T-DXd safety profile in DESTINY-Breast04 trial

Grade ≥3 TEAEs and Exposure-Adjusted Incidence Rates: T-DXd vs. TPC

Grade ≥3 treatment-emergent adverse events (TEAEs) occurred in 202 patients (54.4%) in the T-DXd arm of DESTINY-Breast04, compared with 116 patients (67.4%) in the physician’s choice of treatment (TPC) arm — a difference that becomes even more pronounced when adjusted for the unequal treatment durations across the two groups.

The exposure-adjusted incidence rate (EAIR) for grade ≥3 TEAEs was 0.64 per patient-year for T-DXd versus 1.81 per patient-year for TPC. This is particularly meaningful given that the median treatment duration in the T-DXd arm was 8.2 months, compared with just 3.5 months in the TPC arm. In other words, patients on T-DXd were treated for more than twice as long, yet experienced fewer adverse events per unit of treatment time.

Any-grade TEAE EAIRs followed the same pattern: 1.16 per patient-year for T-DXd versus 2.64 per patient-year for TPC. For serious TEAEs, the gap narrowed but remained meaningful: 0.34 versus 0.69 per patient-year. These data, reported by the DESTINY-Breast04 investigators, indicate that T-DXd does not generate cumulative toxicities with extended follow-up — a finding with direct implications for treatment planning in HER2-low metastatic breast cancer.

The most common grade ≥3 drug-related TEAEs in the T-DXd arm were neutropenia (14.6%), anemia (10.8%), fatigue (9.7%), and leukopenia (7.0%). By contrast, the TPC arm was dominated by neutropenia (42.4%), leukopenia (19.2%), and increased transaminases (13.4%) — a profile consistent with the cytotoxic chemotherapy regimens constituting TPC. Nausea and vomiting, while predominantly any-grade events in the T-DXd arm (76.0% and 40.7%, respectively), underscore the high emetic risk associated with T-DXd and the corresponding need for prophylactic antiemetic management.

ILD/Pneumonitis in DESTINY-Breast04: Incidence, Grading Distribution, and Time to Onset

Adjudicated drug-related interstitial lung disease (ILD) and/or pneumonitis occurred in 45 patients (12.1%) in the T-DXd arm of DESTINY-Breast04, making it the single most clinically consequential adverse event of special interest for this antibody–drug conjugate. In the TPC arm, only one patient (0.6%) experienced a grade 1 adjudicated drug-related ILD/pneumonitis event, associated with eribulin.

The grading distribution reveals that the majority of ILD events were lower-grade: grade 1 in 13 patients (3.5%), grade 2 in 24 patients (6.5%), grade 3 in 4 patients (1.1%), no grade 4 events, and grade 5 (fatal) events in 4 patients (1.1%). The absence of grade 4 events alongside the presence of fatal cases underscores the non-linear severity spectrum of ILD with T-DXd and the imperative for early detection before events escalate.

The median time to onset of ILD/pneumonitis was 129 days (range 26–710 days), indicating that the risk window extends well beyond the first treatment cycle. Critically, no new cases of ILD/pneumonitis were reported after an additional approximately 14 months of follow-up since the primary analysis data cutoff, suggesting that late-onset ILD is uncommon with T-DXd. However, the wide onset range — up to 710 days — means clinicians cannot relax surveillance at any fixed time point.

“ILD/pneumonitis was the most common TEAE associated with drug discontinuation in the T-DXd arm, accounting for 10.2% of patients — underscoring why early detection and established management guidelines are non-negotiable.”

ILD/pneumonitis was the most common TEAE leading to drug discontinuation in the T-DXd arm, affecting 10.2% of patients. This dual burden — both clinical severity and treatment continuity — makes ILD/pneumonitis the defining safety challenge for T-DXd in HER2-low metastatic breast cancer. Regulatory agencies including the FDA and EMA have incorporated ILD monitoring requirements into the approved prescribing information for T-DXd.

Left Ventricular Dysfunction and Cardiac Failure: Event Rates and Clinical Significance

Left ventricular dysfunction occurred in 19 patients (5.1%) in the T-DXd arm of DESTINY-Breast04, with no comparable events observed in the TPC arm. This asymmetry — exclusive to the T-DXd arm — warrants systematic cardiac surveillance throughout treatment, even though the absolute event rate is substantially lower than that of ILD/pneumonitis.

The breakdown of decreased ejection fraction events reveals a predominantly grade 2 profile: 2 patients experienced grade 1, 15 patients grade 2, and 1 patient grade 3 decreased ejection fraction. Cardiac failure was observed in 2 patients (0.5%): 1 grade 2 and 1 grade 3. The investigators noted that these rates of left ventricular dysfunction were consistent with previously observed data for T-DXd, suggesting a class-level signal rather than a study-specific finding.

While the cardiac event rates are lower in absolute terms than ILD/pneumonitis, the clinical implications are significant. Decreased ejection fraction can affect a patient’s eligibility for other cardiotoxic therapies and has direct quality-of-life consequences. Oncology societies including ASCO have published cardio-oncology guidance relevant to HER2-targeted therapies, and the DESTINY-Breast04 data reinforce the need to incorporate cardiac function monitoring into T-DXd treatment protocols from the outset.

Discontinuation, Dose Interruption, and Dose Reduction: Rates and Drivers

Treatment discontinuation due to TEAEs occurred in 62 patients (16.7%) in the T-DXd arm and 14 patients (8.1%) in the TPC arm — a higher absolute rate for T-DXd that is contextualised by the exposure-adjusted analysis. The EAIR for TEAEs associated with discontinuation was 0.20 for T-DXd and 0.22 for TPC, indicating near-parity when treatment duration is accounted for.

Dose interruption due to TEAEs affected 155 patients (41.8%) in the T-DXd arm and 73 patients (42.4%) in the TPC arm — almost identical proportions. However, the EAIRs diverge substantially: 0.49 per patient-year for T-DXd versus 1.14 per patient-year for TPC. Dose reduction due to TEAEs showed a similar pattern: 89 patients (24.0%) for T-DXd versus 65 patients (37.8%) for TPC, with EAIRs of 0.28 and 1.01 per patient-year respectively.

Importantly, disease progression — not toxicity — was the primary overall reason for treatment discontinuation in both arms: 69.0% for T-DXd and 76.7% for TPC. This framing is clinically meaningful: it indicates that the majority of patients stopping T-DXd are doing so because their cancer is progressing, not because the drug is intolerable. Among TEAEs specifically, investigator-assessed ILD/pneumonitis (10.2%) was the leading cause of discontinuation in the T-DXd arm, reinforcing its status as the dominant safety management priority.

Real-World Monitoring Protocols and Clinical Decision-Making for T-DXd in HER2-Low Metastatic Breast Cancer

The long-term safety findings from DESTINY-Breast04 confirm that T-DXd’s overall safety profile is acceptable and manageable, with no evidence of cumulative toxicity accumulating over extended treatment. Translating this into real-world practice requires structured monitoring protocols calibrated to the specific adverse-event hierarchy observed in the trial.

ILD/Pneumonitis: The Priority Surveillance Target

Given that ILD/pneumonitis carries both clinical severity risk (including grade 5 events) and is the leading TEAE cause of treatment discontinuation (10.2%), proactive surveillance is non-negotiable. Real-world protocols should incorporate regular screening for new or worsening respiratory symptoms at every clinical encounter. Chest CT imaging should be considered at baseline and at intervals during treatment, with a low threshold for repeat imaging when patients report dyspnoea, cough, or fever. The median onset of 129 days — but with a range extending to 710 days — means surveillance cannot be safely time-limited. Published management guidelines, including those from ESMO, provide grading-specific algorithms for dose modification and corticosteroid use that should be integrated into institutional protocols.

Gastrointestinal Toxicity: Prophylaxis as Standard Practice

Nausea (76.0%) and vomiting (40.7%) are the most prevalent any-grade drug-related TEAEs in the T-DXd arm, reflecting the drug’s high emetic risk classification. Prophylactic antiemetic regimens — including 5-HT3 antagonists and NK1 receptor antagonists — should be initiated with the first T-DXd infusion and continued proactively rather than reactively. Adequate antiemetic management is directly linked to treatment adherence and patient quality of life, both of which affect the ability to realise the efficacy benefits of T-DXd demonstrated in DESTINY-Breast04.

Hematologic Monitoring: Routine but Essential

Neutropenia (14.6% grade ≥3), anemia (10.8% grade ≥3), and leukopenia (7.0% grade ≥3) in the T-DXd arm require systematic complete blood count monitoring before each treatment cycle. While these rates are substantially lower than in the TPC arm (neutropenia 42.4% grade ≥3), they remain clinically significant and may require dose modification or growth factor support in individual patients.

Cardiac Function: Baseline and On-Treatment Assessment

The 5.1% rate of left ventricular dysfunction — entirely absent in the TPC arm — supports establishing baseline echocardiography or multigated acquisition (MUGA) scanning before initiating T-DXd, with periodic reassessment during treatment. The predominantly grade 2 profile of decreased ejection fraction events suggests that many cases are manageable with dose modification rather than permanent discontinuation, provided they are identified early. According to guidance from WHO and oncology cardiology bodies, patients with pre-existing cardiac dysfunction may require more frequent monitoring intervals.

Synthesising Safety Into Clinical Decision-Making

The DESTINY-Breast04 long-term safety data reinforce T-DXd as a standard of care for HER2-low metastatic breast cancer after prior chemotherapy. The exposure-adjusted analysis demonstrates that the safety burden per unit of treatment time is lower for T-DXd than for TPC across all major TEAE categories. Successful real-world implementation, however, depends on four pillars: structured ILD surveillance with low imaging thresholds; aggressive prophylactic antiemetic management; routine hematologic monitoring; and baseline plus periodic cardiac assessment. Patient education — ensuring individuals can recognise and promptly report early respiratory symptoms — is an equally critical component that clinical teams should formalise at treatment initiation. The PatSnap life sciences intelligence platform and PatSnap Insights provide ongoing analysis of emerging safety signals and regulatory updates relevant to ADC therapies including T-DXd.