WO2021160802A1 - Method for an automatic, semantic-based, functional tissue annotation of histological and cellular features in order to identify molecular features in tissue samples - Google Patents

Abstract

The present invention relates to a method for automated analysis and histological classification of a tissue section. The inventors developed a new imaging pipeline which is a computational multimodal workflow designed to combine molecular imaging data with multiplex immunohistochemistry (IHC). It allows comprehensive and spatially resolved in situ correlation analyses on a cellular resolution. Said pipeline was used to perform an automatic, semantic-based, functional tissue annotation of histological and cellular features in order to identify metabolic profiles. Spatial correlation networks were generated to analyse metabolic heterogeneity associated with cellular features.

Description

METHOD FOR AN AUTOMATIC, SEMANTIC-BASED, FUNCTIONAL TISSUE ANNOTATION OF HISTOLOGICAL AND CELLULAR FEATURES IN ORDER TO IDENTIFY MOLECULAR FEATURES IN

TISSUE SAMPLES

TECHNICAL FIELD OF THE INVENTION

[001] The present invention relates to a method for mass spectrometry analysis and histological classification of a tissue sample and the automated analysis thereof. The inventors developed a new imaging pipeline which is a computational multimodal workflow designed to combine molecular imaging data with multiplex immunohistochemistry (IHC). It allows comprehensive and spatially resolved in situ correlation analyses on a cellular resolution. The imaging pipeline of the present invention was used to perform an automatic, semantic-based, functional tissue annotation of histological and cellular features in order to identify metabolic profiles.

BACKGROUND ART

[002] Histology analysis is generally carried out on a stained tissue sections and concerns the tissue type, differentiation, presence of bacterial or parasitic pathogens, disease statuses, and content of foreign matter, like pesticides or drugs and their metabolites. An automated staining read-out remained illusive for a long time, due to different obstacles like staining intensities and cellular and nuclear overlay within the tissue. The European patent application EP 2 124 192 A1 pertains to a method for histologic classification of a tissue section but reveals no automated analysis thereof.

[003] Matrix-assisted laser desorption and ionization (MALDI) mass spectrometry comprises the ionization of a sample or analyte and has been used for the determination of molecular masses, and for the identification and structural characterization of biological substances, particularly proteins and peptides. Imaging mass spectrometry enables in situ label-free detection of thousands of metabolites from intact tissue samples. However, automated steps for multi-omics analyses and interpretation of histological images have not yet been implemented in mass spectrometry data analysis workflows. The characterization of molecular properties within cellular and histological features is presently done via time- consuming, non-objective, and irreproducible definitions of regions of interest, which are often accompanied by a loss of spatial resolution due to mass spectra averaging.

[004] Accordingly, there is a need for and automated data analysis workflow of histological images. The inventors developed a new automated imaging analysis pipeline which is a computational multimodal workflow designed to combine molecular imaging data with multiplex immunohistochemistry (IHC). The technical problem underlying the present application is thus to comply with the need for and automated data analysis workflow of histological images. [005] The technical problem is solved by providing the embodiments reflected in the claims, described in the description and illustrated in the examples and figures that follow.

SUMMARY OF THE INVENTION

[006] In a first aspect, the present invention relates to a method of automatically selecting regions of interest, ROIs, for analyzing one complete tissue section, wherein the method comprises

(i) generating a first image from mass spectrometry data of the tissue section spatially resolved by a region of the first image represented as a first mask,

(ii) and further obtaining one or more second images comprising any type of optical images of the tissue section used in (i), wherein the one or more second images are in alignment with and scaled to the resolution of the first image; determining one or more second masks associated with respective one or more second images based on predetermined criteria;

(iii) combining masks selected from the first mask and the one or more second masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs;

(iv) extracting mass spectrometry data using said first image and optical image data from said one or more second images, respectively, based on the final ROIs, for analyzing the tissue section; and

(iv) wherein the steps (i) to (iv) result in the automatic definition of ROIs for diagnostic and research application.

[007] Preferably, the method further comprises obtaining one or more custom masks of the one or more second images of (ii), and wherein combining masks in (iii) comprises combining masks selected from the first mask, the one or more second masks, and the one or more custom masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs.

[008] In view of the present invention, the method referred to herein may further comprise determining one or more second masks of (ii) associated with respective one or more second images based on predetermined criteria comprises using predetermined thresholds of color values or ranges of color values of pixels in an image of the one or more second images to determine regions containing pixels of similar color values in said image.

[009] In view of the present invention the mass spectrometry data may be obtained by Matrix-assisted Laser Desorption (MALDI), Desorption Electrospray Ionization (DESI), Laser Ablation Electrospray Ionization (LAESI), Secondary ion Mass Spectrometry (SIMS), Matrix-assisted Laser Desorption Electrospray Ionization (MALDESI), Liquid Extraction Surface Analysis (LESA), Mass spectrometry imaging (MSI), or Inductively Coupled Plasma Mass Spectrometry Imaging (ICP-MSI), respectively. [0010] Preferably, the mass spectrometry data is obtained by any one of MALDI, DESI, LAESI, SIMS, MALDESI, LESA, MSI and ICP-MSI, respectively. Most preferably, the mass spectrometry data is obtained by MALDI.

[0011] It is envisaged, that the images of the tissue section may be selected from images obtained by multiplex immunohistochemical (IHC) staining, histochemical staining, endogeneous or exogeneous fluorescence staining, respectively, fluorescence in situ hybridization (FISH), or autofluorescence.

[0012] In some embodiment, the mass spectrometry data may be represented in imzML format, mis format of fleximaging, slx/sbd format of SCiLS Lab, or a tabular format, respectively.

[0013] It is encompassed by the present invention that the tissue section may comprise one tissue sample obtained by formalin-fixed paraffin-embedding (FFPE) or obtained by any other fixation technique, comprising ethanol fixation and cryofixation, respectively.

[0014] Preferably, determining one or more second masks associated with respective one or more second images comprises determining multiple masks for the same image, said masks pertaining to multiple morphometric data, or data from multiple color components, such as Hematoxylin and eosin (H&E) and other histological stainings.

[0015] Preferably, ranges of color values of pixels are based on a magnitude of expression of a diagnostic feature, such as one or more biomarkers.

[0016] It is also envisaged that in the method of the present invention the biomarker is a) HER2/neu and/or pan-cytokeratin (PC); b) Insulin and/or glucagon; c) PD1 and/or PD-L1 ; or d) Vimentin and/or pan-cytokeratin.

[0017] In some embodiment, the predetermined set expressions may represent semantics of tissue annotation to obtain the final ROIs, based on the first image, the one or more second images and the associated masks of the tissue section.

[0018] It is further envisage that analyzing the tissue section may comprise comparing mass spectrometry data and image data, respectively, derived from the final ROIs of the first image and the one or more second images of the tissue section, by statistical analysis.

[0019] In particular, is is envisaged that mass spectrometry data and image data may be compared, respectively, extracted from the final ROIs comprises assessing of molecular composition and heterogeneity expressed between data extracted from final ROIs, wherein molecular composition comprises metabolites, proteins, drugs, toxic agents, carcinogens, glycanes, and lipids, respectively. [0020] Preferably, the method is performed simultaneously on multiple tissue samples individualized by one or more custom masks, wherein each custom mask is associated with an individual tissue sample.

[0021] In a second aspect the present invention relates to the use of the method for diagnosis or stratification based on a human, animal or plant tissue samples.

[0022] In some embodiment the use of the method may comprise that the samples are human tissue samples used in the diagnosis or stratification of cancer or diabetes.

[0023] In a third aspect the present invention relates to the use of the method for therapy response prediction and prediction of organ rejection based on human or animal tissue.

BRIEF DESCRIPTION OF THE FIGURES

[0024] Figure 1 A: Graphical Abstract SPACIAL - immunohistochemistry-guided in situ metabolics. The inventors developed a new imaging pipeline called Spatial Correlation Image Analysis (SPACIAL), which is a computational multimodal workflow designed to combine molecular imaging data with multiplex immunohistochemistry (IHC). B: Flow chart depicting the SPACIAL work flow. C: Additional information and examples to the SPACIAL work flow.

[0025] Figure 2 Workflow of immunohistochemistry-guided in situ metabolomics using the example of an islet of Langerhans. A: MALDI and IHC workflow starting with matrix application on tissue sections, MALDI imaging and data processing include peak picking and annotation. The matrix is then removed for subsequent multiplex immunohistochemical staining of the very same tissue section using DAPI, glucagon, and insulin. The individual stainings are digitized with a slide scanner. B: The SPACIAL pipeline integrates molecular MALDI data and immunohistochemical data. The IHC images need to be co-registered to the coordinates of the mass spectra per pixel. The MALDI data file is used to generate an image of the measurement region, which can be used for precise co-registration with the tissue image and tissue stainings. Once co-registered, the staining images are scaled to match the resolution of the measurement and color values per pixel are used for the definition of regions or for pixel- accurate analyses of metabolic correlations or heterogeneity.

[0026] Figure 3 Multi-omics data integration via the SPACIAL method. Left: Islet of Langerhans with immunohistochemical staining (glucagon in red, insulin in green). Middle: Spatial distribution of 3-0 Sulfogalactosylceramide (m/z 778.5147). Right: Data integration via SPACIAL, utilizing the IHC stainings to automatically identify alpha cells (semi-transparent, pixelated staining in red) and correlating metabolites. Lateral MALDI resolution (pixel): 15 pm.

[0027] Figure 4 Metabolic heterogeneity within and between islets of Langerhans in a pancreatic tissue section of one mouse (A-E). The column on the left shows multiplex immunostainings after MALDI imaging mass spectrometry. Alpha cells (red) and beta cells (green) are stained with glucagon and insulin, respectively. A tissue fold in the fifth islet (E) was excluded from analyses (dashed). The second and third columns show spatial correlation networks for metabolites in alpha and beta cells, respectively. Nodes and edges represent compounds and their spatial correlation. The networks shown here only include direct neighbors of the glucose 6-phosphate node and edges representing a correlation coefficient of at least 0.7. Scale bar, 150 pm. Abbreviations: adenosine monophosphate (AMP), guanosine monophosphate (GMP), phosphatidic acid (PA), phosphatidylethanolamine (PE), lysophosphatidic acid (LPA), lysophospholipid (LPC), lysophosphatidylinositol (LPI), dihydroxyacetone phosphate (DHAP), glycerophosphoinositol (GroPIns), phosphodimethylethanolamine (P-DME).

[0028] Figure 5 Multiplex immunohistochemistry-guided imaging mass spectrometry on islets of Langerhans to automatically distinguish morphologically indistinguishable cell types (A-E). Alpha and beta cells were stained with glucagon (red) and insulin (green), respectively. The spatial distributions of ADP, cholesterol sulfate and 3-O-sulfogalactosylceramide (sulfatide) are visualized (yellow). Pixel-wise intensity distributions are shown for alpha (red) and beta cells (green), respectively. See the methods section for details about the statistical analysis. Scale bar, 150 pm.

[0029] Figure 6 Image processing workflow to define HER2/neu positive and negative tumor regions. Pan-cytokeratin (green) as an epithelial marker to stain tumor cells. HER2/neu positive cells are shown in red. Both stainings are adjusted to match the lateral imaging mass spectrometry resolution (60 pm) and combined to classify HER2/neu positive and negative tumor cells. Scale bar, 3000 pm.

[0030] Figure 7 Intratumoral heterogeneity of spatially correlating metabolites in three human gastric cancer tissue sections, visualized via spatial correlation networks (A-C). Left: Close-up of the HER2/neu positive (red) and negative (yellow) tumor regions. Middle: Spatial correlation networks for metabolites. Edges represent positive (blue) and negative (red) spatial correlations between metabolites. Line thickness and transparency correspond to the correlation coefficient. Right: Zoom-in to glucose 6- phosphate. Abbreviations: glucose 6-phosphate (G6P), glycerol 3-phosphate (Gly3P), ribose 5-phosphate (R5P), S-adenosylhomocysteine (SAH), glycerophosphoinositol (GroPIns), adenosine diphosphate (ADP), guanosine monophosphate (GMP), adenine (Ade), 5-phosphoribosylamine (PRA), reduce flavin adenine dinucleotide (FADH), D-glutamine (DGN), cysteinyl-methionine (C-M), homocysteine (Hey), phosphatidic acid (PA), phosphatidylglycerol (PG), cyclic phosphatidic acid (CPA), lysophosphatidic acid (LPA), lysophospholipid (LPC), lysophosphatidylinositol (LPI), phosphodimethylethanolamine (P-DME), phosphopantothenate (PPA), dimethyl-2-oxoglutarate (MOG), tetrahydrobiopterin (BH4), O- phosphoethanolamine (PEA), stearic acid (SA). Scale bar, 600 pm.

[0031] Figure 8 Intertumoral heterogeneity of metabolites in five tissue cores from HER2/neu positive patients with gastric cancer, visualized via spatial correlation networks (A-E). Edges represent positive (blue) and negative (red) spatial correlations between metabolites. Line thickness and transparency correspond to the correlation coefficient. Right: Zoom-in to glucose 6-phosphate. Abbreviations: glucose 6-phosphate (G6P), glycerol 3-phosphate (Gly3P), phospholipid (PC), phosphatidylinositol (PI), cyclic phosphatidic acid (CPA), lysophosphatidic acid (LPA), lysophospholipid (LPC), lysophosphatidylinositol (LPI), phosphodimethylethanolamine (P-DME), stearic acid (SA), dihydroxyacetone phosphate (DHAP), alanylglutamine (A-Q), histidinyl-glycine (N-HG). Scale bar, 600 pm.

[0032] Figure 9 Immunohistochemistry staining of mouse pancreas: glucagon (red), insulin (green), DAPI (blue). The highlighted regions were analyzed with MALDI imaging prior to immunostaining.

[0033] Figure 10 Immunohistochemistry staining of mouse pancreas superimposed by molecular distribution of adenosine diphosphate (yellow). Glucagon (red), insulin (green), DAPI (blue). The highlighted regions were analyzed with MALDI imaging prior to immunostaining.

[0034] Figure 11 Immunohistochemistry staining of mouse pancreas superimposed by molecular distribution of cholesterol sulfate (yellow). Glucagon (red), insulin (green), DAPI (blue). The highlighted regions were analyzed with MALDI imaging prior to immunostaining. [0035] Figure 12 Immunohistochemistry staining of mouse pancreas superimposed by molecular distribution of 3-O-sulfogalactosylceramide (yellow). Glucagon (red), insulin (green), DAPI (blue). The highlighted regions were analyzed with MALDI imaging prior to immunostaining.

[0036] Figure 13 Intensity distribution per islet of Langerhans and cell type for adenosine diphosphate. The 25^th percentile, median and 75^th percentile are highlighted. SD: standard deviation; width: 75^th percentile - 25^th percentile.

[0037] Figure 14 Intensity distribution per islet of Langerhans and cell type for cholesterol sulfate. The 25^th percentile, median and 75^th percentile are highlighted. SD: standard deviation; width: 75^th percentile - 25^th percentile.

[0038] Figure 15 Intensity distribution per islet of Langerhans and cell type for 3-O- sulfogalactosylceramide. The 25^th percentile, median and 75^th percentile are highlighted. SD: standard deviation; width: 75^th percentile - 25^th percentile. The heights of the bars that exceed the y-axis limit of 20 is written per bar.

[0039] Figure 16 Immunohistochemistry staining of a human gastric cancer tissue slide: DAPI (blue), pancytokeratin (green), HER2/neu (red). The highlighted regions were analyzed with MALDI imaging prior to immunostaining. Tissue folds were excluded (black).

[0040] Figure 17 HER2/neu positive (red) and negative (yellow) tumor region for the tissue depicted in Figure 16.

[0041] Figure 18 Immunohistochemistry staining of a human gastric cancer tissue slide after MALDI imaging: DAPI (blue), pancytokeratin (green), HER2/neu (red). Tissue folds and artifacts were excluded (black).

[0042] Figure 19 HER2/neu positive (red) and negative (yellow) tumor region for the tissue depicted in Figure 18.

[0043] Figure 20 Immunohistochemistry staining of a human gastric cancer tissue slide after MALDI imaging: DAPI (blue), pancytokeratin (green), HER2/neu (red). Tissue folds, artifacts and normal epithel were excluded (black).

[0044] Figure 21 HER2/neu positive (red) and negative (yellow) tumor region for the tissue depicted in Figure 20.

[0045] Figure 22 Immunohistochemistry staining of human gastric cancer TMA cores after MALDI imaging: DAPI (blue), pancytokeratin (green), HER2/neu (red). Tissue folds and artifacts were excluded (black).

[0046] Figure 23 Spatial correlation networks of metabolites in three human gastric tissue sections with gastroesophageal carcinomas. Order corresponding to images in Figure 14. [0047] Figure 24 Spatial correlation networks of metabolites in five tissue microarray cores human gastric tissue sections with gastroesophageal carcinomas. Order corresponding to images in Figure 15.

[0048] Figure 25 Identification of metabolites detected on islets of Langerhans on consecutive mouse pancreatic tissue section (see MALDI images of Figures 11 and 12) compared to standard compounds: A) ADP, B) cholesterol sulfate, C) 3-O-sulfogalactosylceramide (d18:1/16:0) and D) glucose-6-phosphate. The parent ions were isolated and fragmented by MALDI FTICR on-tissue MS/MS using quadrupole collision-induced dissociation (CID).

[0049] Figure 26 imzML-grid image generated from the imzML file. Each red dot represents one pixel/spectrum

[0050] Figure 27 Tissue overview image (co- registered to Figure 26).

[0051] Figure 28 Immunohistochemical staining: Pan-Cytokeratin (co- registered).

[0052] Figure 29 Immunohistochemical staining: Pan-Cytokeratin (co- registered and downscaled).

[0053] Figure 30 Custom-defined masked regions. Regions in black are discarded.

[0054] Figure 31 Pan-Cytokeratin positive (red) and negative (yellow) regions based on user-defined colour threshold

[0055] Figure 32 A: Measurement of the region e.g. tissue samples, B: masked image, C: automatically identified regions marked by unique colors e.g. separated measurements for tissues samples, and D: manually adjusted regions shown as adjusted colors e.g. tissue samples of separate patients.

[0056] Figure 33 Uniquely colored individual measurement regions from a tissue microarray (TMA). Automatically generated from the imzML file.

[0057] Figure 34 Measurement regions from a tissue microarray re-colorized by patient/group. E.g. control group (orange)

[0058] Figure 35: Excerpts of the SPACIAL pipeline Python source code (Python 3.7).

[0059] Figure 36 Python source code (Python 3.7): The Staining class inherits ROI and comprises additional image data/functions. Stainings defined in the config file are created as shown in figure 36.

[0060] Figure 37 Python source code (Python 3.7): Functions to read, filter and scale images.

[0061] Figure 38 Python source code (Python 3.7): code for evaluating set expressions, which are defined in the config file. DETAILED DESCRIPTION OF THE INVENTION

[0062] The inventors developed a new imaging pipeline called Spatial Correlation Image Analysis (SPACIAL), which is a computational multimodal workflow designed to combine molecular imaging data with multiplex immunohistochemistry (IHC) (Figure 1). SPACIAL allows comprehensive and spatially resolved in situ correlation analyses on a cellular resolution. To demonstrate the method, matrix-assisted laser desorption-ionization (MALDI) FTICR imaging mass spectrometry of metabolites and multiplex IHC staining were performed on the very same tissue section of mouse pancreatic islets and on human gastric cancer tissue specimens (Figure 2 and 3). The SPACIAL pipeline was used to perform an automatic, semantic-based, functional tissue annotation of histological and cellular features in order to identify metabolic profiles (Figure 26 to 34). Spatial correlation networks were generated to analyse metabolic heterogeneity associated with cellular features.

[0063] To demonstrate the new method, the SPACIAL pipeline was used to identify metabolic signatures of alpha and beta cells within islets of Langerhans, which are cell types that would not have been distinguishable via morphology alone (Figure 5, 9 to 15, 25). The semantic-based, functional tissue annotation allows an unprecedented analysis of metabolic heterogeneity via the generation of spatial correlation networks. Additionally, the inventors demonstrate intra- and intertumoral metabolic heterogeneity within HER2/neu positive and negative gastric tumor cells (Figure 6, 7, 8, 16 to 24, 25).

[0064] The inventors developed the SPACIAL workflow to provide immunohistochemistry-guided in situ metabolomics on intact tissue sections. Diminishing the workload by automated recognition of histological and functional features, the pipeline allows comprehensive analyses of metabolic heterogeneity. The multimodality of immunohistochemical staining and extensive molecular information from imaging mass spectrometry has the advantage of increasing both the efficiency and precision for spatially resolved analyses of specific cell types. The SPACIAL method is a stepping stone for the objective analysis of high-throughput, multi-omics data from clinical research and practice, that is required for diagnostics, biomarker discovery or therapy response prediction.

[0065] A combined analysis of mass spectrometry and tissue stainings is thereby a valuable diagnostic tool which enables the precise assessment of for instance the medication of cell classes in different tumor layers or the pollution load of different tissue cells.

Stainings

[0066] It is envisaged, that the images of the tissue section may be selected from images obtained by multiplex immunohistochemical (IHC) staining, histochemical staining, endogeneous or exogeneous fluorescence staining, respectively, fluorescence in situ hybridization (FISH), or autofluorescence.

[0067] Histological tissue stain increases the contrast in the optical image and reveals different cells and tissue structures. A wide variety of histological stains are available which differ in their affinity to certain tissue and cell structures and selectively visualize these structures in the optical image. Hematoxylin and eosin (H&E) staining is most commonly used in routine and general investigations. Histology is usually a morphologic diagnostic method because the histologic classification is done according to the appearance and staining properties of the tissue and cell structures. Another common staining is DAPI, which intercalates into DNA and therefore marks nuclei.

[0068] A classification of histological stainings can be limited to one or more subareas of a tissue section, or can even apply to only one or more individual cells. Histologic classification is generally carried out on a stained tissue section of a few micrometres thickness and concerns the tissue type, differentiation, presence of bacterial or parasitic pathogens, disease statuses, and content of foreign matter like pesticides or drugs and their metabolites. The disease progression of human tissue concerns inflammatory disorders, metabolic diseases and the detection of tumours, especially differentiation between benign and malignant forms of tumours.

[0069] A different type of staining is an immunostaining or immunofluorescent staining, where for instance the distribution of proteins in tissues and cells are visualized by the specific binding of antibodies to certain proteins. As would be understood by those of ordinary skill in the art, the term "antibody" is used herein as a protein comprising one or more polypeptides substantially or partially encoded by immunoglobulin genes or fragments of immunoglobulin genes. A classical immunofluorescence staining is further based on the functional affinity of a second antibody to bind to a first antibody. Said second antibody usually carries a (fluorescent) label to mark the structures recognized by the first antibody. In the context of the present invention, compared to efforts to define regions of interest or distinguish cell types based on molecular distributions, immunostainings represent a method to precisely label specific cell types. Further stainings may be include one or more dye(s) which is/ are excitable through UV or visible light and known in the art.

[0070] An endogenous or exogeneous staining as used in the present invention refers to genetic reporter systems expressing a reporter molecule. Such reporter molecules include genes that induce visually identifiable characteristics including fluorescent and luminescent proteins. Examples include the gene that encodes jellyfish green fluorescent protein (GFP), which causes cells that express it to emitt green light under blue/ UV light, luciferase, which catalyses a reaction with luciferin to produce light, and the red fluorescent protein from the gene dsRed. As outlined herein, any kind of light signal of a tissue sample may be applicable to the method of the present invention. Further examples of endogenous reporter molecules are a fluorescent protein; a bioluminescence-generating enzyme, preferably NanoLuc, NanoKAZ, TurboLuc, Cypridina, Firefly, Renilla luciferase, split luciferase, split APEX2 or mutant derivatives thereof; an enzyme, which is capable of generating a coloured pigment, preferably tyrosinase or an enzyme of a multi-enzymatic process, more preferably the violacein or betanidin synthesis process, a genetically encoded receptor for multimodal contrast agents, preferably Avidin, Streptavidin or HaloTag.

[0071] Traditionally, when morphology alone is not sufficient, even clinical pathologists resort to immunohistochemistry (IHC) to localize proteins or peptides in a single tissue section. The immunohistochemistry technique is currently used to classify tumours or to perform structural tissue analyses to help pathologists establish a diagnosis. With the availability of tissue samples from large- scale clinical cohort studies, manual preprocessing steps can easily take up to weeks or months of work. Furthermore, manually annotating tumor regions is not only time-consuming, non-objective, and irreproducible, but also requires extensive histology knowledge that only expert pathologists possess. Additionally, it only permits the annotation of regions and cell types that are histologically distinguishable, while molecular alterations often do not manifest morphologically.

[0072] Historically, this technique has been performed individually for each marker of interest. However, molecular histopathology has recently been shifting from single-marker immunohistochemistry towards multiplexed marker detection. Thus in the present invention multiplex immunohistochemistry referes to the use of multiple markers. Multiplex immunohistochemistry, also called multiple immunolabeling, or multiplex immunostaining, can maximize the amount of data acquired from an individual sample. This is critical in instances where sample is limited, such as a tumor biopsy or other clinical specimen. Multiplex immunohistochemistry also allows for examination of spatial arrangement of proteins of interest as well as protein interaction/co-localization.

[0073] A different labelling method concerns the so-called in-situ hybridization which employ specific DNA probes (DNA = deoxyribonucleic acid) to hybridize to their counterparts. Immunostaining and in situ hybridization are usually highly specific, so not only morphologic information but also molecular information can be derived. Such nucleotide sequences fluoresce when hybridized to a nucleic acid containing a target or complementary sequence, but are otherwise non-fluorescent when in a non- hybridized state. Such oligonucleotides are disclosed, for example, in U.S. patent application Publication No. 2003/0113765. Such staining or in-situ hybridization methods enable the analysis of tissues regarding diseases, differentiation, infection with pathogens, and distribution of foreign matter as compared to a control of normally differentiated or healthy tissue.

[0074] Alternatively, the staining mixture may be formed using fluorescent polyamides, and more specifically polyamides with a fluorescent label or reporter conjugated thereto. Such labels will fluoresce when bound to nucleic acids. First optical active monomers were based on optical active amino acid cores, such as L-leucine. Examples of polyamides with a fluorescent label or reporter attached thereto include, for example, those disclosed in Best et al., Proc. Natl. Acad. Sci. USA, 100(21): 12063-12068 (2003); Gygi, et al., Nucleic Acids Res., 30(13): 2790-2799 (2002); U.S. Pat. No. 5,998,140; U.S. Pat. No. 6,143,901 ; and U.S. Pat. No. 6,090,947.

[0075] Luminescent, color-selective nanocrystals may also be used to label cells in a staining mixture for the method of the present disclosure. Also referred to as quantum dots, these particles are well known in the art, as demonstrated by U.S. Pat. No. 6,322,901 and U.S. Pat. No. 6,576,29. These nanocrystals have been conjugated to a number of biological materials, including for example, peptides, antibodies, nucleic acids, streptavidin, and polysaccharides, (see, for example, U.S. Pat. Nos. 6,207,392; 6,423,551 ; 5,990,479, and 6,326,144, each of which is hereby incorporated herein by reference), and have been used to detect biological targets (see, for example, U.S. Pat. Nos. 6,207,392 and 6,247,323). Staining work flow

[0076] In general stainings are produced in the following steps:

(a) the tissue is stabilized by chemical fixation or freezing

(b) a section between 2 and 10 micrometres thick is cut with a microtome or vibratome

(c) the tissue section according to protocols known in the art.

[0077] Tissue stabilization means that the tissue structures, the cells of the tissue itself and even intracellular structures (e.g. cell nuclei, endoplasmic reticulum, mitochondria) remain preserved in the tissue section. Common chemicals for tissue stabilization are e.g. paraformaldehyde or ethanol. A further mechanical stabilization of the tissue can be achieved by embedding the tissue into paraffin or gelatine.

[0078] It is encompassed by the present invention that the tissue section may comprise one or more tissue samples obtained by formalin-fixed paraffin-embedding (FFPE) or obtained by any other chemical or physical fixation technique, comprising ethanol fixation, cryofixation and heat fixation, respectively.

[0079] The structures of the stained tissue sections are imaged or scanned by routine techniques with the aid of light-optical microscopes and scanners. Thus, by using these routine methods an image is obtained which means in the context of microscopic images a picture is taken with a microscope.

[0080] Images of tissue samples taken with a laser microscope show autofluorescence due to natural emission of absorbed light by biological structures such as mitochondria and lysozymes.

[0081] All information obtained from a image of a tissue section will be defined in the present invention by the term "light-optical information". A light-optical image of the tissue section can have a spatial resolution of about 250 nanometres, which means that structures of the corresponding size are spatially resolved. As used in the present invention and in terms of digital images, spatial resolution refers to the number of pixels utilized in construction of the image. Images having a higher spatial resolution are composed with a greater number of pixels than those of lower spatial resolution. With electron-optical imaging, or more recent optical fluorescence methods such as STED microscopy (STED = Stimulated Emission Depletion), the spatial resolution of the optical image can be further increased, i.e. even smaller structures can be spatially resolved.

Mass spectrometry

[0082] The present disclosure comprises the use of histology and/or immunostainings as well as in situ hybridization in combination with mass spectrometry data as described elsewhere herein. In one preferred embodiment of the present invention the resulting image information of said stainings and mass spectrometry data may be analysed in combination.

[0083] The present disclosure comprises the use of histology and/or immunostainings, in situ hybridization, fluorescent polyamides, and color-selective nanocrystals in combination with mass spectrometry data as described elsewhere herein. In one preferred embodiment of the present invention the resulting image information of said stainings and/or in situ hybridization, fluorescent polyamides, and color-selective nanocrystals and mass spectrometry data may be analysed in combination.

[0084] Mass spectrometry is a widely known technique used in chemical and biochemical analysis to detect and identify molecules of interest in a sample. In recent years, molecular imaging by mass spectrometry has developed, which allows visualizing the distribution of molecules of interest directly in a sample. Mass spectrometry (ISM or IMS) imaging brings together all imaging technologies that use an ionization source to locate molecular ions from a sample. Multiple sources of ionization, such as laser, ions, gas, liquid, solvent, plasma (single or combined sources), microwaves, electrons, which can be used in imaging mode, can be mentioned such as DESI ("Desorption Electrospray Ionization"), LAESI ("Laser Ablation Electrospray Ionization"), MALDI ("Matrix Assisted Laser Desorption Ionization"), SIMS ("Secondary ion Mass Spectrometry"), Matrix-assisted Laser Desorption Electrospray Ionization (MALDESI) and Liquid Extraction Surface Analysis (LESA), ICP-MSI (Inductively Coupled Plasma Mass Spectrometry Imaging). It is envisaged by the present invention that these methods may produce data combined with the staining and imaging techniques discussed somewhere else herein which is analysed together by using the method of the present invention.

[0085] In some embodiment of the method of the present invention, the mass spectrometry data is obtained by any one of MALDI, DESI, LAESI, SIMS, MALDESI, LESA, MSI and ICP-MSI, respectively. In view of the present invention the mass spectrometry data may be obtained by Matrix-assisted Laser Desorption (MALDI), Desorption Electrospray Ionization (DESI), Laser Ablation Electrospray Ionization (LAESI), Secondary ion Mass Spectrometry (SIMS), Matrix-assisted Laser Desorption Electrospray Ionization (MALDESI), Liquid Extraction Surface Analysis (LESA), Mass spectrometra imaging (MSI) or Inductively Coupled Plasma Mass Spectrometry Imaging (ICP-MSI), respectively.

[0086] In matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, a suitably prepared biological tissue sample is coated with a matrix solution and exposed to laser radiation in a vacuum (see Caprioli RM, Farmer TB, and Gile J., Molecular imaging of biological samples: Localization of peptides and proteins using MALDI-TOFMS, Analytical Chemistry, 69(23):475 1- 4760,1997. doi:10.1021/ac970888i). In this process, biological macromolecules are ionized and extracted from the tissue. The ionized macromolecules typically have a positive charge. The ions are accelerated in an electric field and recorded by a detector. The m/z value, i.e. the mass-to-charge ratio of the ionized molecule, can be determined from the time of flight of the ions from the tissue to the detector, which is itself determined from the ionizing laser pulse and the detector signal. The relative number of recorded ions (spectral intensity), as a function of the m/z value, represents a mass spectrum. Assuming a single positive ionization of the molecules, the m/z value is identical to the mass of the ionized molecule. The mass of the molecules is given in daltons (Da) as a multiple of the atomic mass unit (1 Da=1 amu).

[0087] Matrix-assisted laser desorption and ionization (MALDI) has been used successfully for the determination of molecular masses, and structural characterization of biological substances, e.g. proteins, nucleic acids, lipids, sugars or drugs. If the concentrations of the substances are sufficiently high, the concentration patterns can be detected by a mass spectrometric analysis. Thus, a cell can be characterized by concentration patterns of substances e.g. molecular information. An unusual pattern can result when certain biological substances are under expressed or overexpressed.

[0088] Characteristic concentration patterns can be determined by homogenizing a tissue samples by methods known in the art. In more detail, the substances contained therein are prepared and applied to a sample support together with a solution of a matrix substance. The solvent is evaporated and the matrix substance crystallizes; the biological substances in the matrix crystals crystallize at the same time in the form of widely spaced individual molecules. Bombarding a homogenized sample thus prepared with short laser pulses of sufficient energy causes the matrix substance to explosively vaporize and the biological substances to be ionized.

Imaging mass spectrometric (IMS) workflow and analysis

[0089] Imaging mass spectrometric (IMS) analysis, i.e, acquiring a mass spectrometric image, involves investigating tissue sections instead of homogenized tissue samples. Matrix-assisted laser desorption- ionization (MALDI) imaging mass spectrometry (IMS) can be used for in situ imaging of metabolites from frozen or formalin-fixed, paraffin-embedded (FFPE) tissue samples.

[0090] Mass spectrometry imaging is currently mainly used for the analysis of biological tissues. Indeed, it is possible thanks to the ISM to directly study the molecular composition of a tissue or a section thereof, without fluorescence labeling and without radioactivity. In addition, because of its specificity, ISM makes it possible to discriminate and identify ions detected directly on the sample. Thus, it is now common to use ISM for the study or search for endogenous molecular markers in biological samples of interest. More precisely, it is possible to directly analyze the distribution of a known molecule by targeting an ion or its mass-to-charge ratio (m / z). It is also possible to use statistical tools, and in particular ACP (Principal Component Analysis), PLSA, T-Test, ANOVA, or other, to compare at least two regions of interest and thus identify one or more molecules specific to either of these regions (J. Stauber, et al., J. Proteome Res, 2008. 7 (3): 969-78, D. Bonnel et al., Anal Bioanal Chem, 2011). It is also known to use segmentation methods when analyzing the spectra of an image of a sample, in order to classify the spectra as a function of the intensity of the ions of said sample. Thus, it is known to characterize a biological tissue according to the molecular intensity profile of the detected histological zones (T. Alexandrov et al., J. Cancer Res Clin Oncol, 2012. 139 (1): 85-95; Alexandrov et al., J Proteomics, 2011 . 75 (1): 237-45). It is then possible to classify tissues according to their complete molecular profiles and not only according to the spectral profile of some markers.

[0091] In imaging mass spectrometric (IMS) analysis, a tissue section is placed on an electrically conductive sample support. A suitable method is then employed to apply a matrix solution onto the tissue section. Patent specification DE 10 2006 019 530 B4 elucidates different methods of preparing tissue sections for imaging mass spectrometric analysis. The matrix solution or a recrystallization solution can be applied to the tissue section by pneumatic spraying, nebulizing by vibration or by the nanospotting of droplets, for example. It is no trivial task to apply the matrix solution because (a) lateral smearing of the biological substances must be avoided, (b) the biological substances must preferably be extracted from the tissue section and incorporated into the crystals of the matrix layer, and (c) a favourable ratio of biologically relevant substances to impurities must be achieved. The process of applying the matrix substance to the tissue section, and the effect it has on the tissue section, means that mass spectrometric images of tissue sections are currently limited to a spatial resolution of between twenty and two hundred micrometres. This spatial resolution is more than an order of magnitude worse than that of the optical images used in conventional histology.

[0092] Once the matrix solution has dried, the sample support is introduced into a mass spectrometer. The Caprioli raster scan method (US 5,808,300 A) or stigmatic imaging of a small region of the tissue (Luxembourg et al., Analytical Chemistry, 76(18), 2004, 5339-5344: "High-Spatial Resolution Mass Spectrometric Imaging of Peptide and Protein Distributions on a Surface") can be used for the subsequent imaging mass spectrometric analysis. Both techniques produce a mass spectrometric image of the tissue section, i.e, the molecular information in the mass spectra is spatially resolved. Thus, a mass spectrometry image is obtained when a sample is prepared as mentioned above, a mass spectrometer has scanned the sample and created tissue section image. The types of tissue section images obtained are usually superimposed in a graphical representation, in which the spatially resolved mass spectra are often reduced to individual selected masses or to an assignment of certain classes, based on statistical analysis.

[0093] Providing spatially resolved and label-free detection of hundreds to thousands of molecules within a single tissue section, MALDI imaging has proven to be an invaluable tool for digital histopathology. However, the spatial resolution of molecule distributions is often not fully utilized. For instance, cell and tissue specific structures, such as tumours, are frequently analysed by manually annotating the respective areas on whole-slide images, to mark so-called regions of interest (ROIs). The spatially resolved mass spectra in these regions are lost, because only the mean or representative spectrum of each ROI is used for subsequent calculations. Such segmentation approaches fail to preserve molecular heterogeneity and spatial distribution, while whole-sample-based classifications fail in analysing tissue sections which comprise cells belonging to different classes.

[0094] Three different methods are known for coupling histology based optical images with images based on mass spectrometry (Bruker Application Note #MT-89: "Advances in Molecular Histology with the MALDI Molecular Imager"). Firstly, it is possible to take an optical image of one tissue section and a separate mass spectrometric image of an adjacent tissue section from the same tissue sample. The mechanical tolerances for the production of two tissue sections mean that two adjacent tissue sections are generally not sufficiently congruent, so spatial correlation of the two images is only possible to a very limited extent. The second method is to first acquire an optical image and then a mass spectrometric image of a single tissue section. In this case, staining the tissue section must not influence the extraction of the biological substances and their subsequent ionization. Since most histologic stains do not fulfil these requirements and reduce the information content of the mass spectra, this method is seldom used. Thirdly, a mass spectrometric image can be acquired first and an optical image later on. The matrix layer applied to the tissue section is removed again after the mass spectrometric image has been acquired. Then the tissue section is subjected to routine histologic staining, and an optical image is taken. Automated analysis

[0095] In general image analysis is known to those skilled in the art. Gonzalez, R. et al., 2017 ("Digital Image Processing, Global Edition") is a textbook for studying Image Processing and Computer Vision and Russ, J.C. et al., 2017 ("The Image Processing Handbook, 7th Ed., 2017) provides an overall introduction to computer-based image processing which are both hereby incorporated by reference.

[0096] Computational automation of routine tasks and artificial intelligence guided analyses rapidly gain significance with the increasing amount of data generated from single tissue samples. With the rise of digital pathology, a major objective for precision medicine is the integration of morphological and molecular imaging data from multi-omics studies.

[0097] IHC stainings are commonly used for cell type labelling, but their potential for automated, semantic-based, functional tissue annotation and spatially resolved molecular analyses of heterogeneity is not fully utilized. In recent years, imaging mass spectrometry data and immunohistochemical stainings were successfully combined to increase the resolution of MALDI images, or to characterize individual dissociated cells, but no in situ tissue analysis with automatic identification of regions of interest and data integration has been presented. While there is some software available for tissue image analysis, there currently is no method that integrates and analyses the comprehensive molecular data from imaging mass spectrometry in combination with morphological, proteomic, and genetic information from other omics fields. The translation of imaging mass spectrometry into experimental clinical applications requires time-efficient data post-processing and comprehensive analyses of spatially resolved molecular information by avoiding expensive manual annotations or loss of resolution due to the generation of mean or representative spectra.

[0098] In research, the use of immunostainings in combination with molecular data would represent a significant improvement in scientific quality, by solving the problem of time-consuming and irreproducible user-defined ROIs. In particular, the analysis of metabolic heterogeneity is hampered by pseudoheterogeneity originating from inaccuracies during manual annotation or by the use of consecutive tissue sections. The analysis of metabolic heterogeneity in tissues requires a strict coherence to consistent tissue and data preprocessing which the present invention can offer. In particular, is is envisaged that mass spectrometry data and image data may be compared, respectively, extracted from the final ROIs comprises assessing of molecular composition and heterogeneity expressed between data extracted from final ROIs, wherein molecular composition comprises metabolites, proteins, drugs, toxic agents, carcinogens, glycanes, and lipids, respectively.

[0099] Preferred examples of known anti-cancer drugs are cis-platin, maytansine derivatives, rachelmycin, calicheamicin, docetaxel, etoposide, gemcitabine, ifosfamide, irinotecan, melphalan, mitoxantrone, sorfimer sodiumphotofrin II, temozolmide, topotecan, trimetreate glucuronate, auristatin E vincristine and doxorubicin; and peptide cytotoxins such as ricin, diphtheria toxin, pseudomonas bacterial exotoxin A, DNAase and RNAase; radio-nuclides such as iodine 131 , rhenium 186, indium 111 , yttrium 90, bismuth 210 and 213, actinium 225 and astatine 213; prodrugs, such as antibody directed enzyme pro-drugs; immuno-stimulants, such as IL-2, chemokines such as IL-8, platelet factor 4, melanoma growth stimulatory protein, etc., antibodies or fragments thereof such as anti-CD3 antibodies or fragments thereof, complement activators, xenogeneic protein domains, allogeneic protein domains, viral/bacterial protein domains and viral/bacterial peptides. For the treatment of solid tumours, genes encoding toxic peptides (i.e., chemotherapeutic agents such as ricin, diphtheria toxin and cobra venom factor), tumour suppressor genes, such as p53, genes coding for mRNA sequences, which are antisense to transforming oncogenes, antineoplastic peptides, such as tumour necrosis factor (TNF) and other cytokines, or transdominant negative mutants of transforming oncogenes may be non limiting examples.

Spatial Correlation Image Analysis (SPACIAL)

[00100] In the present invention a Spatial Correlation Image Analysis (SPACIAL) pipeline is presented, a computational multimodal workflow to integrate molecular imaging mass spectrometry data with multiplex IHC stainings from the very same tissue section to provide automated and reliable annotations and allow comprehensive and pixel-accurate correlation analyses of heterogeneity to combine data from multi- omics studies. The pipeline represents a starting point for the objective analysis of high-throughput data from clinical research and practice, which is required for tissue-based diagnostics and research.

[00101] To demonstrate the versatility and analytical power of the SPACIAL method, the inventors deliberately chose two examples of molecular heterogeneity in both a physiological and a pathophysiological application: First, the inventors performed a high-resolution analysis of islets of Langerhans in mouse pancreas. Phenotypic and functional beta cell heterogeneity has been shown to provide pancreatic islets with functional flexibility to adapt to physiological changes in the environment. The metabolomics analysis of islet and islet cell heterogeneity requires in situ analyses of intact islets within tissue slices and it has been insufficiently studied in their natural histological context. With SPACIAL, the inventors distinguish alpha and beta cells and investigate the heterogeneity of different islets within one animal (Example 1 and 2). Second, an analysis of tissue samples from patients with gastric cancer was carried out. The metabolomics, intratumoral, heterogeneous nature of the human epidermal growth factor receptor 2 (HER2/neu) is insufficiently studied in s/fu-especially in relation to gastric cancer-even though it is highly relevant for diagnostics and response to HER2/neu-based treatment. The SPACIAL pipeline was applied on tissue resection specimens and on a tissue microarray to distinguish HER2/neu positive and negative tumor cells and to investigate the molecular intra- and intertumoral heterogeneity (Example 3 and 4). The multimodal approach utilizes pixel-wise molecular information to investigate metabolic heterogeneity via spatial correlation networks from cell populations automatically identified by multiplex immunohistochemical analysis.

SPACIAL Image data processing

[00102] The term high-content analysis is frequently used to describe the combining of approaches from image processing, computer vision, and machine learning to provide fast and objective methods for analyzing large amounts of bioimage data. Spatiotemporal events within a cell can be captured by microscopy and quantified through image processing and machine learning methods to produce meaningful conclusions about the data within the experimental context. The bioimage data for high- content analysis are typically collected using fluorescent tags or stains to identify points of interest within the cells being imaged. By combining high-throughput cell biology and automated image analysis, a much larger number of experiments can be performed, and the subjectivity of the experimental observer can be minimized.

[00103] Image data processing as part of the SPACIAL pipeline comprises in brief defining local signal maxima, correcting against noise level, selecting maxima, matching the selected maxima with the tissue stainings, combining the data and converting it into a numerical matrix, and optionally calculating a correlation network.

[00104] In some embodiment of the method of the present invention, the mass spectrometry data may be represented in imzML format, mis format of fleximaging, slx/sbd format of SCiLS Lab, or a tabular format, respectively.

[00105] In a first step a Bruker fleximaging comprises a software tool for two-dimensional peptide and protein imaging of cells or tissue extracts (Bruker fleximaging (v. 5.0) Software manual) was used to export the mass spectra images raw data as a standardized data format called imzML.

[00106] There are relatively few common Pre-Mass Spectrometry formats (i.e. those not specific to one vendor) for information specifically focused on information prior to MS. There are broadly two categories of such formats: one that describes sample handling, and one that contains formats for mass spectrometer target input. Examples of tools and their related formats are: ProteoWizard mzML, TraML, mzldentML, mzXML, vendor formats, OpenMS mzML, TraML, mzldentML, mzData, mzQuantML, Trans- Proteomic Pipeline (TPP) mzML, mzXML, pepXML, protXML (ProteoWizard) compomics-utilities MSF, tandem, mzML, omx, dat, FASTA, jmzReader mzML, mzXML, mzData, PRIDE XML, dta, MGF, ms2, pkl, jTraML TraML, multiplierz Vendor formats, PEFF Viewer PEFF PRIDE Converter 2 mzTab, PRIDE XML (jmzReader), Mascot & Distiller MGF, mzML, mzXML, mzldentML, vendor formats SpectraST msp, splib, blib, ASF, mzML, mzXML, pepXML, ProHits PSI-MI (TPP formats), Anubis TraML, mzML, mzXML, Proteios TraML, mzML, mzXML, Skyline .sky, .skyd, mzML, mzXML, vendor formats, ATAQS TraML, mzML, mzXML, Corra APML, mzXML, Java MIAPE API PRIDE XML, mzML, mzldentML, GelML (Deutsch 2012, Spectrometry Proteomics, Molecular & Cellular Proteomics, p. 1612-1621).

[00107] Alternative converters may be Thermo: converter for LTQ- and Orbitrap-based instruments is available for download here and an imzML export function is implemented in the new version of ImageQuest by Thermo Fisher Scientific, Waters: imzML files can be generated using an export filter in Waters’ High Definition Imaging (HDI) software, ABI SCIEX: ‘Analyze’ format can be converted by ‘toimzml’ module. Please contact Jean-Pierre Both, Shimadzu: imzML can be exported in Imaging MS Solution 1.1 by Shimadzu, mzML: imzMLConverter (converts mzML to imzML and can stitch multiple imzML files), mspire - a gem for programming language ruby: the mspire executable has a subcommand tojmzml. (commandline interface) and Guide for conversion of high mass resolution data to imzML by METASPACE. [00108] The imzML format consists of two separate files: one for the metadata and one for the MS data. The metadata is saved in an XML file (TimzML). The mass spectral data is saved in a binary file (*.ibd). The connection between the two files is made via links in the XML file which hold the offsets of the mass spectral data in the binary file. Schramm, T. et al., 2012 ("imzML — A common data format for the flexible exchange and processing of mass spectrometry imaging data", Journal of Proteomics, vol. 75, issue 16, pp 5106-5210) describes a common data format developed for flexible and efficient exchange of mass spectrometry imaging data between different instruments and data analysis software.

[00109] A “slx/s bd format” (SCiLS Lab) is a compressed file that contains model information in XML format. If you want to dig into it, you can open the file in say winzip and extract it. I just renamed the file to filename.zip and extracted it and you can see the model in XML format.

[00110] To define signal peaks (maxima) of each mass spectrometric analysis a python 3 pipeline was written to perform pixel-wise and parallelized peak picking. For each coordinate (i.e., spectrum), the peak picking pipeline began by resampling the mass (^Vz) and intensity values between 75 and 1100 Dalton (Da) with a step size of 0.0005 Da. Intensity values were resampled by choosing the maximum intensity per window. Noise levels were estimated for windows of 10 Da and all peaks falling below their respective noise level were filtered. The noise level was calculated as 2.2 times the 85^th percentile of the intensity values within the window. After noise-filtering, only local maxima were kept as preliminary peaks. Preliminary peaks within each spectrum were merged. The merged peaks of all coordinates were then aligned, if their distance did not exceed [(^m/_z) x delta ppm ] ÷ 1 000000 with delta_ppm = 2. Picked peaks were saved as an imzML file. The coordinates from the imzML file were used to generate an image of the measurement region (imzML-grid) (Figure. 26). The resolution of this image is a multiple of the imzML resolution.

[00111] In the next step pictures were matched together or co-registered. The imzML file of picked peaks was used to create a master image of the MALDI measurement region (imzML-grid). A two-step image co-registration can be achieved by first registering the tissue overview image with visible tissue or matrix-specific markers onto the grid-image and then registering all stainings simultaneously onto the tissue overview image. All additional images were precisely co-registered onto this image, allowing an exact integration and correlation of molecular MALDI data with immunostainings. The co-registration was done with the Landmark Correspondences plugin of FIJI ImageJ (v. 1 .52p). Alternatively, co-registration is also feasible with Adobe Photoshop CC 2019 or the GNU Image Manipulation Program (GIMP, v.2.10.8). A grey-scale tissue overview image and measurement points were exported with fleximaging (v. 5.0) and then fitted onto the master image. The integration of mass spectra and image data is done by coregistering the tissue scanned subsequent to MALDI imaging mass spectrometry or mapping the matrix ablation marks to the imzML-grid. The DAPI staining and all other stainings were finally fitted onto the precisely co-registered tissue image. Once all images are co-registered, they have the same resolution (Figure. 27 and Figure. 28) as the x,y-grid from the imzML file (Figure. 29). To integrate the data from all images, they have to be scaled to the exact MALDI measurement resolution by averaging the colour values per x/y-coordinate. Downscaling is done by calculating the mean colour value per square pixel region. Optionally, pixels with no other colour pixel within x±2 and y±2, are removed.

[00112] Thus, in a first aspect, the present invention relates to a method of selecting regions of interest, ROIs, for analyzing a complete tissue section, the method comprises obtaining a first image comprising mass spectrometry data of the tissue section spatially resolved by a region of the first image represented as a first mask, and further obtaining one or more second images comprising any type of images of the tissue section, the one or more second images are in alignment with and scaled to the resolution of the first image, the method also comprises determining one or more second masks associated with respective one or more second images based on predetermined criteria, as well as combining masks selected from the first mask and the one or more second masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs and extracting mass spectrometry data from said first image and image data from said one or more second images, respectively, based on the final ROIs, for analyzing the tissue section.

[00113] In the context of the present invention, and in reference to computer programs a mask is a special value that acts as a data filter. It reveals some parts of the digital information, and conceals or alters others.

[00114] An “expression” refers to the terminology used within and run by a program. A regular expression is a pattern that describes a set of strings. Regular expressions are constructed by using various operators to combine smaller expressions.

[00115] Preferably, determining one or more second masks associated with respective one or more second images comprises determining multiple masks for the same image, said masks pertaining to multiple morphometric data, or data from multiple color components, such as Hematoxylin and eosin (H&E) and other histological stainings. Even more preferably, the method further comprises obtaining one or more custom masks of the one or more second images, and wherein combining masks comprises combining masks selected from the first mask, the one or more second masks, and the one or more custom masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs. In view of the present invention, the method referred to herein may comprise determining one or more second masks associated with respective one or more second images based on predetermined criteria comprises using predetermined thresholds of color values or ranges of color features of pixels in an image of the one or more second images to determine regions containing pixels of similar color features in said image.

[00116] SPACIAL offers the option to create a blacklist image, where the user can manually label regions that should be excluded from subsequent analyses. A black and white image can now be created to mask artefacts (e.g. tissue folds) or split measurement regions. Masked regions were excluded from further analyses (Figure. 30). Such regions may comprise tissue folds, swept away tissue, artefacts, or regions that are generally of no interest. [00117] For each staining, the user can now define a threshold to decide at what colour value a pixel is classified as a positive signal (Figure. 31). First, preliminary regions of interest are defined based on stainings, imaging mass spectrometry measurement regions and custom-generated images. Subsequently, those regions can be combined to define the final regions of interest. When the image mask is imported, it is also converted into coordinates. The previously determined colour thresholds were used to generate two preliminary regions (lists of x,y-coordinates) per staining: e.g. staining X positive and staining X negative pixels (Figure. 31).

[00118] Individual imaging mass spectrometry measurement regions may also be considered as preliminary regions. This is important for the analysis of tissue microarrays, where tissue cores of several of patients are measured on one glass slide. Usually each patient sample is measured in a separate measurement region. Nevertheless, if two patient cores are part of the same measurement region, the measurement region can be split by separating them with a black line in the masking image (see paragraph ‘Image masking’) (Figure. 32).

[00119] Additionally, the user can define custom regions, either by creating one black and white image per region, similar to the masking image, or by creating a multicolour image, where each colour except black is considered as one region. This is used for the analysis of tissue microarrays, where multiple cores may belong to the same patient. First, an image is generated, where each measurement region is coloured uniquely (Figure. 33). The user can then manually colorize regions that belong to the same patient/group (Figure. 34).

[00120] In the context of the present invention the term “custom mask” refers to masks as defined somewhere else herein which are not connected to picture data.

[00121] The previously defined preliminary regions were now combined to generate the final regions of interest. This can be achieved via logical/set expressions with the common operators Union, Intersection, Without/Difference and X or. Examples: islet_ROI = roil P (stain ing2 U staining3) amyloid_ROI = roil P (stainingl P (staining2 D staining3)) beta_ROI = roil P (staining2 \ staining3) alpha_ROI = roil P (staining3 \ staining2)

[00122] In some embodiment, the predetermined set expressions may represent semantics of tissue annotation to obtain the final ROIs, based on the first image, the one or more second images and the associated masks of the tissue section.

[00123] In the present invention the term “annotation” relates to metadata linked to e.g. imaging data. Function annotations of e.g. Python 3.0 allow adding arbitrary metadata to function parameters and return value. The purpose is to have a standard way to link metadata to function parameters and return value. A “tissue annotation” refers to the (meta)data collected by the analysis of a tissue sample by using the method of the present invention. The term “semantic” is to be understood in the sense of computer science as language or set of accepted words.

[00124] The mass spectra of each region of interest can now be extracted from the imzML file and used for subsequent calculations and analyses. The IHC images are then converted into numerical matrices comprised of values corresponding to the lightness values for each pixel. SPACIAL can create images to allow validation of semi-automatically defined ROIs (e.g., HER2/neu positive tumor regions). The user does not have to manually select/circle all the tissue regions belonging to a specific cell type. However, the user still has to select a color threshold to distinguish positive and negative cells.

[00125] In the context of the present invention the term “threshold” refers to segmenting an image. From a grayscale image, thresholding can be used to create binary images. The simplest thresholding methods replace each pixel in an image with a black pixel if the image intensity I_jj is less than some fixed constant T (that is, I_jj <T), or a white pixel if the image intensity is greater than that constant.

[00126] Resulting from the numerical matrices networks can be presented as graphs, that is, a set of vertices (V) connected by edges (E), and consequently can be analyzed using graph theory. Today, graph theory consists of many tens of basic definitions and properties. The understanding of the biological networks lies in the nature of the vertices and edges between them; that is, the vertices may represent one of the components of the three major molecular levels: genes, proteins, or metabolites, while the edges between them represent gene coexpression, protein-protein interactions, or biochemical conversions of metabolites, respectively. However, molecular networks are not limited to illustrate single- level component interactions. They can also show cross-level interactions. Alternatively, and perhaps a little counterintuitive, a network may incorporate vertices representing a set of metabolic reactions, where the connection between a pair of vertices is established if the reactions share one or multiple metabolites used or produced by these reactions.

[00127] Correlation networks were created with Cytoscape (v. 3.7.1). In all networks, nodes represent metabolites with node sizes corresponding to the mean intensity. Edges represent spatial correlations with line thickness and opacity increasing with the correlation coefficient. Nodes were coloured red, if their metabolites take part in glycolysis or they were coloured depending on the molecule super class defined in HMDB [lipids and lipid-like molecules (yellow); nucleosides, nucleotides, and analogues (light red); organic acids and derivatives (green); organoheterocyclic compounds (lime green); alkaloids and derivatives (pink); organic oxygen compounds (blue); benzenoids (violet); phenylpropanoids and polyketides (orange); others (grey)]. The multiplex staining of the complete tissue is shown in Figure 9. All remaining networks were generated by showing metabolites with at least one correlation coefficient larger than 0.5, but without filtering edges. The complete networks are shown in Figures 23 and 24. The multiplex staining of the complete tissues is shown in Figures 16, 17, 20, and 22.

[00128] It is further envisage that analyzing the tissue section may comprise comparing mass spectrometry data and image data, respectively, derived from the final ROIs of the first image and the one or more second images of the tissue section, by statistical analysis. [00129] As refered to herein, a statistical analysis may be applicable for any value correction. Examples may be the correlation networks, which may be analysed by using pairwise Spearman rank-order correlations (Python 3.7, SciPy 1.2.0) were calculated between annotated metabolites using their intensities, and the resulting p-values may be adjusted with Benjamini/Hochberg correction (Python 3.7, StatsModels 0.9.0). For the pancreatic islet cells, circular networks were generated by filtering edges with a coefficient smaller than 0.7. Network metrics may be calculated using Cytoscape’s plugin NetworkAnalyzer. Metabolites may be identified by using the Mann-Whitney U-test (Python 3.7, SciPy 1.2.0). The p-values may be adjusted with Benjamini/Hochberg correction (Python 3.7, StatsModels 0.9.0). The Python 3.7 package NumPy 1.15.4 may be used to calculate statistics for the intensity distributions.

[00130] The inventor’s SPACIAL workflow integrates molecular MALDI imaging mass spectrometry data with IHC stainings to facilitate automatic, reliable, and pixel-accurate annotation of specific cell types. In this context, the phenotypical information provided by immunohistochemistry complements in situ molecular information for cell type specific evaluation. The pipeline was demonstrated for both physiological and pathophysiological applications to investigate metabolic heterogeneity in alpha and beta cells from islets of Langerhans of a mouse model and in HER2/neu positive tumor cells from patients with gastric cancer.

[00131] Glucagon releasing alpha and insulin releasing beta cells of different pancreatic islets within one animal were automatically annotated, demonstrating the basic functionality of the SPACIAL pipeline as a tool for objective immunohistochemistry-guided annotation of otherwise histologically indistinguishable cell types. The pixel-accurate annotation and analysis of metabolites allows previously infeasible assessments of metabolomics heterogeneity between islets of Langerhans.

[00132] Additionally, tissue samples from patients with gastric cancer were chosen to demonstrate the methodological advantages of SPACIAL for the analysis of intra- and intertumoral heterogeneity. The SPACIAL strategy can be extended by integrating other in situ datasets from tissue analytic platforms, since all spatially resolved information of a tissue section can be integrated in this pipeline (e.g., morphometries, fluorescence in situ hybridization, and imaging mass cytometry). Prospectively, the application can also be useful for the automatic readout of regions of interest for metabolite quantification on an absolute, rather than on a relative scale. Quantification is a major topic of investigation in the targeted MALDI IMS field concentrating on the analysis of a subset of metabolites. Furthermore, the workflow was demonstrated to be compatible with both frozen and FFPE tissue samples. With SPACIAL, hundreds of distinct samples within tissue microarrays can be analysed simultaneously. In contrast to the traditional analysis of mean spectra per ROI, SPACIAL allows in-depth and full use of available data without loss of resolution. With the spatial correlation networks of metabolites and the comparative approach to investigate islet cell heterogeneity, the inventors demonstrate one of the many possibilities to utilize MALDI data. Combining the data from multi-omics studies, the pipeline represents an important starting point for the objective analysis of high-throughput data from large-scale clinical cohort studies, which are required for artificial intelligence guided diagnostics, biomarker discovery, or therapy prediction. [00133] Preferably, the tissue section comprises multiple tissue samples individualized by one or more custom masks, wherein each custom mask is associated with an individual tissue sample.

[00134] The method of the present invention may be applied to more than one tissue sample of an individual (e.g. tissue microarray) and tissue regions are managed by automatically generating and subsequently customizing masks. Individual tissue samples are automatically represented in the microarray of figure 32 in color code (C) and the correction thereof is done manually (D).

[00135] In a second aspect the present invention relates to the use of the method for diagnosis or stratification based on a human, animal or plant tissue samples. Preferably, ranges of color values of pixels are based on a magnitude of expression of a diagnostic feature, such as one or more biomarkers. It is also envisaged that in the method of the present invention the biomarker is HER2/neu and/or pan- cytokeratin (PC); Insulin and/or glucagon; PD1 and/or PD-L1 ; or Vimentin and/or pan-cytokeratin. In some embodiment the use of the method may comprise that the samples are human tissue samples used in the diagnosis or stratification of cancer or diabetes. In a third aspect the present invention relates to the use of the method for therapy response prediction and prediction of organ rejection based on human or animal tissue.

[00136] One example of stratification using the method of the invention is the application for prognostic risk stratification of neoadjuvant treated cancer patients, such as esophageal adenocarcinoma patients. Specifically, the methods of the invention can be used to address the question of whether metabolic tumor profiling of neoadjuvantly treated esophageal adenocarcinoma (EAC) can contribute to patient stratification into different prognostic risk groups. It is known that response to neoadjuvant therapy can vary widely between individual patients. Tumor regression grading (TRG) and nodal status are standard clinical-pathological prognostic factors used to predict the survival of esophageal adenocarcinoma (EAC) patients following neoadjuvant treatment and surgery. Neoadjuvant chemoradiotherapy or chemotherapy is associated with a significant survival benefit for patients compared to surgery alone and has become the standard of care for most patients with resectable esophageal and gastroesophageal-junction adenocarcinoma. Preoperative treatment has the effect of tumor and nodal downstaging, which can increase the prospect of complete resection. Despite the advantages achieved through multimodal therapy, the outcome of a significant proportion of patients with advanced esophageal adenocarcinoma (EAC) remains unsatisfactory. Insights in the tumor's constitution and its molecular composition after neoadjuvant treatment are valuable to define risk groups and in order to improve future therapies for patients with a poor response.

[00137] A commonly used method to assess the response to neoadjuvant therapy is histological tumor regression grading (TRG), which, in addition to the presence or absence of lymph node metastases, is an important clinically prognostic indicator of patient survival. With regard to TRG, several systems have been proposed in gastrointestinal malignancies aiming to categorize the extent of regressive changes after cytotoxic treatment by estimating the amount of residual tumor and the degree of therapy-induced fibrosis. Thereby, the regressive tissue changes in resected tumor observed after neoadjuvant therapy can range from complete regression to varying amounts of a vital residual tumor. In addition, it has been suggested that assays based on gene expression profiles or microsatellite instability status could help to further improve prognostic risk stratification of patients. However, no study has addressed metabolites in terms of their prognostic impact to distinguish patients into those who benefit versus not benefit from neoadjuvant EAC treatment. Spatial metabolomics analyses according to the invention have already shown promise as a tool to gain new insights into neoplastic progression to EAC, for the assessment of surgical resection margins, or investigating tumor heterogeneity.

[00138] It is hypothesised that by performing a metabolomics analysis according to the invention on neoadjuvant chemoradiotherapy treated patient resection specimens, different prognostic risk groups can be identified. For this purpose, high-resolution matrix-assisted laser desorption/ionization Fourier- transform ion cyclotron resonance imaging mass spectrometry (MALDI FTICR IMS) can be used to detect the in situ distribution of metabolites with high sensitivity and specificity from tissue sections. To identify prognostic metabolite markers, a spatial correlation image analyses according to the invention could be performed on tumor tissue samples, allowing for identification of metabolite features in tumor and stroma.

[00139] The method of the present invention can be employed for detection, diagnosis, prognosis, prevention and/or as control device in the treatment of diseases or disorders. The term “treatment” in all its grammatical forms includes therapeutic or prophylactic treatment of a subject in need thereof. A “therapeutic or prophylactic treatment” comprises prophylactic treatments aimed at the complete prevention of clinical and/or pathological manifestations or therapeutic treatment aimed at amelioration or remission of clinical and/or pathological manifestations. The term “treatment” thus also includes the amelioration or prevention of diseases.

[00140] As used herein, the term “stratification” refers to the method whereby the sample is subdivided based on the presence or absence of a specific characteristic (unless context dictates otherwise). In this context, a "stratified” sample is a sample obtained by subjecting a source sample to a selection for specific characteristics. A "stratified” sample is therefore in a certain aspect enriched relative to the source sample, to obtain a sample subdivision having these specific characteristic.

[00141] The tissue sample analysed with the method of the present invention may originate from a human, an animal, especially birds and fishes and mammals, or a plant.

[00142] The term “bird” or aves is a group of endothermic vertebrates, characterised by feathers, toothless beaked jaws, the laying of hard-shelled eggs, a high metabolic rate, a four-chambered heart, and a strong yet lightweight skeleton and are to be understood as such in the context of the present invention.

[00143] The term “fish” refers to an animal group of gill-bearing aquatic craniate vertebrates that lack limbs with digits. They form a sister group to the tunicates, together forming the olfactores. Included in this definition are the living hagfish, lampreys, and cartilaginous and bony fish as well as various extinct related groups. [00144] Especially, mammals are a preferred part of the animal kingdom donating tissue for use in the method of the present invention. The term “mammal” includes for instance cats, dogs, horses, pigs, cows, goats, sheep, rodents (e.g. mice or rats, rabbits), primates (e.g. chimpanzees, monkeys, gorillas, and humans) and endangered mammals (such as non-human primates, elephants, rhinos, bears).

Table 1 Network metrics for the glucose 6-phosphate node. Sample identifiers (A, B and C) correspond to the samples in Figure 7. The degree of a node represents the number of neighbors in the network. The average shortest path length is the average, minimum number of edges between glucose 6- phosphate node and any other node. The clustering coefficient is a measure for the connections between neighboring nodes.

Table 2 Network statistics for islets of Langerhans in Figure 28. Statistics were calculated with Cytoscape’s plugin NetworkAnalyzer.

Table 3 Network statistics for gastric cancer tissues in Figure 31. Statistics were calculated with Cytoscape’s plugin NetworkAnalyzer.

Table 4 Network statistics for gastric cancer tissues in Figure 32. Statistics were calculated with Cytoscape’s plugin NetworkAnalyzer.

[00145] Explanation of network metrics in Tables 2-4:

The clustering coefficient (0-1) describes whether nodes in a network tend to form clusters.

The centralization (0-1) of a network describes whether the network has a centre (star shaped). The network’s characteristic path length is the average of the shortest path length between any pair of nodes.

The density (0-1) describes how densely the network is populated with edges.

The heterogeneity (0-1) of a network reflects, whether a network tends to contain hub nodes (i.e. well connected nodes).

[00146] It is noted that as used herein, the singular forms “a”, “an”, and “the”, include plural references unless the context clearly indicates otherwise. Thus, for example, reference to “a reagent” includes one or more of such different reagents and reference to “the method” includes reference to equivalent steps and methods known to those of ordinary skill in the art that could be modified or substituted for the methods described herein.

[00147] The term “about” or “approximately” as used herein means within 20 %, preferably within 10 %, and more preferably within 5 % of a given value or range. It includes, however, also the concrete number, e.g. “about 20” includes 20.

[00148] Unless otherwise indicated, the term “at least” preceding a series of elements is to be understood to refer to every element in the series. The term “at least one” refers, if not particularly defined differently, to one or more such as two, three, four, five, six, seven, eight, nine, ten or more. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the present invention.

[00149] The term “and/or” wherever used herein includes the meaning of “and”, “or” and “all or any other combination of the elements connected by said term”. [00150] The term “less than” or in turn “more than” does not include the concrete number. For example, less than 20 means less than the number indicated. Similarly, more than or greater than means more than or greater than the indicated number, e.g. more than 80 % means more than or greater than the indicated number of 80 %.

[00151] Throughout this specification and the claims which follow, unless the context requires otherwise, the word “comprise”, and variations such as “comprises” and “comprising”, will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integer or step. When used herein the term “comprising” can be substituted with the term “containing” or “including” or sometimes when used herein with the term “having”.

[00152] When used herein “consisting of” excludes any element, step, or ingredient not specified in the claim element. When used herein, “consisting essentially of” does not exclude materials or steps that do not materially affect the basic and novel characteristics of the claim.

[00153] The term “including” means “including but not limited to”. “Including” and “including but not limited to” are used interchangeably.

[00154] The term “about” means plus or minus 10%, preferably plus or minus 5%, more preferably plus or minus 2%, most preferably plus or minus 1%.

[00155] Throughout the description and claims of this specification, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating plurality as well as singularity, unless the context requires otherwise.

[00156] It should be understood that this invention is not limited to the particular methodology, protocols, material, reagents, and substances, etc., described herein and as such can vary. The terminology used herein is for the purpose of describing particular embodiments only, and is not intended to limit the scope of the present invention, which is defined solely by the claims.

[00157] All publications cited throughout the text of this specification (including all patents, patent application, scientific publications, instructions, etc.), whether supra or infra, are hereby incorporated by reference in their entirety. Nothing herein is to be construed as an admission that the invention is not entitled to antedate such disclosure by virtue of prior invention. To the extent the material incorporated by reference contradicts or is inconsistent with this specification, the specification will supersede any such material.

[00158] The content of all documents and patent documents cited herein is incorporated by reference in their entirety. [00159] A better understanding of the present invention and of its advantages will be gained from the following examples, offered for illustrative purposes only. The examples are not intended to limit the scope of the present invention in any way.

The invention of the present application can also be described by the following items:

1. Method of selecting regions of interest, ROIs, for analyzing a complete tissue section, comprising: obtaining a first image comprising mass spectrometry data of the tissue section spatially resolved by a region of the first image represented as a first mask, and further obtaining one or more second images comprising any type of images of the tissue section, wherein the one or more second images are in alignment with and scaled to the resolution of the first image; determining one or more second masks associated with respective one or more second images based on predetermined criteria; combining masks selected from the first mask and the one or more second masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs; and extracting mass spectrometry data from said first image and image data from said one or more second images, respectively, based on the final ROIs, for analyzing the tissue section.

2. The method of item 1 , wherein the mass spectrometry data is obtained by any one of MALDI, DESI, LAESI, SIMS, MALDESI, LESA, MSI and ICP-MSI, respectively.

3. The method of any of the preceding items, wherein the mass spectrometry data are obtained by Matrix-assisted Laser Desorption (MALDI), Desorption Electrospray Ionization (DESI), Laser Ablation Electrospray Ionization (LAESI), Secondary ion Mass Spectrometry (SIMS), Matrix-assisted Laser Desorption Electrospray Ionization (MALDESI), Liquid Extraction Surface Analysis (LESA), Mass spectrometra imaging (MSI), or Inductively Coupled Plasma Mass Spectrometry Imaging (ICP-MSI), respectively.

4. The method of any of the preceding items, wherein the light-optical images of the tissue section are selected from images obtained by multiplex immunohistochemical (IHC) staining, histochemical staining, endogeneous or exogeneous fluorescence staining, respectively, fluorescence in situ hybridization (FISH), or autofluorescence.

5. The method of any one of the preceding items, wherein the mass spectrometry data is represented in imzML format, mis format of fleximaging, slx/sbd format of SCiLS Lab, or a tabular format, respectively.

6. The method of any one of the preceding items, wherein the tissue section comprises one or more tissue samples obtained by formalin-fixed paraffin-embedding (FFPE) or obtained by any other fixation technique, comprising ethanol fixation and cryofixation, respectively.

7. The method of any one of the preceding items, wherein determining one or more second masks associated with respective one or more second images comprises determining multiple masks for the same image, said masks pertaining to multiple morphometric data, or data from multiple color components, such as Hematoxylin and eosin (H&E) and other histological stainings. The method of any of the preceding items, further comprising obtaining one or more custom masks of the one or more second images, and wherein combining masks comprises combining masks selected from the first mask, the one or more second masks, and the one or more custom masks, respectively, based on predetermined set expressions, to obtain final regions of interest, ROIs. The method of any of the preceding items, wherein determining one or more second masks associated with respective one or more second images based on predetermined criteria comprises using predetermined thresholds of color values or ranges of color values of pixels in an image of the one or more second images to determine regions containing pixels of similar color values in said image. The method of the preceding item, wherein ranges of color values of pixels are based on a magnitude of expression of a diagnostic feature, such as one or more biomarkers. The method of item 10, wherein the biomarker is a) HER2/neu and/or pan-cytokeratin (PC); b) Insulin and/or glucagon; c) PD1 and/or PD-L1 ; or d) Vimentin and/or pan-cytokeratin. The method of any of the preceding items, wherein the predetermined set expressions represent semantics of tissue annotation to obtain the final ROIs, based on the first image, the one or more second images and the associated masks of the tissue section. The method of any of the preceding items, wherein analyzing the tissue section comprises comparing mass spectrometry data and image data, respectively, derived from the final ROIs of the first image and the one or more second images of the tissue section, by statistical analysis. The method of the preceding item, wherein comparing mass spectrometry data and image data, respectively, extracted from the final ROIs comprises assessing of molecular composition and heterogeneity expressed between data extracted from final ROIs, wherein molecular composition comprises metabolites, proteins, drugs, toxic agents, carcinogens, glycanes, and lipids, respectively. The method of any one of the preceding items, wherein the tissue section comprises multiple tissue samples individualized by one or more custom masks, wherein each custom mask is associated with an individual tissue sample. Use of the item according to any one of the preceding claims for diagnosis or stratification based on a human, animal or plant tissue samples. The use of item 16, wherein the samples are human tissue samples used in the diagnosis or stratification of cancer or diabetes. Use of the item according to any one of claims 1 to 15 for therapy response prediction or prediction of organ rejection based on human or animal tissue.

EXAMPLES OF THE INVENTION

[00160] The following Examples illustrate the invention, but are not to be construed as limiting the scope of the invention.

[00161] Methods

[00162] Tissue specimens and FTICR MALDI IMS analysis

[00163] Pancreas/islets of Langerhans were obtained from a C57BL/6N mouse and the sample was flash frozen in liquid nitrogen until measurement. The animal was provided ad libitum access to food and water. All animal studies were conducted in accordance with German animal welfare legislation and approved by the government of Upper Bavaria. FFPE tissue patient samples of gastric cancer were collected between 1995 and 2018 at the University of Leipzig and at the Department of Surgery, Klinikum Rechts der Isar, Munich, Germany. The resection specimens were processed in a highly standardized manner, fixed for 12-24 h in 10% neutral buffered formalin, followed by tissue dehydration and paraffin embedding with fully automated systems. The study was approved by the local Ethics Committees. All patients provided informed, signed consent.

[00164] Tissue preparation steps for MALDI imaging analysis was performed as previously described. In brief, frozen (12 pm, Leica Microsystems, CM1950, Germany) and FFPE sections (4 pm, Microm, HM340E, Thermo Fisher Scientific, USA) were mounted onto indium-tin-oxide (ITO)-coated glass slides (Bruker Daltonik, Bremen, Germany) pretreated with 1 :1 poly-L-lysine (Sigma Aldrich, Munich, Germany) and 0.1% Nonidet P-40 (Sigma). The air-dried tissue sections were spray-coated with 10 mg/ml 9- aminoacridine hydrochloride monohydrate matrix (Sigma-Aldrich, Munich, Germany) in 70% methanol using the SunCollect™ sprayer (Sunchrom, Friedrichsdorf, Germany). Prior matrix application, FFPE tissue sections were incubated additionally for 1 h at 70°C and deparaffinised in xylene (2x8 min). Spraycoating of the matrix was conducted in eight passes (ascending flow rates 10 pl/min, 20 pl/min, 30 pl/min for layers 1-3, and layers 4-8 with 40 pl/min), utilizing 2 mm line distance, and a spray velocity of 900 mm/min.

[00165] Metabolites were detected in negative-ion mode on a 7T Solarix XR Fourier-transform ion cyclotron resonance (FTICR) mass spectrometer (Bruker Daltonik) equipped with a dual ESI-MALDI source and a smartbeam-ll Nd:YAG (355 nm) laser. Data acquisition parameters were specified in ftmsControl software 2.2 and fleximaging (v. 5.0) (Bruker Daltonik). Mass spectra were acquired in negative-ion mode covering ^m/_z 75-1100, with a 1M transient (0.367 sec duration), and an estimated resolving power of 49,000 at m/z 200,000. The laser operated at a frequency of 1 ,000 Hz utilizing 200 laser shots per pixel with a pixel resolution of 15 pm (islets of Langerhans) and 60 pm (gastric cancer), respectively. L-Arginine was used for external calibration in the ESI mode. On-tissue MS/MS was conducted on islets of Langerhans from the consecutive mouse pancreatic tissue section using continuous accumulation of selected ions’ mode and collision-induced dissociation (CID) in the collision cell (Figure 25). MS/MS spectra were analysed by Bruker Compass DataAnalysis 5.0 (Build 203.2.3586). [00166] Multiplex fluorescent immunohistochemical staining

[00167] After MALDI IMS analysis, 9-aminoacridine matrix was removed with 70% ethanol for 5 min from tissue sections followed by immunohistochemical staining. Pancreatic islets were analyzed by double staining for insulin [Insulin-monoclonal rabbit anti-insulin (1 :800), catalogue no. 3014, Cell Signaling Technology, Germany; AF750-goat anti-rabbit (1 :100), catalog no. A21039, Thermo Fisher Scientific, US] and glucagon [polyclonal guinea pig anti-glucagon (1 :3000), catalogue no. M182, Takara, USA; biotinylated goat anti-guinea pig IgG (1 :100), catalogue no. BA-7000, Vector Laboratories, US; streptavidin-Cy3, catalogue no. SA1010, Thermo Fisher Scientific]

[00168] Double staining of human gastric cancer tissue specimens and a tissue microarray was performed using HER2 [polyclonal rabbit anti-human c-erbB-2 oncoprotein (1 :300), catalogue no. A0485, DAKO, CiteAb Ltd, UK] and pan-cytokeratin [monoclonal mouse pan cytokeratin plus [AE1/AE3+8/18] (1 :75), catalogue no. CM162, Biocare Medical, US] Signal detection was conducted using fluorescence- labeled secondary antibodies [goat anti-rabbit IgG (H+L)-Cross-Adsorbed Secondary Antibody-DyLight 633 (1 :200), catalogue no. 35563; and goat anti-Mouse IgG (H+L)-Cross-Adsorbed Secondary Antibody- Alexa Fluor 750 (1 :100), catalogue no. A-21037, both Thermo Fisher Scientific] Nuclei were identified with Hoechst 33342 in all stainings. Fluorescence stainings were scanned with an AxioScan.ZI digital slide scanner (Zeiss) equipped with a 20x magnification objective and visualized with the software ZEN 2.3 blue edition (Zeiss). Multi-images were exported as tiff files. Additionally, tissue sections were stained with hematoxylin and eosin after MALDI and IHC for internal visual validation.

[00169] Peak picking

[00170] The Bruker software fleximaging (v. 5.0) was used to export all root mean square normalized mass spectra as processed imzML files. An in-house python 3 pipeline was written to perform pixel-wise and parallelized peak picking. For each coordinate (i.e. , spectrum), the peak picking pipeline began by resampling the mass (^Vz) and intensity values between 75 and 1100 Dalton (Da) with a step size of 0.0005 Da. Intensity values were resampled by choosing the maximum intensity per window. Noise levels were estimated for windows of 10 Da and all peaks falling below their respective noise level were filtered. The noise level was calculated as 2.2 times the 85^th percentile of the intensity values within the window. If fewer than 200 intensities fell within one window, which frequently happens in the higher mass range, then their neighbouring windows were taken into account until at least 200 intensities can be used for the calculation. Since the noise level is expected to increase with the ^m/_z value and to avoid extreme noise level fluctuations, the level of the first and last window were used as upper bounds. After noise-filtering, only local maxima were kept as preliminary peaks. Preliminary peaks within each spectrum were merged as previously described. The merged peaks of all coordinates were then aligned, if their distance did not exceed [(^m/_z) x delta ppm ] ÷ 1000000 with delta ppm = 2. Peaks that occur in less than 0.5% of the spectra were filtered. Picked peaks were saved as an imzML file. Noise levels and the peak pickings were verified by manual inspection of random sample coordinates.

[00171] Metabolite annotation

[00172] The Human Metabolome Database (HMDB, v. 4.0) was used to functionally annotate ^m/_z values. The metabolite XML file was downloaded for offline use and a local PostgreSQL (v. 11) database was set up. Molecules were annotated by allowing M-H, M-H20-H, M+Na-2H, M+CI and M+K-2H as negative adducts with a mass tolerance of 4 ppm. A keyword search was performed on the description text to filter compounds with multiple annotations. Specifically, compounds with indications of being drug-, plant-, food-, or bacteria-specific were filtered stringently.

[00173] Image co-registration [00174] The imzML file of picked peaks was used to create a master image of the MALDI measurement region (imzML-grid). All additional images were precisely co-registered onto this image, allowing an exact integration and correlation of molecular MALDI data with immunostainings. The co-registration was done with the Landmark Correspondences plugin of FIJI ImageJ (v. 1.52p). Alternatively, co-registration is also feasible with Adobe Photoshop CC 2019 or the GNU Image Manipulation Program (GIMP, v.2.10.8). A grey-scale tissue overview image and measurement points were exported with fleximaging (v. 5.0) and then fitted onto the master image. The integration of mass spectra and image data is done by coregistering the tissue scanned subsequent to MALDI imaging mass spectrometry and mapping the matrix ablation marks to the imzML-grid. The DAPI staining and all other stainings were finally fitted onto the precisely co-registered tissue image.

[00175] Region inclusion/exclusion criteria

[00176] After co-registration, all images had the exact same dimension and resolution. SPACIAL now offers the option to create a blacklist image, where the user can manually label regions that should be excluded from subsequent analyses. Such regions may comprise tissue folds, swept away tissue, artefacts, or regions that were generally of no interest.

[00177] To integrate the data from all images, they have to be scaled to the exact MALDI measurement resolution by averaging the colour values per x/y-coordinate. The IHC images were then converted into numerical matrices comprised of values corresponding to the lightness values for each pixel. SPACIAL can create images to allow validation of semi-automatically defined ROIs (e.g., HER2/neu positive tumor regions).

[00178] Pixel-accurate definition of HER2/neu positive tumor regions

[00179] FFPE tissue sections of human gastric cancer samples were used to analyse the metabolic heterogeneity within HER2/neu positive tumor regions. Tumor cells were annotated via the pan- cytokeratin staining. They were then classified as HER2/neu positive, if they also exhibited a positive signal in the HER2/neu staining. Otherwise, they were classified as HER2/neu negative.

[00180] Networks

[00181] Correlation networks were created with Cytoscape (v. 3.7.1). In all networks, nodes represent metabolites with node sizes corresponding to the mean intensity. Edges represent spatial correlations with line thickness and opacity increasing with the correlation coefficient. Nodes were coloured red, if their metabolites take part in glycolysis or they were coloured depending on the molecule super class defined in HMDB [lipids and lipid-like molecules (yellow); nucleosides, nucleotides, and analogues (light red); organic acids and derivatives (green); organoheterocyclic compounds (lime green); alkaloids and derivatives (pink); organic oxygen compounds (blue); benzenoids (violet); phenylpropanoids and polyketides (orange); others (grey)]. All networks were either visualized using the yFiles circular layout or edge weighted spring embedded layout using the absolute value of the correlation coefficient. For the pancreatic islet cells, circular networks were generated by filtering edges with a coefficient smaller than 0.7 and by only visualizing direct neighbours of glucose 6-phostphate. All islets were located on the same tissue slide and were analysed concurrently. The multiplex staining of the complete tissue is shown in Figure 9. All remaining networks were generated by showing metabolites with at least one correlation coefficient larger than 0.5, but without filtering edges. The complete networks are shown in Figures 23 and 24. The multiplex staining of the complete tissues is shown in Figures 16, 17, 20, and 22. [00182] Statistical analyses

[00183] For the networks, pairwise Spearman rank-order correlations (Python 3.7, SciPy 1.2.0) were calculated between annotated metabolites using their intensities, and the resulting p-values were adjusted with Benjamini/Hochberg correction (Python 3.7, StatsModels 0.9.0). For the pancreatic islet cells, circular networks were generated by filtering edges with a coefficient smaller than 0.7. Network metrics (Tables 2-4) were calculated using Cytoscape’s plugin NetworkAnalyzer.

[00184] Metabolites localized predominantly on alpha or beta cells in Islets of Langerhans were identified by using the Mann-Whitney U-test (Python 3.7, SciPy 1.2.0). The p-values were adjusted with Benjamini/Hochberg correction (Python 3.7, StatsModels 0.9.0). The number of cell-type specific pixels per islet ranges between 59 and 194 for alpha cells and between 112 and 228 for beta cells. The Python 3.7 package NumPy 1.15.4 was used to calculate statistics for the intensity distributions of ADP, cholesterol sulfate and 3-O-sulfogalactosylceramide in the islets of Langerhans (Figures 13-15).

[00185] Example 1 : The SPACIAL workflow for immunohistochemistrv-quided imaging mass spectrometry

[00186] The SPACIAL pipeline comprises a series of MALDI data and image processing steps to combine molecular data with morphological and immunophenotypic information from immunohistochemistry (IHC) or other imaging data. Immunostaining following MALDI imaging has previously been shown to be feasible, hence the entire workflow works on the very same tissue section. In this example the inventors demonstrate that even multiplex immunostainings were entirely possible after MALDI imaging of the very same tissue section, which allows automatic data integration of morphological and spatially resolved in situ data of thousands of molecules via the SPACIAL method. The entire tissue and data pre-processing workflow preceding the application of the SPACIAL algorithm includes matrix coating of tissue sections, MALDI imaging, peak picking, matrix removal, IHC staining and image digitalization, which is shown schematically for an islet of Langerhans with glucagon, insulin and DAPI stainings (Figure 2A). SPACIAL then uses MALDI imaging files to create a reference image for subsequent co-registration of the molecular data with other image information (Figure 2B). The digitized and co-registered immunostaining images were scaled to match the exact MALDI resolution and then converted into numerical data without loss of spatial resolution. This ultimately allows pixel-accurate, objective tissue annotations based on semantics and function, which is shown here as an example for alpha and beta cells stained with glucagon (red) and insulin (green), respectively, and one metabolite (yellow) co-localizing with alpha cells (Figure 3). The SPACIAL pipeline paves the way for further statistical calculations and for the analysis of tissue heterogeneity and previously infeasible molecular in situ analyses of cell subpopulations within intact tissue sections. To illustrate the versatility and analytical power of the SPACIAL pipeline, it was applied on two datasets; i.e. , a physiological and a pathophysiological use case.

[00187] Example 2: SPACIAL analysis of metabolic heterogeneity within and between islets of Langerhans

[00188] To demonstrate the SPACIAL pipeline, it was applied on islets of Langerhans in the pancreas of a wild type mouse to distinguish the glucagon releasing alpha and insulin releasing beta cells and to investigate the heterogeneity of different islets within one animal. Previous studies highlight heterogeneity as a fundamental characteristic of pancreatic islets. Beta cells were functionally heterogeneous and display different activity patterns in response to glucose stimulation or the ability to secret insulin. The metabolic heterogeneity within automatically detected alpha and beta cells was analysed in detail for the glucose metabolism. The islets of Langerhans - originating from one tissue section (Figure 9-12) - were imaged with both high lateral (15 pm) and high mass resolution. Correlation networks were created to identify functional relationships of metabolites with glucose 6-phosphate and to assess metabolic heterogeneity within and between individual islets of Langerhans (Figure 4, Table 2). Glucose 6- phosphate was chosen as a relevant example, because it is an important intermediate in the glycolysis, gluconeogenesis and pentose phosphate pathways.

[00189] Clear differences regarding network size were found between the islets and islet cell populations, reflecting differential metabolic states (Figure 4). For example, the alpha and beta cell network of the islet in Figure 4D indicates a low dependency on glucose metabolism with only two metabolites showing a significant correlation to glucose 6-phosphate. Within networks from other islets, a variety of metabolites including lipids, nucleotides, amino acid and analogues correlate with glucose 6- phosphate (Figure 4). The highest number of correlations were found in beta and alpha cell populations of the islets A and E, respectively, indicating a high dependency on glucose metabolism. The spatial distribution of lipid-associated compounds, such as palmitic acid, stearic acid, lysophosphatidylinositol (LPI), and lysophosphatidic acid (LPA) were found to be correlated almost consistently. Other compounds, such as phosphodimethylethanolamine (P-DME) or glycerophosphoinositol (GroPIns), were found to inconsistently correlate with glucose 6-phosphate.

[00190] Metabolic signatures related to specific cell types and subpopulations can now easily be extracted with SPACIAL. Alpha and beta cells were defined automatically as ROIs and metabolic differences between alpha and beta cells were assessed. Significant differences were detected for adenosine diphosphate (ADP), cholesterol sulfate and 3-O-sulfogalactosylceramide (Figure 5). The presence of ADP, cholesterol sulfate and 3-O-sulfogalactosylceramide was validated via MALDI FTICR on-tissue MS/MS using quadrupole collision-induced dissociation and comparison to standard compounds (Figure 25). Interestingly, not all islets reveal similar significant changes, also reflecting inter- and intra-islet metabolic heterogeneity. For instance, across four of the five measured islets, significantly higher ADP levels were detected in beta cells in comparison to alpha cells. Thus the SPACIAL pipeline paves the way for in situ analyses of individual energy conditions of alpha and beta cells in each islet due to adenine nucleotide measurements. Cholesterol sulfate was found abundantly in beta cells, but it also exhibits a strong heterogeneous distribution between islets and even within cells of the same islet. Cholesterol sulfate is a component of the cell membrane and in pancreatic beta cells, elevated intracellular cholesterol levels have been associated with reduced insulin secretion in mice. Correlating with alpha cells, the inventors found the 3-O-sulfogalactosylceramide. Sulfatides were glycosphingolipids which have been described in pancreatic islets with different abundancies in alpha and beta cells.

[00191] Finally, pronounced molecular heterogeneity, both within single and between different islets of Langerhans, is reflected by a varying distribution of metabolite abundances (Figure 13-15). Between islets and between cell types, the standard deviation of metabolite intensities differs by a factor of between 2.28 and 14.25 (ADP 0.32-0.73, cholesterol sulfate 0.5-1.52; 3-O-sulfogalactosylceramide 0.04-0.57). Hence, even individual cells within one islet exhibit different metabolite compositions, possibly reflecting different metabolic or cell differentiation states.

[00192] The in situ analysis of metabolic heterogeneity within pancreatic islets is just one potential field of application for the SPACIAL pipeline. Metabolic data together with detailed spatial information can be exploited to assess the extent and modulation of alpha and beta cells in situ. Considering that both insulin and glucagon were dysregulated in pathophysiological conditions such as diabetes, the inventor’s pipeline is valuable for future studies. The analysis of different subpopulations of islets of Langerhans can help to illuminate underlying phenotypic mechanisms in order to expand the knowledge of cell function and to develop new therapeutic strategies.

[00193] Example 3: Intratumoral metabolic heterogeneity in gastric cancer [00194] The SPACIAL strategy has been shown to be powerful for close-to single-cell analyses of the metabolome in tissues of animal models, but it is also valuable for clinically relevant tissue analyses regarding diagnostics, prognosis, and therapy response prediction. For this reason, the inventors applied the SPACIAL pipeline for the analysis of intra- and intertumoral heterogeneity in gastric cancer. While the inventors used glucagon and insulin to stain alpha and beta cells within a frozen pancreatic tissue section, here the inventors used pan-cytokeratin as an epithelial marker to stain tumor cells and HER2/neu for tumor cell classification within human FFPE tissue sections.

[00195] In gastric cancer, intratumoral HER2/neu heterogeneity is frequent, but its clinical significance remains open in terms of treatment with trastuzumab-chemotherapy. The investigation of HER2/neu heterogeneity in gastric cancer in relation to the metabolic state of tumours is completely unexplored and may contribute to the improvement of treatment success. The SPACIAL pipeline was applied on tissue samples from three patients with gastric cancer to evaluate metabolic heterogeneity depending on the HER2/neu state. SPACIAL automatically determines the HER2/neu positive and negative tumor regions in a standardized way by evaluating expression values both in quantity on the basis of pixel intensity and localization by pixel co-localization (Figure 6). Regions displaying both pan-cytokeratin and HER2/neu signals were defined as HER2/neu positive tumor regions, while regions displaying only a pan-cytokeratin signal were classified as HER2/neu negative. Whole slide immunohistochemical stainings and regions defined as HER2/neu positive (red) and negative (yellow) were shown in Figures 16-21. The pixel- accurate annotation allows an unprecedented analysis of metabolic heterogeneity within tumor cells based on metabolic correlation networks that were calculated for annotated metabolites detected and stringently filtered from gastric cancer tissue sections (Figure 7 and 8). For visualization purposes, a zoom-in of HER2/neu positive and negative tumor regions of Figures 17, 19, and 21 is shown.

[00196] Since glucose plays a major role in altered energy metabolism in cancer, focusing on captured glucose as glucose 6-phosphate in the analysis of correlation networks provides insight into the complexity of the tumor biology regarding HER2/neu status. The spatial correlation networks comprise 67 to 171 metabolites (Table 3) in HER2/neu positive or negative tumor regions, revealing intratumoral heterogeneity (Figure 7A-C). Although for the networks from sample A and C a similar number of correlating metabolites was identified, the number and strength of pairwise spatial correlations is different, leading to different network structures - particularly the network density is higher in C by a factor of approximately four (Table 3). A majority of the correlating metabolites belong to the class of lipid and lipidlike molecules. In HER2/neu positive regions of all patients, lysophosphatidylinositole (LPI) abundance correlates positively with glucose 6-phosphate. LPI is a bioactive lipid produced by the phospholipase A family, which is believed to play an essential role in several physiological and pathological processes. As a ligand for the G-protein-coupled receptor GPR55, LPI may increase the glycolytic activity, since a GPR55 antagonist was shown to decrease glycolytic activity in cancer cell lines. In one sample, glucose 6-phosphate forms a cluster together with numerous lipids (Figure 7C), while in the other two samples, the neighbouring nodes belong to different metabolic classes - including carbohydrates, dipeptides and glycosylamines (Figure 7A and B). The intratumoral heterogeneity is most prominent in tumor sample C, reflected by the difference in degree, average shortest path length and clustering coefficient of HER2/neu positive and negative metabolic networks (Figure 7, 8 and Table 1). Overall, the degree and clustering coefficient of the glucose 6-phosphate node varies more strongly between patient samples, than between HER2/neu positive and negative tumor regions within individual patient samples - reflecting intertumoral heterogeneity (Table 1).

[00197] Example 4: Intertumoral metabolic heterogeneity in gastric cancer [00198] To additionally demonstrate the compatibility of SPACIAL for high-throughput multiplex phenotyping, metabolic correlation networks were created for gastric cancer patient tissues from an FFPE tissue microarray (Figure 8A-E). Networks on the extracted HER2/neu positive tumor regions of five gastric cancer patients comprise 30 to 39 metabolites and exhibit diverse correlation patterns. Similar to the results from whole gastric cancer resection specimens, most of the correlating metabolites were lipids. An altered lipid metabolism has been described previously in a HER2/neu positive breast cancer model. Thus, a changed lipid metabolism may be associated with a high positive correlation of individual lipids to glucose 6-phosphate in human gastric cancer patients.

[00199] The diversity between metabolic correlation networks in individual patients demonstrate high intertumoral heterogeneity of HER2/neu positive gastric cancer tissue (also see network metrics in Table 4). The novel pipeline is a starting point for intra- and intertumoral heterogeneity analyses, enabling simultaneous analysis of distinct tissue and cellular compartments. The spatially resolved information from the molecular analysis has been used in the inventors study to generate correlation networks between metabolites within ROIs that were semi-automatically defined by immunohistochemical staining.

[00200] Example 5: Data preprocessing and image co-registration

[00201] The coordinates from the imzML file were used to generate an image of the measurement region (imzML-grid) (Figure. 26). The resolution of this image is a multiple of the imzML resolution. A two- step image co-registration can be performed via existing software (e.g. Fiji ImageJ, Gimp or Photoshop) and can be achieved by first registering the tissue overview image with visible tissue or matrix-specific markers onto the grid-image and then registering all stainings simultaneously onto the tissue overview image. Once all images were co-registered, they all have the same resolution (Figure. 27 and Figure. 28).

[00202] The immunostaining images were then scaled to the exact same resolution as the x,y-grid from the imzML file (Figure. 29). Downscaling is done by calculating the mean colour value per square pixel region. Optionally, pixels with no other colour pixel within x±2 and y±2, can be removed.

[00203] Image masking

[00204] A black and white image can now be created to mask artefacts (e.g. tissue folds) or split measurement regions (see paragraph ‘Regions of interest’). Masked regions were excluded from further analyses (Figure. 30).

[00205] Colour threshold definition

[00206] For each staining, the user can now define a threshold to decide at what colour value a pixel is classified as a positive signal (Figure. 31).

[00207] Regions of interest

[00208] First, preliminary regions of interest were defined based on stainings, IMS measurement regions and custom-generated images. Subsequently, those regions can be combined to define the final regions of interest.

[00209] Regions generated from staining images

[00210] When the image mask is imported, it is also converted into coordinates. The previously determined colour thresholds were used to generate two preliminary regions (lists of x,y-coordinates) per staining: e.g. staining X positive and staining X negative pixels (Figure. 31).

[00211] IMS measurement regions

[00212] Individual IMS measurement regions may also be considered as preliminary regions. This is important for the analysis of tissue microarrays, where tissue cores of several of patients were measured on one glass slide. Usually each patient sample is measured in a separate measurement region. Nevertheless, if two patient cores are part of the same measurement region, the measurement region can be split by separating them with a black line in the masking image (see paragraph ‘Image masking’) (Figure. 32).

[00213] Custom regions

[00214] Additionally, the user can define custom regions, either by creating one black and white image per region, similar to the masking image, or by creating a multicolour image, where each colour except black is considered as one region. This is used for the analysis of tissue microarrays, where multiple cores may belong to the same patient. First, an image is generated, where each measurement region is coloured uniquely (Figure. 33). The user can then manually colorize regions that belong to the same patient/group (Figure. 34).

[00215] Final regions of interest

[00216] The previously defined preliminary regions were combined to generate the final regions of interest. This can be achieved via logical/set expressions with the common operators Union, Intersection, Without/Difference and X or.

[00217] Examples: islet_ROI = roil P (staining2 U staining3) amyloid_ROI = roil P (stain ing1 P (staining2 D staining3)) beta_ROI = roil P (staining2 \ staining3) alpha_ROI = roil P (staining3 \ staining2)

[00218] Data integration

[00219] The mass spectra of each region of interest can further be extracted from the imzML file and used for subsequent calculations and analyses. REFERENCES:

- Ahmed, M., Broeckx, G., Baggerman, G., Schildermans, K., Pauwels, R, Craenenbroeck, A.H. Van., et al., 2019. Next-generation protein analysis in the pathology department: 1-6, Doi: 10.1136/jclinpath- 2019-205864.

Aeffner, F., Zarella, M., Buchbinder, N., Bui, M., Goodman, M., Hartman, D., et al., 2019. Introduction to digital image analysis in whole-slide imaging: A white paper from the digital pathology association. Journal of Pathology Informatics 10(1): 9, Doi: 10.4103/jpi.jpi_82_18.

Aichler, M., Borgmann, D., Krumsiek, J., Buck, A., MacDonald, PE., Fox, J.E.M., et al., 2017. N-acyl Taurines and Acylcarnitines Cause an Imbalance in Insulin Synthesis and Secretion Provoking b Cell Dysfunction in Type 2 Diabetes. Cell Metabolism 25(6): 1334-1347. e4, Doi: 10.1016/j.cmet.2017.04.012.

Brieu, N., Caie, R, Gavriel, C., Schmidt, G., Harrison, D.J., 2018. Context-based interpolation of coarse deep learning prediction maps for the segmentation of fine structures in immunofluorescence images (March 2018): 24, Doi: 10.1117/12.2292794.

Buck, A., Ly, A., Balluff, B., Sun, N., Gorzolka, K., Feuchtinger, A., et al., 2015. High-resolution MALDI- FT-ICR MS imaging for the analysis of metabolites from formalin-fixed, paraffin-embedded clinical tissue samples. Journal of Pathology 237: 123-32, Doi: 10.1002/path.4560.

Carrano, A.C., Mulas, F, Zeng, C., Sander, M., 2017. Interrogating islets in health and disease with single-cell technologies. Molecular Metabolism 6(9): 991-1001 , Doi: 10.1016/j.molmet.2017.04.012. Farack, L, Golan, M., Egozi, A., Dezorella, N., Bahar Halpern, K., Ben-Moshe, S., et al., 2019. Transcriptional Heterogeneity of Beta Cells in the Intact Pancreas. Developmental Cell 48(1): IIS- 125. e4, Doi: 10.1016/j.devcel.2018.11 .001 .

Feuchtinger, A., Stiehler, T, Jotting, U., Marjanovic, G., Luber, B., Langer, R., et al., 2014. Image analysis of immunohistochemistry is superior to visual scoring as shown for patient outcome of esophageal adenocarcinoma. Histochemistry and Cell Biology 143(1): 1-9, Doi: 10.1007/s00418-014- 1258-2.

Ly, A., Buck, A., Balluff, B., Sun, N., Gorzolka, K., Feuchtinger, A., et al., 2016. High-mass-resolution MALDI mass spectrometry imaging of metabolites from formalin-fixed paraffin-embedded tissue. Nature Protocols 11 (8): 1428-43, Doi: 10.1038/nprot.2016.081 .

Roscioni, S.S., Migliorini, A., Gegg, M., Lickert, H., 2016. Impact of islet architecture on b-cell heterogeneity, plasticity and function. Nature Reviews Endocrinology 12(12): 695-709, Doi: 10.1038/nrendo.2016.147.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V, Longair, M., Pietzsch, T, et al., 2012. Fiji: An open-source platform for biological-image analysis. Nature Methods 9(7): 676-82, Doi: 10.1038/nmeth.2019.

Wishart, D.S., Feunang, Y.D., Marcu, A., Guo, A.C., Liang, K., Rosa, V, et al., 2018. HMDB 4.0 : the human metabolome database for 2018. Nucleic Acids Research 46(November 2017): 608-17, Doi: 10.1093/nar/gkx1089.

EP2124192A1 included by reference

Gonzalez, R. et al.: "Digital Image Processing, Global Edition", Pearson 2018 Russ, J.C. et al.: "The Image Processing Handbook, 7th Ed., CRC Press 2016 Bruker fleximaging (v. 5.0) Software manual Schramm, T. et al., 2012: "imzML — A common data format for the flexible exchange and processing of mass spectrometry imaging data", Journal of Proteomics, vol. 75, issue 16, pp 5106-5210.

Claims

1. Method of automatically selecting regions of interest, ROIs, for analyzing one complete tissue section, comprising:

(i) generating a first image from mass spectrometry data of the tissue section spatially resolved by a region of the first image represented as a first mask,

(ii) and further obtaining one or more second images comprising any type of optical images of the tissue section used in (i), wherein the one or more second images are in alignment with and scaled to the resolution of the first image; determining one or more second masks associated with respective one or more second images based on predetermined criteria;

(iii) combining masks selected from the first mask and the one or more second masks, respectively, based on predetermined set expressions, to automatically obtain final regions of interest, ROIs;

(iv) extracting mass spectrometry data using said first image and optical image data from said one or more second images, respectively, based on the final ROIs, for analyzing the tissue section; and

(v) wherein the steps (i) to (iv) result in the automatic definition of ROIs for diagnostic and research application.

2. The method of claim 1 , further comprising obtaining one or more custom masks of the one or more second images of (ii), and wherein combining masks in (iii) comprises combining masks selected from the first mask, the one or more second masks, and the one or more custom masks.

3. The method of any of the preceding claims, wherein determining one or more second masks of (ii) associated with respective one or more second images based on predetermined criteria comprises using predetermined thresholds of color values or ranges of color values of pixels in an image of the one or more second images to determine regions containing pixels of similar color values in said image.

4. The method of any of the preceding claims, wherein the mass spectrometry data are obtained by Matrix-assisted Laser Desorption (MALDI), Desorption Electrospray Ionization (DESI), Laser Ablation Electrospray Ionization (LAESI), Secondary ion Mass Spectrometry (SIMS), Matrix- assisted Laser Desorption Electrospray Ionization (MALDESI), Liquid Extraction Surface Analysis (LESA), Mass spectrometra imaging (MSI), or Inductively Coupled Plasma Mass Spectrometry Imaging (ICP-MSI), respectively.

5. The method of claim 4, wherein the mass spectrometry data is generated by MALDI,.

6. The method of any of the preceding claims, wherein the optical images of the tissue section are selected from images obtained by multiplex immunohistochemical (IHC) staining, histochemical staining, endogeneous or exogeneous fluorescence staining, respectively, fluorescence in situ hybridization (FISH), or autofluorescence.

7. The method of any one of the preceding claims, wherein the mass spectrometry data is represented in imzML format, mis format of fleximaging, slx/sbd format of SCiLS Lab, or a tabular format, respectively.

8. The method of any one of the preceding claims, wherein the tissue section comprises one tissue sample obtained by formalin-fixed paraffin-embedding (FFPE) or obtained by any other fixation technique, comprising ethanol fixation and cryofixation, respectively.

9. The method of any one of the preceding claims, wherein determining one or more second masks associated with respective one or more second images comprises determining multiple masks for the same image, said masks pertaining to multiple morphometric data, or data from multiple color components, such as Hematoxylin and eosin (H&E) and other histological stainings.

10. The method of the preceding claim, wherein ranges of color values of pixels are based on a magnitude of expression of a diagnostic feature, such as one or more biomarkers.

11. The method of claim 10, wherein the biomarker is a) HER2/neu and/or pan-cytokeratin (PC); b) Insulin and/or glucagon; c) PD1 and/or PD-L1 ; or d) Vimentin and/or pan-cytokeratin.

12. The method of any of the preceding claims, wherein the predetermined set expressions represent semantics of tissue annotation to obtain the final ROIs, based on the first image, the one or more second images and the associated masks of the tissue section.

13. The method of any of the preceding claims, wherein analyzing the tissue section comprises comparing mass spectrometry data and image data, respectively, derived from the final ROIs of the first image and the one or more second images of the tissue section, by statistical analysis.

14. The method of the preceding claim, wherein comparing mass spectrometry data and image data, respectively, extracted from the final ROIs comprises assessing of molecular composition and heterogeneity expressed between data extracted from final ROIs, wherein molecular composition comprises metabolites, proteins, drugs, toxic agents, carcinogens, glycanes, and lipids, respectively.

15. The method of any one of the preceding claims, wherein the method is performed simultaneously on multiple tissue samples individualized by one or more custom masks, wherein each custom mask is associated with an individual tissue sample.

16. Use of the method according to any one of the preceding claims for diagnosis or stratification based on a human, animal or plant tissue samples.

17. The use of claim 16, wherein the samples are human tissue samples used in the diagnosis or stratification of cancer or diabetes.

18. Use of the method according to any one of claims 1 to 15 for therapy response prediction or prediction of organ rejection based on human or animal tissue.