US 2021/0279442 A1 · TISSUE POTENCY DETERMINATION THROUGH QUANTITATIVE HISTOMORPHOLOGY ANALYSIS

Abstract

Systems and methods for performing quantitative histopathology analysis for determining tissue potency are disclosed. According to some embodiments, a method training a tissue classifier is provided. According to the method, training the tissue classifier includes generating feature fingerprints of detected nuclei within slide images in a control library and clustering the slide images based on their corresponding feature fingerprints. According to some embodiments, a method for utilizing the trained tissue classifier is provided. According to the method, the trained tissue classifier determines whether tissue in an unknown slide image corresponds to slide images clustered during the training of the tissue classifier.

Claims61

Dependent claims

1-60. (canceled)

method of training a tissue classifier for performing quantitative histopathological assessment, comprising:

converting a slide image into a binary slide image, wherein the slide image is selected from a library of control images, wherein each slide image in the library of control images is associated with a pass classification or a fail classification;
detecting one or more nuclei within the binary slide image;
for each detected nucleus, extracting a feature from the detected nucleus, wherein the feature represents a property of the detected nucleus within the binary slide image and comprises at least one of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, or a circularity of the detected nucleus;
for each detected nucleus, generating, based on the feature, a feature fingerprint associated with the binary slide image, wherein the feature fingerprint is a numerical value calculated from processing the feature;
incorporating the binary slide image into a cluster wherein the cluster comprises a plurality of images, wherein each of the plurality of images is associated with a corresponding feature fingerprint, and wherein the incorporating is based on comparing the feature fingerprint with the corresponding feature fingerprint;
applying a cutoff height to a cluster dendrogram to form a plurality of groups, wherein the cutoff height minimizes a number of groups within the plurality of groups based on multivariate analysis of variance analysis of the cluster;
categorizing a first group within the plurality of groups as a positive control group if the first group comprises first slide images associated with the pass classification;
and
categorizing a second group within the plurality of groups as a negative control group if the second group comprises second slide images associated with the fail classification.

6 more elements

The method according to claim 1, wherein the tissue classifier is capable of determining potency or the quality of the tissue for transplantation into a subject.

The method according to claim 2, wherein the tissue is thymus tissue.

The method according to claim 3, wherein the tissue is allogeneic cultured postnatal thymus tissue-derived product slices for implantation into a human subject.

The method according to claim 4, wherein the human subject is suffering from complete DiGeorge syndrome associated with 22q11.2 deletion;

coloboma, heart defect, choanal atresia, growth or mental retardation, genital hypoplasia and ear anomalies or deafness syndrome (CHARGE), or athymia associated with forkhead box protein N1 (FOXN1) deficiency.

The method according to claim 2, wherein the tissue classifier determines potency or quality of a tissue for transplantation into a subject of an unknown tissue for transplantation by generating feature fingerprints of detected nuclei within slide images in a control library and clustering the slide images based on their corresponding feature fingerprints.

The method according to claim 6, wherein the unknown tissue is thymus tissue.

The method according to claim 7, where the thymus tissue has been subjected to a culturing process in a thymus organ medium for a period of time to partially deplete the thymus tissue of thymocytes.

The method according to claim 8, wherein the period of time is up to 21 days.

The method according to claim 8, wherein the period of time is from about 12 to about 21 days.

The method according to claim 8, wherein the period of time is from about 5 to about 9 days.

The method according to claim 8, wherein the culturing process preserves the functional architecture of the thymic stroma.

The method according to claim 12, wherein the thymic stroma comprises thymic epithelial cells and fibroblasts.

The method according to claim 1, wherein the tissue is selected from the group consisting of vascular tissue, skin tissue, hepatic tissue, pancreatic tissue, neural tissue, urogenital tissue, gastrointestinal tissue, skeletal tissue including bone and cartilage, adipose tissue, connective tissue including tendons and ligaments, amniotic tissue, chorionic tissue, dura, pericardia, muscle tissue, glandular tissue, facial tissue, ophthalmic tissue.

The method according to claim 1, wherein the feature fingerprint is generated from measurements comprising numerical values of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, and a circularity of the detected nucleus.

A method for performing quantitative histopathological assessment of an unclassified slide image of a tissue, comprising:

etecting a hematoxylin channel from the unclassified slide image, wherein the hematoxylin channel is associated with a cellular nucleus within the unclassified slide image of a tissue;
extracting a feature from the detected hematoxylin channel, wherein the feature represents a property of the nucleus within the binary slide image and comprises at least one of an area of the nucleus, a perimeter of the nucleus, an integrated density of the nucleus, or a circularity of the nucleus;
generating, based on the feature, a feature fingerprint associated with the unclassified slide image, wherein the feature fingerprint is a numerical value calculated from processing the feature;
co-clustering the feature fingerprint with a first group of fingerprints categorized as a positive control group or a second group of fingerprints categorized as a negative control group, wherein:
the first group and the second group are formed by applying a cutoff height to a cluster dendrogram, and
the positive control group comprises a first set of slide images associated with a pass classification and the negative control group comprises a second set of slide images associated with a fail classification;
and
determining, based on the co-clustering, if the feature fingerprint is associated with the pass classification or with the fail classification.

5 more elements

The method according to claim 16, wherein the feature fingerprint is capable of determining potency or the quality of the tissue for transplantation into a subject.

The method according to claim 17, wherein the tissue is thymus tissue.

The method according to claim 18, wherein the tissue is allogeneic cultured postnatal thymus tissue-derived product slices for implantation into a human subject.

The method according to claim 19, wherein the human subject is suffering from complete DiGeorge syndrome associated with 22q11.2 deletion;

coloboma, heart defect, choanal atresia, growth or mental retardation, genital hypoplasia and ear anomalies or deafness syndrome (CHARGE), or athymia associated with forkhead box protein N1 (FOXN1) deficiency.

The method according to claim 17, wherein the feature fingerprint determines potency or quality of the tissue for transplantation into a subject of an unknown tissue by generating feature fingerprints of detected nuclei within slide images in a control library and clustering the slide images based on their corresponding feature fingerprints.

The method according to claim 21, wherein the unknown tissue is thymus tissue.

The method according to claim 22, where the thymus tissue has been subjected to a culturing process in a thymus organ medium for a period of time to partially deplete the thymus tissue of thymocytes.

The method according to claim 23, wherein the period of time is up to 21 days.

The method according to claim 23, wherein the period of time is from about 12 to about 21 days.

The method according to claim 23, wherein the period of time is from about 5 to about 9 days.

The method according to claim 23, wherein the culturing process preserves the functional architecture of the thymic stroma.

The method according to claim 27, wherein the thymic stroma comprises thymic epithelial cells and fibroblasts.

The method according to claim 16, wherein the tissue is selected from the group consisting of vascular tissue, skin tissue, hepatic tissue, pancreatic tissue, neural tissue, urogenital tissue, gastrointestinal tissue, skeletal tissue including bone and cartilage, adipose tissue, connective tissue including tendons and ligaments, amniotic tissue, chorionic tissue, dura, pericardia, muscle tissue, glandular tissue, facial tissue, ophthalmic tissue.

The method according to claim 17, wherein the feature fingerprint is generated from measurements comprising numerical values of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, and a circularity of the detected nucleus.

A system for classifying objects within digital images of tissue, comprising:

means for converting a slide image into a binary slide image, wherein the slide image is selected from a library of control images, wherein each slide image in the library of control images is associated with a pass classification or a fail classification;
means for detecting one or more nuclei within the binary slide image;
means for extracting, for each detected nucleus, a feature, wherein the feature represents a property of the detected nucleus within the binary slide image and comprises at least one of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, or a circularity of the detected nucleus;
means for generating, based on the feature, a feature fingerprint associated with the binary slide image, wherein the feature fingerprint is a numerical value calculated from processing the feature;
means for incorporating the binary slide image into a cluster wherein the cluster comprises a plurality of images, wherein each of the plurality of images is associated with a corresponding feature fingerprint, and wherein the incorporating is based on comparing the feature fingerprint with the corresponding feature fingerprint;
means for applying a cutoff height to a cluster dendrogram to form a plurality of groups, wherein the cutoff height minimizes a number of groups within the plurality of groups based on multivariate analysis of variance analysis of the cluster;
means for categorizing a first group within the plurality of groups as a positive control group if the first group comprises first slide images associated with the pass classification;
and
means for categorizing a second group within the plurality of groups as a negative control group if the second group comprises second slide images associated with the fail classification.

6 more elements

The system according to claim 31, wherein the system is capable of determining potency or the quality of a tissue for transplantation into a subject.

The system according to claim 32, wherein the tissue is thymus tissue.

The system according to claim 33, wherein the thymus tissue is allogeneic cultured postnatal thymus tissue-derived product slices for implantation into a human subject.

The system according to claim 34, wherein the human subject is suffering from complete DiGeorge syndrome associated with 22q11.2 deletion;

coloboma, heart defect, choanal atresia, growth or mental retardation, genital hypoplasia and ear anomalies or deafness syndrome (CHARGE), or athymia associated with forkhead box protein N1 (FOXN1) deficiency.

The system according to claim 31, wherein the system classifies potency or quality of a tissue for transplantation into a subject of an unknown tissue by generating feature fingerprints of detected nuclei within slide images in a control library and clustering the slide images based on their corresponding feature fingerprints.

The system according to claim 36, wherein the unknown tissue is thymus tissue.

The system according to claim 37, where the thymus tissue has been subjected to a culturing process in a thymus organ medium for a period of time to partially deplete the thymus tissue of thymocytes.

The system according to claim 38, wherein the period of time is up to 21 days.

The system according to claim 38, wherein the period of time is from about 12 to about 21 days.

The system according to claim 38, wherein the period of time is from about 5 to about 9 days.

The system according to claim 38, wherein the culturing process preserves the functional architecture of the thymic stroma.

The system according to claim 42, wherein the thymic stroma comprises thymic epithelial cells and fibroblasts.

The system according to claim 31, wherein the tissue is selected from the group consisting of vascular tissue, skin tissue, hepatic tissue, pancreatic tissue, neural tissue, urogenital tissue, gastrointestinal tissue, skeletal tissue including bone and cartilage, adipose tissue, connective tissue including tendons and ligaments, amniotic tissue, chorionic tissue, dura, pericardia, muscle tissue, glandular tissue, facial tissue, ophthalmic tissue.

The system according to claim 36, wherein the feature fingerprint is generated from measurements comprising numerical values of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, and a circularity of the detected nucleus.

A classifier, comprising:

means for detecting a hematoxylin channel from the unclassified slide image, wherein the hematoxylin channel is associated with a cellular nucleus within the unclassified slide image of a tissue;
means for extracting a feature from the detected hematoxylin channel, wherein the feature represents a property of the nucleus within the binary slide image and comprises at least one of an area of the nucleus, a perimeter of the nucleus, an integrated density of the nucleus, or a circularity of the nucleus;
means for generating, based on the feature, a feature fingerprint associated with the unclassified slide image, wherein the feature fingerprint is a numerical value calculated from processing the feature;
means for co-clustering the feature fingerprint with a first group of fingerprints categorized as a positive control group and a second group of fingerprints categorized as a negative control group, wherein:
the first group and the second group are formed by applying a cutoff height to a cluster dendrogram, and
the positive control group comprises a first set of slide images associated with a pass classification and the negative control group comprises a second set of slide images associated with a fail classification;
and
means for determining, based on the co-clustering, if the feature fingerprint is associated with the pass classification or with the fail classification.

5 more elements

The classifier according to claim 46, wherein the system is capable of determining potency or the quality of a tissue for transplantation into a subject.

The classifier according to claim 47, wherein the tissue is thymus tissue.

The classifier according to claim 48, wherein the thymus tissue is allogeneic cultured postnatal thymus tissue-derived product slices for implantation into a human subject.

The classifier according to claim 49, wherein the human subject is suffering from complete DiGeorge syndrome associated with 22q11.2 deletion;

coloboma, heart defect, choanal atresia, growth or mental retardation, genital hypoplasia and ear anomalies or deafness syndrome (CHARGE), or athymia associated with forkhead box protein N1 (FOXN1) deficiency.

The classifier according to claim 46, wherein the system classifies potency of an unknown tissue by generating feature fingerprints of detected nuclei within slide images in a control library and clustering the slide images based on their corresponding feature fingerprints.

The classifier according to claim 51, wherein the unknown tissue is thymus tissue.

The classifier according to claim 52, where the thymus tissue has been subjected to a culturing process in a thymus organ medium for a period of time to partially deplete the thymus tissue of thymocytes.

The classifier according to claim 53, wherein the period of time is up to 21 days.

The classifier according to claim 53, wherein the period of time is from about 12 to about 21 days.

The classifier according to claim 53, wherein the period of time is from about 5 to about 9 days.

The classifier according to claim 53, wherein the culturing process preserves the functional architecture of the thymic stroma.

The classifier according to claim 57, wherein the thymic stroma comprises thymic epithelial cells and fibroblasts.

The classifier according to claim 46, wherein the tissue is selected from the group consisting of vascular tissue, skin tissue, hepatic tissue, pancreatic tissue, neural tissue, urogenital tissue, gastrointestinal tissue, skeletal tissue including bone and cartilage, adipose tissue, connective tissue including tendons and ligaments, amniotic tissue, chorionic tissue, dura, pericardia, muscle tissue, glandular tissue, facial tissue, ophthalmic tissue.

The classifier according to claim 46, wherein the feature fingerprint is generated from measurements comprising numerical values of an area of the detected nucleus, a perimeter of the detected nucleus, an integrated density of the detected nucleus, and a circularity of the detected nucleus.

Dependencies are inferred from the claim text.

Tissue potency determination through quantitative histomorphology analysis

Life of this patent

Tissue potency determination through quantitative histomorphology analysis

Life of this patent

Abstract

Claims61