Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Biological or biochemical
Reexamination Certificate
2003-04-16
2004-12-21
McElheny, Jr., Donald (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Biological or biochemical
Reexamination Certificate
active
06834237
ABSTRACT:
The present invention relates to a method of training a classification system for characterising a biological sample, a diagnostic classification system, as well as a method of characterising a condition in an animal or a human being by using parameters obtained from the sample.
BACKGROUND
A need for a fast and reliable primary diagnostic tool providing information indicative of a disease or a group of diseases has existed for years.
In U.S. Pat. No. 4,755,684 (Leiner et al.) a method for tumor diagnosis by means of serum tests is disclosed. The method includes excitation of the serum by an excitation radiation at least of a wavelength between 250 nm and 300 nm, and its fluorescence intensity is measured at predetermined emission wavelengths. From deviations of these measuring values, a conclusion may be drawn with respect to the presence of a neoplastic disease. Measurements at one or two excitation wavelengths are suggested. Up to three emission wavelengths are determined for each excitation wavelength and an intensity value is determined. Since very little information from the fluorescence spectroscopy is used the diagnosis is very rough and insecure. Only about 60% are diagnosed correctly and the diagnosis is limited to a yes or no.
In WO 96/30746 and WO 98/24369 fluorescence spectra are used to screen tissue samples in situ, wherein the tissue suspected to be dysplastic tissue is directly subjected to fluorescence spectroscopy. The methods are used to distinguish between dysplastic cervical tissue and normal cervical tissue. In O'Brien K. M. et al “Development and evaluation of spectral classification algorithms for fluorescence guided laser angioplasty”, IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, vol. 36, No. 4, April 1989, pages 424-4430, fluorescence spectroscopy is used to distinguish normal arterial tissue from atherosclerotic tissue. None of these methods allows a specific diagnosis to be made based on analysis of spectra from tissue or bodufluids not directly related to the diseased tissue.
In U.S. Pat. No. 5,734,587 a method of analyzing sample liquids by generating infrared spectra of dried samples and evaluating using a multivariate evaluation procedure is disclosed. In the evaluation procedure the samples are assigned to classes. The evaluation procedure is trained with samples of known classes to adjust the parameters of the evaluation procedures, such that samples of unknown classification can be assigned to known classes. The samples analysed are clinically relevant liquid samples, that have to be dried before generating the infrared spectra of the samples due to the nature of infrared spectra.
Most organic compounds absorb light in the visible or ultraviolet part of the electro-magnetic spectrum. Many molecules emit the absorbed excitation energy in the form of fluorescence. A fluorescence spectrum is obtained by transmitting light to the sample (excitation light) and determining the spectral distribution of the light emitted from the sample. In the case where only one fluorescent compound is present in a weakly absorbing solution, the spectral profile of the fluorescence will be invariant with respect to the excitation wavelength. Only the intensity of the fluorescence will vary with the wavelength of the excitation light in accordance with the absorption spectrum.
If more than one fluorescent compound is present in the solution the relation between excitation and emission intensities will rapidly increase to a very high level of complexity. The individual compounds will absorb differently for each excitation wavelength, the intensity and distribution of the fluorescence will vary with excitation wavelength, and reabsorption of emitted photons might occur.
When a series of fluorescence spectra using different excitation wavelengths are recorded, the spectra collected represents an emission-excitation-matrix (EEM), which can be displayed as a 3-dimensional landscape (FIG.
1
). The EEM is specific for the specific mixture of compounds and the conditions under which it is measured.
SUMMARY
It has been an object of the present invention to provide a method capable of classifying samples with unknown properties in a system not requiring any drying, enrichment, separation or concentration of the sample before determining the class, to which the sample belongs.
This has been possible by subjecting the sample to fluorescence spectroscopy or a variant thereof, whereby liquid as well as solid samples may be classified.
Thus, in a first aspect the present invention relates to a method of training a classification system for characterising a biological sample with respect to at least one condition, comprising
a) obtaining a biological sample from an animal, including a human, wherein said biological sample is selected from body fluids and/or tissue, wherein the tissue sample is not associated with said condition(s),
b) obtaining characterisation information related to each biological sample,
c) exposing the sample to excitation light within a predetermined range of wavelength,
d) determining physical parameter(s) of light emitted from the sample,
e) repeating step a) to d) until the physical parameters of all training samples have been determined,
f) optionally performing a data handling of the obtained physical parameters obtaining data variables,
g) optionally performing a multivariate data analysis of the data variables obtaining model parameters describing the variation of the data variables,
h) classifying the biological samples into at least two different classes correlated to the characterisation information, obtaining a trained classification system.
In a preferred embodiment the method comprises the steps of:
a) obtaining a biological sample from an animal, including a human, wherein said biological sample is selected from body fluids and/or tissue, wherein the tissue sample is not associated with said condition(s),
b) obtaining characterisation information related to each biological sample,
c) exposing the sample to excitation light within a predetermined range of wavelength,
d) determining physical parameter(s) of light emitted from the sample,
e) repeating step a) to d) until the physical parameters of all training samples have been determined,
f) performing a data handling of the obtained physical parameters obtaining data variables,
g) optionally performing a multivariate data analysis of the data variables obtaining model parameters describing the variation of the data variables,
h) classifying the biological samples into at least two different classes correlated to the characterisation information, obtaining a trained classification system.
In another preferred embodiment the method comprises the steps of:
a) obtaining a biological sample from an animal, including a human, wherein said biological sample is selected from body fluids and/or tissue, wherein the tissue sample is not associated with said condition(s),
b) obtaining characterisation information related to each biological sample,
c) exposing the sample to excitation light within a predetermined range of wavelength,
d) determining physical parameter(s) of light emitted from the sample,
e) repeating step a) to c) until the physical parameters of all training samples have been determined,
f) performing a data handling of the obtained physical parameters obtaining data variables,
g) performing a multivariate data analysis of the data variables obtaining model parameters describing the variation of the data variables,
h) classifying the biological samples into at least two different classes correlated to the characterisation information, obtaining a trained classification system.
In another aspect the present invention relates to a classification system for characterising a biological sample, said system comprising:
a) a sample domain for comprising a biological sample,
b) light means for exposing the sample to excitation light in the sample domain,
c) a detecting means recording the physical parameter(s) of light emitted from the sample,
d) optionally computing means for performing data h
Albrechtsen Morten
Bro-Joergensen Rasmus
Harrit Niels
Noergaard Lars
Olsen Ole Ingemann
McElheny Jr. Donald
Medicometrics APS
Merchant & Gould P.C.
LandOfFree
Method and system for classifying a biological sample does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for classifying a biological sample, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for classifying a biological sample will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3301887