Notch gene signature for neoadjuvant chemotherapy response in triple-negative breast cancer
Notch-based gene signature for predicting the response to neoadjuvant chemotherapy in triple-negative breast cancer
Background: Neoadjuvant chemotherapy (NACT) is widely accepted as a treatment for triple-negative breast cancer (TNBC), but not all patients benefit from it. There are currently no validated biomarkers to predict the response to NACT, and previous attempts to create predictive classifiers using gene expression data have lacked clinical relevance. However, models that integrate biological constraints have shown improved robustness and performance compared to agnostic classifiers.
Methods: We utilized preoperative transcriptomic profiles from 298 TNBC patients to train and evaluate a rank-based classifier, k-top scoring pairs, designed to predict whether a patient will achieve pathological complete response (pCR) or have residual disease (RD) after NACT. To minimize overfitting and enhance the signature’s interpretability, we focused on genes involved in the Notch signaling pathway during the training process. The model was then tested on two independent patient cohorts, comprising 75 and 71 patients, respectively. Additionally, we examined the signature’s prognostic value by investigating its correlation with relapse-free survival (RFS) through Kaplan‒Meier (KM) survival estimates and a multivariate Cox proportional hazards model.
Results: The final signature, consisting of five gene pairs, can predict NACT response based on their relative ordering. The model demonstrated robust performance in predicting pCR in TNBC patients, with an area under the ROC curve (AUC) of 0.76 and 0.85 in the first and second test cohorts, respectively. It outperformed other gene signatures developed for similar purposes. Furthermore, the signature was significantly associated with RFS in an independent TNBC cohort, even after adjusting for factors such as T stage, patient age at diagnosis, type of surgery, and menopausal status.
Conclusion: We present a robust gene signature that predicts pathological complete response (pCR) in TNBC patients. This signature uses easily interpretable, rank-based decision rules focused on genes regulated by the Notch signaling pathway, which plays a key role in breast cancer chemoresistance. The signature’s strong predictive and prognostic capabilities make it a promising tool for clinical use, assisting in the stratification of TNBC patients undergoing NACT.
