Multimodal esophageal cancer imaging: establishing data processing techniques and assessing diagnostic sensitivity

Area of Science:

• Biomedical Optics
• Cancer Imaging
• Medical Diagnostics

Background:

• Multimodal optical imaging offers promise for early cancer detection by integrating complementary data.
• Challenges exist in optimizing multimodal combinations and computational analysis for clinical application.
• Esophageal cancer remains a significant global health challenge, necessitating improved diagnostic tools.

Purpose of the Study:

• To evaluate multimodal optical imaging for esophageal cancer detection.
• To determine the optimal combination of imaging modalities for enhanced discrimination.
• To identify suitable computational methods for analyzing high-dimensional multimodal data.

Main Methods:

• Acquisition of autofluorescence, hyperspectral, polarized light imaging (PLI), and optical coherence tomography (OCT) data from human esophageal tissue samples.
• Comparative analysis of individual modality performance in differentiating healthy from cancerous tissues.
• Investigation of data integration strategies and computational approaches for multimodal analysis.

Main Results:

• Polarized light imaging (PLI) demonstrated the highest discriminatory power among the individual modalities.
• Combining PLI with optical coherence tomography (OCT) showed potential for improved diagnostic accuracy.
• Specific computational methods were assessed for their suitability in handling complex multimodal datasets.

Conclusions:

• Polarized light imaging is a promising modality for esophageal cancer detection.
• Multimodal approaches, particularly combining PLI and OCT, may enhance diagnostic capabilities.
• Further research into computational strategies is needed to fully leverage multimodal optical imaging in oncology.