The success of molecular subtyping leads to numerous prognostic algorithms for personalized treatment of breast cancer patients. However, until now, immunohistochemistry (IHC)-based subtyping method still plays a dominant role in daily clinical practice. The discordance between mRNA based and protein based method is well recognized in the field. In this study, we aimed to reconcile these two types of subtyping methods by measuring ER, PR, Her2 and Ki67 as absolute and continuous variables in 1048 Formalin Fixed Paraffin Embedded (FFPE) specimens.
Total protein from 2X15 mm FFPE slices provided by local hospitals sequentially and non-selectively were extracted and used for Quantitative Dot Blot (QDB) analysis using the same antibodies for IHC analysis, including SP1 for ER, 1E2 for PR, EP3 and 4b5 for Her2, and MIB1 and UMAB107 for Ki67. Purified recombinant proteins, either prepared in house, or acquired commercially, were used as protein standard to achieve absolute quantification of these biomarkers.
When using absolute and quantitative levels of ER, PR and Her2 as X, Y, and Z axes, we found the samples naturally segregated into three distinct groups. Those samples with high Her2 levels were exclusively wrapped around Z axis (Her2 group), while samples with ER or PR overexpression were spread at the ER and PR floor (HR group). This pattern persisted until a small block with ER < 0.2 nmole/g, PR < 0.8 nmole/g and Her2 <0.3 nmole/g can be observed with samples distributed inside randomly. We named this small block as corner group. Interestingly, when using 0.3 nmole/g as cutoff for Her2, we observed 93.6% concordance between results from QDB and 786 IHC scores provided from local hospitals. The corner group was also found to be most heterogeneous, with those samples with highest Ki67 levels exclusively accumulated within this group.
We believe what we observed of the 3D distribution of samples by ER, PR and Her2 levels is the projection at protein levels of various molecular subtypes. Thus, the absolute and quantitative measurement of biomarkers at protein level may provide a simple, reliable 3D prognostic tool to incorporate what we have learned at both protein and mRNA level to be used in personalized treatment of breast cancer patients.
Yantai Quanticision Diagnostics, Inc.
Quanticision Diagnostics, Inc & Binzhou Medical University of China.
J. Zhang, J. Lv Y. Zhang, F. Tang: Employee: Yantai Quanticision Diagnostics, Inc & Quanticision Diagnostics, Inc. who owned or has filed patent application for QDB plate and QDB method to measure biomarkers in FFPE specimens. All other authors have declared no conflicts of interest.