Intracranial efficacy and safety of trastuzumab deruxtecan in HER2-positive and HER2-low metastatic breast cancer with central nervous system involvement: A systematic review and proportional meta-analysis.

Abstract

1053 Background: Central nervous system (CNS) metastases and leptomeningeal disease (LMD) are major drivers of morbidity and mortality in metastatic breast cancer (MBC). Trastuzumab deruxtecan (T-DXd) has demonstrated intracranial activity across multiple studies; however, evidence remains fragmented across heterogeneous CNS subgroups. We conducted a systematic review and meta-analysis to assess intracranial efficacy and survival outcomes, along with interstitial lung disease (ILD) risk with T-DXd monotherapy in CNS-involved HER2-positive and HER2-low MBC. Methods: Following PRISMA guidelines, prospective trials and retrospective cohorts reporting CNS outcomes with T-DXd monotherapy were included. Fifteen unique cohorts met eligibility criteria. The primary endpoint was intracranial objective response rate (iORR). Secondary endpoints included progression-free survival (PFS) and overall survival (OS). Safety analysis included ILD incidence. Random-effects proportional meta-analyses and prespecified subgroup and sensitivity analyses were performed (RevMan 5.4.1). Results: In patients with active brain metastases (11 cohorts), pooled iORR was 0.61 (95% CI 0.55–0.66; I²=0%). Similarly, in stable brain metastases (6 cohorts), pooled iORR was 0.61 (95% CI 0.50–0.71; I²=75%); exclusion of an influential cohort increased pooled iORR to 0.65 (95% CI 0.55–0.75) and reduced heterogeneity. In LMD (3 cohorts), pooled CNS response was 0.53 (95% CI 0.35–0.70; I²=0%). HER2-low cohorts demonstrated a pooled iORR of 0.51 (95% CI 0.33–0.69), which did not differ significantly from HER2-positive CNS subgroups (p=0.75). Where comparative data were available, treatment with T-DXd was associated with improved PFS (HR 0.33, 95% CI 0.19–0.58) and overall survival (log HR −0.32, 95% CI −0.55 to −0.09; ≈HR 0.73). Safety analyses showed a pooled incidence of ILD (any grade) of 0.12 (95% CI 0.09–0.15), with higher rates observed in high-risk (0.16) and intermediate-risk (0.13) cohorts compared with low-risk cohorts (0.04). Conclusions: Across 15 cohorts, T-DXd demonstrates consistent and clinically meaningful intracranial activity in active and stable brain metastases, substantial activity in LMD, and a survival signal where comparative data exist, with quantifiable ILD risk. These findings support T-DXd as a key systemic therapy for CNS-involved HER2-positive and HER2-low MBC and underscore the need for standardized CNS endpoints and ILD mitigation strategies in future studies.