Method for observing a sample in the infrared range

Abstract

A method is provided for observing a biological sample between a light source and a pixelated image sensor, the light emitting an incident light beam, which propagates to the sample along a propagation axis and at an emission wavelength, the method including: illuminating the sample with the source; and acquiring an image of the sample with the sensor, no image-forming optic being placed between the sample and the sensor, the sample absorbing some of the beam, such that the acquired image is representative of an absorption of the beam by the sample at the emission wavelength, the source illuminates an area of the sample larger than 1 mm.sup.2, the image acquired of the sample by the sensor corresponds to an area of sample larger than 1 mm.sup.2, and pixels of the sensor define a detection plane, the sample being placed at a distance from the plane smaller than 1 mm.

Claims

1. A method for observing a biological sample, the sample being placed between a light source and a pixelated image sensor, the light source emitting an incident light beam, which propagates to the sample along a propagation axis, and at an emission wavelength comprised between 1 μm and 20 μm, the method comprising: a) illuminating the sample with the light source; and b) acquiring an image of the sample with the pixelated image sensor, no image-forming optic being placed between the sample and the image sensor, wherein: the sample absorbs some of the incident light beam, such that the acquired image is representative of an absorption of the incident beam by the sample, at the emission wavelength, the light source illuminates an area of the sample larger than 1 mm.sup.2 or larger than 5 mm.sup.2, the image acquired of the sample by the image sensor corresponds to an area of sample larger than 1 mm.sup.2 or larger than 5 mm.sup.2, and pixels of the image sensor define a detection plane, the sample being placed at a distance from the detection plane smaller than 1 mm.

2. The method of claim 1, further comprising: c) illuminating the image sensor with the light source, at the emission wavelength, with no sample between the image sensor and the light source, so as to obtain a background image; and d) comparing the image acquired in b) and the background image acquired in c) to obtain an image of the absorbance of the sample at the emission wavelength.

3. The method of claim 2, wherein the emission wavelength is an absorption wavelength of an analyte, corresponding to an absorption peak of the analyte, and the method further comprising mapping an amount of the analyte in the sample on the basis of the image of the absorbance at the emission wavelength.

4. The method of claim 2, wherein the emission wavelength is an absorption wavelength of an analyte, corresponding to an absorption peak of the analyte, so as to obtain an image of the absorbance of the sample at the absorption wavelength, and the method further comprising: e) illuminating the sample at a base wavelength, at which the absorption of the analyte is lower than the absorption of the analyte at the absorption wavelength; f) acquiring an image of the sample with the pixelated image sensor; g) illuminating the image sensor with the light source, at the base wavelength, with no sample between the image sensor and the light source, so as to obtain a background image at the base wavelength; h) comparing the image acquired in 0 and the background image acquired in g) so as to obtain an image of the absorbance of the sample at the base wavelength; and i) subtracting the absorbance images of the sample at the absorption wavelength and at the base wavelength, respectively, so as to obtain an image of absorbance due to the analyte.

5. The method of claim 4, wherein steps a) to d) are repeated so as to successively illuminate the sample at: a first absorption wavelength, corresponding to an absorption wavelength of a first analyte, and a second absorption wavelength, corresponding to an absorption wavelength of a second analyte, so as to obtain images of the absorbance of the sample at the first absorption wavelength and at the second absorption wavelength, respectively, and wherein steps e) to h) are repeated so as to successively illuminate the sample at: a first base wavelength, at which the absorption of the first analyte is lower than the absorption of the first analyte at the first absorption wavelength, and a second base wavelength, at which the absorption of the second analyte is lower than the absorption of the second analyte at the second absorption wavelength, so as to obtain images of the absorbance of the sample at the first base wavelength and at the second base wavelength, respectively, the method further comprising: subtracting the images of absorbance of the sample at the first absorption wavelength and at the first base wavelength, respectively, so as to obtain an image of absorbance due to the first analyte; subtracting the images of absorbance of the sample at the second absorption wavelength and at the second base wavelength, respectively, so as to obtain an image of absorbance due to the second analyte; comparing the image of absorbance due to the first analyte and the image of absorbance due to the second analyte.

6. The method of claim 1, wherein the sample is a slide of biological tissue.

7. The method of claim 1, further comprising, on the basis of the acquired image, determining at least one region of interest of the sample.

8. The method of claim 1, wherein the sample is held by a carrier, the carrier comprising at least one of: silicon, and/or germanium, and/or calcium fluoride, and/or barium fluoride.

9. The method of claim 1, wherein the pixelated image sensor is formed from a matrix-array of bolometers.

Description

FIGURES

(1) FIG. 1 shows a device allowing the invention to be implemented.

(2) FIG. 2A is a visible image of a first sample. FIG. 2B shows images of the first sample, obtained by implementing the invention. FIG. 2B is representative of an absorption of the first sample at a wavelength of 7.35 μm.

(3) FIG. 3A schematically shows an absorption peak of an analyte. FIG. 3B contains the main steps of one embodiment of the invention. FIG. 3C is an image of the absorbance of a second sample at a wavenumber of 1081 cm.sup.−1 (i.e. λ=9.2 μm). FIG. 3D is an image of the absorbance of the second sample at a wavenumber of 1654 cm.sup.−1 (i.e. λ=6 μm).

DESCRIPTION OF PARTICULAR EMBODIMENTS

(4) FIG. 1 shows an example of a device allowing the invention to be implemented. A light source 11 is configured to emit a light beam 12, called the incident light beam, that propagates in the direction of a sample 10. The incident light beam reaches the sample by propagating along a propagation axis Z.

(5) In this example, the sample 10 is a biological sample that it is desired to characterize. It may notably be a tissue slide intended for histological analysis, or pathology slide, comprising a small thickness of tissue deposited on a transparent slide 15 that acts as a sample carrier. By small thickness, what is meant is a thickness preferably smaller than 100 μm, and preferably smaller than 10 μm, typically of a few microns. The sample lies in a plane P.sub.10, called the sample plane. The sample plane is preferably perpendicular or substantially perpendicular to the propagation axis Z. The term substantially perpendicular means perpendicular to within an angular tolerance of a few degrees, smaller than 20° or 10°.

(6) The tissue slide 10 is obtained according to known preparation methods, from a tissue sample extracted by biopsy or a swab. The sample is then prepared so as to be in the form of a small thickness deposited on the transparent slide 15. Such methods are known in the field of histology. They for example comprise a slice of frozen tissue, or sampled tissue embedded in a paraffin-wax matrix. Preferably, the sample has not been labeled beforehand, by means of an exogenous marker added to the sample before its analysis.

(7) The sample may comprise an analyte, a spatial distribution of which in the sample it is desired to evaluate. By analyte, what is meant is, for example, a molecule or a portion of a molecule or a molecular bond.

(8) The slide 15 is transparent to the incident beam 12. It may comprise or consist of materials such as silicon, germanium, calcium fluoride (CaF.sub.2) and barium fluoride (BaF.sub.2).

(9) The distance Δ between the light source and the sample, along the axis Z, is preferably larger than 1 cm. It is preferably comprised between 2 and 30 cm. Preferably, the light source, seen by the sample, is considered to be point-like. This means that its diameter (or its diagonal) is preferably smaller than one tenth, and better still one hundredth, of the distance between the sample and the light source. Thus, the light preferably reaches the sample in the form of plane waves, or waves that may be considered as such.

(10) The light source 11 is a source that emits in the infrared. It may be a question of the short wavelength infrared (SWIR), which lies between 1 and 3 μm, or of the medium wavelength infrared (MWIR), which lies between 3 and 5 μm, or even of the long wavelength infrared (LWIR), which lies between 8 and 20 μm. Thus, generally, the incident beam 12 is emitted at a wavelength λ lying between 1 μm and 20 μm, this corresponding to a wavenumber (1/λ) comprised between 500 cm.sup.−1 and 10000 cm.sup.−1. In the example shown, the device comprises a reflector 13, for reflecting the light beam 12 emitted by the light source 11 toward the sample. Preferably, the wavelength is comprised between 5 μm and 20 μm.

(11) The light source 11 is preferably a laser source. It may notably be a laser source the wavelength of which is tunable, such as for example a quantum cascade laser (QCL), and in particular an external-cavity laser. The width of the emission spectral band of the light source is preferably smaller than 50 nm, or even than 10 nm, or even than 5 nm. A light source may comprise a plurality of elementary QCL laser sources, respectively emitting in various spectral bands.

(12) The sample 10 is placed between the light source 11 and an image sensor 20. The latter preferably lies parallel, or substantially parallel, to the transparent slide 15 carrying the sample. The term “substantially parallel” means that the two elements may not be rigorously parallel, an angular tolerance of a few degrees, smaller than 20° or 10°, being acceptable. The sample is placed on a holding element, configured to hold the sample, and its carrier, between the light source and the image sensor.

(13) Under the effect of illumination by the light beam 12, propagating along the propagation axis Z to the sample, the latter transmits a light wave 14, called the transmitted light wave. The transmitted light wave 14 propagates, parallel to the axis Z, to an image sensor 20. The sample absorbs some of the light beam 12. Thus, the transmitted light wave 14 corresponds to a portion of the light beam 12 not absorbed by the sample.

(14) The image sensor 20 is able to form an image of the transmitted light wave 14 in a detection plane P.sub.20. In this example, the image sensor is formed by a matrix-array of bolometers, each bolometer of the matrix-array having a detection spectral band comprised between 5 μm and 20 μm. Each bolometer forms a pixel. In the examples described below, each pixel is formed by one vacuum-encapsulated bolometer. Conventionally, a dark image, corresponding to the noise of each bolometer in the absence of illumination, may be acquired. The dark image is then subtracted from each acquired image. When the sample is placed between the image sensor and the light source, the image acquired by the image sensor is representative of the absorption of the incident beam 12 by the sample 10.

(15) When the image sensor 20 comprises non-functional pixels, or dead pixels, the intensity of each non-functional pixel is replaced by an average of the intensities measured by the pixels adjacent to the non-functional pixel.

(16) The distanced between each pixel and the sample 10 is preferably smaller than 5 mm. The smaller it is, the better the spatial resolution of the image acquired by the image sensor. Also, it is advantageous for the distance d to be smaller than 1 mm, or even smaller than 500 μm. The sample may be deposited in direct contact with the pixels of the image sensor 20, in a shadowgraph configuration such as described in US20050190286.

(17) Due to the absence of image-forming optic between the sample 10 and the image sensor 20, the field of observation of the image sensor is defined by the size of the image sensor and the size of the incident beam. The field of observation may be larger than 1 mm.sup.2, or even larger than 5 mm.sup.2 or 10 mm.sup.2. Therefore, the acquisition of a single image allows an intensity of the light wave 14 transmitted by several mm.sup.2 of the sample, typically at least 5 or 10 mm.sup.2 of the sample, to be obtained simultaneously.

(18) Preferably, the transparent slide 15 comprises an anti-reflective coating. For example, when the transparent slide is made of silicon, it may comprise a thin layer of germanium or of zinc sulfide (ZnS). This allows the appearance of interference fringes in the images formed by the image sensor to be limited. In the absence of a thin anti-reflective layer, the transparent slide 15 behaves like a Fabry-Perot cavity, this leading to the formation of undesirable interference fringes in the image acquired by the image sensor 20.

(19) A processing unit, for example taking the form of a microprocessor 22, is configured to perform, on the basis of the images acquired by the image sensor 20, image-processing operations such as described below. The microprocessor 22 is connected to a memory 23, which contains instructions relating to the image-processing operations to be carried out. It may be connected to a screen 24.

(20) FIG. 2A shows a first sample that was examined using a device such as illustrated in FIG. 1. It is a question of a section of muscle tissue taken from a mouse and then frozen. The sampled tissue contained a tumor as a result of injection of CAL33 tumor cells. It was sliced to obtain a sample of thickness equal to 4 μm. The sample was then deposited on a silicon slide or on a slide of CaF.sub.2 (calcium fluoride).

(21) The experimental set-up was: light source: QCL laser source emitting at the wavelength of 7.35 μm; diameter of the incident beam: 1.5 mm; image sensor: matrix-array of 80×80 bolometers of 17 μm side length, with a center-to-center distance between each pixel equal to 30 μm, giving a field of observation of about 2.4×2.4 mm.sup.2.

(22) The slide 15 was mounted on a translation stage, so as to be translated in two orthogonal directions X and Y perpendicular to the propagation axis Z. A scan allowed a matrix-array of 150 images arranged in 25 rows and 6 columns to be obtained. The field covered by the matrix-array of images corresponds to a sample area 3 mm wide and 12.5 mm long. The observed area of the sample has been framed by a box in FIG. 2A. The central region of the sample, designated by the letter T, corresponds to a malignant tumor whereas the peripheral region, designated by the letter H, corresponds to healthy tissue. FIG. 2B shows the images acquired by the image sensor.

(23) It may be seen that the images corresponding to the cancerous region T have a dense appearance, whereas the images corresponding to the healthy region H have a spongy appearance. Thus, an infrared image, acquired using a lensless imaging method, allows the presence of a cancerous region to be detected and viewed. It allows a region of interest corresponding to the cancerous region or to the healthy region to be defined in the sample.

(24) FIGS. 3A to 3D correspond to a second embodiment, in which a plurality of images of a sample are produced, by modulating the wavelength λ of the illumination beam 12.

(25) It is known that the spectral light transmittance of a sample varies as a function of the composition of the latter, because of the presence of absorption peaks corresponding to modes of vibration of the molecules from which the sample is composed. The presence of absorption peaks is the basis of vibrational spectrometry methods such as infrared spectroscopy or Raman spectrometry.

(26) By transmittance tr.sub.λ.sub.i, what is meant is a ratio between an intensity i.sub.λ.sub.i of the light wave 14 transmitted by the sample, and detected by the image sensor, at the wavelength λ.sub.i, to an intensity of the light wave detected by the image sensor, at the same wavelength, in the absence of a sample.

(27) Thus, according to the Beer-Lambert law:

(28) $\begin{array}{cc}{\mathrm{tr}}_{{\lambda }_i}=\frac{i_{{\lambda }_i}}{i_{0,{\lambda }_i}}& \left(1\right)\end{array}$
where i.sub.0,λ.sub.i is the intensity detected by the image sensor in the absence of a sample.

(29) The absorbance abs.sub.λ.sub.i at the wavelength λ.sub.i is obtained using the expression:

(30) $\begin{array}{cc}ab{s}_{{\lambda }_i}=-\ln \left({\mathrm{tr}}_{{\lambda }_i}\right)=-\ln \left(\frac{i_{{\lambda }_i}}{i_{0,{\lambda }_i}}\right).& \left(1^{\prime }\right)\end{array}$

(31) FIG. 3A shows an example of the absorption spectrum of an analyte, the x-axis representing the wavelength. An absorption peak, indicated by an arrow, at an absorption wavelength λ.sub.a may be seen. On either side of the absorption peak lies a baseline, identified by a dashed line. The baseline corresponds to base wavelengths λ.sub.b, lying on either side of the absorption peak. Thus, a base wavelength is a wavelength located outside the absorption peak of the analyte. Preferably, a base wavelength bounds the absorption peak. At the base wavelength, the absorption of the incident light beam by the analyte is lower than the absorption of the incident beam by the analyte in the absorption peak.

(32) The approach proposed by the inventors consists in determining a map of an absorbance resulting from the presence of an analyte, the absorption wavelength λ.sub.a of which is known. To do this, the method consists in: acquiring an image, representative of the absorption of the light beam 12 by the sample, corresponding to an image I.sub.λ.sub.a acquired by the image sensor when the illumination beam is emitted at the absorption wavelength λ.sub.a of the analyte. Such an image is designated by the term “absorption image”. acquiring an image, called the background image, corresponding to an image I.sub.0,λ.sub.a acquired by the image sensor in the absence of a sample, at the absorption wavelength λ.sub.a.

(33) Comparison of the absorption image I.sub.λ.sub.a and background image I.sub.0,λ.sub.a may make it possible to obtain an absorbance image Abs.sub.λ.sub.a of the sample, at the absorption wavelength λ.sub.a, such that:

(34) $\begin{array}{cc}Ab{s}_{{\lambda }_a}\left(x,y\right)=-\ln \left(\frac{I_{{\lambda }_a}\left(x,y\right)}{I_{0,{\lambda }_a}\left(x,y\right)}\right)& \left(2\right)\end{array}$
where Abs.sub.λ.sub.a (x, y) is the value of the absorbance image Abs.sub.λ.sub.a at the coordinates (x, y); I.sub.λ.sub.a(x, y) and I.sub.0,λ.sub.a(x, y) are the intensities of the pixels of the absorption and background image at the coordinate (x, y), respectively. These images may be corrected using the dark image of the image sensor.

(35) The coordinates (x, y) are defined in the detection plane P.sub.20. Since the latter is parallel to the sample plane P.sub.10, the coordinates (x, y) also correspond to coordinates in the sample plane P.sub.10.

(36) On the basis of the absorbance image Abs.sub.λ.sub.a, it is possible to estimate an amount Q (x, y) of analyte at each coordinate (x, y), with

(37) $\begin{array}{cc}Q\left(x,y\right)=-\frac{Ab{s}_{{\lambda }_a}\left(x,y\right)}{{\mu }_{{\lambda }_a}\mathrm{.Math.}\left(x,y\right)},& \left(3\right)\end{array}$
where: ε(x, y) is the thickness of the sample, along the propagation axis Z, at the coordinates (x, y); μ.sub.λ.sub.a is the absorption coefficient of the analyte, per unit length, at the wavelength λ.sub.a.

(38) These steps are summarized in FIG. 3B:

(39) Step 100: acquiring an absorption image I.sub.λ.sub.a, when the sample is illuminated by a light source at the absorption wavelength λ.sub.a.

(40) Step 110: acquiring a background image I.sub.0,λ.sub.a, with no sample between the image sensor and the light source, at the absorption wavelength λ.sub.a.

(41) Step 120: computing a ratio between the absorption image and the background image, to obtain an absorbance image of the sample Abs.sub.λ.sub.a, at the absorption wavelength λ.sub.a. According to one variant, the sample is illuminated at a base wavelength λ.sub.b, at which the absorption of the incident beam by the sample is lower than the absorption at the absorption wavelength λ.sub.a. The method then comprises the following steps:
Step 130: acquiring an absorption image I.sub.λ.sub.b, when the sample is illuminated by a light source at the base wavelength λ.sub.b.
Step 140: acquiring a background image I.sub.0,λ.sub.b, with no sample between the image sensor and the light source, at the base wavelength λ.sub.b.
Step 150: computing a ratio between the absorption image and the background image, to obtain an absorbance image of the sample Abs.sub.λ.sub.b, at the base wavelength λ.sub.b.
Step 160: subtracting the absorbance image of the sample at the base wavelength from the absorbance image of the sample at the absorption wavelength, so as to obtain an image I, representative of the absorbance due to the analyte. I=Abs.sub.λ.sub.a−Abs.sub.λ.sub.b.

(42) However, it may be difficult to estimate amounts of analyte quantitatively. The inventors believe that it may be preferable to perform comparisons between absorbance images resulting from various analytes, and for example from various biomarkers.

(43) It is known that cancerous activity may be characterized by a morphological indicator, representative of a ratio between the volume of the nucleus and the volume of the cytoplasm in the sample. Specifically, it is known that cancer cells have a higher metabolic activity than healthy cells. Thus, they tend to have a nucleus the volume of which is larger than that of healthy cells. Therefore, the ratio of nuclear volume to cytoplasm volume is an indicator used by histopathologists to establish the malignancy of a tumor.

(44) The publication Amrania H “Digistain: a digital staining instrument for histopathology”, Optics Express 7299, Vol. 20, No. 7, 26 Mar. 2012, describes a method based on a comparison of the absorbance due to PO.sub.2.sub.− groups, representative of phosphodiester bonds inside the cell nucleus, and of the absorbance due to Amide bonds, the latter being representative of peptide bonds, inside the cytoplasm. By making a comparison between the absorbance due to phosphodiester bonds and the absorbance due to peptide bonds, a morphological indicator representing a nuclear volume/cytoplasmic volume ratio may be obtained.

(45) The inventors were inspired by this method. Specifically, they successively illuminated a sample, such as described with reference to FIG. 2A, at various wavelengths λ.sub.a,1, λ.sub.a,2, λ.sub.b,1 and λ.sub.b,2, such that: 1/λ.sub.a,1=1081 cm.sup.−1, which corresponds to an absorption peak of PO.sub.2.sub.−; 1/λ.sub.a,2=1654 cm.sup.−1, which corresponds to an absorption peak of an Amide group, in a spectral band commonly referred to as Amide I, which corresponds to a vibration of the C═O bond. 1/λ.sub.b,1=951 cm.sup.−1, which corresponds to the baseline about the absorption peak of PO.sub.2.sub.−; 1/λ.sub.b,2=1491 cm.sup.−1, which corresponds to the baseline about the Amide I absorption peak.

(46) The light source used included 4 QCL lasers, respectively emitting in the following spectral ranges: 1949 cm.sup.−1 to 1706 cm.sup.−1; 1712 cm.sup.−1 to 1410 cm.sup.−1; 1464 cm.sup.−1 to 1149 cm.sup.−1, 1218 cm.sup.−1 to 896 cm.sup.−1.

(47) At each wavelength λ, two images were acquired: a background image I.sub.0,λ, with no sample between the light source and the image sensor; an image of absorption I.sub.λ, the sample being placed between the light source and the image sensor.

(48) By calculating a ratio between the absorption image I.sub.λ and the background image I.sub.0,λ, absorbance images Abs.sub.λ were obtained at each wavelength λ.

(49) Thus, for each analyte, in this case for PO.sub.2.sub.− and the Amide bond, the following were obtained: an image of absorbance in each absorption peak; these images being denoted Abs.sub.λ.sub.a,1 and Abs.sub.λ.sub.a, 2; an image of absorbance at the baseline level lying on either side of each absorption peak, these images being denoted Abs.sub.λ.sub.b,1 and Abs.sub.λ.sub.b,2.

(50) FIGS. 3C and 3D show an image of absorbance at the wavelengths 1081 cm.sup.−1 and 1654 cm.sup.−1.

(51) The absorbance image obtained at a base wavelength was then subtracted from the absorbance image obtained at each absorption wavelength, so as to obtain an image representing an absorbance due of each analyte. In this example, the analytes in question are PO.sub.2.sub.− and an amide group.

(52) The images I.sub.1 and I.sub.2 of the absorbances due to each analyte, corresponding to PO.sub.2.sub.− and to the amide bond, respectively, are such that:
I.sub.1=Abs.sub.λ.sub.a,1−Abs.sub.λ.sub.b,1  (4)
I.sub.2=Abs.sub.λ.sub.a,2−Abs.sub.λ.sub.b,2  (4′)

(53) This embodiment amounts to performing steps 100 to 160 described with reference to FIG. 3B, a first time for the wavelengths λ.sub.a,1 and λ.sub.b,1, and a second time for the wavelengths λ.sub.a,2 and λ.sub.b,2.

(54) A ratio

(55) $\frac{I_1}{I_2}$
between two images I.sub.1 and I.sub.2 may then be calculated, so as to obtain a map of the nuclear volume/cytoplasmic volume ratio. Depending on the ratio

(56) $\frac{I_1\left(x,y\right)}{I_2\left(x,y\right)},$
regions of interest corresponding to cancerous regions are defined in the sample.

(57) As shown by the preceding examples, the invention allows, without labeling, regions of interest of a sample, liable to have a pathological character, to be defined. The lensless imaging configuration allows a large field of observation to be addressed.