Nomograms for predicting survival in patients with HER2 negative breast cancer.

Abstract

e12528 Background: It is increasingly recognized that there is considerable heterogeneity in HER2-negative breast cancer. The clinicopathological characteristics differed between HER2-low and HER2-0 breast cancer evaluated by immunohistochemistry (IHC) or in situ hybridization (ISH). HER2-low breast cancer may benefit from novel antibody-drug conjugates. However, the immunohistochemical agreement of HER2-low is poor. We aimed to construct nomograms to predict the survival of HER2-negative breast cancer and to evaluate the effects of HER2 on survival at the RNA transcript level. Methods: A retrospective analysis was performed on breast cancer patients who were evaluated HER2-negative by IHC or ISH in the TCGA database. Univariate and multivariate cox regression analysis was used to evaluate whether the HER2 RNA transcript level and clinical parameters were independent predictors for overall survival (OS), progression-free survival (PFS), and Disease-Specific Survival (DSS), and nomograms were generated for the prediction of 1-, 2-, 3-, 5-year OS, PFS, and DSS based on significant parameters. Survival curves were drawn using the Kaplan-Meier methods to compare mortality outcomes according to HER2 RNA transcript level (75% vs. 25%). Results: A total of 769 HER2-negative patients were included in the analysis. The median age of patients was 56 years (range 26-30), and the majority were White (80.3%). Univariate and multivariate cox regression analysis showed that age, pN_stage, and pM_stage were independent predictors for OS, but HER2 was not. The C-index of a nomogram based on these three parameters was 0.723 (95%CI, 0.6555-1). HER2, pT_stage, pN_stage, and pM_stage were analyzed as independent predictors of PFS. HER2, pN_stage, and pM_stage were independent predictors of DSS. The C-indexes of established nomograms were 0.734 (95%CI, 0.672-1) and 0.775 (95%CI, 0.686-1) respectively. PFS and DSS were significantly better for patients with high HER2 RNA transcript levels (PFS, log-rank p = 0.00446, HR = 0.352; DSS, log-rank p = 0.0119, HR = 0.247). Conclusions: In the TCGA database, the HER2 RNA transcript level was an independent predictor for PFS and DSS in HER2-negative breast cancer. Stratification of patients at the RNA level can help clinicians optimize treatment for HER2-negative breast cancer.