Employing Raman Spectroscopy and Machine Learning for the Identification of Breast Cancer

Area of Science:

• Biomedical Engineering
• Spectroscopy
• Machine Learning

Background:

• Breast cancer is a major global health concern, necessitating accurate intraoperative tissue identification.
• Distinguishing cancerous from non-cancerous mammary tissue is vital for effective tumor removal during surgery.

Purpose of the Study:

• To evaluate machine learning algorithms for classifying breast cancer tissues using Raman spectroscopy.
• To compare the efficacy of Random Forest, Support Vector Machine, and Convolutional Neural Network models in this classification task.

Main Methods:

• Raman spectroscopy was employed to analyze normal and cancerous murine mammary tissues.
• Three machine learning models were developed and tested: Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN).
• Specific spectral peaks correlating with tissue types were identified.

Main Results:

• The study achieved high classification accuracies: 94.47% (RF), 96.76% (SVM), and 97.58% (CNN).
• This research represents the first comparison of these three machine learning techniques for breast cancer classification using Raman spectra.
• Key spectral biomarkers distinguishing cancerous from normal tissues were identified.

Conclusions:

• The integration of machine learning and Raman spectroscopy provides a rapid, non-invasive diagnostic method for breast cancer.
• This approach shows significant potential for improving intraoperative surgical guidance and patient outcomes.