Methods for measuring tissue dynamics

Abstract

The invention concerns methods for predicting changes in a tissue, comprising obtaining snap shot information on individual cells of at least one cell type in a section of a tissue biopsy, wherein the information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state at a specific point in time; calculating cell division probabilities for each of said individual cells; determining cell population dynamics based on the cell division probabilities and predicting future changes in the tissue based thereon. Particularly, the method concerns predicting tissue dynamics in cancer biopsies relevant for prognosis, and assessment of response to therapy.

Description

[0001] Methods for Measuring Tissue Dynamics

[0002] TECHNOLOGICAL FIELD

[0003] The present invention generally relates to methods for predicting changes in the cellular composition in a solid tissue biopsy, for example in a tumor biopsy.

[0004] BACKGROUND ART

[0005] References considered to be relevant as background to the presently disclosed subject matter are listed below:

[0006] Adler et al., Annu. Rev. Cell Dev. Biol. 39, 67-89 (2023)

[0007] Danenberg et al., Nature Genetics 54, 660-669 (2022)

[0008] Entenberg et al., Nat. Rev. Cancer 23(1) : 25-42 (2023)

[0009] Gianni et al. Annals of Oncology 33(5): 534-543 (2022)

[0010] Horns et al., Cell 186 (17), 3642-3658 (2023)

[0011] Jerby-Amon and Regev, Nat. Biotechnol. 40(10): 1467-1477 (2022)

[0012] Kim et al., Nat. Rev. Mol. Cell Biol. 21 (10) : 571-584 (2020)

[0013] Kirschenbaum et al., Cell 187, 1, 149-165. e23 (2024)

[0014] Kokkinos et al., Nature Research Scientific Reports 11 : 1944 (2021)

[0015] Leung et al., Nature Reviews Methods primers 2 :33 (2022)

[0016] Wang et al., Nature 621, 868-876, (2023)

[0017] Acknowledgement of the above references herein is not to be inferred as meaning that these are in any way relevant to the patentability of the presently disclosed subject matter.

[0018] BACKGROUND

[0019] Physiological and pathological processes such as inflammation or cancer emerge from the interactions between cell populations that change over time (Adler et al., 2023). Some cell populations expand, and others are removed, as occurs in development and in the immune response to pathogens. A clinically important example of changing cell populations is the evolution of the cancer microenvironment, in which the growing tumor recruits stromal and immune cells that are crucial for tumor survival. Understanding cell dynamics and the underlying cell-cell communication circuits is a major goal of tissue biology and can enable new treatment strategies based on sculpting cell populations in desired ways.

[0020] However, measuring the dynamics of cell populations in human tissues is currently very difficult. Biopsies provide a single snapshot, and taking multiple biopsies is unfeasible and does not provide longitudinal evidence from the same cells.

[0021] Current strategies to measure tissue dynamics do not apply to human biopsies. For example, cell lines, mice, organ-on-a-chip (Leung et al., 2022) or organoid models (Kim et al., 2020) are treated as replicas and are analyzed at different time points. Ex-vivo tissues can be followed over time but lack the native immune context (Kokkinos et al., 2021). Intravital fluorescence microscopy has made it possible to follow living cells within animal models (Entenberg et al., 2023). Recent advances employ synthetic biology to engineer cells to record their activity or lineage (Horns et al., 2023). These approaches are limited to animal models or in vitro settings, limiting their ability to capture the full complexity of in vivo dynamics and the diversity of human tumors.

[0022] The emergence of single-cell technologies offers new opportunities for studying tissues at high resolution. Notable approaches use single-cell data for understanding cellcell communication at a single time point (Jerby-Amon and Regev, 2022). Other approaches attempt to infer dynamics of processes such as transcription within individual cells on a timescale of hours, such as RNA velocity, ergodic rate analysis and Zman-seq (Kirschenbaum et al., 2024). These methods do not address the challenge of understanding how cell populations change on the tissue level, processes that could take days to weeks. Thus, methods to follow cell populations over time within the native context of human tissue are lacking.

[0023] GENERAL DESCRIPTION

[0024] The present invention provides a method for estimating cell-level division or death rates based on measurements of the cell surroundings. Furthermore, the method allows for predicting tissue composition in future points in time) based on the estimated cell division and death rates. Predictions can be done either at the level of single cells, by sampling division or death events based on each cell’s surroundings, or rather at the level of tissues sections based on the inferred cell-level models. Accordingly, in one aspect, the present disclosure provides a method for predicting changes in a tissue, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; and c. determining cell population dynamics and predicting said changes based thereon.

[0025] In another aspect, the present invention provides an ex vivo method for estimating the progression of a solid tumor in a patient, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a solid tumor biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; c. determining cell population dynamics of each type of cells; and d. estimating potential tumor progression based on the cell population dynamics.

[0026] In one embodiment, the ex vivo method for estimating the progression of a solid tumor further comprises providing disease prognosis.

[0027] In another aspect, the present invention provides an ex vivo method for assessing the responsiveness of a solid tumor to an antitumor therapy administered to a patient, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a solid tumor biopsy taken at different times comprising one or more times during the antitumor therapy and optionally one or more times prior to the antitumor therapy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) estimating tumor progression based on the cell population dynamics; and c. based on changes in the estimated tumor progression between the sections determining responsiveness of the solid tumor to the antitumor therapy.

[0028] In one embodiment, the ex vivo method for assessing the responsiveness of a solid tumor to an antitumor therapy further comprises providing assessment of the response of the patient to the antitumor therapy.

[0029] In another aspect, the present invention provides an ex-vivo method for estimating the progression of a disease in a patient, comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; c. determining cell population dynamics of each type of cells; and d. estimating potential disease progression based on the cell population dynamics.

[0030] In one embodiment, the ex vivo method for estimating the progression of a disease further comprises providing disease prognosis.

[0031] In another aspect, the present invention provides an ex vivo method for assessing the responsiveness to therapy of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a tissue biopsy taken at different times comprising one or more times during therapy and optionally one or more times prior to the therapy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells and (iii) estimating disease progression based on the cell population dynamics; and c. based on changes in the estimated disease progression between the sections determining responsiveness of the patient to the therapy.

[0032] In another aspect, the present invention provides a method of treatment of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) identifying points of intervention that drive cell populations towards desired densities; and c. administering a therapeutic agent to the patient at the identified points of intervention.

[0033] In another aspect, the present invention provides a method for assessing the efficacy of a therapeutic agent for treating a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a tissue biopsy, taken at different times comprising one or more times after administration of the therapeutic agent and optionally one or more times prior to administration of the therapeutic agent, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) estimating the pathological condition of the tissue based on the cell population dynamics; and c. based on changes in the pathological condition of the tissue between the sections determining the efficacy of the therapeutic agent.

[0034] In some embodiments the tissue biopsy is taken from an animal model of the disease or from an organoid grown ex vivo.

[0035] The present invention also provides a computer software product, comprising a computer-readable medium in which program instructions are stored, which instructions, when read by a data processor, configure the data processor to obtain information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; and to (2) execute the methods of the invention.

[0036] The present invention also provides a system for predicting changes in a tissue, comprising: an input utility for obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; and a data processor configured for analyzing said data for executing the methods of the invention.

[0037] In one embodiment, said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process and said calculating comprises calculating cell death probabilities.

[0038] In one embodiment, the methods comprise obtaining information on individual cells of at least two different cell types in the tissue section, or the solid tumor section.

[0039] In one embodiment, said information is obtained at a single time point.

[0040] In one embodiment, said information comprises data obtained by one or more of multiplexed protein imaging, spatial-transcriptomics, methylation analysis, fluorescent based cell staining, mass cytometry, and X-ray based imaging.

[0041] In one embodiment, the information on spatial localization and cell type is obtained by (x) staining the section with cell type specific markers, (y) determining location of individual cells within the tissue section and (z) quantifying the number of at least one cell type, or at least two different cell types, within a spatial perimeter surrounding the individual cell.

[0042] In one embodiment, the methods comprise staining the sections with cell divisionspecific dyes or markers.

[0043] In one embodiment, the methods comprise staining the cells with cell death- or cell death process-specific dyes or markers.

[0044] In one embodiment, said calculating cell division probabilities comprises: bl. Defining a neighborhood around each of said individual cells in the section; and b2. calculating the division probability for each of the individual cells as a function of cell composition and optionally spatial organization in each defined neighborhood.

[0045] In one embodiment, said neighborhood comprises cells surrounding each of said individual cells.

[0046] In one embodiment, said neighborhood comprises cells in a predefined radius.

[0047] In one embodiment, said step (b2) is performed using a statistical inference model.

[0048] In one embodiment, said statistical inference model is logistic regression. In one embodiment, said step (b2) is performed using a machine learning algorithm.

[0049] In one embodiment, said step of determining cell population dynamics comprises inferring probabilistic rate equations based on the cell division probabilities.

[0050] In one embodiment, the methods comprise deriving a deterministic differential equation from the probabilistic rate equation, predicting the direction of change that would have been expected in the tissue section had it been still within the body.

[0051] In one embodiment, the different cell types comprise epithelial cells, mesenchymal cells, endothelial cells, antigen presenting cells, leukocytes, lymphoid cells, or myeloid cells.

[0052] In a specific embodiment, the different cell types comprise fibroblasts and macrophages.

[0053] In another specific embodiment, the different cell types comprise T cells and B cells.

[0054] In one embodiment, the tissue biopsy is from a solid tumor.

[0055] In another aspect, the present invention provides a method for optimizing immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken at one or more times during the immunotherapy; b. for each section (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell population dynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. adjusting the frequency of immunotherapy sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or d. administering an additional therapeutic intervention to the patient, thereby optimizing immunotherapy.

[0056] In another aspect, the present invention provides a method of treatment of a solid tumor by immunotherapy, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken from a patient at one or more times during the immunotherapy; b. for each section (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell population dynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. administering immunotherapy at a frequency of sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or administering an additional therapeutic intervention to the patient.

[0057] In another aspect, the present invention provides an ex vivo method for assessing the responsiveness of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; wherein a reduction in tumor proliferation near T cells indicates that the solid tumor is responsive to the immunotherapy, and wherein an increase in tumor proliferation near T cells indicates that the solid tumor is not responsive to immunotherapy.

[0058] In one embodiment, if the solid tumor is not responsive to immunotherapy, the method further comprises halting the immunotherapy administered to the patient.

[0059] In another aspect, the present invention provides an ex vivo method for predicting the likelihood of response of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at one or more times prior to the immunotherapy; wherein a low level of tumor proliferation near T cells indicates that the solid tumor will be responsive to the immunotherapy, and wherein a high level of tumor proliferation near T cells indicates that the solid tumor will not be responsive to immunotherapy.

[0060] In one embodiment, if the prediction is that the solid tumor will not be responsive to immunotherapy, the method further comprises refraining from administering immunotherapy to the patient.

[0061] In another aspect, the present invention provides a method for preventing hyper progression of a solid tumor in response to immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; b. determining whether there is an increase in tumor proliferation near T cells; and c. If an increase in tumor proliferation near T cells is detected halting the administration of the immunotherapy to the patient, thereby preventing the hyper progression of the solid tumor in said patient.

[0062] In one embodiment, the solid tumor is breast cancer. In another embodiment, the solid tumor is triple negative breast cancer.

[0063] BRIEF DESCRIPTION OF THE DRAWINGS

[0064] To better understand the subject matter that is disclosed herein and to exemplify how it may be carried out in practice, embodiments will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:

[0065] Figure 1A is a graphical representation showing the translation of the number of cells in a neighborhood to a location in a state-space graph. The left graph shows a spatial tissue section indicating individual cells (filled dots) and a defined neighborhood (encircled). On the right is a graph of the state-space representing the number of two types of cells (Y axis represents the number of one cell type and the X axis represents the number of a second cell type) within the encircled neighborhood of an individual cell.

[0066] Figure IB is a graphical representation showing a spatial tissue section indicating individual cells (filled dots) and a defined neighborhood (encircled). Black arrows point to dividing cells (based on KI67 levels above threshold): 2 cells of type A and 1 cell of type B are dividing (lefthand graph). On the right is a graph of the state-space representing the number of the two types of cells (Y axis represents the number of one cell type and the X axis represents the number of a second cell type). Black arrows represent the direction of change based on the number of dividing cells for each cell type.

[0067] Figure 1C is a graphical representation showing spatial distribution of cells in an exemplary breast-cancer tissue section (left) from a dataset by Danenberg et al. (2022). light-colored cells represent fibroblasts and dark-colored cells represent macrophages. Gray circle marks a neighborhood in the tissue. The neighborhood is mapped to its corresponding location in a phase portrait estimated by using OSDR (right).

[0068] Figure ID is a graphical representation showing a simulation of stochastic future trajectories of fibroblast and macrophage cell populations within a tissue, based on probabilistic division and death events for each cell. The graph on the left shows the number of cells (log2) as a function of time. The inset shows the dynamic change caused by division and death events.

[0069] Figure 2A is a graphical representation showing KI67 levels as a function of time (hours). The dotted line represents the KI67 threshold above which a division event is defined. The peak in KI67 levels represents a cell at the G2 / M stage which marks a division event.

[0070] Figure 2B is a scheme illustrating the cellular composition of a cell’s defined neighborhood, according to an embodiment of the present disclosure. A neighborhood of radius r around each cell is defined (marked by a dotted line) and the number of each cell type within this area is tabulated. The tabulated counts are used to predict the division event, as defined in Figure 2 A.

[0071] Figure 2C is a graph showing a phase portrait of macrophage density vs. fibroblast density. The enlarged square shows the direction of change in tissue composition from a starting density of fibroblasts and macrophages. The direction and rate of change (solid black vector) is a sum of the expected rate of change in fibroblast density and expected rate of change in macrophage density (dotted vectors). Figures 3A-3D are graphs showing reconstructed cell dynamics representing four two-cell phase portrait topologies: One stable point (Fig. 3 A), two stable points (Fig. 3B), three stable points (Fig. 3C), and four stable points (Fig. 3D). The graphs at the left-hand side display the ground-truth model used during simulation (“known”). Arrows indicate directions of change and lines are nullclines in which only one cell type changes. Black dots and white dots are stable and unstable fixed points. The graphs at the right-hand side display the estimated fixed points from 10 simulations (“Inferred”). Each Inferred figure also displays the OSDR arrows from the first simulation of the 10.

[0072] Figures 4A-4D are graphs showing on the x-axis the average predicted probability for a cell calculated using the method of the present disclosure and the y-axis is the experimental fraction of dividing cells, as determined by Ki67 > threshold (“true probability”). Fig. 4A represents a fibroblast model, Fig. 4B represents a macrophage model, Fig. 4C represents a T-cell model, and Fig. 4D represents a B-cell model. Each dot in the graph corresponds to a subset of 4000 cells.

[0073] Figure 4E is a graph showing KI67 marker level above threshold in fibroblasts as a function of neighborhood composition (macrophage density and fibroblast density) based on data from the MIBI breast cancer dataset. Arrow points to the line which separates inferred regions of rising and falling cell numbers (nullcline). Black contours represent density.

[0074] Figure 4F is a graph showing KI67 marker level above threshold in macrophages as a function of neighborhood composition (macrophage density and fibroblast density) based on data from the MIBI breast cancer dataset. Arrow points to the dashed line which separates inferred regions of rising and falling cell numbers (nullclines). Black contours represent density.

[0075] Figure 4G is a graph showing OSDR inferred fibroblast-macrophage phase portrait for breast cancer associated fibroblasts and macrophages. Black and white circles are stable and unstable fixed points respectively. Arrows point to the lines which separate inferred regions of rising and falling cell numbers (nullclines), as presented in Fig. 4E for fibroblasts and in Fig. 4F for macrophages.

[0076] Figure 5A is a graph showing OSDR T and B-cell inferred phase-portrait which indicates an excitable system. The highlighted trajectory displays a large pulse of adaptive immune cells. Figure 5B is a graph showing the number of B cells and T cells as a function of time.

[0077] Figure 5C is a graph showing the number of B cells (dark line) and T cells (lightcolored line) as a function of time, and in response to simulations of T cell additions at various time points (marked by vertical arrows).

[0078] Figures 6A-6L are graphs showing longitudinal predictions of the number of cells over time for 6 cell types. OSDR models were fitted separately to responders and nonresponders in each treatment arm. The tumor cell population collapses in responders (solid line) but is kept stable in non-responders (dashed line). Figs 6A-6F represent the patients receiving chemotherapy. Figs 6G-6L represent the patients receiving immunotherapy. To isolate the effect of the tumor environment (TME) dynamics, as opposed to the initial tissue composition, trajectories were computed starting from the composition of all biopsies taken at the beginning of treatment. Plots display the average cell count over all trajectories.

[0079] Figures 7A-7B are graphs showing a comparison of tumor proliferation rates in immunotherapy responders and non-responders. 7A - distribution of proliferation rates (referred to as tumor proliferation rate) in all neighborhoods. 7B - distribution of tumor proliferation in neighborhoods with a high T-cell density (top 1%).

[0080] Figures 8A-8B are graphs showing a comparison of pre-treatment tumor proliferation rates in immunotherapy responders and non-responders. 8A - distribution of pre-treatment proliferation rates (referred to as tumor division rate) in all neighborhoods. 8B - distribution of pre-treatment tumor proliferation in neighborhoods with a high T-cell density (top 1%).

[0081] Figures 9A-9B are graphs showing immunotherapy treatment response dynamics, namely the number of cells as a function of time. Fig. 8A shows the T-cell density which increases with time in both responders (solid line) and non-responders (dashed line). Fig. 8B shows the tumor cell population which collapses with time only in responders.

[0082] DETAILED DESCRIPTION OF EMBODIMENTS

[0083] The present disclosure concerns the assessment of temporal tissue dynamics based on a spatial biopsy snapshot, also referred to herein as one-shot tissue dynamics reconstruction (OSDR). OSDR is a novel approach to estimate a dynamical model of cell populations based on a single histological evaluation of a tissue sample. OSDR uses spatial cellular (e.g., proteomics) data, using for example cell-cycle markers, to infer cell division rate as a function of the composition of cellular neighborhoods. Based on an individual cell’s cellular neighborhood / environment the model predicts the division probability of the cell, providing a dynamic model of cell population change over time. Thus, division rates measured at the level of single cells are used to produce dynamics at the level of tissues.

[0084] As shown in the examples below, OSDR was applied to analyze spatial proteomics snapshot data of human breast cancer tumor microenvironment samples (data obtained from Danenberg et al., 2022) and two fixed points of fibroblasts and macrophage circuit interactions were reconstructed. The estimated fibroblast-macrophage dynamics agreed with dynamics measured directly from co-culture experiments in breast cancer conditioned medium, showing a flow towards fixed points of hot and cold fibrosis, as well as self-sustaining macrophages. OSDR was also used to discover an inferred pulsegenerating excitable circuit of T cells and B cells in a tumor microenvironment, suggesting that cancer surveillance in the tumor microenvironment operates in temporal flares of adaptive anti-cancer responses, as opposed to the steady-state picture implicit in current literature. OSDR enables the inference of a wide range of tissue and cell population dynamics based on a single snapshot of patient biopsies.

[0085] Furthermore, as shown in the Examples below, and as a non-limiting example, OSDR was used to analyze the dynamics of response to chemotherapy and immunotherapy in triple negative breast cancer (data obtained from Wang et al., 2023).

[0086] The methods of the present disclosure allow for the generation of an individualized dynamical model for a given patient based on a single biopsy.

[0087] Although a tissue biopsy typically comprises multiple cell types, it was discovered in accordance with the present disclosure that OSDR can accurately predict cell population dynamics based on the analysis of two different cell populations. The present disclosure thus refers to the modelling of at least two cell populations but also encompasses modelling of three, four or even more types of cell populations.

[0088] Accordingly, the present disclosure provides a method for predicting changes in a tissue, also referred to herein as cell population dynamics, said method comprising: a. Producing a spatial cellular snapshot of a tissue section taken from a biopsy; b. Performing a dynamical model estimation; and c. Determining cell population dynamics. Cell population dynamics are also referred to herein as tissue dynamics reconstruction.

[0089] The present disclosure thus provides a method for inferring a dynamical model of cell populations based on a single tissue sample, allowing the understanding of the dynamics of a tissue microenvironment, for example a tumor microenvironment, e.g., breast cancer.

[0090] In a first of its aspects, the present invention provides a method for predicting changes in a tissue, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; and c. determining cell population dynamics and predicting said changes based thereon.

[0091] As used herein the term “tissue” refers to a group of cells that are interconnected and function together in the body as a unit.

[0092] As used herein the term “cellular snapshot” refers to visualization of the cell type, spatial location in the tissue, and division state of individual cells in a tissue section taken from the biopsy.

[0093] As used herein the term “biopsy” refers to a tissue sample comprising cells taken from the body, e.g., a human body sample. In some embodiments the biopsy is a tumor biopsy. In some embodiments, the biopsy comprises at least 5,000 cells, or at least 10,000 cells, or more of a specific cell type. In some embodiments, the size of the biopsy is at least 0.1xl02cm, e.g. 0.2xl02cm. In an embodiment the biopsy is cut into tissue sections using methods well known in the art, in a non-limiting example the size of the tissue section is about 0.25mm2to 0. 5mm2. Multiple sections of a single tissue biopsy may be used for the analysis.

[0094] The methods of the present disclosure can be applied to any experimental modality that provides measurements of cell location, type, and division marker levels. Cell density can be inferred from the measurements of cell location and type. Possible modalities range from standard quantitative immunostaining to state-of-the-art spatial transcriptomics and proteomics. In some embodiments, the method of the present disclosure is performed with biopsy sections stained for proteins or nucleic acids (e.g., RNA) that include cell type markers and cell-division markers and visualized under a suitable microscope.

[0095] This data may be acquired by any technology that enables classification and spatial localization of discrete cells, together with robust measurement of markers for cell division (and optionally cell death or a cell death process).

[0096] As used herein the term “cell death process” encompasses apoptosis, necrosis, ferroptosis and necroptosis.

[0097] Cell division and cell death may be visualized using specific markers as specified below, and they may also be identified based on morphological features such as, but not limited to, a visible mitotic spindle as an indicator of cell division, or blebbing as an indicator of apoptosis. Cell death may also be identified by the proximity of macrophages that clear apoptotic cells.

[0098] Cell type markers are specific proteins, or any other factor whose presence is typical to a specific cell type. Such other factors may include, but are not limited to specific RNA molecules, or specific methylation patterns. Each cell type in the body is characterized by a set of markers that allow the unique staining and identification of the cell type in histological analysis of sections of tissue samples or biopsies, for example using specific antibodies. For example, typical markers for staining epithelial cells include, but are not limited to, pan-cytokeratin (Pan-CK), CK8-18, and CK5, typical markers for staining mesenchymal cells include, but are not limited to, CD31-vWF, FSP1, SMA, Caveolin-1, Podoplanin, and PDGFRB, (a list of typical markers for staining fibroblasts can be found for example in Muhl et al., Nature Communications 11, Article No. 3953, 2020), typical markers for staining endothelial cells include, but are not limited to, ACE / CD143, CD93, VE-Cadherin, CC chemokine receptor D6, CD31 / PECAM-1, CD34, CD36 and CD151, typical markers for staining cells involved in antigen processing and presentation include, but are not limited to, HLA-ABC, B2M, and HLA- DR, typical markers for staining leukocytes include, but are not limited to, CD45RO, and CD45RA, typical markers for staining lymphoid cells include, but are not limited to, CD3, CD20, CD4, CD8, CD57, CD38, and FOXP3, typical markers for staining myeloid cells include, but are not limited to, CD 16, CD68, CD 163, CD 11c, and CD15. Cell division markers are specific proteins, or other factors whose presence in actively growing and dividing cells can serve as an indicator for such cells. Non-limiting examples include proliferating cell nuclear antigen (PCNA), MCM-2, EdU, and Ki-67.

[0099] A non-limiting example for a marker indicating apoptosis / cell death is c- Caspase3.

[0100] Cancer cell markers are specific proteins, or other factors, typical to cancer cells. Multiple cancer cell markers are known in the art. The following is a non-limiting list of markers that can be used to identify breast cancer cells: ER, HER2, CD15, CD57, and CXCL12.

[0101] Any cell-type or cell division marker can be used in accordance with the present disclosure for staining of the tissue sections, according to known methods. From this staining data a list of cell locations and division status for each cell type is generated. The cell location and division status data are referred to herein as the snapshot data. This snapshot data is modeled as a dynamical system. The variables are cell concentrations, such as fibroblasts per unit area. In such a system, the change in cell counts is a balance between division, death, and migration. In accordance with the present disclosure, this dynamical system is inferred from snapshot data and is used to predict the future changes in cell populations.

[0102] In an embodiment, the dynamic model is constructed based on cell division and death. In other embodiments, the model may further incorporate migration or influx of cells to or from external sources. This dynamical system is used for predicting future changes in cell populations.

[0103] In some embodiments, the magnitude of error due to unmodeled terms such as cell migration, is estimated for example, by applying mass continuity equation to the statespace of neighborhoods, for example as shown in Example 3 below.

[0104] Accordingly, the dynamical equations for the rate of change of each cell type are modelled as its division rate minus removal rate. This model captures the dynamics resulting from growth-factor signaling circuits within a tissue, since growth factors are an important modulator of cell division rate. However, the model also indirectly captures the effects of other elements such as hypoxia (e.g., through the assessment of endothelial cells density), and inflammation (e.g., through the assessment of cancer cells or immune cells density). The effect of external sources of flux can be simulated by adding flux terms to the estimated model. Spatial omics snapshot is thus used to estimate the division and removal rates of cells by means of molecular markers of division (and if available, removal), as a function of the composition of the cells neighborhood.

[0105] To estimate the effect of each cell type on the growth and removal of other cells, it is assumed that the local neighborhood surrounding each cell observed in a biopsy slide contains growth factors and other signals secreted by nearby cells. Thus, individual cells in a tissue section are tabulated, indicating their type, as well as the number and types of cells within the neighborhood.

[0106] As used herein the term “neighborhood” refers to the cellular environment surrounding an individual cell. A neighborhood may be 2-dimentional, i.e., as seen on the surface of a section taken from a tissue biopsy, or 3-dimentional, i.e., occupying a cubical space around the individual cell.

[0107] In one embodiment, for implementing the methods of the present disclosure, the neighborhood has a predefined spatial perimeter having a fixed radius “r” . Different values of r may be defined in accordance with the present disclosure. These values correlate with the typical ranges of cellular communication within a tissue. Non-limiting examples include values of “r” ranging from 20 microns to 200 microns, or from 30 microns to 150 microns, or from 60 microns to 140 microns.

[0108] In other embodiments, the neighborhood does not have a predefined radius, and for implementing the methods of the present disclosure, the spatial image of the whole stained tissue section or parts thereof is considered a neighborhood.

[0109] Each possible neighborhood composition is represented as a point in a state space whose axes are the numbers of cells of each type (Fig 1 A). The rate of change of the cell concentrations in each neighborhood is deduced from division markers. By using data on many different neighborhoods, available from the spatial heterogeneity of the samples, the probability of division is obtained as a function of neighborhood cell composition. This provides the direction of change for each neighborhood composition, represented by an arrow in the state space (Fig IB). Fig. IB shows how cell division rates and cell death rates, inferred from markers, determine the direction of change in state-space. It is noted that the figure shows only division events (and not cell death events) merely for simplifying the illustration. Taking all the data together provides the phase portrait, which shows the entire dynamics (Fig 1C). The phase portrait allows detection of the system’s stable fixed points and predicts cell population dynamics from any initial concentration (Fig ID).

[0110] As used herein the term “phase portrait” refers to the directional behavior of a system of ordinary differential equations. The phase portrait can indicate the stability of the system.

[0111] As used herein the term “nullcline” refers to a set of points in the phase plane so that dx / dt = 0, or dy / dt =0. Geometrically, these are the points where the vectors are either straight up or straight down.

[0112] The method of the invention can be tailored to diverse spatial omics technologies. Namely, any method that allows the identification of a specific cell type based on the detection of specific cell markers or features, including protein-based, RNA-based, or methylation-based cell recognition methods. Non-limiting examples include multiplexed protein imaging, for example by mass cytometry (Giesen et al., Nature Methods 11, 417- 422, 2014), multiplexed ion beam imaging (Angelo et al., Nature Medicine 20, 436-442, 2014), multiplexed fluorescence microscopy methods (Gerdes et al., PNAS, vol. 110, no. 29, 11982-11987, 2013), tissue-based cyclic immunofluorescence (t-CyCIF) method for highly multiplexed imm uno-fluorescence imaging (Lin et al., Elife 7:e31657, 2018), DNA barcoding-based imaging, e.g., CODEX (Goltsev et al., Cell 174, 968-981, 2018), barcode-based spatial-transcriptomics (e.g., Visium 10X spatial transcriptomics), fluorescent based approaches e.g., MERFISH or MIBI-TOF, and methylation analysis.

[0113] Classical staining methods can also be used, e.g., for analyzing specific pairs of cells. For example, for analyzing fibroblasts and macrophages, four markers can be used: fibroblast and macrophage markers, a cell nucleus marker, and a cell division marker, e.g., KI67.

[0114] Next, a suitable software tool is applied to analyze the spatial proteomic data that includes at least one cell division marker, e.g., KI67. The analysis can be performed using, for example, standard open-source resources or a specialized software tool. In accordance with this embodiment, division events are defined using KI67 (Uxa et al., Cell Death and Differentiation 28, 3357-3370, 2021) (Fig 2A). KI67 levels peak towards the G2 / M phase of the cell cycle. Threshold KI67 levels are determined for example as shown in the examples below, and cell division is defined by the normalized KI67 above a division threshold Td, namely, first a KI67 value above the typical magnitude of noise (Tn) in the measurement is selected, then the noise is subtracted, and the value is divided by the standard deviation. Before standardization the values depend on the measurement magnitudes and have different ranges per cell type. A KI67 threshold defining a division event may range from 0.0 to 1.0, for example above 0.2, or 0.3, or 0.4, or 0.5 or 0.6, or 0.8, etc. There is no fixed cutoff value above which the model fails to perform. A person skilled in the art would appreciate that the higher the chosen threshold the fewer division events are registered and the harder it is to statistically estimate the relation between neighborhood and division rate. Thus, a high threshold renders the number of cells a limiting factor, while a low threshold increases the fraction of false division events.

[0115] For modelling, it is assumed that death rate is a constant for each cell type, namely that removal rate is not affected by the composition of the cell's neighborhood.

[0116] Accordingly, the removal rate is approximated by the mean division rate for each cell type - an assumption that is exact in the case of steady state tissues in which mean removal and division rates are equal.

[0117] In one embodiment, at least one additional marker for cell death is used.

[0118] To estimate the division probability as a function of neighboring cell composition, neighborhoods of radius r around each cell are considered, and the division (e.g., KI67 > threshold) is tabulated as a function of neighborhood composition (Fig 2B). The dependence of the probability is then modelled to divide on composition.

[0119] Any statistical inference model or machine learning model can be used next, for example logistic regression.

[0120] The number of cells of a given type in a tissue section can change by division or death, moving in or out of the section (flux) or by trans-differentiation. If the effect of trans-differentiation and flux are minimized, for example by modeling cell-types who do not transdifferentiate at appreciable rates such as T-cells and Macrophages or analyze tissues in which flux is negligible relative to rates of death or division, or add flux terms post-hoc, the rate of cell division and death practically determine the rate of change of cells in the examined tissue section.

[0121] In accordance with the present disclosure, static observations of cell division or death in a tissue are translated into rates, for example according to the following principle: if a marker for cell division (or death), remains above a defined threshold in each cell division for a period dt, then all observed divisions occurred within the last dt hours.

[0122] The exemplary rate equations shown below are non-limiting examples of tools for visualizing and interpreting the estimated dynamics. Thus, in one embodiment, the following equation is used to determine the rate of change of a population of cells of type i, also referred to as deterministic rate equation:

[0123] # deaths dr

[0124] A cell’s rate of division or death is influenced by the signals that it receives from its environment, its access to nutrients, its genetics, and factors such as physical contact with other cells. This complete set of factors is referred to herein as the cell’s “neighborhood” or “environment” . This definition sets an ideal, while the unique set of features used to approximate this ideal is denoted as N(x), whereby x is an individual cell in the tissue.

[0125] If cells with identical neighborhoods are considered, a fraction of them will be in a dividing state. This fraction is higher if the neighborhood induces a high rate of division and vice versa. Thus, the observations of division or death are viewed as random events whose probabilities are determined by the cell’s neighborhood.

[0126] Therefore, according to some embodiments, the estimated models are used to construct probabilistic rate equations, e.g., a stochastic equation for division minus death rate.

[0127] Thus, for cell x of type i, the distribution of the observation Oi,t(x) of a division or death event is modeled as follows:

[0128] 0 remaining

[0129] Where pi+, pf are the statistical inference-models for division or death, and dt+, dt" are the durations of observed division or death markers respectively. The value is divided by durations so that the following simple formula is obtained for the (stochastic) change in the number of cells in each timestep:

[0130] Tissues are heterogenous, and the diverse cellular compositions in different regions result in various directions of change. To analyze the change at a certain state, rather than the change in complete tissues, the expected change is computed with respect to an initial condition where all cells have the same neighborhood (this can be seen as placing several cells at the exact same location, in a setting without migration):

[0131] In an embodiment, the number of neighboring cells of each type within a certain predefined radius is used as features. The choice of these features is based on a separation of timescales, namely the half-life of typical signaling proteins (by endocytosis or degradation) is much shorter than the rates of cell division and death. As a result, the signaling proteins reach a steady state before the density of cells changes considerably. This steady state is then proportional to the density of secreting cells. The neighborhoods can thus be represented as a vector of cell densities, whereby each coordinate in the state vector is the density of one cell:

[0132] As a result, the models can be interpreted as components of a set of ordinary differential equations defined over a state-space of cell densities:

[0133] A deterministic ordinary differential equation can be derived from the stochastic models using the expected direction of change: Plotting the expected direction of change for each possible neighborhood composition produces a phase portrait as shown in Fig 2C for two exemplary cell types, macrophages, and fibroblasts.

[0134] It should be noted that the method of the present disclosure is inference-model- agnostic, while logistic regression is exemplified here, any other statistical inference models can be adopted within this framework.

[0135] In an embodiment, the following model inference algorithm is used: Input:

[0136] Cell level annotations: (B£, Tt, xit) for cell i = 1, ... IV. Where:

[0137] • Bi - an ID for the biopsy from which the sample cell i came from. is used to formalize that a cell can only influence cells from the same tissue.

[0138] • TtE T - cell type for cell i (e.g., Tt= “Fibroblast”) and let T be the set of types.

[0139] • For any tET let ntbe the total number of cells of type t. Some order on the cells of type t is also assumed, such that: t[j] = i for any index and the original index iE{l, . . N}.

[0140] • x[ E R2- 2D (or possibly 3D) spatial coordinates for cell i.

[0141] • Oi E {0,1} - binary observation of division

[0142] Algorithm:

[0143] For each cell-type tET,:

[0144] 1. Compute a table of features XtE RntxT.

[0145] Here t[i] is the original index of the i’th cell of type t. The value at row i and column t', is a count of all cells that came from the same tissue biopsy as cell t[i], that have type t' and are within a distance r. Row i in Xtis the representation for the “neighborhood” of cell t[i] - the counts of all cell-types in its proximity. Denote this vector by TV (t [i]) : =XT. [i, : ]

[0146] 2. Perform feature transformations on Xt, such as adding interaction terms, (transformations are selected through a separate process of cross validation)

[0147] 3. Define the binary labels ytE {0, l}nt: yt[i] = ot[£]

[0148] 4. Fit a multivariate logistic regression model p+tfor division rate based on the features and labels Xt, yt.

[0149] 5. Define death rate: p~ = mean^yt)

[0150] Output:

[0151] Enforcing Maximal Density:

[0152] Assuming there cannot be a net positive flux at the maximal observed density of a cell type, a correction is applied in the form of:

[0153] Where di is the density of cell i, and the power n controls the steepness - high n implies the correction applies primarily to higher densities (S IF). The constant term Ci is defined as the minimal correction ensuring non-positive flux at max density (Di). Division minus death rates is computed at the maximal density of one cell, over all possible values of the second cell:

[0154] In practice, the state-space region near maximal density for both cell types are unpopulated by cells, so values up to 95% quantile density are used for each type. This is more robust to artifacts due to extrapolation in areas with minimal data and produces smaller corrections. For example, in the fibroblast- macrophages (FM) model shown in the Examples below this correction is trivial because the estimated net flux was negative at high densities, for the T cells-B cells (TB) model shown in the Examples below the correction applies only to T-cells. The present disclosure further provides ex vivo methods for diagnosing a disease, for providing prognosis and for assessing the responsiveness to therapy of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of at least one tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division, and optionally cell death or cell death process state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cell, and (iii) estimating disease progression based on the cell population dynamics; and

[0155] Based on the disease progression estimates, the method provides a diagnosis or prognosis of the disease and to determine responsiveness of the patient to therapy.

[0156] In an embodiment, the patient is a human patient.

[0157] As used herein the term “disease” encompasses any pathologic condition that is manifested, among others, by the cellular composition of a tissue. This term encompasses, but is not limited to cancer (i.e., solid tumors), fibrosis, inflammation, stroke, atherosclerosis, vascular stiffness, kidney disease (e.g., glomerular kidney disease or an interstitial kidney disease), keloid or hypertrophic scar, pterygium, or autoimmune disease (e.g., scleroderma).

[0158] Non limiting examples of a solid tumor include pancreatic cancer, lung cancer, colon cancer, esophageal cancer, breast cancer (including triple negative breast cancer) or ovarian cancer.

[0159] As used herein the term “triple negative breast cancer” (TNBC) refers to a tumor in which the cancer cells are estrogen receptor negative, progesterone receptor negative and HER2 negative.

[0160] Non limiting examples of fibrosis include liver fibrosis, e.g., non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH), and primary sclerosis cholangitis (PSC), idiopathic pulmonary fibrosis (IPF), post intubation fibrosis, postsurgery fibrosis, fibrosis resulting from radiation therapy, or cardiac fibrosis.

[0161] The response to cancer treatment is often heterogeneous. Some patients respond while others do not. Treating a patient with ineffective treatment may not only cause unnecessary side-effects but also prevents or defers the administration of alternative treatments that could be more effective. Thus, it is important to know as early as possible if the administered therapy is beneficial.

[0162] As shown in the Examples below, and in accordance with the present disclosure, OSDR was used to analyze data from a triple-negative breast cancer clinical trial where biopsies were taken at the start and end of treatment. The trial included two treatment arms where patients received either chemotherapy or a combination of chemotherapy and immunotherapy. Using biopsies collected at an early stage of treatment (week 3 of 24) OSDR accurately predicted the collapse of the tumor cell population in responders. This result was reproduced in both treatment arms and was robust to the subset of patients used for fitting the model.

[0163] Accordingly, in one of its aspects, the present invention provides an ex vivo method for assessing the responsiveness of a solid tumor to an antitumor therapy administered to a patient, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a solid tumor biopsy taken at different times comprising one or more times during the antitumor therapy and optionally one or more times prior to the antitumor therapy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) estimating tumor progression based on the cell population dynamics; and c. based on changes in the estimated tumor progression between the sections determining responsiveness of the solid tumor to the antitumor therapy.

[0164] The effects of any antitumor therapy can be assessed in accordance with the present disclosure including but not limited to irradiation therapy, chemotherapy, immunotherapy, treatment with biologicals, cell therapy and combinations thereof.

[0165] As shown in the Examples below (see e.g., Example 5), initially, there was a low density of T-cells, and the proliferation rate of tumor cells was similar in both responders and non-responders. The difference between the two groups became evident only when the number of T-cells increased. In responders, tumor proliferation dropped near T-cells. As a result, when T-cell levels rise the tumor cell population begins shrinking.

[0166] Accordingly, in accordance with the present disclosure a drop in tumor proliferation near T-cells serves as a functional marker of response to immunotherapy. Furthermore, as shown in the Examples below, in non-responders, tumor proliferation increases near T-cells. As a result, an increase in T-cells following immunotherapy increases tumor proliferation rate. The change in tumor cell proliferation as a function of the number of T cells in the vicinity of the tumor cells can therefore serve as a simple, easy to apply biomarker of response to immunotherapy.

[0167] Thus, in one of its aspects, the present invention provides an ex vivo method for assessing the responsiveness of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; wherein a reduction in tumor proliferation near T cells indicates that the solid tumor is responsive to the immunotherapy, and wherein an increase in tumor proliferation near T cells indicates that the solid tumor is not responsive to immunotherapy.

[0168] In addition, even when measured prior to commencement of treatment tumor proliferation in neighborhoods with large numbers of T cells was lower in responders than in non-responders to immunotherapy.

[0169] Thus, in another one of its aspects, the present invention provides an ex vivo method for predicting the likelihood of response of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at one or more times prior to the immunotherapy; wherein a low level of tumor proliferation near T cells indicates that the solid tumor will be responsive to the immunotherapy, and wherein a high level of tumor proliferation near T cells indicates that the solid tumor will not be responsive to immunotherapy.

[0170] As used herein the term “high level” of tumor proliferation refers to a high level relative to the mean proliferation rate over the entire tissue, or a high level relative to neighborhoods without T-cells. As used herein the term “low level” of tumor proliferation refers to a low level relative to the mean proliferation rate over the entire tissue, or a low level relative to neighborhoods without T-cells.

[0171] In other words, in accordance with the invention, a model was fit that predicts the tumor proliferation rate based on the number of cells in its neighborhood, including T- cells and based on this model, a negative correlation was found between tumor proliferation and the number of T-cells in responders to immunotherapy, and a positive correlation was found in non-responders to immunotherapy. Thus, a negative correlation is typical of response (i.e., proliferation drops with an increase in the number of T-cells), while a positive correlation is more typical of non-response.

[0172] Moreover, an increase of tumor proliferation near T-cells can serve as an early marker for hyper-progression. As used herein the term “hyper progression” refers to a subset of patients (also termed “hyper-progressors”) which are harmed by immunotherapy. In these patients, the tumor proliferates more rapidly following immunotherapy.

[0173] The present invention thus provides a method for preventing hyper-progression. The method comprises obtaining a biopsy before treatment and / or at an early stage during treatment and visualizing T cells and tumor cells in the biopsy. If tumor proliferation increases near T-cells, treatment is halted before the TME changes sufficiently to induce hyper-progression.

[0174] Thus, in another aspect, the present invention provides a method for preventing hyper progression of a solid tumor in response to immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; b. determining whether there is an increase in tumor proliferation near T cells; c. If an increase in tumor proliferation near T cells is detected, halting administration of the immunotherapy to the patient. In an embodiment, the solid tumor biopsy is taken prior to the administration of immunotherapy.

[0175] In an embodiment, the solid tumor biopsy is taken at an early stage of immunotherapy. An early stage of treatment can be determined by the skilled artisan and may vary between types of solid tumors and types of the immunotherapy. In non-limiting examples, an early stage is between one week and 6 weeks of therapy, between 2 weeks and 4 weeks of therapy, or after 3 weeks of therapy.

[0176] In an embodiment a three-marker staining procedure is employed comprising staining the biopsies with markers for (i) tumor cells, (ii) T-cells and (iii) a proliferation marker (e.g., Ki67).

[0177] As shown in the Examples section below, the estimated dynamics provided by the methods of the current disclosure can be used to identify points of potential therapeutic intervention that drive cell populations towards desired densities.

[0178] For example, if the methods of the present disclosure show that cell X increases density of cell Y which has pathological ramifications; this understanding can facilitate therapeutic decisions; for example, taking a decision to block the interaction between cell X and cell Y by administering specific agents that neutralize growth factors secreted by cell X and which affect cell Y.

[0179] In another aspect, the present invention concerns timing the treatment based on the estimated population dynamics.

[0180] The frequency and dose of immunotherapy treatments is commonly established through large clinical trials. Thus, each patient is treated with the same dose, and same frequency. One of the significant limitations of this approach is that treatment must be stopped in approximately 30% of patients due to severe side-effects. On the other hand, there are large fractions of patients who do not respond. Moreover, currently, physicians do not have a reliable way of timing immunotherapy.

[0181] As shown in the Examples below (see e.g., Example 4) the T-cell and B-cell dynamics in the tumor microenvironment are characterized by flares / pulses. If T-cells cross a certain density level, a flare is triggered where both cell types rapidly proliferate. Then, when the level of B-cells rises, both cell populations collapse. Notably, adding T- cells (or activating T-cells using immunotherapy) while B-cells are still abundant, will not trigger another flare. This implies that the determination of immunotherapy frequency should consider the native duration of immune flares. Accordingly, it may be advisable to defer applying a subsequent immunotherapy session until the B cell levels drop, because the model of the invention infers that they inhibit T-cell proliferation. Reducing frequency could potentially reduce toxic side effects.

[0182] Thus, in accordance with the invention, OSDR is applied to analyze the duration of T cell and B cell flares in patients’ tumors. Treatment frequency of these patients may then be adjusted according to the pulsed T cell and B cell dynamics, e.g., frequency of treatment sessions may be reduced for some patients to allow B-cells to drop before the next dose. In addition, therapeutic interventions may be applied to modulate the T cells and B cells pulses such that they match the immunotherapy protocol.

[0183] Accordingly, in another aspect, the present invention provides a method for optimizing immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken at one or more times during the immunotherapy; b. for each section, (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell population dynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. adjusting the frequency of immunotherapy sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or d. administering an additional therapeutic intervention to the patient, thereby optimizing immunotherapy

[0184] In another aspect, the present invention provides a method of treatment of a solid tumor in a patient by immunotherapy, said method comprising a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken at one or more times during the immunotherapy; b. for each section, (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell population dynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. administering immunotherapy to the patient at a frequency of sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or d. administering an additional therapeutic intervention to the patient.

[0185] Depending on the specific T cells and B cells dynamics in the individual patient’s tumor, the therapeutic intervention may either be directed to stimulating the immune system, i.e., increasing T cells activation, or suppressing the immune system.

[0186] In an embodiment, determining cell population dynamics of T cells and B cells in the solid tumor comprises, but is not limited to, providing a predicted peak density of T cells, determining the threshold T cell density for a flare, or determining an estimated duration of flare.

[0187] For example, if the predicted peak density of T cells for a patient is low compared with peaks observed in responders to immunotherapy, a stimulating agent can be administered. Similarly, immune activation can be applied if the threshold density for a flare is higher than the typical threshold in responders to immunotherapy, or if the estimated duration of flares is too short, e.g., shorter than the duration of flares in responders to immunotherapy (thus allowing the tumor sufficient time to recover).

[0188] And conversely, if the predicted peak density of T cells for a patient is high compared with peaks observed in responders to immunotherapy, an immune suppressive agent can be administered. Similarly, immune suppression can be applied if the threshold density for a flare is lower than the typical threshold in responders to immunotherapy, or if the estimated duration of flares is too long, e.g., longer than the duration of flares in responders to immunotherapy.

[0189] In one embodiment, the therapeutic intervention comprises immune stimulation. In one embodiment, the immune stimulation is affected by increasing the dose of checkpoint inhibitors (e.g., pembrolizumab, ipilimumab) that are administered to the patient. In other embodiments, immune stimulation is affected by administering inflammatory cytokines (e.g., IL-2), or agents that cause B-cell killing (e.g., using anti lymphoma drugs such as Rituximab (anti CD20)). In another embodiment, the therapeutic intervention comprises immune suppression. In some embodiments, immune suppression is affected by reducing the dose of checkpoint inhibitors (e.g., pembrolizumab, ipilimumab) that are administered to the patient. In other embodiments, immune suppression is affected by administering immunosuppressive drugs such as corticosteroids or anti IL2 antibodies, or by chemotherapy.

[0190] In another embodiment, the therapeutic intervention comprises behavioral intervention such as adapting an anti-inflammatory nutrition (e.g., fasting), sleep modulation or meditation techniques.

[0191] Furthermore, the methods provided by the present disclosure can be utilized within a control system loop that includes continuous feedback. For example, it may be connected to a remote device (e.g., a computer) which controls a monitored drug delivery system, and which receives feedback information that is fed into the model of the present disclosure.

[0192] Accordingly, the present disclosure also provides methods of treatment of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) identifying points of intervention that drive cell populations towards desired densities; and c. administering a therapeutic agent to the patient at the identified points of intervention.

[0193] The term "about" as used herein indicates values that may deviate up to 1%, more specifically 5%, more specifically 10%, more specifically 15%, and in some cases up to 20% higher or lower than the value referred to, the deviation range including integer values, and, if applicable, non-integer values as well, constituting a continuous range. Disclosed and described, it is to be understood that this invention is not limited to the examples, methods steps, and compositions disclosed herein as such methods steps and compositions may vary somewhat. It is also to be understood that the terminology used herein is used for the purpose of describing specific embodiments only and not intended to be limiting since the scope of the present invention will be limited only by the appended claims and equivalents thereof.

[0194] It must be noted that, as used in this specification and the appended claims, the singular forms “a”, “an” and “the” include plural referents unless the content clearly dictates otherwise.

[0195] Throughout this specification and the Examples and 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 integers or steps.

[0196] EXAMPLES

[0197] Example 1

[0198] Inferring cellular population dynamics from a spatial omics snapshot

[0199] To test the feasibility of the OSDR approach, simulated data was used first and then experimental data on breast cancer (MIBI spatial proteomics samples (Keren et al., Sci. Adv. 5 (10), eaax5851, 2019), was analyzed. Simulation of dynamical models:

[0200] Spatial data was simulated based on a dynamical model (i.e., functions p+and p") using the following procedure:

[0201] 1. A random initial number of cells was sampled for each cell type.

[0202] 2. A random spatial position in the tissue was sampled for each cell.

[0203] 3. For n steps:

[0204] For each cell the probability of division or death was computed based on its current neighborhood.

[0205] An event of division, death or none was sampled based on the computed probabilities.

[0206] Dead cells were removed, and a new cell was placed next to each dividing cell. The location of the new cells was sampled uniformly within the neighborhood. Optionally, knowledge about cell motility was incorporated at this stage, if available.

[0207] This simulation produces a tissue section whose density and spatial distribution is biased by the dynamics, as is the case for real tissues. The procedure was repeated 100 times from various initial densities and sampled with replacement 50,000 cells evenly from all tissues. From this pool 1,000, 5,000, 10,000 or 25,000 cells were sampled for the model fits.

[0208] Since empirical distributions in the data had tails towards lower cell-densities, the initial cell densities were sampled from a beta distribution biased towards lower densities (parameters 2, 4 scaled to a maximal value of 7). Model parameters were selected to produce division rates in ranges that resemble real data (mostly within l%-6%). The fraction of divisions was approximately 2% for all models. The number of simulation time steps was selected as the number that is required to produce distributions qualitatively similar to real data.

[0209] To evaluate model fits it was tested whether the correct location and type of stable or semi-stable fixed points were recovered. To account for discretization error, if a stable point on an axis was recovered, and an unstable point was located within 1 cell from that point, it was counted as a semi-stable. Such a point effectively has no basin of attraction. The same precaution should apply to interpretation of model fits in general.

[0210] Using these simulations, the number of cells that are required to reconstruct a known dynamical system was determined. In each simulation a specific dynamical system was stipulated in which cells affect each other's growth rate. Then experimental spatial data was simulated by running the dynamics from various initial conditions of cell density, for a period that results in distributions of spatial cell concentrations that resemble experimental data. This creates a variety of cell compositions.

[0211] OSDR was used to reconstruct the phase portraits which were then compared to the actual phase portrait of the known dynamic system used in each simulation. All possible phase portrait topologies of two cell types were studied, where each cell type can be stable on its own, or stable only in the presence of the other cell type. This results in 4 phase portrait topologies (Fig 3). As shown in Fig. 3, the OSDR method of the current disclosure accurately reconstructed dynamics from simulations of known dynamical circuits based on simulated data of 10,000 cells.

[0212] It was found that a few thousand cells are typically sufficient to reconstruct the fixed-point structure reliably. Some aspects were recovered with very few cells (as little as 1000), such as a stable point off-axis or the presence of any kind of stability on an axis. Discerning between types of stability on the axes requires more cells. At 25,000 cells all aspects were recovered with high accuracy. Accordingly, the OSDR method of the present disclosure can reconstruct simple 2D dynamics from simulated cell data and recover the underlying phase portrait given enough cells.

[0213] Example 2

[0214] OSDR reconstructs breast cancer fibroblast and macrophage dynamics in agreement with in-vitro dynamic experiments.

[0215] Next the OSDR approach was tested using spatial proteomics data. The breast cancer microenvironment was selected as a test case. The breast cancer microenvironment includes interactions between cancer associated fibroblasts and macrophages. Fibroblasts and macrophages communicate via growth factors in a paracrine manner. In addition, growth factors secreted by each of these cell types also participate in autocrine regulation. Established in vitro dynamics for fibroblasts and macrophages in a breast cancer medium (Mayer et al., Nature Communications 14, Article No. 5810, 2023) were used as a reference phase portrait for the cell population dynamics.

[0216] In Mayer et al., the in-vitro phase portrait was obtained by seeding mice mammary fibroblasts (F) and bone-marrow derived macrophages (M) at different initial concentrations in co-cultures and following the changes in the cell populations over several days. This is a measurement of the population level dynamics. In the presence of breast-cancer conditioned medium, F and M cells showed a phase portrait with several distinct fixed points. Fibroblasts and macrophages supported each other in a fixed point called ‘hot fibrosis’ - in the sense that both fibroblasts and macrophages coexist. Fibroblasts alone could support themselves in a ‘cold fibrosis’ fixed point. Macrophages were induced by the cancer conditioned medium to form a macrophage-only fixed point at high macrophage densities.

[0217] To test the ODSR method of the current disclosure, a MIBI spatial proteomics dataset of human breast cancer samples (Danenberg et al., 2022) was analyzed. The dataset includes samples from 693 treatment-naive primary breast tumor biopsies totaling 1.12 million cells. The samples are from patients recruited to the METABRIC study (Pereira et al., Nat. Commun. 7: 11479, 2016) and include various breast-cancer genotypes and stages. Each sample is approximately a 500pm x 500pm square of tissue, and all samples were imaged on the same slide. Cell types were identified using standard markers and included fibroblasts, macrophages, endothelial cells, adaptive immune cells, and epithelial cells. The data includes the KI67 marker.

[0218] Since OSDR infers division probability (KI67>threshold) based on the identity of neighboring cells, it was first tested whether variation in KI67 is indeed explained by the composition of the cell’s neighborhoods in the data. To compensate for potential unmeasured elements that may affect the dynamics, such as zones with different levels of metabolites, hypoxia, inflammation, and other factors, that stem from proximity to blood vessels or tumor mass, the density of endothelial and / or tumor cells were also included in the model. Indeed, the estimated dynamics were preserved across such zones. Similarly, the estimated dynamics were also preserved across various patient subgroups defined by external factors such as tumor genetics or stage.

[0219] Logistic regression was performed with cell counts as features (number of neighboring cells of each type) and KI67 as target (KI67 over threshold).

[0220] Ki67 Thresholds: previous data was used to establish a KI67 threshold for cell division events. Uxa et al., 2021 demonstrated that KI67 levels peak towards the G2 / M phase of the cell cycle, with preserved kinetics (up to scale) across two human and one mouse cell line. KI67 levels were therefore adjusted to correct for different scaling and division rates between cell lines by selecting KI67 values above a noise threshold Tn, subtracting Tn and dividing by the KI67 standard deviation. Tn=0.5 mean isotopic counts were chosen because this was the typical magnitude of experimental noise in this dataset. This produced distributions with similar shapes in accordance with the cell line results of Uxa et al. A cell division is thus defined by the normalized KI67 value above a division threshold Td. The resulting model estimates are robust to Td values between 0 and 1 on this adjusted scale.

[0221] Based on this threshold and using OSDR to perform statistical inference on MIBI data from human breast cancer biopsies, it was found that neighbor cell-counts accurately capture a wide range of division rates (Fig 4A-D). All fits were very significant (p-value < 10'13). It is therefore concluded that the counts of neighboring cells are a strong predictor of division rates.

[0222] Next OSDR was used to obtain a phase portrait for cancer-associated fibroblasts and macrophages, as described above. Only neighborhoods with these two cell-types, along with epithelial and endothelial cells were analyzed. Neighborhoods with adaptive immune cells were excluded since these cells affect the dynamics, and were not present in the in vitro experiments performed by Mayer et al. The samples included a total of 69,874 fibroblasts with 796 divisions (1.1%), and 3,761 macrophages with 79 divisions (2.1%).

[0223] The OSDR approach used the division probabilities of fibroblasts (Fig 4E) and macrophages (Fig 4F) as a function of neighborhood composition to yield the phase portrait shown in Fig 4G.

[0224] Notably, the reconstructed phase portrait shares most of the key features of the in- vitro portrait presented by Mayer et al. There is a stable fixed point in which fibroblasts and macrophages support each other, namely a hot fibrosis fixed point. There is a cold fibrosis fixed point with fibroblasts only. There is a third semi stable fixed point with high numbers of macrophages without fibroblasts.

[0225] The fixed-point structure of the phase portrait inferred by OSDR was found to be robust to resampling cells and samples. It was also consistently found in a wide range of parameter choices (e.g., neighborhood radius, KI67 threshold), perturbations in the death rates, inclusion of endothelial or tumor cells in the statistical model, namely inclusion of a third cell type in the model, or when considering only patient subgroups with specific cancer stage, tumor size, breast-cancer genotype, patient survival time or type of treatment.

[0226] It is thus concluded that the reconstructed phase portrait captures the salient features of the in-vitro phase portrait. Whereas the in-vitro phase portrait used dynamical measurements of cell populations over days, the reconstructed portrait uses a single snapshot with cell density and division information.

[0227] Example 3

[0228] Estimating magnitude of error due to cell migration

[0229] After estimating the cell division and death dynamics using markers, an independent source of information available from the same data - the empirical distribution of tissue states (i.e., density of each type) can be used to quantify the magnitude of error due to unmodeled terms. This can be done for example by using the mass continuity equation.

[0230] The following is a non-limiting example for a mass continuity equation that can be applied in accordance with an embodiment of the present disclosure for estimating magnitude error due to an unmodeled feature. Each neighborhood is analogous to a particle moving through a space whose coordinates are defined by the densities of each cell type. For example, x, y, could be the density of fibroblasts and macrophages. The velocity through this space is determined by the rate of change in each cell type. This velocity is the sum of v(x, y), the division minus removal rates that are estimated from the data, and u(x, y), which includes unmodeled terms such as cell migration and differentiation from stem cells.

[0231] Notation: p(x,y) - density of neighborhoods at location x, y. v(x, y) + u(x,y) - the velocity of a neighborhood in the state-space. v(x, y)- division minus death rate field. u(x, y) - velocity field due to unmodeled terms such as cell migration

[0232] For large samples of patients, it is assumed that sampling the same tumors within a few days would produce approximately the same distribution. This implies that Thus:

[0233] Accordingly, the divergence of the cell migration field can be computed:

[0234] This is sufficient for identifying the regions in state-space where cell migration would produce a net increase or decrease in density, p is estimated by computing the location of each patient’s tissue in the state-space followed by kernel density estimation. The divergence V(p • v) is then numerically evaluated. By plotting V(p • v) the negative divergence of the migration field is obtained. This produces the sources (negative 7 (p • v)) and sinks (positive 7 (p • v)) of the cell migration field. For interpretability the rates are divided by the maximal p. Accordingly, units are in [fractions of maximal density / At], The fibroblast-macrophage model is provided herein as a non-limiting example: For the Fibroblast-Macrophage model the residual field produces a net effect that disperses cells from the peak density. This form of error could be explained by unmodeled stochasticity; accordingly, the dynamics are modeled using a Fokker-Plack equation. Let X be the location of a neighborhood in state-space, then the velocity is given by the following stochastic differential equation:

[0235] Where e is a standard Brownian motion.

[0236] The Fokker-Planck equation states that the density of neighborhoods changes according to:

[0237] Assuming = 0, the cell migration field is modeled as:

[0238] D that best explains the error of the deterministic model is selected using ordinary least squares regression with D as a single parameter:

[0239] Example 4

[0240] An excitable pulse-generating circuit of T and B cells in the breast cancer microenvironment

[0241] Next, two other cell types from the breast cancer MIBI dataset were considered, T cells and B cells. These cells are components of the adaptive immune system which play a major role in the cancer microenvironment. T and B cells interact in the tumor microenvironment in autocrine and paracrine manners via secreted cytokines or by direct contact.

[0242] All neighborhoods which contained at least one T and / or one B cell were analyzed. This included a total of 74,006 T cells and 27,643 B cells, with 1069 and 213 division events respectively. Thus, about 1.4% of the T cells and 0.7% of the B cells showed division events, consistent with previous breast cancer data (Wu et al., Nature Genetics 53, 1334-1347, 2021).

[0243] The dynamics between T and B cells were estimated using the OSDR method of the current disclosure. The inferred phase portrait was found to have a striking feature called excitability. It showed a stable fixed point at zero cells (Fig 5A). However, if T cells are raised above a threshold, they generate a pronounced pulse in which T cells rise, followed by B cells, and then both populations are reduced to zero (Fig 5B). This pulse is similar to pulses observed in autoimmune diseases such as relapsing-remitting MS (Lebel et al., iScience 26, 108084, 2023).

[0244] As for the fibroblast-macrophage analysis shown in Example 2, the estimated dynamics were robust to resampling at the level of both cells and patients and were consistent under a wide range of death rates, neighborhood sizes and KI67 thresholds, namely to neighborhood sizes with a radius ranging from 20 microns to 200 microns and to KI67 thresholds ranging from 0.0 to 1.0, and even higher. The dynamics were not confounded by unmodeled cell types, namely when T cells and B cells were modeled together with a third cell type (e.g., tumor cells, macrophages, endothelial cells, or fibroblasts) the dynamics remained consistent. Furthermore, the dynamics remained consistent for different patient subgroups; namely in patients with short survival or long survival, in patients with differing tumor size (i.e., below or above the median tumor size) or tumor stage (stages 0, 1 as compared with stages 2, 3, and 4), in patients who were ER negative or positive, or HER2 negative or positive, in patients who received hormone therapy or not, or in patients that underwent mastectomy or a breast conserving surgery.

[0245] The pulsatile adaptive immune response in the inferred model is shown in Fig 5B, where T cells rise, followed by B cells that inhibit them, and then both cell populations decline. The model further indicated that after a pulse there is a refractory period, where B cells are still high and T cells are low, and new pulses cannot be triggered. As evidenced in the simulations, during this period, even when additional T cells are added a9shown in the figure as small vertical arrows) a new pulse cannot be generated due to the inhibitory effects of the B cells (Fig 5C). A new pulse is possible only after the recovery period, namely only when T cells are introduced after the B cells from the previous pulse have declined (Fig 5C). Example 5:

[0246] OSDR predicts the collapse of the tumor-cell population in treatment responders based on early-treatment biopsies

[0247] Cancer patients can go through many months of treatment without knowing if the tumor was growing or shrinking. To adjust treatment when it is ineffective, early signs of response are needed. To address this challenge OSDR was applied to a clinical trial which provided three longitudinal biopsies - the NeoTRIPtrial (Gianni et al., 2022). A cohort of 279 triple negative breast cancer (TNBC) patients were randomly assigned to neoadjuvant chemotherapy (n=141) or chemotherapy + immunotherapy (n=138). Biopsies were collected at three timepoints: before treatment (defined as “baseline”), 3 weeks into treatment (first day of the second treatment cycle, defined as “early treatment”) and posttreatment (at surgical excision of the tumor following eight 3 -week treatment cycles, namely at ~24 weeks). Pathologists labeled post-treatment biopsies as pathological complete response (pCR) or residual disease (RD). Data on the spatial staining of the biopsies was taken from Wang et al (2023).

[0248] OSDR was applied to week-3 biopsies from responders and non-responders in each treatment arm (total of four groups). The dynamics of six cell types were modelled - fibroblasts, macrophages, tumor, endothelial, T- and B-cells. The composition of the tissue over time was then predicted. To detect differences resulting from dynamics, as opposed to differences resulting from initial tissue composition or tumor burden, both responder and non-responder models were applied to the same initial states - namely, both models were applied to every early-treatment biopsy. Figs 6A-6L show the predicted celldensity over time, averaged over all starting states. Remarkably, in both treatment arms, OSDR predicts the collapse of the tumor-cell population in responders, but not in non- responders. Non-tumor cell types followed similar trajectories.

[0249] This analysis shows that the collapse of tumor cells in responders is not a result of the initial tissue state, but rather a result of the dynamics (since the models were applied to all initial states). Moreover, it was found that the average proliferation rate of tumor cells at week 3 does not distinguish between responders and non-responders. In fact, the division rate of tumor cells in responders was higher than non-responders in the immunotherapy treatment arm. Thus, the tumor cell population collapses because of dynamics and not because of initial tissue composition or average proliferation rates. A striking feature of the predicted trajectories in immunotherapy is a sharp rise in T-cells (Fig 6D and 6J). This rise appears in both responders and non-responders receiving immunotherapy but not in patients in the chemotherapy arm. To further investigate this correlation, tumor division rate in all neighborhoods compared with tumor division rate in high T-cell neighborhoods was plotted. For each tumor cell the number of T-cells in its neighborhood (80-micron radius) was counted. The data presented in Fig 7A-7B concerns tumor cells whose number of T-cell neighbors was at the top 1% of all tumor cells. In responders, tumor division rate drops near T-cells, while in non- responders, tumor division rate increases near T-cells.

[0250] Furthermore, even when analyzing biopsies taken prior to treatment, comparing pre-treatment tumor division rate in all neighborhoods with tumor division rate in high T-cell neighborhoods (Fig 8A-8B), in responders, pre-treatment tumor division rate drops near T-cells, while in non-responders, tumor division rate increases near T-cells.

[0251] A dynamical perspective is essential for predicting the collapse of the tumor cell population. At the beginning of treatment there is a low density of immune cells. As a result, the tumor population grows in both groups. When the density of immune cells increases a difference emerges between responders and non-responders (Fig 9A-9B). A difference is detected between responders and non-responders by focusing on the division rate of tumor cells at the state the TME is moving towards.

Claims

CLAIMS:

1. A method for predicting changes in a tissue, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; and c. determining cell population dynamics and predicting said changes based thereon.

2. The method of claim 1, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

3. The method of any one of claims 1 or 2, wherein said method comprises obtaining information on individual cells of at least two different cell types in the tissue section.

4. The method of any one of claims 1 to 3, wherein said information is obtained at a single time point.

5. The method of any one of claims 1 to 4, wherein said information comprises data obtained by one or more of multiplexed protein imaging, spatial-transcriptomics, methylation analysis, fluorescent based cell staining, mass cytometry, and X-ray based imaging.

6. The method of any one of claims 1 to 5, wherein the information on spatial localization and cell type is obtained by (x) staining the section with cell type specific markers, (y) determining location of individual cells within the tissue section and (z) quantifying the number of at least one cell type within a spatial perimeter surrounding the individual cell.

7. The method of any one of claims 1 to 6, comprising staining the sections with cell division-specific dyes or markers.

8. The method of any one of claims 1 to 7, comprising staining the cells with cell death- or cell death-specific dyes or markers.

9. The method of any one of claims 1 to 8, wherein said calculating cell division probabilities comprises:bl. Defining a neighborhood around each of said individual cells in the section; and b2. calculating the division probability for each of the individual cells as a function of cell composition and optionally spatial organization in each defined neighborhood.

10. The method of claim 9 wherein said neighborhood comprises cells surrounding each of said individual cells.

11. The method of any one of claims 9 or 10 wherein said neighborhood comprises cells in a predefined radius.

12. The method of any one of claims 9 to 11, wherein said step (b2) is performed using a statistical inference model.

13. The method of claim 12 wherein said statistical inference model is logistic regression.

14. The method of any one of claims 9 to 13 wherein said step (b2) is performed using a machine learning algorithm.

15. The method of any one of claims 1 to 14, wherein said step of determining cell population dynamics comprises inferring probabilistic rate equations based on the cell division probabilities.

16. The method of claim 15, comprising deriving a deterministic differential equation from the probabilistic rate equation, predicting the direction of change that would have been expected in the tissue section had it been still within the body.

17. The method of any one of claims 1 to 16, wherein the different cell types comprise epithelial cells, mesenchymal cells, endothelial cells, antigen presenting cells, leukocytes, lymphoid cells, or myeloid cells.

18. The method of any one of claims 1 to 17, wherein the different cell types comprise fibroblasts and macrophages.

19. The method of any one of claims 1 to 17, wherein the different cell types comprise T cells and B cells.

20. The method of any one of claims 1 to 19, wherein the tissue biopsy is from a solid tumor.

21. An ex vivo method for estimating the progression of a solid tumor in a patient, said method comprising:a. obtaining information on individual cells of at least one cell type in a section of a solid tumor biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; c. determining cell population dynamics of each type of cells; and d. estimating potential tumor progression based on the cell population dynamics.

22. The method of claim 21, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

23. The method of any one of claims 21 or 22 further comprising providing disease prognosis.

24. An ex vivo method for assessing the responsiveness of a solid tumor to an antitumor therapy administered to a patient, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a solid tumor biopsy taken at different times comprising one or more times during the antitumor therapy and optionally one or more times prior to the antitumor therapy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) estimating tumor progression based on the cell population dynamics; and c. based on changes in the estimated tumor progression between the sections determining responsiveness of the solid tumor to the antitumor therapy.

25. The method of claim 24, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

26. The method of any one of claims 24 or 25 further comprising providing assessment of the response of the patient to the antitumor therapy.

27. The method of any one of claims 21 to 26, wherein said method comprises obtaining information on individual cells of at least two different cell types in the solid tumor sections.

28. The method of claim 27, wherein the different cell types comprise tumor cells and non-tumor cells.

29. The method of any one of claims 21 to 28, wherein said solid tumor is pancreatic cancer, lung cancer, colon cancer, esophageal cancer, or ovarian cancer.

30. An ex vivo method for estimating the progression of a disease in a patient, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. calculating cell division probabilities for each of said individual cells; c. determining cell population dynamics of each type of cells; and d. estimating potential disease progression based on the cell population dynamics.

31. The method of claim 30, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

32. The method of any one of claims 30 or 31 further comprising providing disease prognosis.

33. An ex vivo method for assessing the responsiveness to therapy of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a tissue biopsy taken at different times comprising one or more times during therapy and optionally one or more times prior to the therapy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells and (iii) estimating disease progression based on the cell population dynamics; andc. based on changes in the estimated disease progression between the sections determining responsiveness of the patient to the therapy.

34. The method of claim 33, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

35. The method of any one of claims 30 to 34, wherein said disease is fibrosis, inflammation, stroke, atherosclerosis, vascular stiffness, kidney disease, keloid or hypertrophic scar, pterygium, or autoimmune disease.

36. The method of claim 35 wherein said fibrosis is liver fibrosis, idiopathic pulmonary fibrosis (IPF), post intubation fibrosis, post-surgery fibrosis, fibrosis resulting from radiation therapy, or cardiac fibrosis.

37. The method of claim 35 wherein said autoimmune disease is scleroderma,38. The method of claim 35 wherein said kidney disease is a glomerular kidney disease or an interstitial kidney disease.

39. The method of any one of claims 30 to 38, comprising obtaining information on individual cells of at least two different cell types in the tissue biopsy sections.

40. A method of treatment of a patient suffering from a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) identifying points of intervention that drive cell populations towards desired densities; and c. administering a therapeutic agent to the patient at the identified points of intervention.

41. The method of claim 40, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

42. The method of any one of claims 40 or 41, comprising obtaining information on individual cells of at least two different cell types in the tissue biopsy sections.

43. A method for assessing the efficacy of a therapeutic agent for treating a disease, said method comprising: a. obtaining information on individual cells of at least one cell type in sections of a tissue biopsy, taken at different times comprising one or more times after administration of the therapeutic agent and optionally one or more times prior to administration of the therapeutic agent, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; b. for each section (i) calculating cell division probabilities for each of said individual cells (ii) determining cell population dynamics of each type of cells, and (iii) estimating the pathological condition of the tissue based on the cell population dynamics; and c. based on changes in the pathological condition of the tissue between the sections determining the efficacy of the therapeutic agent.

44. The method of claim 43, wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

45. The method of any one of claims 43 or 44, comprising obtaining information on individual cells of at least two different cell types in the tissue biopsy sections.

46. The method of any one of claims 43 to 45, wherein the tissue biopsy is taken from an animal model of the disease.

47. The method of any one of claims 43 to 45, wherein the tissue biopsy is taken from an organoid grown ex vivo.

48. A method for optimizing immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken at one or more times during the immunotherapy; b. for each section (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell populationdynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. adjusting the frequency of immunotherapy sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or d. administering an additional therapeutic intervention to the patient, thereby optimizing immunotherapy.

49. A method of treatment of a solid tumor by immunotherapy, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual B cells in sections of a solid tumor biopsy taken from a patient at one or more times during the immunotherapy; b. for each section (i) calculating cell division probabilities for each of said individual T cells and individual B cells, (ii) determining cell population dynamics of T cells and B cells in the solid tumor, and (iii) determining time points in which a drop in B cell population is expected; c. administering immunotherapy at a frequency of sessions such that an immunotherapy session is administered when a drop in B cell population is expected, and / or administering an additional therapeutic intervention to the patient.

50. A computer software product, comprising a computer-readable medium in which program instructions are stored, which instructions, when read by a data processor, configure the data processor to obtain information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and (iii) cell division state; and to (2) execute the method according to any one of claims 1-49.

51. A system for predicting changes in a tissue, comprising: an input utility for obtaining information on individual cells of at least one cell type in a section of a tissue biopsy, said information comprises: (i) spatial localization, (ii) cell type and(iii) cell division state; and a data processor configured for analyzing said data for executing the method according to any one of claims 1-49.

52. The computer software product of claim 50, or the system of claim 51 wherein said information comprises (iv) whether cells are dead or alive or (v) undergoing a cell death process, and wherein said calculating comprises calculating cell death probabilities.

53. An ex vivo method for assessing the responsiveness of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; wherein a reduction in tumor proliferation near T cells indicates that the solid tumor is responsive to the immunotherapy, and wherein an increase in tumor proliferation near T cells indicates that the solid tumor is not responsive to immunotherapy.

54. The method of claim 53, wherein if the solid tumor is not responsive to immunotherapy, the method further comprises halting the immunotherapy administered to the patient.

55. An ex vivo method for predicting the likelihood of response of a solid tumor to immunotherapy administered to a patient, said method comprising: obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at one or more times prior to the immunotherapy; wherein a low level of tumor proliferation near T cells indicates that the solid tumor will be responsive to the immunotherapy, and wherein a high level of tumor proliferation near T cells indicates that the solid tumor will not be responsive to immunotherapy.

56. The method of claim 55, wherein if the prediction is that the solid tumor will not be responsive to immunotherapy, the method further comprises refraining from administering immunotherapy to the patient.

57. A method for preventing hyper progression of a solid tumor in response to immunotherapy administered to a patient, said method comprising: a. obtaining information on (i) spatial localization and (ii) cell division state of individual T cells and individual tumor cells in sections of a solid tumor biopsy taken at different times, comprising one or more times during the immunotherapy and optionally one or more times prior to the immunotherapy; b. determining whether there is an increase in tumor proliferation near T cells; and c. If an increase in tumor proliferation near T cells is detected halting the administration of the immunotherapy to the patient, thereby preventing the hyper progression of the solid tumor in said patient.

Images