-DXd use across BMs status, with individualization versus tucatinib-based regimens based on efficacy and toxicity considerations.

T-DXd produced substantial and durable overall and intracranial clinical activity in previously treated HER2+ metastatic breast cancer, irrespective of the presence of stable or active brain metastases, with safety consistent with prior reports and no new signals. ILD/pneumonitis remains an important identified risk requiring vigilant monitoring and management. These results support T-DXd as a treatment option for HER2+ mBC patients with or without BMs. Future work should further define CNS efficacy in HER2-low disease and other HER2+ solid tumors and assess patient-reported and neurocognitive outcomes and longer-term durability.

Open-label, single-arm design with no comparator limits causal inference and cross-trial comparisons. Data were immature at cutoff with no planned long-term follow-up. In the non-BMs cohort, ORR was the primary endpoint and PFS was not assessed; brain imaging was not regularly performed, restricting CNS analyses to symptomatic metastases. Prior CNS radiotherapy in some patients (especially stable and previously treated/progressing BMs) may have impaired target lesion assessments. ILD/pneumonitis was not adjudicated by a formal committee, complicating comparisons to studies with adjudication. Patients with leptomeningeal metastases were excluded. Black and Asian patients were underrepresented. Some endpoints (e.g., PFS2, time to progression, DOR) were immature at analysis.