Physiological model using diffuse reflectance spectroscopy for nonmelanoma skin cancer diagnosis

Area of Science:

• Biomedical Optics
• Dermatology
• Medical Imaging

Background:

• Nonmelanoma skin cancer diagnosis relies on accurate tissue characterization.
• Diffuse reflectance spectroscopy (DRS) offers a noninvasive method for optical tissue analysis.
• Existing diagnostic methods may lack the specificity for early-stage detection.

Purpose of the Study:

• To evaluate a physiological computational model for diagnosing nonmelanoma skin cancer using DRS.
• To extract physiologically relevant parameters from skin tissue using DRS.
• To assess the diagnostic performance of extracted parameters for classifying skin lesions.

Main Methods:

• Applied a Monte Carlo lookup table inverse model to a clinical DRS dataset.
• Extracted scattering parameters, blood volume fraction, oxygen saturation, and vessel radius.
• Utilized extracted parameters to train classifiers for distinguishing between cancer types and normal tissue.

Main Results:

• The physiological model successfully captured relevant skin cancer information.
• Classification performance using extracted physiological parameters was comparable to Principal Component Analysis (PCA).
• DRS revealed distinct physiological characteristics of cancerous versus normal skin.

Conclusions:

• DRS, coupled with a physiological model, can provide valuable insights into skin tissue optics.
• This approach offers a flexible alternative to purely statistical analysis for skin cancer diagnosis.
• The extracted physiological parameters show promise for improving nonmelanoma skin cancer detection.