Diagnosing benign and malignant lesions in breast tissue sections by using IR-microspectroscopy

Area of Science:

• Biomedical Optics
• Spectroscopy
• Cancer Research

Background:

• Infrared (IR) microspectroscopy utilizes intrinsic spectral features for contrast.
• This technique offers a novel visualization approach for biological tissues.

Purpose of the Study:

• To evaluate the potential of IR-microspectroscopy in breast cancer research.
• To develop an automated classifier for distinguishing benign and malignant breast lesions.

Main Methods:

• Acquisition of IR-microspectroscopic data from breast tissue sections using point microscopy (30-40 micrometer spot size).
• Definition and storage of four distinct breast tissue spectral classes: fibroadenoma, ductal carcinoma in situ, connective tissue, and adipose tissue.
• Application of artificial neural network (ANN) analysis for supervised pattern recognition and automated classification.

Main Results:

• ANN classifier achieved 100% accuracy in correctly classifying all spectra from fibroadenomas.
• The classifier correctly identified 93% of spectra from the in situ component of invasive ductal carcinomas.
• Successful validation using independent external datasets demonstrated the robustness of the developed classifier.

Conclusions:

• IR-microspectroscopy is a promising tool for the diagnosis of breast tissue lesions.
• The developed automated classifier shows high potential for accurate differentiation of benign and malignant breast tissues.
• This spectroscopic methodology can significantly contribute to advancements in breast cancer research and diagnostics.