Molecular Profiling Thyroid Cancer Using 2T2D-FTIR Spectroscopy Integrated With Machine Learning Models.

Area of Science:

• Biomedical Engineering
• Molecular Spectroscopy
• Oncology

Background:

• Thyroid cancer is the most common endocrine malignancy, necessitating advanced diagnostic tools.
• Two-trace two-dimensional Fourier transform infrared spectroscopy (2T2D-FTIR) offers detailed molecular insights for cancer research.
• Distinguishing tumor types and stages is crucial for effective thyroid cancer management.

Purpose of the Study:

• To investigate the efficacy of 2T2D-FTIR combined with machine learning (ML) for diagnosing thyroid cancer across diverse population groups.
• To explore the capability of 2T2D-FTIR in detecting subtle molecular alterations between healthy and malignant thyroid tissues.
• To assess the potential of this integrated approach for early cancer detection.

Main Methods:

• Utilized 2T2D-FTIR spectroscopy to analyze tissue samples.
• Employed machine learning algorithms for data analysis and classification.
• Stratified the study population into three age and sex-defined groups (G1, G2, G3).

Main Results:

• The combination of 2T2D-FTIR and ML demonstrated effectiveness in diagnosing thyroid cancer.
• The technique successfully detected molecular changes differentiating healthy from malignant tissues in a two-dimensional spectral space.
• Despite a limited sample size, the ML approach enhanced diagnostic accuracy.

Conclusions:

• 2T2D-FTIR coupled with ML shows significant promise for early thyroid cancer detection.
• This methodology could potentially be extended to the diagnosis of other cancer types.
• Larger datasets are expected to further improve the performance of the ML-based diagnostic model.