Résumé
Abstract Background Lymphopenia has been associated with poor outcomes in metastatic breast cancer (BC), but its prognostic relevance in early-stage disease remains unclear. This study aimed to evaluate the prognostic value of baseline lymphopenia in patients with non-metastatic BC using data from the prospective French CANTO cohort (NCT01993498). Methods 10,854 patients with baseline absolute lymphocyte count (ALC) were analysed. Lymphopenia was defined as ALC < 1.0 × 10⁹ cells/L, measured before any cancer treatment. Relapse-free survival (RFS), overall survival (OS), and treatment-related toxicities were assessed using Kaplan-Meier estimates, piecewise-Cox regression models, and cause-specific hazard analyses considering pre-specified time periods. Results Baseline lymphopenia was observed in 342 patients (3.1%). It was associated with ECOG performance status > 0, hypoalbuminemia, and prior neoplasia. Lymphopenia correlated with inferior 5-year RFS (89.1% vs. 92.9%) and OS (92.5% vs. 96.4%). In multivariable analyses, lymphopenia was associated with increased risk of early relapse during the first 24 months (HR = 2.11; 95%CI: 1.28–3.48; p = 0.003), but not beyond, and no independent prognostic impact on OS was observed. No significant differences were found in surgical complications or chemotherapy dose intensity, though granulocyte colony-stimulating factor use was higher among lymphopenic patients. Conclusion Baseline lymphopenia is uncommon in early BC but may serve as a marker of early relapse risk. Although it does not independently predict long-term survival, its presence could reflect underlying tumour aggressiveness or patient frailty. These findings support further investigation of lymphocyte subpopulations to refine prognostic stratification in early BC. Trial registration Approved by French ethics committee in 2011 (ID-RCB:2011-A01095- 36,11–039 /NCT01993498 [20FEB2012]).