Platform for co-culture imaging to characterize in vitro efficacy of heterotypic effector cellular therapies in cancer
A method for characterizing cancer organoid response to an immune cell based therapy, includes providing a panel of different combinations of cancer organoid cells and immune cells to culturing wells and culturing the different combination under conditions that support organoid growth. Brightfield and corresponding fluorescence images of the culturing wells are captured and provided to one or more trained machine learning algorithms that identify and distinguish cancer organoid cells from immune cells and characterize cancer organoid morphology changes caused by an immune cell based therapies, from which an analytical report including a characterization of cancer organoid cell death caused by the immune cell based therapy is provided.
1 . A method for characterizing cancer organoid response to an immune cell based therapy, the method comprising:
providing a first combination to a first culturing well,
wherein the first combination comprises a mixture of one or more cancer organoid cells with one or more immune cells;
culturing the first combination under conditions which support growth of the one or more cancer organoid cells, wherein culturing the first combination creates cultured cancer organoid cells and cultured immune cells;
capturing, at different time points, a plurality of brightfield images of the first culturing well comprising the first combination;
capturing, at different time points and corresponding to the plurality of brightfield images, a plurality of fluorescent images of the first culturing well comprising the first combination;
providing the plurality of brightfield images and the plurality of fluorescent images to a trained model;
distinguishing the one or more cancer organoid cells from the one or more immune cells within the plurality of fluorescent images by:
using the trained model, generating from the plurality of brightfield images a plurality of masking images of segmented organoids;
applying the plurality of masking images to the plurality of fluorescent images to identify the segmented organoids in the plurality of fluorescent images;
characterizing a response of the one or more cancer organoid cells to the one or more immune cells by comparing at least one of the plurality of fluorescent images to another one of the plurality of fluorescent images;
and
generating an analytical report including the cancer organoid response to the immune cell based therapy, wherein the analytical report includes a characterization of (i) cancer organoid morphology change over time caused by the immune cell based therapy and/or (ii) cell death caused by the immune cell based therapy.
2 . The method of claim 1 , further comprising:
performing a stacking process on the plurality of fluorescent images to generate flattened fluorescent images of organoids within the first culturing well for each different time point; and
wherein providing the plurality of fluorescent images to the trained model comprises providing the flattened fluorescent images to the trained model.
3 . The method of claim 1 , wherein the plurality of fluorescent images are three-dimensional (3D) fluorescence images, the method further comprising:
performing a flattening process on the 3D fluorescence images to generate 2D fluorescence images of the first combination within the first culturing well for each different time point; and
wherein providing the plurality of fluorescent images to the trained model comprises providing the 2D fluorescence images to the trained model.
4 . The method of claim 1 , wherein the plurality of fluorescent images are three-dimensional (3D) fluorescence images, the method further comprising:
performing a slice on the 3D fluorescence images to generate 2D fluorescence images of the first combination within the first culturing well for each different time point; and
wherein providing the plurality of fluorescent images to the trained model comprises providing the 2D fluorescence images to the trained model.
5 . The method of claim 1 , further comprising, prior to providing the first combination of the one or more cancer organoid cells and the one or more immune cells to the first culturing well:
selecting the first combination of the one or more cancer organoid cells and the one or more immune cells based on cancer cell characteristics and the immune cell based therapy.
6 . The method of claim 1 , further comprising:
providing a second combination of the one or more cancer organoid cells and the one or more immune cells to a second culturing well, wherein the second combination represents a different concentration of the one or more cancer organoid cells to the one or more immune cells than the first combination.
7 . The method of claim 6 , wherein generating the analytical report comprises:
providing a comparison of immune cell based therapy efficacies between the first combination and the second combination.
8 . The method of claim 1 , wherein characterizing the cancer organoid morphology change caused by the immune cell based therapy comprises: characterizing cell death, cytokine secretion, and/or a shape change.
9 . The method of claim 1 , wherein characterizing the cancer organoid morphology change comprises: quantifying a number of cancer organoid cell deaths corresponding to each respective periodic capture of the plurality of brightfield images and the corresponding plurality of fluorescent images.
10 . The method of claim 1 , wherein characterizing the cancer organoid morphology change comprises:
determining (i) a number and/or percentage of the one or more cancer organoid cells undergoing a phenotypic change over the different time points, (ii) proliferation of the one or more cancer organoid cells over the different time points, and/or (iii) a change in morphology of the one or more cancer organoid cells over the different time points.
11 . The method of claim 1 , further comprising:
identifying, using the trained model, the one or more cancer organoid cells that are surrounded by the one or more immune cells;
identifying, using the trained model, a spatial relationship between the one or more cancer organoid cells resistant to the immune cell based therapy;
characterizing, using the trained model, migration and/or chemotaxis of the one or more immune cells, and/or
characterizing, using the trained model, co-localization of the one or more immune cells and the one or more cancer organoid cells.
12 . The method of claim 1 , further comprising:
quantifying a number of immune cell deaths in each of the plurality of brightfield images using the corresponding plurality of fluorescent images.
13 . The method of claim 1 , wherein the trained model comprises an organoid trained segmentation model for identifying cancer organoid cells in the plurality of brightfield images.
14 . The method of claim 1 , wherein the trained model comprises an immune cell trained segmentation model for identifying immune cells in the plurality of brightfield images.
15 . The method of claim 1 , wherein identifying and distinguishing the cultured cancer organoid cells and/or the cultured immune cells comprises:
using the trained model, generating from the plurality of brightfield images a second plurality of masking images of the one or more immune cells or a combination of the one or more cancer organoid cells and the one or more immune cells.
16 . The method of claim 1 , wherein the plurality of masking images comprises a binary mask of segmented cancer organoid cells.
17 . The method of claim 15 , wherein the second plurality of masking images comprises a binary mask of the one or more immune cells.
18 . The method of claim 15 , wherein the plurality of masking images comprise a categorical mask differentiating between the one or more cancer organoid cells and one or more types of immune cells, and wherein the categorical mask comprising a plurality of boundary types, each identifying a different one of the one or more cancer organoid cells and the one or more types of immune cells.
19 . The method of claim 1 , wherein the trained model comprises a machine learning algorithm trained using a plurality of training images, wherein at least some of the plurality of training images include annotations identifying one or more of organoid targets, organoids of different morphology, organoids of different locations, and organoids generated by different culturing methods.
20 . The method of claim 1 , wherein the trained model comprises a machine learning algorithm trained using a plurality of training images, wherein at least some of the plurality of training images include annotations identifying one or more of fluorescence dye regions, different cell types, immune cell therapies, and degrees of immune cell therapy response.
21 . The method of claim 15 , wherein the trained model comprises a machine learning algorithm trained using a plurality of training images, wherein the plurality of training images are obtained from a pre-trained segmentation model, and wherein the pre-trained segmentation model removed the one or more cancer organoid cells, leaving the one or more immune cells.
22 . The method of claim 1 , wherein the trained model comprises a machine learning algorithm trained using a plurality of training images, wherein the plurality of training images are obtained from a pre-trained segmentation model, and wherein the pre-trained segmentation model removed the one or more immune cells, leaving the one or more cancer organoid cells.
23 . The method of claim 1 , wherein the trained model comprises a machine learning algorithm trained using a plurality of training images obtained from a pre-trained segmentation model that is segmented out from a series of input images of the one or more immune cells and the one or more cancer organoid cells.
24 . The method of claim 1 , wherein the trained model comprises a segmentation model trained to detect the one or more cancer organoid cells and having a first machine learning algorithm trained using a plurality of training images.
25 . The method of claim 15 , wherein the trained model comprises a segmentation model trained to detect immune cells and having a first machine learning algorithm trained using a plurality of training images and a plurality of corresponding training fluorescence images having the one or more immune cells labeled.
26 . The method of claim 1 , wherein generating the analytical report comprises:
generating a time-lapse imaging of the one or more cancer organoid cells and/or the one or more immune cells at the different time points to identify cell death.
27 . The method of claim 1 , wherein generating the analytical report comprises:
identifying the one or more cancer organoid cells resistant to the immune cell based therapy and/or identifying the one or more cancer organoid cells susceptible to the immune cell based therapy based on changes in the one or more cancer organoid cells over the different time points.
28 . The method of claim 1 , wherein the analytical report further includes a structured file of intensity values, pixel location, size, index, and/or death of the one or more cancer organoid cells within the plurality of brightfield images and the corresponding plurality of fluorescent images.
29 . The method of claim 1 , further comprising: applying an intensity normalization process to each of the plurality of fluorescent images by determining a background intensity of each respective image and normalizing the background intensity by subtracting the background intensity from each respective image.
30 . The method of claim 1 , wherein providing the plurality of brightfield images and the plurality of fluorescent images to the trained model comprises: combining the plurality of brightfield images and the plurality of fluorescent images into one or more multi-channel images to be provided to the trained model.
31 . The method of claim 1 , further comprising: quantifying metabolic activity of the one or more cancer organoid cells and/or the one or more immune cells.
32 . The method of claim 1 , further comprising:
characterizing a presence or amount of one or more biomarkers; and
generating the analytical report to further include the characterized presence or amount of the one or more biomarkers.
33 . The method of claim 1 , further comprising:
quantifying a proportion of specific cell types from the first combination of the one or more cancer organoid cells and the one or more immune cells using fluorescence-activated cell sorting (FACS); and
generating the analytical report to further include the quantified proportion of specific cell types.
34 . The method of claim 1 , wherein the one or more immune cells are peripheral blood mononuclear cells (PBMCs).
35 . The method of claim 1 , wherein the one or more immune cells are lymphocytes, monocytes, and/or dendritic cells.
36 . The method of claim 1 , wherein the one or more immune cells are T cells and/or natural killer (NK) cells.
37 . The method of claim 1 , wherein the one or more immune cells are neutrophils, eosinophils, basophils, and/or macrophages.
38 . The method of claim 1 , wherein the one or more cancer organoid cells are derived from at least one of anal cancer, a basal cell skin cancer, a squamous cancer, a benign cancer, a brain cancer, a glioblastoma, a breast cancer, a bladder cancer, a cervical cancer, a colon cancer, a colorectal cancer, an endometrial cancer, an esophageal cancer, a head and neck cancer, a liver cancer, a hepatobiliary cancer, a kidney cancer, a renal cancer, a gastric cancer, a gastrointestinal cancer, a lung cancer, a non-small cell lung cancer (NSCLC), a mesothelial cancer of the pleural cavity, a mesothelioma, an ovarian cancer, a pancreatic cancer, a prostate cancer, a rectal cancer, a lymphoma, a melanoma, a skin cancer, a meningioma, a sarcoma, and a thymus cancer.