Diagnosis of disease using laser-induced breakdown spectroscopy and machine learning
Diagnosing a pathology using LIBS and biological fluids includes focusing light on to a sample on a substrate to cause ablation of the sample and formation of a plasma, collecting optical emission from the plasma, and providing the optical emission to a spectroscopic acquisition component that provides information on spectral data of the plasma. The spectral data is provided to a machine learning algorithm to diagnose a pathology in the sample, where the algorithm is trained on a training set that includes spectral features in a difference spectrum derived from differences between a first LIBS optical emission spectrum collected from one or more samples of the biological fluid that have the pathology or known progress of the pathology and a second LIBS optical emission spectrum collected from one or more samples of the predetermined biological fluid that do not have the pathology or the known progress of the pathology.
1. A method for screening patients with, diagnosing, or monitoring progress of a pathology using laser induced breakdown spectroscopy (LIBS) and biological fluids, the method comprising:
focusing light from a laser light source on a sample of biological fluid deposited on a substrate, wherein an energy and pulse length of the light focused on the sample cause ablation of the sample and formation of a plasma;
collecting optical emission from the plasma;
providing the collected optical emission to a spectroscopic acquisition component, wherein the spectroscopic acquisition component provides information on spectral data of the plasma;
providing the spectral data from the collected optical emission to a processing component, wherein the processing component includes one or more processors; and
using a machine learning algorithm and the one or more processors to screen a patient with, diagnose, or monitor progress of the pathology in the sample, wherein the machine learning algorithm is trained on a training set that includes spectral features in a difference spectrum derived from differences between a first LIBS optical emission spectrum collected from one or more samples of the biological fluid that have the pathology or known progress of the pathology and a second LIBS optical emission spectrum collected from one or more samples of the biological fluid that do not have the pathology or the known progress of the pathology,
wherein using the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample includes determining polarities of spectral features in the spectral data of the plasma when compared to a model difference spectrum and classifying the sample based on the number of positive polarity spectral features in the spectral data of the plasma compared to the number of negative polarity spectral features in the spectral data of the plasma.
2. The method of claim 1 , wherein the biological fluid includes blood or blood plasma and wherein the pathology is Alzheimer's disease.
3. The method of claim 1 , wherein the machine learning algorithm includes a Quadratic Discriminant Analysis (QDA) algorithm.
4. The method of claim 1 , wherein the spectral features used to train the machine learning algorithm are weighted based on their ability to correctly classify a sample.
5. The method of claim 1 , wherein each spectral feature in the spectral data of the plasma used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample is labelled as indicating or contra-indicating the pathology or the known progress of the pathology based on its polarity in the difference spectrum.
6. The method of claim 5 , wherein the first LIBS optical emission spectrum and second LIBS optical emission spectrum are normalized for their total emission intensities.
7. The method of claim 5 , wherein spectral features used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample include only features that are absent in a spectrum of the substrate.
8. The method of claim 5 , wherein spectral features used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample include only features that are absent in a spectrum of the substrate or that, in a spectrum of the substrate, had an intensity lower than 50% of an average intensity in the LIBS optical emission spectra from multiple GWI positive samples.
9. The method of claim 1 , wherein providing the spectral data from the collected optical emission to the processing component includes generating difference spectra to enhance spectral differences between samples obtained from patients having the pathology and samples obtained from healthy controls.
10. A system for screening patients with, diagnosing, or monitoring progress of a pathology using laser induced breakdown spectroscopy (LIBS) and biological fluids, using laser induced breakdown spectroscopy (LIBS) and biomedical fluids, the system comprising:
a substrate configured to support a sample of a biological fluid;
a laser light source;
an optical subsystem configured to receive light from the laser light source and focus the received light on the sample on the substrate, wherein an energy and a pulse length of the laser light source are configured to cause ablation of the sample and substrate and formation of a plasma;
a light collection optical subsystem configured to collect optical emission from the plasma;
a spectroscopic acquisition component configured to receive collected optical emission from the light collection optical subsystem and to provide spectral data, the spectroscopic acquisition component including a spectrometer and a detector; and
a processing component including one or more processors and configured to receive the spectral data from the spectroscopic acquisition component, wherein the one or more processors are configured use a machine learning algorithm to screen a patient with, diagnose, or monitor progress of the pathology in the sample, wherein the machine learning algorithm is trained on a training set that includes spectral features in a difference spectrum derived from differences between a first LIBS optical emission spectrum collected from one or more samples of the biological fluid that have the pathology or known progress of the pathology and a second LIBS optical emission spectrum collected from one or more samples of the biological fluid that do not have the pathology or the known progress of the pathology,
wherein using the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample includes determining polarities of spectral features in the spectral data of the plasma when compared to a model difference spectrum and classifying the sample based on the number of positive polarity spectral features in the spectral data of the plasma compared to the number of negative polarity spectral features in the spectral data of the plasma.
11. The system of claim 10 , wherein the biological fluid includes blood or blood plasma and the pathology is Alzheimer's disease.
12. The system of claim 10 , wherein the machine learning algorithm includes a Quadratic Discriminant Analysis (QDA) algorithm.
13. The system of claim 10 , wherein the spectral features used to train the machine learning algorithm are weighted based on their ability to correctly classify a sample.
14. The system of claim 10 , wherein each spectral feature in the spectral data of the plasma used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample is labelled as indicating or contra-indicating the pathology or the known progress of the pathology based on its polarity in the difference spectrum.
15. The system of claim 14 , wherein the first LIBS optical emission spectrum and second LIBS optical emission spectrum are normalized for their total emission intensities.
16. The system of claim 14 , wherein spectral features used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample include only features that are absent in a spectrum of the substrate.
17. The system of claim 14 , wherein spectral features used in the machine learning algorithm and the one or more processors to screen a patient with the pathology, diagnose the pathology, or monitor progress of the pathology in the sample include only features that are absent in a spectrum of the substrate or that, in a spectrum of the substrate, had an intensity lower than 50% of an average intensity in the LIBS optical emission spectra from multiple GWI positive samples.
18. The system of claim 10 , wherein providing the spectral data from the collected optical emission to the processing component includes generating difference spectra to enhance spectral differences between samples obtained from patients having the pathology and samples obtained from healthy controls.