Abstract P6-10-30: Developing 3D subtyping model for breast cancer patients based on absolute quantitation or ER, PR and Her2 at protein level

Abstract

Abstract The concept of intrinsic subtypes of breast cancer have been well accepted with several predictor applications based on this concept achieved success in clinical practice. However, all these predictor applications are based on gene expression profiling, and the surrogate assay at protein level is based on immunohistochemistry (IHC), which is known to be associated with inherent problems of subjectivity and inconsistency. Consequently, subtyping of breast cancer patients in clinical practice is limited either by the cost or the inaccuracy (surrogate assay). In this study, we proposed the first predictor model at protein level based on the three-dimensional (3D) distribution of samples using ER, PR and Her2 protein levels as coordinates. Method: Quantitative dot blot (QDB) method was used as a universal platform to achieve absolute quantitative of ER, PR, Her2 and Ki67 protein levels using total tissue lysates prepared from 2 × 15 um Formalin Fixed Paraffin Embedded (FFPE) slices. Total of 1248 FFPE slices were provided from local hospitals sequentially and non-selectively. The measured tissue biomarker levels were plotted in 3D scatterplot for analysis, in combination with provided clinicopathological parameters and overall survival data. Results: We observed natural segregation of these samples into three distinct groups to mimic “balls falling from ceiling corner” distribution pattern. This distribution pattern disappeared when the scatterplot was narrowed in a small box of ER<0.2 nmole/g, PR <0.8 nmole/g and Her2<0.3 nmole/g. When the observed Her2 level at 0.3 nmole/g was used as cutoff to convert samples dichotomously, we achieved overall concordance with IHC at 94.8%. In addition, 260 out 271 Her2 positive samples were either with ER<0.2 nmole/g or PR<0.8 nmole/g or both, with 8 more samples at the vicinity of corner group. The model was further evaluated with both the overall recurrence and survival of the patients. Conclusion: We developed the first 3D predictor model for subtyping of breast cancer patients based on the absolute and quantitative levels of ER, PR and Her2 at protein level. We propose this model the projection of intrinsic types at protein level. The adoption of this predictor model in clinical practice should significantly lower the barrier to allow practicing subtyping of breast cancer patients in daily clinical practice worldwide. Citation Format: Jiandi Zhang. Developing 3D subtyping model for breast cancer patients based on absolute quantitation or ER, PR and Her2 at protein level [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P6-10-30.