Accurate Estimation of Methemoglobin and Oxygen Saturation in Skin Tissue Using Diffuse Reflectance Spectroscopy and...

Area of Science:

• Biomedical Optics
• Medical Diagnostics
• Machine Learning in Healthcare

Background:

• Methemoglobinemia requires accurate monitoring.
• Noninvasive methods for assessing blood oxygenation and methemoglobin are needed.
• Diffuse reflectance spectroscopy (DRS) shows potential for tissue analysis.

Purpose of the Study:

• To develop and validate a noninvasive method for accurate methemoglobin concentration estimation.
• To utilize artificial neural networks (ANNs) for analyzing diffuse reflectance spectra.
• To differentiate methemoglobin from hemoglobin using spectral data.

Main Methods:

• Simulated 66 skin tissue spectra using a four-layer model with varied oxygen saturation and methemoglobin levels.
• Employed a multifiber probe-based DRS setup in the visible and near-infrared range.
• Developed a novel artificial neural network model for spectral analysis.

Main Results:

• Achieved a mean absolute error (MAE) of 0.0392% for methemoglobin concentration.
• Achieved an MAE of 0.0273% for oxygen saturation on simulated data.
• Experimentally verified the method using DRS spectra from human subjects.

Conclusions:

• Broadband DRS combined with ANNs enables accurate, noninvasive methemoglobin estimation.
• The method can differentiate subtle changes in methemoglobin and hemoglobin levels.
• This technique offers a promising approach for clinical diagnostics.