Abstract
PURPOSE: Predicting response to neoadjuvant therapy (NAT) remains a clinical challenge in patients with HER2-positive breast cancer (BC). Systemic inflammatory and immune-nutritional biomarkers have emerged as potential predictors of treatment response; however, their value in patients receiving dual anti-HER2 therapy is not well defined. PATIENTS AND METHODS: This retrospective study included patients with HER2-positive BC treated with neoadjuvant dual anti-HER2 therapy between January 2023 and February 2025. A total of 136 patients were included. Pathological complete response (pCR) and radiological response assessed by positron emission tomography/computed tomography (PET/CT) were the primary outcomes. Routinely available inflammatory and immune-nutritional indices, including lymphocyte-to-monocyte ratio (LMR), systemic immune-inflammation index (SII), and C-reactive protein-to-albumin ratio (CAR), were evaluated using receiver operating characteristic analysis and multivariable logistic regression. Exploratory RF models were constructed to contextualize regression-based findings, with feature importance assessed using permutation importance and the Gini index. These machine learning analyses were conducted as exploratory, hypothesis-generating tools to support and contextualize regression-based findings rather than to establish standalone predictive models. RESULTS: Among 136 patients, 74% had locally advanced disease; pCR was achieved in 52.9%, and radiological response in 84.6%. Higher LMR (≥2.98) was independently associated with increased odds of pCR, whereas elevated SII and CAR were associated with reduced response. For radiological response, LMR, CAR, baseline CA 15-3 levels, and intermediate Ki-67 expression (20-30%) remained independently associated with outcomes. Exploratory machine-learning analyses consistently identified inflammatory and immune-nutritional biomarkers among the most influential predictors. CONCLUSION: Routinely available systemic inflammatory and immune-nutritional biomarkers, particularly LMR, SII, and CAR, are independently associated with pathological and radiological response to neoadjuvant dual anti-HER2 therapy in HER2-positive BC. These findings support the potential role of host-related biomarkers in treatment response prediction, pending prospective validation.