Development and validation of a novel HER2RI assay for predicting the risk of recurrence and survival in HER2-positive early breast cancer

Abstract

Abstract De-escalated or escalated systemic therapy remains to be a controversial topic for HER2-positive early breast cancer. Here we presented a novel assay for predicting the relapse risk and survival in HER2-positive early breast cancer. The prognostic models based on the 20-gene signature (HER2RI model) and 20-gene signature combined with nodal stage (HER2RIclin model) were built using random forest, respectively. Totally 254 patients were used for training the models, and two independent cohorts for validating the accuracy. The HER2RI presented a better predictive accuracy than the HER2RIclin. Compared with high-risk patients, the low-risk patients stratified by the HER2RI showed a better 7-year relapse-free survival (RFS) rate. Multivariate analysis further confirmed the high risk stratified by the HER2RI was a significant risk factor for RFS. Overall, the HER2RI assay is predictive of the relapse risk in HER2-positive early breast cancer and can help the patient risk stratification to aid the treatment decisions.