Abstract
We aimed to develop and internally validate a nomogram to estimate axillary pathological complete response (pCR, ypN0) after neoadjuvant systemic therapy (NAST) in clinically node-positive (cN1-2) breast cancer. In a single-center retrospective cohort of 144 consecutive patients treated with NAST (anti-HER2 as indicated), all underwent standardized pre- and post-NAST 18F-FDG PET/CT and axillary staging (sentinel lymph node biopsy [SLNB], targeted axillary dissection [TAD], or axillary lymph node dissection [ALND]). Axillary pCR occurred in 51.4% (74/144). In a multivariable analysis, independent positive determinants of ypN0 included the triple-negative subtype, Modified PERCIST (SUVmax-based) reduction ≥ 80.70%, pre-NAST tumor-to-axilla SUVmax ratio ≥ 1.21, and residual breast tumor size < 0.5 mm; conversely, conglomerate/matted nodal morphology at diagnosis was inversely associated. The model showed good internal discrimination (AUC 0.857, 95% CI 0.797-0.917) and acceptable calibration (Hosmer-Lemeshow p = 0.425). Exploratory, subtype-restricted signals were observed for inflammatory indices within Luminal B (HER2+) but were not retained in the final model. The resulting nomogram-combining tumor biology, PET/CT response, and pre-NAST nodal features-may support risk stratification for axillary de-escalation after NAST; however, prospective external validation-ideally embedded in ongoing de-escalation frameworks-remains essential before clinical implementation, and the tool should currently be regarded as hypothesis-generating rather than a stand-alone decision aid for routine practice.