Infrared micro-spectroscopy coupled with multivariate and machine learning techniques for cancer classification in...

Area of Science:

• Biomedical Engineering
• Spectroscopy
• Computational Pathology

Background:

• Cancer diagnosis relies on visual tissue analysis, which is challenging and time-consuming.
• Computational methods are increasingly used to aid histological cancer diagnoses.
• Vibrational spectroscopy, particularly Infrared (IR) spectroscopy, analyzes tissue chemical composition.

Purpose of the Study:

• To review and highlight the potential of infrared microscopy techniques for improving cancer diagnostic accuracy.
• To discuss the application of Fourier Transform Infrared Spectroscopy (FTIR) and Quantum Cascade Laser Infrared Spectroscopy (QCL) for earlier detection of human neoplastic disease.
• To provide an overview of cancer tissue detection and classification using FTIR spectroscopy with multivariate and Machine Learning (ML) techniques.

Main Methods:

• Utilizing vibrational patterns from IR irradiation to detect and grade cancerous tissues.
• Applying multivariate and Machine Learning (ML) techniques for data analysis.
• Employing Fourier Transform Infrared Spectroscopy (FTIR) and Quantum Cascade Laser Infrared Spectroscopy (QCL).

Main Results:

• Infrared spectroscopy, combined with ML, shows potential for detecting and classifying cancerous tissues.
• The F1-Score is used as a metric for comparing the performance of different models.
• Data handling techniques and pre-processing protocols are discussed for future studies.

Conclusions:

• Infrared microscopy techniques offer a promising avenue for enhancing diagnostic accuracy in oncology.
• These methods can lead to earlier detection of human neoplastic diseases.
• Standardized reporting and pre-processing protocols are suggested for future research in this field.