Digital de-waxing on FTIR images.

Area of Science:

• Biomedical Optics
• Spectroscopy
• Computational Pathology

Background:

• Paraffin embedding is standard for histological analysis but introduces spectral artifacts in FTIR hyperspectral imaging.
• These paraffin-related spectral signatures can obscure true biological information, hindering accurate data interpretation.
• Developing methods to digitally remove paraffin is crucial for advancing quantitative FTIR-based tissue analysis.

Purpose of the Study:

• To present a robust digital procedure for neutralizing paraffin's contribution in FTIR hyperspectral images.
• To provide guidance on parameter selection for reproducible de-waxing.
• To validate the effectiveness of the de-waxing procedure on diverse human colon biopsy samples.

Main Methods:

• A paraffin model was constructed using Principal Component Analysis (PCA).
• Extended Multiplicative Signal Correction (EMSC) normalization was applied.
• Two spectral quality control techniques were implemented.
• The procedure was tested on 71 hyperspectral images from 55 human colon biopsies.

Main Results:

• The digital de-waxing procedure effectively neutralized paraffin spectral contributions.
• De-waxed FTIR hyperspectral images were suitable for pattern recognition.
• Analysis of normal, ulcerative colitis, and adenocarcinoma samples demonstrated successful paraffin removal.

Conclusions:

• The proposed digital de-waxing method significantly improves FTIR hyperspectral image quality for biomedical research.
• This technique enables reliable application of pattern recognition algorithms like k-means clustering and PCA-LDA on de-waxed colon tissue data.
• The procedure offers a standardized approach to mitigate paraffin artifacts in hyperspectral imaging of biological tissues.