Discrimination of tumor from normal tissues in a mouse model of breast cancer using CARS spectroscopy combined with PC-DFA methodology

Xi Huang, Ye Yuan, Timothy A. Bielecki, Bhopal C. Mohapatra, Haitao Luan, Edibaldo Silva-Lopez, William W. West, Vimla Band, Yongfeng Lu, Hamid Band, Tian C. Zhang

Research output: Contribution to journalArticlepeer-review

9 Scopus citations

Abstract

Objective methodologies to discriminate tumor from normal tissue in biopsies and resection specimens are of great interest as complementary approaches to existing pathological diagnosis of tumors. In the present study, coherent anti-Stokes Raman scattering (CARS) spectroscopy was applied as an approach to discriminate resected tumor from normal mammary tissue in murine mammary tumor virus-Wnt-1 transgenic mouse model of breast cancer. Due to the dense CH molecular vibration in the range from 2500 to 3100 cm−1, the classification was performed by using principal component-discriminant function analysis to discriminate tumor from the normal tissue. A total of 240 training and 40 testing CARS spectra were acquired. The overall accuracy of CARS, based on cross-validation and external validation method, was 98% and 95%, respectively. The present study demonstrates a diagnostic method with a 1-s spectral acquirement rate, using a CARS spectroscopic technique. Our results suggest that CARS combined with the principal component-discriminant function analysis is a potentially useful tool for identification and classification of breast cancer tissues.

Original languageEnglish (US)
Pages (from-to)1166-1170
Number of pages5
JournalJournal of Raman Spectroscopy
Volume48
Issue number9
DOIs
StatePublished - Sep 2017

Keywords

  • CARS spectroscopy
  • PC-DFA methodology
  • Raman spectroscopy
  • tumor classification

ASJC Scopus subject areas

  • General Materials Science
  • Spectroscopy

Fingerprint

Dive into the research topics of 'Discrimination of tumor from normal tissues in a mouse model of breast cancer using CARS spectroscopy combined with PC-DFA methodology'. Together they form a unique fingerprint.

Cite this