Method and apparatus for testing a progression of neoplasia including cancer chemoprevention testing

• US 6,031,930 A
• Filed: 08/23/1996
• Issued: 02/29/2000
• Est. Priority Date: 08/23/1996
• Status: Expired due to Term

• Current Assignee: Evident Scientific Incorporated (Olympus Corporation)
• Original Assignee: Bacus Laboratories Incorporated (Olympus Corporation)
• Inventors: Bacus, James W., Bacus, James V.
• Primary Examiner(s): Chang, Jon
• Assistant Examiner(s): Patel, Jayanti K.
• Application Number:
  US08/701,974
• Time in Patent Office: 1,285 Days
• Field of Search: 250/461.2, 348/79, 348/588, 356/39, 364/922, 364/922.2, 382/128, 382/133, 382/134, 382/284, 382/132, 382/130, 382/131, 382/171, 382/172, 382/173, 382/189, 382/209, 382/218, 382/282, 382/286, 382/299, 382/308, 395/924, 435/509, 435/515-17, 435/129-131, 435/1.1, 345/115, 345/117, 424/572, 424/573
• US Class Current: 382/133
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G06T 3/4038 Image mosaicing, e.g. composing plane images from plane sub-images
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation
• CPC Class Codes: Y10S 128/922 including image analysis

Abstract

A system is provided for analysis of neoplasia in tissue at a very early stage as well as later stages of its progression and for reporting the progression or regression of the neoplasia. The system performs multi-parametric measurements of the morphological structure and texture of the tissue structure and correlates these measurements on a common morphological grading scale. The system performs certain tissue measurements which are more highly discriminating for one kind of neoplasia and performs other tissue measurement, which are more highly discriminating, for another kind of neoplasia. Diverse tissue measurements may be made on diverse tissue types such as breast, colon, or cervix tissue, etc. from animals or humans which tissue has subjected to different carcinogens or chemopreventive agents. The measurements are made in different units and on different scales; and then these measurements are combined and reported on a valid, objective, common, universal scale. Microscopic images of stained neoplastic tissue sections are optically and microscopically scanned to provide electronic tissue sample images that are electronically recorded. Then, morphometric features of tissue sample images are measured in first unit vales and texture measurements of the tissue samples, such as a Markovian texture measurement. Usually, normal tissue samples and abnormal cancerous tissue samples are analyzed using the same morphometric and texture measurements and their respective results are reported onto a grading common scale so that progression of the cancer can be ascertained relative to the normal tissue.

Claims

1. A method of analyzing a tissue sample having a layer of adjacent, connected tissue cells for pre-invasive neoplasia, comprising:

optically and microscopically scanning the tissue sample having adjacent, connected tissue cells to provide, multiple image fields and electronically recording these image fields to provide spatially adjacent multiple image fields of the tissue sample larger than a single view of the adjacent, connected tissue cells;

forming a reconstituted image of the suspected neoplasia tissue from the multiple, electronically recorded, spatially adjacent image fields larger than a single field of view and having adjacent connected tissue cells in the reconstituted image;

displaying to the viewer the reconstituted tissue image comprised of multiple, spatially adjacent image fields larger than a single field of view;

determining and selecting from the reconstituted, tissue image a region of adjacent connected tissue cells for analysis at a resolution; and

analyzing tissue sample images of the selected region at the resolution for the analysis of pre-invasive neoplasia.

2. A method in accordance with claim 1 wherein the step of analyzing the selected region includes:

measuring morphometric features of tissue sample images in first unit values;

measuring texture features of tissue sample images in other units or scale values; and

combining the measured units and scale values into a common scale.

3. A method in accordance with claim 1 wherein the step of analyzing the selected region includes:

optically and microscopically scanning normal tissue samples and analyzing these normal tissue samples using a combination of morphometric and texture measurements; and

optically and microscopically scanning abnormal tissue samples and analyzing these abnormal tissue samples using the same combination of morphometric and texture measurements as used for the normal tissue.

4. A method in accordance with claim 3 including the steps of:

determining the mean and standard deviation for the measurements of the normal tissue; and

subtracting the mean for normal tissue measurements from the abnormal tissue measurements and dividing by the normal tissue'"'"'s standard deviation to provide a morphometric score.

5. A method in accordance with claim 4 including wherein the recording step comprises recording the morphometric scores on a graph with a morphometric scale on the ordinate and a time scale on the abscissa to illustrate any progression of neoplasia.

6. A method in accordance with claim 1 wherein the tissue includes a basal layer of tissue cells and the analysis is for neoplasia, said method including the steps of:

addressing a plurality of the tissue sample images adjacent the basal layer;

orienting the addressed tissue sample images to form a tiled, combined, reconstituted image of a portion of the scanned tissue sample; and

the determining and selecting step comprises editing the reconstituted image of the basal layer to delete portions thereof from the subsequent analysis.

7. A method in accordance with claim 6 including the steps of:

scanning the tissue sample at a first magnification resolution; and

decreasing the resolution for the reconstituted image.

8. A method in accordance with claim 7 including the steps of:

interactively viewing the recorded tissue sample images; and

interactively editing to leave substantially only the basal layer of cells and cells evolving therefrom.

9. A method in accordance with claim 6 wherein the editing step comprises:

displaying a tissue sample image selected from the reconstituted image at a higher resolution than the reconstituted image; and

erasing from the displayed tissue sample image, one or more portions thereof having limited analysis value.

10. A method in accordance with claim 9 comprising replacing the selected image of the reconstituted image with the edited version of the selected tissue sample image.

11. A method in accordance with claim 1 including the step of selecting, from a menu of measurements, a first set of morphometric and texture measurements which are highly discriminating for a first kind of tissue sample;

and

selecting a second set of different morphometric and texture measurements which are highly discriminating for a second kind of tissue sample.

12. A method in accordance with claim 1 wherein the analyzing step comprises analyzing a tissue sample image for a predetermined texture feature and the method comprises:

identifying a mean value of the result of analyzing normal tissue images for the texture feature;

determining the mean value of the result of analyzing the tissue sample image for the predetermined texture feature;

subtracting the mean value identified for the particular texture feature of normal tissue from the mean value determined for the particular texture feature of the analyzed tissue sample image; and

recording the result of the subtracting step.

13. A method in accordance with claim 12 comprising identifying the standard deviation of the results of analyzing normal tissue images for the predetermined texture feature and dividing the result of the subtraction step by the identified standard deviation.

14. A method of claim 1 comprising presenting the recorded results in a manner which normalizes the results of different analyses.

15. A method of analysis of a tissue sample having adjacent, connected tissue cells comprising the steps of:

optically and mircroscopically scanning the tissue sample to provide multiple tissue sample images having adjacent, connected tissue cells therein, and electronically recording the tissue sample images to provide multiple tissue sample images of the tissue sample;

reassembling the recorded tissue sample images to form a reconstituted, tissue sample image comprised of multiple spatially adjacent, microscopic, tissue sample images from the tissue sample;

measuring sections of the reconstituted tissue sample image with respect to area;

measuring sections of the reconstituted tissue sample image with respect to optical density;

measuring the morphological texture of adjacent, connected tissue cells of sections of the reconstituted tissue sample image;

analyzing the results of the measured area, optical densities and morphological texture of adjacent, connected tissue cells of the selected sections of the reconstituted tissue sample image of adjacent, connected tissue cells to provide morphological data; and

providing the morphological data for an analysis of the reconstituted tissue sample image of adjacent, connected tissue cells.

16. A method of analysis in accordance with claim 15 wherein the step of analyzing the texture of the tissue sample image comprises the step of measuring run length which includes an analysis of neighboring cells in the tissue sample image for optical densities above a predetermined threshold.

17. A method of analysis in accordance with claim 16 wherein the tissue sample has a basal layer extending in a given direction;

and

making the run length configurable by analyzing neighboring cells in the given direction of the basal layer.

18. A method in accordance with claim 15 wherein the texture measuring step comprises making at least one of a valley, slope or peak measurement.

19. A method in accordance with claim 18 wherein a coarseness measurement of the texture is obtained by subtracting from the slope measurements, the peak and valley measurements.

20. A method in accordance with claim 15 wherein the texture measuring step comprises the making of a Markovian Texture measurement.

21. A method in accordance with claim 20 including the steps of analysis of human cervix tissue sample images wherein the step of measuring morphological texture includes the making of Markovian Texture measurements.

22. A method of analyzing a basal layer sample of tissue having adjacent, connected tissue cells evolving from the basal layer for neoplasia, said method comprising the steps of:

optically and microscopically scanning the adjacent, connected tissue cells along the basal layer over a plurality of tissue sample fields to provide microscopic, multiple spatially adjacent, image fields of the tissue sample;

recording these image fields and reassembling these image fields to form a reconstituted, magnified image comprised of multiple image fields of the tissue and having the basal layer thereon;

tracing the basal layer for analysis;

analyzing the traced basal layer of the reconstituted tissue sample image with respect to area and optical density;

analyzing the traced basal layer of the reconstituted tissue sample image with respect to texture by making measurements including the step of an analysis of texture by examining adjacent, neighboring cell images with respect to values above or below a predetermined threshold; and

providing the texture analysis of adjacent, connected tissue cells for use in evaluation of neoplasia.

23. A method in accordance with claim 22 including the step of configuring the run length to extend in the direction that the basal layer extends.

24. A method in accordance with claim 22 including the step of analyzing adjacent tissue cells to provide at least one of valley, slope and peak measurements.

25. A method in accordance with claim 22 including the step of measuring the tissue cell images using a Markovian Texture analysis.

26. A method in accordance with claim 22 for analysis of human cervix tissue including the steps of:

measuring the summed optical densities from the cervix tissue cell images; and

measuring the slope between adjacent human cervix cell images.

27. A method in accordance with claim 22 including the steps of:

analyzing an animal tissue to provide discriminating data comprising the steps of;

measuring the average optical densities of the animal tissue cell images;

measuring the valley and slope from adjacent cell images of the animal tissue; and

measuring Markovian Textures of the animal cell images.

28. A method in accordance with claim 22 wherein the texture measurements of animal tissue includes using at least one of the following steps:

measuring the run length of adjacent animal tissue cell images;

measuring the run length which is configured in direction and the spacing between analyzed cell images;

measuring the valley, slope and peak between adjacent cell images;

measuring the coarseness which includes the step of subtracting the peak and valley from the slope; and

measuring Markovian Textures of the analyzed cell images.

29. A method in accordance with claim 22 including the step of recording the results of the respective measurements in a manner which normalizes the results of the morphometric and texture analyses.

30. A method of analyzing histological tissue samples of different kinds of tissues having adjacent connected tissue cells having inherent variabilities for pre-invasive cancer comprising the steps of:

for each of the histological tissue samples, optically scanning the tissue sample at a plurality of non-overlapping tissue sample fields and electronically recording a high resolution tissue sample image representing selected fields for each of the different kinds of tissues;

analyzing each of the respective tissue sample images using a combination of morphometric and texture analyses of histological tissue predetermined to be indicative of the progression of cancer development for each of the kinds of tissues;

normalizing results of selected analyses performed in the analyzing step for the respective kinds of tissues to a common standard; and

deriving values for the respective kinds of tissues so that a comparison can be made indicative of the progression of their respective cancer developments from the normalized results.

31. A method in accordance with claim 30 including the step of editing the recorded tissue sample images to delete portions of the image not representing the basal layer and the evolving tissue cells to create edited tissue sample images and wherein the analyzing step comprises analyzing the edited tissue sample images.

32. A method of analyzing the efficacy of a chemopreventive agent on first and second precancerous tissues of adjacent, connected tissue cells, said method comprising the steps of:

optically and microscopically scanning the first precancerous tissue to provide image fields and electronically recording the image fields to provide multiple image fields of the first precancerous tissue;

integrating the recorded image fields into a reconstituted, first tissue sample image comprised of multiple image fields representing the first tissue;

analyzing tissue regions of the first reconstituted tissue sample image using a first set of morphometric and texture features highly discriminating for the first precancerous tissue;

optically and microscopically scanning the second precancerous tissue to provide image fields and electronically recording the image fields to provide multiple image fields of the second precancerous tissue;

integrating the recorded image fields into a reconstituted, second tissue sample image comprised of multiple image fields representing the second tissue;

analyzing tissue regions of the second reconstituted tissue sample image using a second set of morphometric and texture features highly discriminating for the second precancerous tissue;

combining the morphometric and texture feature measurements of the tissue regions from the first precancerous tissue to a common scale; and

combining the morphometric and texture feature measurements of the tissue regions from the second precancerous tissue to the same common scale so that progression of cancer of the first and second tissues can be compared on the same basis.

33. A method in accordance with claim 32 wherein there is provided as first precancerous tissue a cervical tissue sample and wherein there is provided as the second precancerous tissue an esophageal tissue.

34. A method in accordance with claim 32 including the step of taking sequentially in time the first and second precancerous tissue samples from the same part of the body and using thereon the first and second sets of morphometric and texture features for analysis.

35. A method in accordance with claim 32 wherein the first and second precancerous tissues are taken from first and second parts of a body.

36. A method in accordance with claim 32 wherein the first and second precancerous tissues are taken from different animals.

37. A method in accordance with claim 32 including the steps of analyzing the first and second precancerous tissue over a plurality of tissue fields.

38. A method in accordance with claim 32 including the step of normalizing the measurements recording the scores on a scale that has plus and minus numbers below twenty (20).

39. A method in accordance with claim 32 including the step of measuring at least some of the first morphometric or texture features with scales or units different from the scales or units used to measure the second morphometric or texture features.

40. Apparatus for analyzing a tissue sample having a layer of adjacent, connected tissue cells for pre-invasive neoplasia, comprising:

a scanning device for optically and microscopically scanning the tissue sample to provide magnified multiple image fields of the tissue sample;

a recorder for electronically recording the magnified, multiple image fields;

apparatus for forming a reconstituted, low magnification image comprised of the multiple image fields of the suspected neoplasia tissue;

a display for displaying to a viewer a reconstituted, low magnification tissue image comprised of multiple image fields;

a selector allowing the viewer to select a region from the reconstituted, low magnification image for analyses at a higher resolution; and

an analyzer for analyzing the selected tissue sample image fields of the selected region using at least one morphometric and texture measurement for adjacent, connected tissue cells predetermined to be indicative of the progression of pre-invasive neoplasia development in the tissue sample.

41. Apparatus in accordance with claim 40 wherein the analyzer for analyzing the texture of the selected tissue sample image fields comprises a means for measuring texture which includes an analysis of neighboring cells for optical densities above a predetermined threshold.

42. Apparatus in accordance with claim 41 wherein the cells are part of a tissue sample having a basal layer extending in a given direction;

and the means for measuring the run length is configurable to analyzing neighboring cells in the given direction of the basal layer.

43. Apparatus in accordance with claim 40 wherein the analyzer for measuring texture comprises means for making at least one of a valley, slope or peak measurement.

44. Apparatus in accordance with claim 43 comprising means for measuring a coarseness of the texture by subtracting from the slope measurements, the peak and valley measurements.

45. Apparatus in accordance with claim 40 wherein the analyzer for measuring texture comprises means for making a Markovian Texture measurement.

46. Apparatus in accordance with claim 45 wherein the analyzed cells are human cervix tissue cells images and the means for measuring morphological texture includes means for the making of Markovian Texture measurements.