Spectral imaging
View Patent ↗Apparatus and methods are provided for the imaging of structures in deep tissue within biological specimens, using spectral imaging to provide highly sensitive detection. By acquiring data that provides a plurality of images of the sample with different spectral weightings, and subsequent spectral analysis, light emission from a target compound is separated from autofluorescence in the sample. With the autofluorescence reduced or eliminated, an improved measurement of the target compound is obtained.
1. A method for analyzing spectrally resolved information about light coming from different spatial locations of a biological sample in response to an illumination of the sample, wherein the light includes contributions from different components in the sample, said method comprising using one or more electronic processors to perform the following:
deriving an estimate of a pure spectrum for at least a first one of the components in the sample from at least part of the spectrally resolved information corresponding to a first set of one or more of the different spatial locations and a spectral estimate of a pure spectrum for a second one of the components, wherein each pure spectrum estimate corresponds to the spectrally resolved information that would result if only the component corresponding to that estimate contributes to the light and wherein each pure spectrum estimate comprises a plurality of values corresponding to different spectral weighting functions;
decomposing at least some of the spectrally resolved information for each of at least some of the different spatial locations into a contribution from the spectral estimate of the pure spectrum for at least a selected one of the components in the sample; and
constructing an image of the sample based on the decomposition to preferentially show the selected one of the components.
2. The method of claim 1 , wherein the sample comprises deep tissue, tissue slices, cells, subdermal tissue, or a microscope slide carrying biological material.
3. The method of claim 1 , wherein the decomposing comprises decomposing the spectrally resolved information for each of the at least some of the different spatial locations into contributions from estimates of pure spectra for multiple components in the sample.
4. The method of claim 1 , wherein the decomposition is a linear decomposition.
5. The method of claim 4 , wherein the decomposition comprises solving at least one component of a matrix equation in which one matrix is based on the spectrally resolved information and another matrix is based on the estimate of the pure spectrum for the selected component and an estimate for the pure spectrum of at least one additional component.
6. The method of claim 1 , wherein the spectral estimate of the pure spectrum for the second component is provided independently of the spectrally resolved information.
7. The method of claim 1 , wherein the spectral estimate of the pure spectrum for the second component is determined from the spectrally resolved information.
8. The method of claim 7 , wherein the spectral estimate of the pure spectrum for the second component is determined from the spectrally resolved information by associating a region of one or more of the spatial locations with the second component.
9. The method of claim 1 , wherein the spatial locations in the first set are spatial locations in which the light includes contributions from more than one of the components.
10. The method of claim 1 , wherein deriving the spectral estimate of the pure spectrum for the first component comprises calculating a remainder spectrum based on the spectrally resolved information from the first set and the spectral estimate for the second component, and wherein for each location in the first set, the remainder spectrum is calculated according to
R a (λ)= S (λ) −aF (λ)
where R a (λ) is the remainder spectrum, S(λ) is the spectrally resolved information corresponding to the location, F(λ) is the spectral estimate of the pure spectrum for the second component, a is a parameter, and λ is a wavelength of the spectrally resolved information.
11. The method of claim 10 , wherein the remainder spectrum is calculated at each of one or more of the spatial locations in the first set of spatial locations.
12. The method of claim 1 , wherein deriving the spectral estimate of the pure spectrum for the first component comprises adjusting values corresponding to the spectrally resolved information for the first set of spatial locations to remove a contribution from the second component based on the spectral estimate for the second component.
13. The method of claim 12 , wherein the removed contribution is a maximal contribution.
14. The method of claim 1 , wherein the spectrally resolved information comprises information about a set of images in which the light coming from the sample is spectrally filtered, wherein the spectral filtering for each image corresponds to a different spectral weighting function.
15. The method of claim 14 , wherein each of the different spatial locations corresponds to a set of pixels in each of the images corresponding to a common region of the sample.
16. The method of claim 15 , wherein the information about the set of images comprises a series of values at each of the pixels, wherein each value is related to an intensity of the light coming from the sample with respect to a corresponding one of the spectral weighting functions.
17. The method of claim 14 , wherein the different spectral weighting functions correspond to different spectral bands.
18. The method of claim 14 , wherein the set of images comprises four or more images.
19. The method of claim 1 , wherein the spectrally resolved information for each spatial location comprises information corresponding to at least four different spectral weighting functions.
20. The method of claim 1 , wherein the spectrally resolved information comprises a spectral image cube.
21. The method of claim 1 , wherein the light coming from the sample comprises fluorescence from the sample.
22. The method of claim 1 , wherein at least one of the components relates to autofluorescence.
23. The method of claim 1 , wherein at least one of the components comprises a target compound.
24. The method of claim 23 , wherein the selected component is the component comprising the target compound.
25. The method of claim 23 , wherein the target compound comprises a fluorescent protein or a quantum dot.
26. The method of claim 1 , further comprising illuminating the sample and collecting the spectrally resolved information.
27. The method of claim 1 , wherein collecting the spectrally resolved information comprises using a liquid crystal tunable spectral filter, an acousto-optical tunable spectral filter, a set of spectral filters, or a spectral filter wheel.
28. The method of claim 1 , wherein the image of the selected component comprises an image in which signal from the other components is reduced relative to signal from the selected component.
29. The method of claim 1 , further comprising constructing a second image of the sample based on the decomposition to preferentially show another one of the components in the sample different from the selected component.
30. An apparatus comprising a non-transitory computer-readable medium storing a program that causes a processor to carry out the steps of claim 1 .
31. A method for analyzing a set of images of spectrally filtered radiation emitted from a biological sample in response to an illumination, the method comprising:
providing the set of images, wherein the sample comprises a component supporting a target compound and the emitted radiation comprises emission from the target compound and emission from one or more additional components in the sample, and wherein each image corresponds to a different spectral weighting function for a common set of pixels; and
using one or more electronic processors to process the images of the spectrally filtered radiation to construct an output image of the sample in which signal from the additional components is reduced relative to signal from the target compound,
wherein the processing comprises calculating a remainder spectrum for one or more pixels in the set of images based on an estimate for a pure spectrum of at least one of the components;
wherein, for each of the at least one of the components, the estimate for the pure spectrum corresponds to the spectrally filtered radiation that would result if only that component contributed to the emitted radiation, and the estimate for the pure spectrum comprises a plurality of values corresponding to different spectral weighting functions; and
wherein for each of the one or more pixels, the remainder spectrum is calculated according to
R a (λ)= S (λ) −aF (λ)
where R a (λ) is the remainder spectrum, S(λ) is the spectrally filtered radiation corresponding to the pixel, F(λ) is the estimate of the pure spectrum of the at least one of the components, a is a parameter, and λ is a wavelength of the spectrally filtered radiation.