An Explicit Estimated Baseline Model for Robust Estimation of Fluorophores Using Multiple-Wavelength Excitation Fluorescence Spectroscopy

Neurology Analytical Chemistry Chemistry Optics Medical Laboratory Science Medicine Physics Bioengineering Oncology Medical Imaging Life Sciences Neurosurgery
fluorescence spectroscopy multiple-wavelength excitation fluorophores baseline model non-linear optics neuro-oncology glioma

Research Background

Fluorescence spectroscopy is a widely used method for identifying and quantifying fluorescent substances (fluorophores). However, quantifying the fluorophores of interest becomes challenging when the material contains other fluorophores (baseline fluorophores), especially when the emission spectrum of the baseline is not well-defined and overlaps with the emission spectrum of the target fluorophores. To accurately distinguish and quantify these fluorescent substances, researchers have proposed a new method based on multi-wavelength excitation fluorescence spectroscopy. The main goal of this study is to address the issue of baseline fluorescence interference and to provide a robust estimation algorithm without prior assumptions.

Paper Source

The paper titled “An Explicit Estimated Baseline Model for Robust Estimation of Fluorophores Using Multiple-Wavelength Excitation Fluorescence Spectroscopy” is co-authored by A. Gautheron, M. Sdika, M. Hébert, and B. Montcel. The authors are affiliated with Université de Lyon, INSA-Lyon, Université Claude Bernard Lyon 1, CNRS, and Université Jean Monnet Saint-Etienne. This paper was published in the January 2024 issue of the IEEE Transactions on Biomedical Engineering journal.

Research Process

a) Detailed Research Process

The research consists of several key steps. Firstly, it focuses on estimating the fluorescence signal of two forms of protoporphyrin IX (PPIX), a fluorophore used to distinguish healthy tissue from tumor tissue during neurosurgery. Fluorescence signals excited by multiple wavelengths are calibrated through digital simulation and validated with clinical and experimental data. The specific steps are as follows:

  1. Model Establishment: Propose a basic model without prior assumptions, acquiring fluorescence signals through multiple excitation wavelengths. Although the model is nonlinear, the authors derived a closed-form solution for the least squares estimation.

  2. Experimental Design: Use PPIX as the fluorophore to estimate its contribution in the fluorescence signal. Focus particularly on the identification of brain tumor boundaries during neurosurgery.

  3. Data Simulation: Use a digital simulation model calibrated with clinical and experimental data to verify the accuracy and robustness of the new method.

  4. Parameter Estimation: Estimate the baseline fluorescence signal through measurements at multiple excitation wavelengths and derive an analytical expression for the algorithm.

  5. Result Analysis: Compare the performance of the new method with existing methods, focusing on the accuracy of PPIX contribution estimation and the classification of healthy and tumor tissues.

b) Main Results

Through digital simulation and clinical data validation, the new method outperforms existing methods in high-variation baseline conditions, achieving an accuracy of 87% in distinguishing healthy from tumor tissues, while existing methods have an accuracy close to 0. The specific results are as follows:

  1. Baseline Estimation: Without prior models, the new method successfully estimated the baseline fluorescence signal under different excitation wavelengths, addressing the overlap of baseline spectral bands.

  2. Classification Accuracy: Tested with digital simulation models, the new method shows high classification accuracy under various baseline variation conditions, especially when the baseline overlaps with the target fluorophore’s spectral bands.

  3. Computational Efficiency: Compared with existing blind source separation methods, the new method avoids iterative solutions and achieves efficient calculations with only analytical expressions.

c) Conclusion and Significance

The new method provides a robust and assumption-free fluorophore estimation method, significantly improving the classification accuracy of healthy and tumor tissues. Its scientific value lies in the absence of complex baseline modeling while maintaining high computational efficiency. It has substantial application value, particularly in tumor boundary identification, providing more accurate information for clinical surgery.

d) Research Highlights

  1. High Accuracy: In simulated clinical environments, the new method shows an accuracy of up to 87% in classification tasks.
  2. No Prior Assumptions: The new method does not require assumptions about the shape of the baseline spectrum, providing high adaptability.
  3. High Computational Efficiency: The derived analytical expressions significantly improve computational efficiency, suitable for real-time clinical applications.

e) Other Valuable Information

From the perspective of clinical practical application, the new method can provide more accurate fluorescence information in medium and low-density tumor tissue areas, effectively supporting tumor resection surgeries. Further research can explore the application of the method with more than two excitation wavelengths and its application to other clinically interesting fluorophores, such as exploring the relative amounts of protein-bound and free forms of NADH.

Summary

This study proposes a new method based on multi-wavelength excitation, effectively solving the problem of fluorescence spectroscopy baseline interference. It improves the classification accuracy and robustness of tumor boundary identification, presenting significant scientific value and clinical application potential.