Development of a nomogram for predicting axillary pathologic complete response after neoadjuvant therapy in clinically node-positive breast cancer

Abstract

Background Reliable and accessible tools for estimating axillary pathologic complete response (ax-pCR) after neoadjuvant therapy (NAT) remain limited. This study evaluated whether routinely available biomarkers measured before and after NAT provide complementary information for ax-pCR prediction. Methods We conducted a retrospective study of patients with clinically node-positive, non-metastatic breast cancer who underwent NAT followed by surgery. Clinicopathological characteristics, treatment information, serum albumin, and inflammatory indices measured before and after NAT were evaluated. Biomarker cutoffs were selected using the maximum Youden index. Variables with P<0.05 in univariable analyses were entered into multivariable logistic regression. A nomogram was constructed from variables that remained significant in the initial multivariable model. Discrimination, calibration, clinical net benefit and optimism-corrected performance were assessed using bootstrap resampling and decision-curve analysis. Results Among 158 patients, ax-pCR occurred in 74 (46.8%). In the initial multivariable model, HER2 positivity and pre-NAT albumin ≥43.95 g/L were associated with higher odds of ax-pCR, whereas clinical stage III and post-NAT NLR ≥4.58 were associated with lower odds. The four-variable nomogram achieved an apparent AUC of 0.782 (95% CI, 0.708–0.851). At the maximum-Youden-index threshold, sensitivity was 74.3% (95% CI, 63.3%–82.9%) and specificity was 73.8% (95% CI, 63.5%–82.0%). After bootstrap correction, the AUC was 0.766, the calibration slope was 0.908 and the Brier score was 0.200. Decision-curve analysis suggested potential net benefit across a range of threshold probabilities. Conclusion Clinical characteristics and blood-based biomarkers measured before and after NAT provided complementary information for ax-pCR estimation. The nomogram showed moderate discrimination and reasonable internal calibration. External validation is required before clinical application.