Spectroscopic system and method for predicting outcome of disease
A system and method to predict the progression of disease of a test sample. A group of known biological samples is provided. Each known biological sample has an associated known outcome including a non-diseased sample or a diseased sample. A Raman data set is obtained for each known biological sample. Each Raman data set is analyzed to identify a diseased or non-diseased reference Raman data set depending on whether respective biological sample is the non-diseased sample or the diseased sample. A first database is generated where the first database contains reference Raman data sets for all diseased samples. A second database is generated where the second database contains reference Raman data sets for all non-diseased samples. A test Raman data set of a test biological sample is received, where the test biological sample has an unknown disease status. A diagnostic is provided as to whether the test sample is a non-diseased sample or a diseased sample. The diagnostic is obtained by comparing the test Raman data set against the reference Raman data sets in the first and the second databases using a chemometric technique. A prediction of the progression of disease may be then provided.
1. A method comprising:
providing a database containing a plurality of reference Raman data sets wherein each said reference Raman data set has an associated known biological sample and an associated known metabolic state and at least one of the following associated with it: a corresponding reference Raman image, and a corresponding reference non-Raman image;
using said reference Raman image, identifying one or more regions of interest of corresponding known biological sample, wherein said one or more regions of interest contain at least one of the following: an epithelium tissue, a stroma tissue, and a nuclei tissue
irradiating a test biological sample with substantially monochromatic light to thereby generate scattered photons;
collecting a test Raman data set based on said scattered photons; and
comparing said test Raman data set to said plurality of reference Raman data sets using a chemometric technique; and based on said comparing, predicting a metabolic state of the test biological sample.
2. The method of claim 1 , wherein one of said reference Raman data sets comprises: a plurality of reference Raman spectra obtained from said one or more regions of interest of said known biological sample.
3. A method comprising:
providing a database containing a plurality of reference Raman data sets wherein each said reference Raman data set has an associated known biological sample and an associated known metabolic state and at least one of the following associated with it: a corresponding reference Raman image, and a corresponding reference non-Raman image, wherein said reference non-Raman image is at least one of the following: a brightfield image;
a polarized light image; and a UV-induced autofluorescence image;
irradiating a test biological sample with substantially monochromatic light to thereby generate scattered photons;
collecting a test Raman data set based on said scattered photons; and
comparing said test Raman data set to said plurality of reference Raman data sets using a chemometric technique; and based on said comparing, predicting a metabolic state of the test biological sample.
4. A method comprising:
providing a database containing a plurality of reference Raman data sets, each reference Raman data set having an associated known biological sample and an associated known metabolic state;
irradiating a test biological sample with substantially monochromatic light to thereby generate scattered photons;
collecting a test Raman data set based on said scattered photons, wherein said test Raman data set comprises at least one of the following associated therewith: a corresponding test Raman image and a corresponding test non-Raman image;
using said test Raman image, identifying one or more regions of interest of said test biological sample, wherein said one or more regions of interest contain at least one of the following: an epithelium tissue, a stroma tissue, and a nuclei tissue of said test biological sample;
comparing said test Raman data set to said plurality of reference Raman data sets using a chemometric technique; and
based on said comparing, predicting a metabolic state of the test biological sample.
5. The method of claim 4 , wherein said test Raman data set comprises: a plurality of test Raman spectra obtained from said one or more regions of interest of said test biological sample.
6. A method comprising:
providing a database containing a plurality of reference Raman data sets, each reference Raman data set having an associated known biological sample and an associated known metabolic state;
irradiating a test biological sample with substantially monochromatic light to thereby generate scattered photons;
collecting a test Raman data set based on said scattered photons, wherein each test Raman data set has at least one of the following associated therewith: a corresponding test Raman image, and a corresponding test non-Raman image, wherein said test non-Raman image is at least one of the following: a brightfield image; a polarized light image; and a UV-induced autofluorescence image;
comparing said test Raman data set to said plurality of reference Raman data sets using a chemometric technique; and
based on said comparing, predicating a metabolic state of the test biological sample.
7. A method further comprising:
providing a database containing a plurality of reference Raman data sets, each reference Raman data set having an associated known biological sample and an associated known metabolic state;
irradiating a test biological sample with substantially monochromatic light to thereby generate scattered photons;
collecting a test Raman data set based on said scattered photons;
selecting a pre-determined vector space that mathematically describes said plurality of reference Raman data sets;
transforming the test Raman data set into said pre-determined vector space; and analyzing a distribution of transformed data in the pre-determined vector space;
comparing said test Raman data set to said plurality of reference Raman data sets using a chemometric technique; and
based on said comparing, predicting a metabolic state of the test biological sample.
8. The method of claim 7 , wherein said analyzing is performed by using a classified scheme.
9. The method of claim 8 , wherein said classification scheme is at least one of the following: Mahalanobis distance, Adaptive subspace detector, Band target entropy method, Neural network, and support vector machines.
10. The method of claim 9 , wherein the classification scheme is Mahalanobis distance, and wherein said method further comprising: calculating a Mahalanobis distance, and wherein said method further comprising: calculating a Mahalanobis distance between the test Raman data set transformed into said vector space and the plurality of reference Raman data sets in said pre-determined vector space.