Abstract
We developed and temporally internally validated a pretreatment nomogram for predicting pathological complete response (pCR) in human epidermal growth factor receptor 2 (HER2)-positive breast cancer and examined cross-sectional associations among systemic inflammation, model-derived pCR probability, and baseline psychosocial burden. This ambispective single-center cohort included women receiving neoadjuvant chemotherapy plus HER2-directed therapy followed by surgery. Patients treated from January 2019 to December 2023 formed the training cohort (n=198), and those treated from January to August 2024 formed a same-center temporal validation cohort (n=48). Psychosocial support need was defined as Distress Thermometer score ≥4, Hospital Anxiety and Depression Scale-Anxiety score ≥8, or Hospital Anxiety and Depression Scale-Depression score ≥8. Among 315 screened patients, 246 had complete data. pCR occurred in 72 training patients (36.36%) and 17 validation patients (35.42%). The final nomogram retained tumor size, clinical nodal status, hormone receptor status, and standardized neutrophil-to-lymphocyte ratio (NLR). Discrimination was acceptable in the training cohort (AUC=0.776; optimism-corrected AUC=0.758) and temporal validation cohort (AUC=0.769), although validation estimates were imprecise. Overall, 125 patients (50.81%) met the psychosocial support-need endpoint. Lower NLR-containing predicted pCR probability was associated with support need (OR=1.49 per 0.10 decrease), but this association attenuated after direct NLR adjustment (OR=1.13), indicating limited incremental psychosocial value beyond inflammation. The nomogram showed acceptable pretreatment pCR prediction performance, whereas psychosocial associations should be interpreted as non-causal clinical-inflammatory correlations requiring prospective multicenter validation.