DESTINY-Breast03
In HER2-positive metastatic breast cancer previously treated with trastuzumab and a taxane, does trastuzumab deruxtecan improve outcomes versus T-DM1?
Trial design and population
524 patients with HER2-positive unresectable/metastatic breast cancer previously treated with trastuzumab and a taxane; active brain metastases were excluded at enrolment.
Trastuzumab deruxtecan 5.4 mg/kg IV every 3 weeks vs T-DM1 3.6 mg/kg IV every 3 weeks.
524
PFS by blinded independent central review; OS key secondary endpoint
Key results
At median follow-up 28.4 months, median PFS was 28.8 vs 6.8 months (HR 0.33, 95% CI 0.26–0.43); OS HR 0.64 (95% CI 0.47–0.87). Grade ≥3 treatment-emergent AEs were 56% vs 52%. Drug-related ILD/pneumonitis occurred in 15% vs 3%, with no grade 4/5 ILD in that analysis.
Source-reported outcome visualization
Median progression-free survival · Updated analysis
Source-reported endpoint and unit. Median survival is not converted into an NNT.
Absolute Effect
Median PFS difference 22 months is not an NNT. ILD/pneumonitis is the distinctive serious harm and requires surveillance.
NNT: Not calculated; median PFS difference is not a valid NNT. Do not convert median survival differences into NNT.
NNH: Not calculated from aggregate ILD rates because severity/time-to-event handling differs; report incidence and grade explicitly. Drug-related ILD/pneumonitis about 15% vs 3%, mostly low grade in the cited analysis; promptly investigate cough, dyspnoea or radiographic change and permanently discontinue for higher-grade ILD.
Safety and treatment burden
Drug-related ILD/pneumonitis about 15% vs 3%, mostly low grade in the cited analysis; promptly investigate cough, dyspnoea or radiographic change and permanently discontinue for higher-grade ILD.
Follow-up, maturity and limitations
Updated analysis median follow-up 28.4 months.
Applies to HER2-positive disease after trastuzumab/taxane, not HER2-low populations. Active brain metastases were excluded at baseline; later CNS data are separate evidence.
Clinical interpretation
Current-practice context · reviewed 20 October 2026
Historical evidence is separated from current sequencing. Confirm current guidelines, local regulatory status, biomarkers, prior therapies and patient-specific contraindications. Newer treatments may have changed the preferred sequence; where NNT/NNH cannot be defensibly quantified, no surrogate calculation is made.