Method for probabilistically classifying tissue in vitro and in vivo using fluorescence spectroscopy

• US 7,236,815 B2
• Filed: 10/17/2003
• Issued: 06/26/2007
• Est. Priority Date: 03/14/1995
• Status: Expired due to Fees

• Current Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Original Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Inventors: Mahadevan-Jansen, Anita, Richards-Kortum, Rebecca, Follen, Michele, Ramanujam, Nirmala, Utzinger, Urs
• Primary Examiner(s): Smith; Ruth S.
• Application Number:
  US10/688,152
• Publication Number:
  US 20040162489A1
• Time in Patent Office: 1,348 Days
• Field of Search: 600/407, 600/473, 600475-478, 436/63, 436/64, 436/813, 356/301, 356/432, 250/461.2, 250/459.1
• US Class Current: 600/407
• CPC Class Codes: A61B 5/0071 by measuring fluorescence emission
• CPC Class Codes: A61B 5/0075 by spectroscopy, i.e. measuring spectra, e.g. Raman spectroscopy, infrared absorption spectroscopy A61B5/0071 takes precedence
• CPC Class Codes: A61B 5/0084 for introduction into the body, e.g. by catheters A61B5/1459 takes precedence
• CPC Class Codes: A61B 5/4331 of the cervix A61B5/435 takes precedence
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G01N 2021/6417 Spectrofluorimetric devices
• CPC Class Codes: G01N 2021/6419 Excitation at two or more wavelengths
• CPC Class Codes: G01N 2021/6484 Optical fibres
• CPC Class Codes: G01N 21/274 Calibration, base line adjustment, drift correction
• CPC Class Codes: G01N 21/6486 Measuring fluorescence of biological material, e.g. DNA, RNA, cells G01N21/6428 takes precedence
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: Y10S 436/813 Cancer

Abstract

Fluorescence spectral data acquired from tissues in vivo or in vitro is processed in accordance with a multivariate statistical method to achieve the ability to probabilistically classify tissue in a diagnostically useful manner, such as by histopathological classification. The apparatus includes a controllable illumination device for emitting electromagnetic radiation selected to cause tissue to produce a fluorescence intensity spectrum. Also included are an optical system for applying the plurality of radiation wavelengths to a tissue sample, and a fluorescence intensity spectrum detecting device for detecting an intensity of fluorescence spectra emitted by the sample as a result of illumination by the controllable illumination device. The system also include a data processor, connected to the detecting device, for analyzing detected fluorescence spectra to calculate a probability that the sample belongs in a particular classification. The data processor analyzes the detected fluorescence spectra using a multivariate statistical method. The five primary steps involved in the multivariate statistical method are (i) preprocessing of spectral data from each patient to account for inter-patient variation, (ii) partitioning of the preprocessed spectral data from all patients into calibration and prediction sets, (iii) dimension reduction of the preprocessed spectra in the calibration set using principal component analysis, (iv) selection of the diagnostically most useful principal components using a two-sided unpaired student'"'"'s t-test and (v) development of an optimal classification scheme based on logistic discrimination using the diagnostically useful principal component scores of the calibration set as inputs.

Claims

1. A method of probabilistically classifying a sample of tissue of a mammalian anatomical structure, tissues of which may have various morphological and biochemical states and are classifiable in accordance therewith, comprising:

illuminating the tissue sample with electromagnetic radiation of a first wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a first classification thereof;

detecting a first fluorescence intensity spectrum from the tissue sample resulting from the first wavelength illuminating step; and

calculating a first probability that the tissue sample belongs in the first classification from a data set comprising the first fluorescence intensity spectrum.

2. The method of claim 1, wherein the first wavelength is within one of the ranges of 317-357 nm, 360-400 nm and 440-480 nm.

3. The method of claim 1 wherein the first fluorescence intensity spectrum comprises emission wavelengths of about 410 nm, about 460 nm, about 510 nm and about 580 nm when the first wavelength is about 337 nm;

about 460 nm, about 510 nm, about 580 nm, about 600 nm and about 640 nm when the first wavelength is about 380 nm; and

about 510, about 580 nm, about 600 nm, about 620 nm, about 640 nm and about 660 nm when the first wavelength is about 460 nm.

4. A method as in claim 1, further comprising:

illuminating the tissue sample with electromagnetic radiation of a second wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a second classification thereof;

detecting a second fluorescence intensity spectrum from the tissue sample resulting from the second wavelength illuminating step;

calculating a second probability that the tissue sample belongs in the second classification from a data set comprising the second fluorescence intensity spectrum; and

classifying the tissue sample in the second classification if the first and second probabilities exceed respective thresholds.

5. The method of claim 4 wherein the first probability calculating step comprises:

providing a statistically significant first plurality of tissue samples, at least some of which are tissues known to belong to the first classification;

illuminating the first plurality of tissue samples with the first wavelength electromagnetic radiation;

detecting a first plurality of fluorescence intensity spectra from the first plurality of tissue samples;

generating first vectors that account for variation in the first plurality of fluorescence intensity spectra and that are indicative of the first classification;

calculating the first probability from the data set comprising the first fluorescence intensity spectrum, with use of the first vectors;

and wherein the second probability calculating step comprises;

providing a statistically significant second plurality of tissue samples, at least some of which are tissues known to belong to the second classification;

illuminating the second plurality of tissue samples with the second wavelength electromagnetic radiation;

detecting a second plurality of fluorescence intensity spectra from the second plurality of tissue samples;

generating second vectors that account for variation in the second plurality of fluorescence intensity spectra and that are indicative of the second classification; and

calculating the second probability from the data set comprising the second fluorescence intensity spectrum, with use of the second vectors.

6. A method as in claim 1, further comprising:

illuminating the tissue sample with electromagnetic radiation of a second wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a second classification thereof;

detecting a second fluorescence intensity spectrum from the tissue sample resulting from the second wavelength illuminating step;

calculating a second probability that the tissue sample belongs in the second classification from a data set comprising the second fluorescence intensity spectrum;

illuminating the tissue sample with electromagnetic radiation of a third wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a third classification thereof;

detecting a third fluorescence intensity spectrum from the tissue sample resulting from the third wavelength illuminating step;

calculating a third probability that the tissue sample belongs in the third classification from a data set comprising the third fluorescence intensity spectrum; and

classifying the tissue sample in the second classification if the third, first and second probabilities exceed respective thresholds.

7. A method as in claim 6, wherein:

the third classification is SIL as distinguished from normal squamous, and the wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of the third classification thereof is selected for cervical tissues from 337 nm and 460 nm;

the first classification is SIL as distinguished from normal columnar and inflammation, and the wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of the first classification thereof is 380 nm for cervical tissues; and

the second classification is high grade SIL as distinguished from low grade SIL, and the wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of the first classification thereof is 460 nm for cervical tissues.

8. The method of claim 6 wherein the first probability calculating step comprises:

providing a statistically significant first plurality of tissue samples, at least some of which are tissues known to belong to the first classification;

illuminating the first plurality of tissue samples with the first wavelength electromagnetic radiation;

detecting a first plurality of fluorescence intensity spectra from the first plurality of tissue samples;

generating first vectors that account for variation in the first plurality of fluorescence intensity spectra and that are indicative of the first classification;

calculating the first probability from the data set comprising the first fluorescence intensity spectrum, with use of the first vectors;

wherein the second probability calculating step comprises;

providing a statistically significant second plurality of tissue samples, at least some of which are tissues known to belong to the second classification;

illuminating the second plurality of tissue samples with the second wavelength electromagnetic radiation;

detecting a second plurality of fluorescence intensity spectra from the second plurality of tissue samples;

generating second vectors that account for variation in the second plurality of fluorescence intensity spectra and that are indicative of the second classification; and

calculating the second probability from the data set comprising the second fluorescence intensity spectrum, with use of the second vectors;

and wherein the third probability calculating step comprises;

providing a statistically significant third plurality of tissue samples, at least some of which are tissues known to belong to the third classification;

illuminating the third plurality of tissue samples with the third wavelength electromagnetic radiation;

detecting a third plurality of fluorescence intensity spectra from the third plurality of tissue samples;

generating third vectors that account for variation in the second plurality of fluorescence intensity spectra and that are indicative of the second classification; and

calculating the third probability from the data set comprising the third fluorescence intensity spectrum, with use of the third vectors.

9. A method as in claim 1 further comprising:

illuminating the tissue sample with electromagnetic radiation of a second wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a first classification thereof; and

detecting a second fluorescence intensity spectrum from the tissue sample resulting from the second wavelength illuminating step;

wherein the calculating step comprises calculating the first probability from a data set comprising the first and second fluorescence intensity spectrum.

10. A method as in claim 9 further comprising:

illuminating the tissue sample with electromagnetic radiation of a third wavelength selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a first classification thereof; and

detecting a third fluorescence intensity spectrum from the tissue sample resulting from the second wavelength illuminating step;

wherein the calculating step comprises calculating the first probability from a data set comprising the first, second and third fluorescence intensity spectrum.

11. A method as in claim 10 wherein the electromagnetic radiation of the first, second and third wavelengths further is selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a second classification thereof, further comprising calculating a second probability that the tissue sample belongs in the second classification from a data set comprising the first, second and third fluorescence intensity spectrum.

12. A method as in claim 11 wherein the electromagnetic radiation of the first, second and third wavelengths further is selected to stimulate in tissues of the mammalian anatomical structure a fluorescence having spectral characteristics indicative of a third classification thereof, further comprising calculating a third probability that the tissue sample belongs in the third classification from a data set comprising the first, second and third fluorescence intensity spectrum.

13. A method as in claim 12, wherein:

the first wavelength is about 337 nm;

the second wavelength is about 380 nm;

the third wavelength is about 460 nm;

the third classification is SIL as distinguished from normal squamous;

the first classification is SIL as distinguished from normal columnar; and

the second classification is high grade SIL as distinguished from low grade SIL.

14. The method of claim 10 wherein the step of calculating the first probability from a data set comprising the first, second and third fluorescence intensity spectrum comprises:

providing a statistically significant plurality of additional tissue samples, at least some of which are tissues known to belong to the first classification;

illuminating the additional tissue samples with the first wavelength electromagnetic radiation;

detecting a first plurality of fluorescence intensity spectra from the additional tissue samples illuminated in the first wavelength illuminating step;

illuminating the additional tissue samples with the second wavelength electromagnetic radiation;

detecting a second plurality of fluorescence intensity spectra from the additional tissue samples illuminated in the second wavelength illuminating step;

illuminating the additional tissue samples with the third wavelength electromagnetic radiation;

detecting a third plurality of fluorescence intensity spectra from the additional tissue samples illuminated in the second wavelength illuminating step;

generating vectors that account for variation in the first, second and third pluralities of fluorescence intensity spectra and that are indicative of the first classification; and

calculating the first probability from the data set comprising the first fluorescence intensity spectrum, with use of the vectors.

15. The method of claim 9 wherein the step of calculating the first probability from a data set comprising the first and second fluorescence intensity spectrum comprises:

providing a statistically significant plurality of additional tissue samples, at least some of which are tissues known to belong to the first classification;

illuminating the additional tissue samples with the first wavelength electromagnetic radiation;

detecting a first plurality of fluorescence intensity spectra from the additional tissue samples illuminated in the first wavelength illuminating step;

illuminating the additional tissue samples with the second wavelength electromagnetic radiation;

detecting a second plurality of fluorescence intensity spectra from the additional tissue samples illuminated in the second wavelength illuminating step;

generating vectors that account for variation in the first and second pluralities of fluorescence intensity spectra and that are indicative of the first classification; and

calculating the first probability from the data set comprising the first fluorescence intensity spectrum, with use of the vectors.

16. A method as in claim 1 wherein the calculating step comprises calculating a probability from the first fluorescence intensity spectrum that the tissue is SIL versus normal squamous.

17. A method as in claim 16 wherein the illuminating step comprises illuminating the tissue sample with electromagnetic radiation having a wavelength of about 337 nm.

18. A method as in claim 16 wherein the illuminating step comprises illuminating the tissue sample with electromagnetic radiation having a wavelength of about 460 nm.

19. A method as in claim 1 wherein the calculating step comprises calculating a probability from the first fluorescence intensity spectrum that the tissue is SIL versus normal columnar and inflammation.

20. A method as in claim 19 wherein the illuminating step comprises illuminating the tissue sample with electromagnetic radiation having a wavelength of about 380 nm.

21. A method as in claim 1 wherein the calculating step comprises calculating a probability from the first fluorescence intensity spectrum that the tissue is high grade SIL versus low grade SIL.

22. A method as in claim 21 wherein the illuminating step comprises illuminating the tissue with electromagnetic radiation having a wavelength of about 460 nm.

23. The method of claim 1 wherein the illuminating step is performed in vivo.

24. The method of claim 1 wherein the illuminating step is performed in vitro.

25. The method of claim 1 wherein the calculating step comprises:

providing a statistically significant plurality of additional tissue samples, at least some of which are tissues known to belong to the first classification;

illuminating the additional tissue samples with the first wavelength electromagnetic radiation;

detecting a plurality of additional fluorescence intensity spectra from the additional tissue samples;

generating vectors that account for variation in the additional fluorescence intensity spectra and that are indicative of the first classification; and

calculating the first probability from the data set comprising the first fluorescence intensity spectrum, with use of the vectors.

26. The method of claim 25 wherein the first fluorescence intensity spectrum comprises emission wavelengths of about 410 nm, about 460 nm, about 510 nm and about 580 nm when the first wavelength is about 337 nm;

about 460 nm, about 510 nm, about 580 nm, about 600 nm and about 640 nm when the first wavelength is about 380 nm; and

about 510, about 580 nm, about 600 nm, about 620 nm, about 640 nm and about 660 nm when the first wavelength is about 460 nm.

27. A method of assigning a probability that a tissue sample belongs to a particular tissue category, comprising:

providing a first tissue sample;

illuminating the first tissue sample with electromagnetic radiation having at least one wavelength known to excite tissue into producing a fluorescence intensity spectra containing information about whether tissue belongs to the particular tissue category;

detecting a fluorescence intensity spectra from the first tissue sample; and

calculating from the fluorescence intensity spectra from the first tissue sample a probability that the tissue sample belongs to the particular tissue category.

28. A method as in claim 27 wherein the calculating step comprises:

providing a statistically significant plurality of second tissue samples, at least some of which are tissues known to belong to the particular tissue category;

illuminating the second tissue samples with the electromagnetic radiation;

detecting a plurality of fluorescence intensity spectra from the second tissue samples, respectively;

calculating from the fluorescence intensity spectra from the second tissue samples a probability distribution for the second tissue samples belonging to the particular tissue category; and

calculating the probability that the tissue sample belongs to the particular tissue category using the fluorescence intensity spectra from the first tissue sample and the probability distribution for the second tissue samples.

29. A method as in claim 28 wherein the probability distribution calculating step comprises:

generating a set of first vectors that account for variation in the fluorescence intensity spectra from the second tissue samples; and

selecting from the first vectors a set of second vectors that are indicative of the particular tissue category, the second vectors containing first indicia of the probability distribution for the second tissue samples belonging to the particular tissue category.

30. A method as in claim 29 wherein:

the first vector generating step comprises principle component analysis;

the second vector generating step comprises a student'"'"'s t-test; and

the step of calculating the probability that the tissue sample belongs to the particular tissue category using the fluorescence intensity spectra from the first tissue sample and the probability distribution for the second tissue samples comprises logistic discrimination.

31. A method as in claim 27 wherein:

the illuminating step comprises illuminating the tissue sample with electromagnetic radiation having at least a first wavelength known to excite tissue into producing a fluorescence intensity spectra containing information about whether tissue belongs to a first tissue category, and a second wavelength known to excite tissue into producing a fluorescence intensity spectra containing information about whether tissue belongs to a second tissue category that is a refinement of the first tissue category;

the detecting step comprises detecting first and second fluorescence intensity spectra from the illuminating step to obtain respective first and second spectral data; and

the calculating step comprises calculating from the first spectral data a first probability that the tissue sample belongs to the first tissue category, calculating from the second spectral data a second probability that the tissue sample belongs to the first tissue category, and assigning the tissue sample a probability of belonging to the second tissue category from the first and second probabilities.

32. A method as in claim 27 wherein:

the illuminating step comprises illuminating the tissue sample with electromagnetic radiation having a first wavelength of about 337 nm, a second wavelength of about 380 nm, and a third wavelength of about 460 nm; and

the detecting step comprises detecting first, second and third fluorescence intensity spectra from the illuminating step to obtain respective first, second and third spectral data; and

calculating from the first, second and third spectral data a probability that the tissue sample belongs to the particular tissue category.