Background and purpose: Dual-target therapy may increase the incidence of adverse events and cause economic burden to patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer. It is necessary to identify the patients who could benefit greatly from a single-target neoadjuvant therapy in order to avoid overtreatment of patients.
Patients and methods: The baseline transcriptome data and clinical characteristics of patients with HER2-positive breast cancer who received neoadjuvant trastuzumab therapy were obtained from the Gene Expression Omnibus database. Least absolute shrinkage and selection operator (LASSO) regression analyses were used to construct the predictive model for pathologic complete response (pCR).
Results: A 10-gene signature model for predicting pCR rate after neoadjuvant trastuzumab therapy was constructed by LASSO regression. The areas under the receiver operating characteristics (ROC) curves in the training set and validation set were 0.896 (95% confidence interval [CI], 0.8165-0.9758) and 0.775 (95% CI, 0.5402-1), respectively. The result of logistic regression analysis showed that the risk score calculated by the 10-gene signature model was a potential predictor for pCR. Among the 10-gene signature, TFAP2B, SUSD2, AQP3, MUCL1, and ANKRD30A were found to be predictors for worse relapse-free survival (RFS) in patients with HER2-positive breast cancer, whereas MGP, YIF1B, ANKRD36BP2, and FBXO6 were found to be predictors for favorable RFS.
Conclusion: A novel 10-gene signature that could predict the response of neoadjuvant anti-HER2 therapy in patients with HER2-positive breast cancer was developed, and the risk score of the 10-gene signature could be calculated to guide the selection of anti-HER2 therapy regimens.
Keywords: Anti-HER2 therapy; Bioinformatic analysis; LASSO regression; Pathologic complete response; Relapse-free survival.
Copyright © 2021. Published by Elsevier Inc.