Prognostic Value of Baseline Neutrophil-to-Lymphocyte Ratio in Advanced Breast Cancer Patients Treated with Trastuzumab Deruxtecan: A Real-World Study

Background: Trastuzumab deruxtecan (T-DXd) has redefined the treatment of HER2-positive advanced breast cancer (HER2+ ABC), yet responses remain heterogeneous and accessible biomarkers are lacking. The neutrophil-to-lymphocyte ratio (NLR) is an inexpensive, routinely available systemic inflammatory marker with established prognostic value in HER2-positive disease treated with anti-HER2 monoclonal antibodies; its prognostic significance in patients receiving the antibody–drug conjugate T-DXd remains undetermined. Methods: This two-centre real-world study screened 146 consecutive patients with HER2+ ABC who received at least one cycle of T-DXd between January 2023 and June 2025 (data cutoff 16 October 2025) and enrolled 92 who met all eligibility criteria. The prespecified eligibility rule requiring a baseline complete blood count within 7 days before the first T-DXd dose was applied uniformly to every screened candidate; 34 candidates failed this rule (count drawn after the first dose, n = 16; drawn more than 7 days before, n = 17; collection date not recorded, n = 1) and one candidate was excluded because a baseline lymphocyte count at the assay lower limit rendered the derived ratio unreliable. The primary analysis evaluated baseline NLR as a continuous variable in Cox proportional-hazards regression for progression-free survival (PFS). Prespecified secondary analyses comprised median-split Kaplan–Meier analysis, quartile analysis with trend testing, individual blood components, and multivariable models. Prespecified sensitivity analyses, proportional-hazards testing, restricted cubic spline assessment of linearity, nested-model likelihood-ratio testing of the incremental value of NLR over the absolute lymphocyte count, and a centre-stratified analysis were performed. Exploratory subgroup analyses included interaction testing with false-discovery-rate correction. Results: Of 146 screened patients, 92 were enrolled (54 excluded); 39 PFS events occurred. Median follow-up by reverse Kaplan–Meier was 7.6 months; overall median PFS was 9.82 months, with 6- and 12-month PFS rates of 73.7% (95% CI 62.3–82.1) and 37.1% (22.6–51.6). Each one-unit increase in baseline NLR was associated with a 37% higher hazard of progression or death (HR 1.37, 95% CI 1.12–1.69, p = 0.0028; C-index 0.672). At the cohort median cut-off (2.157), high NLR conferred shorter PFS than low NLR (median 7.66 vs 11.50 months; HR 2.30, 95% CI 1.21–4.38; log-rank p = 0.0096). Quartile analysis showed a significant dose–response trend (HR per quartile 1.58, 95% CI 1.16–2.13, p for trend = 0.0033), with Q4 versus Q1 HR 4.14 (1.52–11.26, p = 0.0053). Among individual components, only the lymphocyte count showed a significant median-split association (log-rank p = 0.011), and NLR provided prognostic information beyond the lymphocyte count (likelihood-ratio p = 0.039) and beyond a clinically driven multivariable model (likelihood-ratio p = 0.014; ΔC-index +0.077). In the clinically driven multivariable model (n = 92, 7.8 events per variable), NLR remained an independent prognostic factor (HR 1.36, 95% CI 1.09–1.70, p = 0.0062), and estimates were consistent across a minimal model, two extended models and a ridge-penalised model. The association was consistent across both centres (interaction p = 0.73). Model assumptions were satisfied (Schoenfeld p = 0.75; non-linearity p = 0.46). No subgroup interaction remained significant after false-discovery-rate correction across 17 tests. Conclusions: Elevated baseline NLR is an independent adverse prognostic marker for PFS among HER2+ ABC patients treated with T-DXd, and carries prognostic information beyond its individual blood components and beyond standard clinical variables. Because no comparator arm exists, NLR should be interpreted as a prognostic—not predictive—biomarker in this setting; prospective validation is required before any clinical application.