Characterization of Hormone Receptor and HER2 Status in Breast Cancer Using Mass Spectrometry Imaging

Abstract

Immunohistochemical evaluation of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor-2 status stratify the different subtypes of breast cancer and define the treatment course. Triple-negative breast cancer (TNBC), which does not register receptor overexpression, is often associated with worse patient prognosis. Mass spectrometry imaging transcribes the molecular content of tissue specimens without requiring additional tags or preliminary analysis of the samples, being therefore an excellent methodology for an unbiased determination of tissue constituents, in particular tumor markers. In this study, the proteomic content of 1191 human breast cancer samples was characterized by mass spectrometry imaging and the epithelial regions were employed to train and test machine-learning models to characterize the individual receptor status and to classify TNBC. The classification models presented yielded high accuracies for estrogen and progesterone receptors and over 95% accuracy for classification of TNBC. Analysis of the molecular features revealed that vimentin overexpression is associated with TNBC, supported by immunohistochemistry validation, revealing a new potential target for diagnosis and treatment.

Keywords: breast cancer; histopathology; mass spectrometry imaging; proteomics; tissue typing.

Figure 1

Comparison between the average spectra of the different classes used in the classification models: (A) ER-positive versus ER-negative; (B) PR-positive versus PR-negative; (C) HER2-positive versus HER2-negative; and (D) TNBC versus luminal. ER—estrogen receptor; PR—progesterone receptor; HER2—human epidermal growth factor receptor-2.

Figure 2

Example of the on-tissue distribution of a few molecular features obtained by ROC-AUC calculation. Histological annotation of the tumor regions (TNBC in green, non-TNBC in orange) could be correlated with the distribution intensity of the peptide fragments selected for TNBC.

Figure 3

Exemplary images of vimentin IHC and corresponding H&E: (A,D) TNBC with strong positivity within the tumor cells as well as the tumor stroma; (B,E) TNBC with intermediate positivity within the tumor cells and strong positivity within the tumor stroma; (C,F) non-TNBC with no vimentin staining within the tumor cells and strong positivity within the tumor stroma. Scale bar—200 μm.

Figure 4

Sample processing for matrix-assisted laser desorption/ionization mass spectrometry imaging: (1) The resected specimen is fixed in formalin and embedded into paraffin as per standard sample treatment for histopathological evaluation and clinical diagnosis. From the blocks, a section is cut and adhered to an indium-tin-oxide (ITO) slide. (2) The sample is then subjected to enzymatic digestion, followed by matrix application with a solution sprayer. (3) The samples are measured using a MALDI-TOF system. (4) The acquired data is subjected to pre-processing where the spectra are normalized and re-sampled. (5) The sample is stained with hematoxylin and eosin to facilitate histological annotation of the regions of interest, followed by co-registration with the mass spectrometry measurements. The data is then subjected to statistical analyses.