Page 23 - Vibrational Spectroscopic Imaging for Biomedical Applications

CHAPTER 1

Toward Automated

Breast Histopathology

Using Mid-IR

Spectroscopic Imaging

F. Nell Pounder and R. Bhargava

Department of Bioengineering and Beckman Institute for
Advanced Science and Technology
The University of Illinois at Urbana-Champaign
Illinois, USA

ecent technological developments in Fourier transform infra-
red (FT-IR) spectroscopic imaging, high-throughput tissue
Rmicroarray (TMA) sampling, and multivariate data analysis
have greatly accelerated efforts toward automated and reproducible
cancer diagnosis. While several studies indicate the potential of tis-
sue analysis by FT-IR imaging for clinical applications, vigorously
validated protocols with rapid data acquisition and highly accurate
classification are needed. Here, we report progress toward that goal
by the development of a protocol for breast cancer histopathology.
We first employ FT-IR imaging to acquire data from human breast
TMAs. This TMA sampling permits rapid acquisition of spectral
images from large sets of patients to select potentially useful spectral
and spatial features, termed metrics, for subsequent classification.
These metrics are applied to develop a robust classification system
and results are extensively validated. This multivariate segmenta-
tion approach, based on a modified Bayesian classifier, demonstrates

1