High Definition Fourier Transform Infrared (FT-IR) spectroscopic imaging for breast histopathology

a) Classified image of the breast TMA with a four class random forest model. b) H&E stained image (to the left) and the classified image (to the right) of a Dysplasia (on the top) sample and a malignant (on the bottom) tissue sample.

High definition FTIR spectroscopic imaging enables visualization of tissue constituents, label free extraction of biochemical information and is a platform to obtain diagnostic markers for disease classification. This methodology, along with fast and efficient data analysis, is a potential tool for the development of quantitative and automated pathology. In this manuscript, we demonstrate a combination of high definition Fourier transform infrared (FTIR) spectroscopic imaging of tissue microarrays (TMAs) with data analysis algorithms. This automated histological tool is applied to two hundred biological samples representing various disease states. i.e. hyperplasia, dysplasia, malignant and normal. The information obtained with multiple immunohistochemical stains can be achieved in just one classified FT-IR spectroscopic image. Finally, we identify various cell types which could act as biomarkers for breast cancer detection and differentiate between them using statistical pattern recognition tools i.e. random forest and Bayesian algorithms. Disease prognosis can also be studied based on the quantitative distribution of these biomarkers in the tissue sample.

Guidance for Performing Multivariate Data Analysis of Bioprocessing Data: 
Pitfalls and Recommendations
Flowchart illustrating the proposed MVDA approach for analysis of bioprocessing data.
Flowchart illustrating the proposed MVDA approach for analysis of bioprocessing data.