Imaging analysis device and imaging data analysis method

Abstract

One embodiment of the imaging analysis device of the present invention comprises: an analysis execution unit (1) for collecting imaging data by performing analysis based on a prescribed analysis method on a plurality of measurement points set in a measurement area on a sample; a reference image acquisition unit (2) for acquiring a reference image for the measurement area; a regression analysis execution unit (36) for performing a prescribed regression analysis operation on the imaging data and the reference image obtained for the same sample, with the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to acquire a regression model; and a predicted image generation unit (37) for generating a predicted image based on a simulated regression analysis result by applying the imaging data obtained by the analysis execution unit to the regression model for a sample different from the sample when the regression analysis is performed.

Description

Technical Field

[0001] The present invention relates to an imaging analysis device such as an imaging mass analyzer, and an analysis method for performing analysis processing using imaging data in the imaging analysis device. Background Art

[0002] Imaging mass spectrometers are devices that can simultaneously observe the surface morphology of a sample, such as a biological tissue section, using an optical microscope and measure the two-dimensional intensity distribution of ions with a specific mass-to-charge ratio (m / z) on the same sample surface. By using an imaging mass spectrometer to observe two-dimensional intensity distribution images (MS imaging images) of ions derived from compounds characteristic of specific diseases, such as cancer, it is possible to understand the spread of the disease. Consequently, in recent years, research using imaging mass spectrometers has been gaining momentum for analyzing drug dynamics in biological tissue sections, differences in compound distribution across organs, and differences in compound distribution between pathological sites, such as cancer, and normal sites.

[0003] Imaging mass spectrometers generate mass spectral data showing signal intensities across a wide range of mass-to-charge ratios at each of the multiple measurement points on a sample. Consequently, the amount of data generated for a single sample is enormous, and multivariate analysis is widely used to extract meaningful information from this vast amount of data.

[0004] For example, Non-Patent Document 1 discloses data analysis software that compares a reference image, such as a stained image of a sample such as a biological tissue section, with an MS imaging image obtained by an imaging mass spectrometer, which shows the signal intensity distribution at each mass-to-charge ratio (m / z). The software then extracts mass-to-charge ratios that exhibit a two-dimensional distribution similar to that of the reference image and displays the MS imaging image at the extracted mass-to-charge ratios. As described in Patent Document 1 and other documents, the search for similar images can utilize partial least squares (PLS) regression, which uses the data constituting the reference image as the target variable Y and the mass spectrometry imaging data as the explanatory variable X.

[0005] In the PLS regression described above for imaging quality analysis, regression coefficients (PLS scores) are calculated for each mass-to-charge ratio of the explanatory variable X, i.e., the mass imaging data, to produce a regression coefficient matrix. A model image is then generated for each measurement point within a two-dimensional range on the specimen being analyzed for imaging quality, based on the result of multiplying the signal intensity value at each mass-to-charge ratio by the regression coefficient corresponding to that mass-to-charge ratio. This image is then displayed on the display. This model image is a distribution image in which the errors (residuals) in the PLS regression have been removed. If the regression is performed well, the distribution of the tissue of interest (e.g., a specific lesion) can be accurately displayed.

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: International Publication No. 2017 / 002226

[0009] Non-patent literature

[0010] Non-Patent Document 1: "Making Data Analysis for IMAGEREVEAL™ MS Mass Spectrometry Imaging Simple and Flexible," [Online], Shimadzu Corporation, [retrieved July 3, 2019], URL<URL:https: / / www.an.shimadzu.co.jp / bio / imagereveal / index.htm> Summary of the Invention

[0011] Technical problem to be solved by the invention

[0012] In the aforementioned conventional data analysis software or imaging mass spectrometers that utilize such software, PLS regression cannot be performed, and therefore model images cannot be calculated, without a reference image such as a stained image (i.e., an optical image) covering the same range as the MS image. However, when analyzing multiple samples, performing imaging mass analysis and optical imaging of the same two-dimensional range for each sample is extremely cumbersome and time-consuming.

[0013] Furthermore, the same problem is not limited to imaging mass spectrometers, but is also common to imaging analysis devices that use various measurement methods such as Raman spectroscopic imaging, fluorescence imaging, and FTIR imaging.

[0014] The present invention is completed to solve the above-mentioned technical problems, and its purpose is to provide an imaging analysis device and an imaging data analysis method that can generate an accurate model image even when it is impossible to prepare a reference image of a two-dimensional range on a sample on which imaging analysis such as imaging quality analysis is performed.

[0015] Solutions for solving the above technical problems

[0016] In order to solve the above-mentioned technical problems, one embodiment of the imaging analysis device of the present invention comprises:

[0017] an analysis execution unit that performs analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data;

[0018] A reference image acquiring unit, configured to acquire a reference image for the measurement area;

[0019] a regression analysis execution unit that executes a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model;

[0020] The predicted image generating unit generates a predicted image based on a simulated regression analysis result by applying the imaging data obtained by the analysis executing unit to the regression model for a sample different from the sample when the regression analysis is executed.

[0021] In addition, in order to solve the above-mentioned technical problems, one solution of the imaging data analysis method of the present invention comprises the following steps:

[0022] an analysis execution step of performing analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data;

[0023] a reference image acquisition step of acquiring a reference image for the measurement area;

[0024] a regression analysis execution step of executing a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model;

[0025] The predicted image generating step applies the imaging data obtained by the analysis executing step to the regression model for a sample different from the sample when the regression analysis is performed in the regression analysis executing step, thereby generating a predicted image based on the simulated regression analysis result.

[0026] The prescribed analysis method includes mass spectrometry, Raman spectroscopy, infrared spectroscopy (including Fourier transform infrared spectroscopy), fluorescence spectroscopy, and the like. Furthermore, the reference image can be an image obtained by analyzing the sample using an analysis method different from the one selected as the prescribed analysis method among the spectroscopy methods exemplified above. Alternatively, the reference image can be an image obtained using a conventional optical microscope.

[0027] Effects of the Invention

[0028] As an example, the aforementioned analysis method is mass spectrometry, and the reference image is an optical image of the measurement area on the sample. This optical image is an image corresponding to the purpose of the analysis, such as a stained image obtained by staining a substance associated with a specific lesion.

[0029] In the above-described aspect of the imaging analysis device of the present invention, the regression analysis execution unit performs regression analysis using mass analysis imaging data of a sample as an explanatory variable and image data constituting an optical image (reference image) of the same measurement area of ​​the same sample as a target variable. The regression coefficients are calculated for each mass-to-charge ratio to obtain a regression model, and the regression model is pre-stored. When the analysis execution unit analyzes a sample for which a reference image has not been obtained, thereby obtaining mass analysis imaging data, the predicted image generation unit applies the mass analysis imaging data to the regression model to perform a simulated regression analysis, i.e., to perform computational processing that yields results substantially identical to those obtained when the regression analysis is actually performed, thereby generating a predicted image that would have been obtained if the regression analysis had been performed correctly.

[0030] According to one embodiment of the imaging analysis device and one embodiment of the imaging data analysis method of the present invention, even when a reference image for a target sample to be analyzed is not available, a regression model generated using another sample can be used to generate a model image that approximates the results of a regression analysis performed on the imaging data of the target sample. Thus, by pre-generating an accurate regression model, information such as the distribution of tissues with a high probability of a specific lesion, such as cancer, can be obtained based on the imaging data of the target sample. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a schematic diagram of the configuration of an imaging mass spectrometer as one embodiment of the imaging analysis device of the present invention.

[0032] Figure 2 FIG. 1 is a schematic diagram of data processing performed in the imaging mass spectrometer of this embodiment. DETAILED DESCRIPTION

[0033] An imaging mass analyzer as one embodiment of the imaging analyzer of the present invention will be described with reference to the accompanying drawings.

[0034] [Configuration of the Imaging Mass Spectrometer of the Present Embodiment]

[0035] Figure 1 This is a schematic diagram of the configuration of the imaging mass spectrometer according to this embodiment.

[0036] The imaging mass spectrometer comprises an imaging mass spectrometer unit 1 for analyzing a sample using imaging mass spectrometry, an optical microscope unit 2 for capturing an optical image (fluorescence image) on the sample, a data processing unit 3, an input unit 4 serving as a user interface, and a display unit 5.

[0037] The imaging mass analysis unit 1 includes, for example, a MALDI ion trap time-of-flight mass analyzer, which performs mass analysis on multiple measurement points (micro areas) within a two-dimensional measurement area on a sample such as a biological tissue slice, and obtains mass analysis data for each measurement point. In the following description, the mass analysis data is mass spectrum data over a specified mass-to-charge ratio range, but it can also be MS data for a specific precursor ion. n Spectrum (product ion scanning) data The optical microscope imaging unit 2 is an optical microscope with an imaging unit added thereto, and is used to obtain an optical image of a two-dimensional region on the surface of a sample.

[0038] The data processing unit 3 receives mass spectrum data at each measurement point collected by the imaging mass analyzer 1 and optical image data input from the optical microscopic imaging unit 2 and performs predetermined processing. The data processing unit 3 includes a data collection unit 31, a data storage unit 32, an image generation unit 33, an image alignment processing unit 34, a regression analysis execution unit 35, a regression model storage unit 36, a predicted image generation calculation unit 37, and a display processing unit 38 as functional blocks. The data storage unit 32 includes a spectrum data storage area 321 for storing data collected through measurements by the imaging mass analyzer 1 and an optical image data storage area 322 for storing optical image data collected through measurements (photographs) by the optical microscopic imaging unit 2.

[0039] In general, the data processing unit 3 is implemented as a personal computer (or a higher-performance workstation), and is configured to implement the functions of the aforementioned functional blocks by running dedicated software installed on the computer. In this case, the input unit 4 is a pointing device such as a keyboard and mouse, and the display unit 5 is a monitor.

[0040] [Operation of the Imaging Mass Spectrometer of the Present Embodiment]

[0041] Next, refer to Figure 2 The following describes the sample-based analysis and analysis results of the imaging mass spectrometer of this embodiment. The sample to be analyzed is, for example, a slice of a specific biological tissue, such as an organ or brain, taken from an experimental animal. When performing the same analysis on multiple samples of the same biological tissue slices taken from different organisms (individuals), a regression model (described below) is first generated based on the analysis results of a single sample (hereinafter referred to as sample A).

[0042] The steps to generate the regression model are as follows.

[0043] The operator places the specimen A at a predetermined measurement position on the optical microscopic imaging unit 2 and performs a predetermined operation via the input unit 4. The optical microscopic imaging unit 2, in response to the operation, images the surface of the specimen A and stores the optical microscopic image data in the optical image data storage area 322. Furthermore, the image generation unit 33 generates an optical image, and the display processing unit 38 displays the image on the screen of the display unit 5. The operator uses the input unit 4 to indicate a measurement area for the entire specimen or a portion of the specimen on this image.

[0044] The operator temporarily removes sample A from the apparatus and allows the MALDI matrix to adhere to the surface of sample A. Sample A, with the matrix attached, is then placed at a specified measurement position in the imaging mass spectrometer 1 and a specified operation is performed via the input unit 4. The imaging mass spectrometer 1 then performs mass analysis at multiple measurement points within the measurement area indicated above on sample A, acquiring mass analysis data across the specified mass-to-charge ratio range. At this point, the data acquisition unit 31 performs so-called profile acquisition, collecting profile spectrum data that forms a continuous waveform in the mass-to-charge ratio direction within the specified mass-to-charge ratio range and storing it in the spectrum data storage area 321 of the data storage unit 32.

[0045] Furthermore, if the sample surface appearance (eg, boundaries between different tissues) can be observed relatively clearly even when the matrix is ​​attached to the sample surface, the optical microscope imaging unit 2 may be used to perform imaging after the matrix is ​​attached to the sample surface.

[0046] As described above, in a state where the mass spectrometry imaging data and the optical image data regarding the sample A are stored in the data storage unit 32 , data processing is performed as follows.

[0047] The image generation unit 33 reads the profile data about the above-mentioned sample A from the spectrum data storage area 321 of the data storage unit 32, calculates the signal intensity in a predetermined plurality of target mass-to-charge ratios at each measurement point, and generates an MS imaging image showing a two-dimensional distribution of the signal intensity according to each of the mass-to-charge ratios.

[0048] Specifically, a profile spectrum is generated based on the profile data, peaks are detected on the profile spectrum, and each detected peak is subjected to a centroid conversion process to thereby determine the correct peak position (mass-to-charge ratio value). Then, if the mass-to-charge ratio value of the centroid peak exists within the prescribed mass-to-charge ratio range centered on the specified mass-to-charge ratio, it is considered that the centroid peak is the peak corresponding to the target mass-to-charge ratio. Further, in the profile spectrum, the signal intensity values ​​within the prescribed mass-to-charge ratio range (the range of the mass accuracy of the mass analyzer) centered on the centroid peak are integrated and set to the signal intensity value relative to the target mass-to-charge ratio. Since the profile data in each microregion are subjected to the same processing, a two-dimensional distribution of the signal intensity values ​​in the target mass-to-charge ratio can be obtained. Therefore, if it is imaged, an MS imaging image in a target mass-to-charge ratio can be obtained.

[0049] Furthermore, the image generation unit 33 reads optical image data for the same sample from the optical image data storage area 322 of the data storage unit 32 to generate a single optical image. While the spatial resolution of the optical microscopic imaging unit 2 is generally determined by the resolution of the imaging camera, the resolution of the MS imaging image is determined by the spot diameter of the laser beam irradiating the sample for ionization. Therefore, the resolution of the MS imaging image is often lower than that of the optical image. Therefore, if the spatial resolution of the optical image and the MS imaging image differ, the image alignment processing unit 34 performs resolution adjustment processing to align the spatial resolutions.

[0050] A simple method for achieving resolution consistency is to reduce the resolution of the higher-resolution image to match it with the lower-resolution image. For example, merging can be used as such. Alternatively, the resolution of the lower-resolution image can be increased to match it with the higher-resolution image. To do this, after upsampling the lower-resolution image to find the number of pixels and roughly matching them, interpolation using the values ​​of multiple pixels adjacent to or close to a pixel is performed to calculate and fill in the pixel values ​​of the newly inserted pixels through upsampling.

[0051] After aligning the spatial resolutions, the image alignment processing unit 34 appropriately deforms the optical image to roughly align the positions of the MS imaging image and the optical image on a pixel-by-pixel basis. Specifically, for example, the MS imaging image is scaled up, reduced, rotated, and translated, using the optical image as a reference, and then deformed according to a prescribed algorithm to roughly align the positions of the specimens in the two images. This processing enables pixels in the same two-dimensional position to be associated with each other between the optical image and the MS imaging image. The thus-processed optical image serves as the reference image. Alternatively, instead of using the optical image directly, a two-dimensional distribution image of the brightness values ​​of specific color components, either pre-specified or selected as a result of automatic processing, can be used as the reference image.

[0052] Afterwards, the regression analysis execution unit 35 uses the matrix generated based on the data constituting the processed MS imaging image, with the signal intensity value of each mass-to-charge ratio in each pixel as an element, as the explanatory variable X. Similarly, the matrix generated based on the reference image, with the brightness value of each pixel as an element, is used as the target variable Y to perform the PLS regression operation. PLS regression is a well-known statistical analysis method that can be calculated using a variety of generally available software, so a detailed description is omitted. Through PLS regression, the parameters of the explanatory variable X, that is, the regression coefficient for each mass-to-charge ratio, can be calculated. The relationship between the target variable Y and the explanatory variable X is represented by the following regression formula (regression model).

[0053] Y=Bpis·X+B0 …(1)

[0054] In formula (1), Bpis is the regression coefficient matrix. In addition, B0 is the error (residual) generated in the regression. The higher the confidence level of the fit in the linear regression formula, the smaller the residual.

[0055] The regression analysis execution unit 35 stores the regression model or regression coefficient matrix calculated as described above in the regression model storage unit 36. Furthermore, once the regression model is obtained, the predicted image generation calculation unit 37 applies the regression model to the explanatory variable X, i.e., the signal intensity value for each mass-to-charge ratio value in each pixel of the mass spectrometry imaging data, to generate a predicted image based on the regression analysis results. This predicted image is obtained by the following equation (2) by removing the residual B0 from equation (1).

[0056] Y'=Bpis·X …(2)

[0057] If the accuracy of the regression model is high, that is, if the regression model is a model that well describes the target variable based on the explanatory variables, the two-dimensional distribution of the predicted image Y' will become similar to the two-dimensional distribution of the reference image Y.

[0058] The following processing is performed on target samples other than Sample A. Similarly to Sample A, the operator performs mass analysis on Sample B at multiple measurement points within a specified measurement area, acquiring mass analysis data across the specified mass-to-charge ratio range. The resulting profile data is then subjected to the aforementioned data processing to collect mass analysis imaging data. Sample B is not imaged by the optical microscopic imaging unit 2, and therefore no optical image data is generated.

[0059] The predicted image generation calculation unit 37 uses the matrix generated based on the data constituting the MS imaging image of the processed sample B, whose elements are the signal intensity values ​​for each mass-to-charge ratio value in each pixel, as the explanatory variable X, and applies the regression model stored in the regression model storage unit 36 ​​to generate a predicted image. This can be understood as a simulated PLS regression analysis using a regression model generated based on another sample.

[0060] To generate a predicted image for a particular sample, it is generally preferred to use a regression model derived by performing PLS regression on a reference image and mass spectrometry imaging data obtained for that sample. However, under conditions where the same biological tissue of the same organism (e.g., the same experimental animal) is analyzed for the same purpose (e.g., estimating the same type of cancerous tissue), a regression model generated for one sample can be applied to other samples to generate a substantially accurate predicted image. The display processing unit 38 displays the predicted images generated for each sample on the screen of the display unit 5, for example, as a list or individually.

[0061] As described above, the imaging mass analyzer of this embodiment can generate a predicted image as a simulated regression analysis result based on the mass analysis imaging data and present it to the operator, even when a reference image is not available or when the regression coefficients for PLS regression analysis are not actually calculated. This saves the effort and time required to obtain a reference image and also reduces the time required to perform the PLS regression calculations.

[0062] Alternatively, multiple reference images may be acquired for a single sample in advance, and different regression models may be generated corresponding to each of these reference images. For example, by pre-generating regression models corresponding to different diseases and lesions that occur in the same biological tissue, multiple predicted images corresponding to different diseases and lesions may be generated from a single sample.

[0063] Furthermore, if the reference image is a color image, the color can be separated into components of the three primary colors of light to obtain reference images of three different color components. Since the two-dimensional distributions in these three reference images should be different, regression models can be calculated for each of the three reference images using the above steps and pre-stored in the regression model storage unit 36. In this way, by calculating multiple regression models from a reference image of a single color, generating predicted images for each regression model, and integrating them into a single predicted image, the accuracy of the predicted image can be improved.

[0064] In addition, in the imaging mass analysis device of the above-mentioned embodiment, the optical microscope image is set as the reference image, but the imaging image of the same sample obtained by other measurement methods other than the imaging mass analysis method can also be used as a reference image, such as Raman spectroscopic imaging, infrared spectroscopic imaging, X-ray analysis imaging, surface analysis imaging using particle beams such as electron beams or ion beams, or surface analysis imaging using probes such as scanning probe microscopes (SPM).

[0065] Furthermore, PLS regression can be performed using data such as Raman spectroscopic data or infrared spectroscopic data for each measurement point within the measurement area, obtained using Raman spectroscopic imaging or infrared spectroscopic imaging, as explanatory variables. In other words, the present invention can be applied not to imaging mass analyzers but to Raman spectroscopic imaging devices, infrared spectroscopic imaging devices, fluorescence spectroscopic imaging devices, and the like.

[0066] The above-described embodiment is merely an example of the present invention, and in addition to the various modifications described above, changes, corrections, and additions may be made as appropriate within the scope of the gist of the present invention, which are naturally also encompassed by the scope of the claims of the present application.

[0067] [Various options]

[0068] An embodiment of the present invention has been described above with reference to the drawings. Finally, various aspects of the present invention will be described.

[0069] (Item 1) One embodiment of the imaging analysis device of the present invention comprises:

[0070] an analysis execution unit that performs analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data;

[0071] A reference image acquiring unit, configured to acquire a reference image for the measurement area;

[0072] a regression analysis execution unit that executes a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model;

[0073] The predicted image generating unit generates a predicted image based on a simulated regression analysis result by applying the imaging data obtained by the analysis executing unit to the regression model for a sample different from the sample when the regression analysis is executed.

[0074] (Item 5) One embodiment of the imaging data analysis method of the present invention comprises:

[0075] an analysis execution step of performing analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data;

[0076] a reference image acquisition step of acquiring a reference image for the measurement area;

[0077] a regression analysis execution step of executing a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model;

[0078] The predicted image generating step applies the imaging data obtained by the analysis executing step to the regression model for a sample different from the sample when the regression analysis is performed in the regression analysis executing step, thereby generating a predicted image based on the simulated regression analysis result.

[0079] According to the imaging analysis device described in item 1 and the imaging data analysis method described in item 5, even when a reference image for the target sample to be analyzed is not available, a regression model generated using another sample can be used to generate a model image similar to an image obtained by regression analysis based on imaging data of the target sample. Thus, by pre-generating an accurate regression model, information such as the distribution of biological tissues with a high probability of a specific lesion, such as cancer, can be obtained based on imaging data of the target sample.

[0080] (Item 2) In the imaging analysis device described in Item 1, the analysis method may be mass spectrometry, and the imaging data may be mass spectrum data within a predetermined mass-to-charge ratio range obtained at each measurement point.

[0081] (Item 6) Furthermore, in the imaging data analysis method described in Item 5, the analysis method may be mass spectrometry, and the imaging data may be mass spectrum data within a predetermined mass-to-charge ratio range obtained at each measurement point.

[0082] (Item 3) In the imaging analysis device described in Item 1, it can be set that the analysis method is Raman spectroscopy or infrared spectroscopy, and the imaging data is spectral data of a specified wavelength range or wavenumber range obtained at each measurement point.

[0083] (Item 7) In the imaging data analysis method described in Item 5, it can be set that the analysis method is Raman spectroscopy or infrared spectroscopy, and the imaging data is spectral data of a specified wavelength range or wavenumber range obtained at each measurement point.

[0084] (Item 4) In the imaging analysis device described in Item 1, it can be set that the regression analysis is a partial least squares regression analysis.

[0085] (Item 8) Furthermore, in the imaging data analysis method described in Item 5, the regression analysis may be a partial least squares regression analysis.

[0086] According to the imaging analysis device described in item 4 and the imaging data analysis method described in item 8, good regression can be performed and a highly accurate predicted image with small residual error can be generated.

[0087] Description of Reference Numerals

[0088] 1 Imaging Quality Analysis Department

[0089] 2 Optical microscopy unit

[0090] 3 Data Processing Department

[0091] 31 Data Collection Department

[0092] 32 Data storage unit

[0093] 321 Spectrum data storage area

[0094] 322 Optical image data storage area

[0095] 33 Image Generation Unit

[0096] 34 Image alignment processing unit

[0097] 35 Regression Analysis Execution Department

[0098] 36 Regression model storage unit

[0099] 37 Prediction image generation calculation unit

[0100] 38 Display Processing Unit

[0101] 4 Input section

[0102] 5. Display unit.

Claims

1. An imaging analysis device, characterized in that: have: an analysis execution unit that performs analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data; A reference image acquiring unit, configured to acquire a reference image for the measurement area; a regression analysis execution unit that executes a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model; The predicted image generating unit generates a predicted image based on a simulated regression analysis result by applying the imaging data obtained by the analysis executing unit to the regression model for a sample different from the sample when the regression analysis is executed.

2. The imaging analysis device according to claim 1, wherein The analysis method is mass spectrometry, and the imaging data is mass spectrum data within a predetermined mass-to-charge ratio range obtained at each measurement point.

3. The imaging analysis device according to claim 1, wherein The analysis method is Raman spectroscopy or infrared spectroscopy, and the imaging data is spectrum data in a predetermined wavelength range or wavenumber range obtained at each measurement point.

4. The imaging analysis device according to claim 1, wherein The regression analysis is a partial least squares regression analysis.

5. A method for analyzing imaging data, characterized in that: have: an analysis execution step of performing analysis based on a predetermined analysis method on each of a plurality of measurement points set within a measurement area on the sample to collect imaging data; a reference image acquisition step of acquiring a reference image for the measurement area; a regression analysis execution step of executing a predetermined regression analysis operation on the imaging data and the reference image obtained for the same sample, using the imaging data as an explanatory variable and the data constituting the reference image as a target variable, to obtain a regression model; The predicted image generating step applies the imaging data obtained by the analysis executing step to the regression model for a sample different from the sample when the regression analysis is performed in the regression analysis executing step, thereby generating a predicted image based on the simulated regression analysis result.

6. The imaging data analysis method according to claim 5, wherein: The analysis method is mass spectrometry, and the imaging data is mass spectrum data within a predetermined mass-to-charge ratio range obtained at each measurement point.

7. The imaging data analysis method according to claim 5, wherein: The analysis method is Raman spectroscopy or infrared spectroscopy, and the imaging data is spectrum data in a predetermined wavelength range or wavenumber range obtained at each measurement point.

8. The imaging data analysis method according to claim 5, wherein: The regression analysis is a partial least squares regression analysis.

Images