Implicit and explicit prior information in near-infrared spectral imaging: accuracy, quantification and diagnostic...

Area of Science:

• Biomedical Optics
• Medical Imaging
• Spectroscopy

Background:

• Near-infrared spectroscopy (NIRS) quantifies tissue properties using measurement data and assumptions.
• Prior knowledge (tissue shape, spectral data, biophysical models) is crucial for NIRS accuracy.
• Current NIRS development often uses empirical prior information integration.

Purpose of the Study:

• To present a general framework for incorporating prior information into NIRS imaging.
• To analyze how different prior information types and integration stages impact accuracy and diagnostic value.
• To determine the optimal stage for incorporating prior information in the NIRS workflow.

Main Methods:

• Developed a general framework for NIRS imaging with prior information integration.
• Described the NIRS workflow: data acquisition, pre-processing, forward model, inversion/reconstruction, post-processing, and interpretation.
• Considered various types of prior information: tissue shape, spectral constituents, parameter limits, demographic data, and biophysical models.

Main Results:

• Prior information can be incorporated at any stage of the NIRS process.
• Integration in the inversion/reconstruction stage often addresses mathematical challenges.
• The most beneficial stage for prior information inclusion requires comprehensive analysis for maximizing diagnostic accuracy.

Conclusions:

• Prior information is critical for advancing NIRS tissue quantification and medical diagnosis.
• A systematic approach to integrating prior information across the NIRS workflow is needed.
• Further research should focus on identifying optimal prior information and integration strategies for specific NIRS applications.