US 2018/0121709 A1 · Descriptive Measurements and Quantification of Staining Artifacts for In Situ Hybridization

Abstract

Immunohistochemistry (IHC) and in situ hybridization (ISH) have the aim of detecting, localizing and quantifying certain analytes for various diagnostic purposes. The quality of the stains which are analyzed may deviate for various reasons. Therefore, the present invention provides a method and system for assessing the stain quality and for establishing objective criteria for assessing the stain quality for application in the fields of in-situ hybridization and immunohistochemistry. In one possible embodiment, the invention comprises the steps of unmixing multi-spectral image data of a tissue specimen to obtain analyte intensity images, each analyte intensity image comprising signals from a single stain, computing metrics based on the analyte intensity images, wherein the metrics are uniformity, distribution and/or dispersion of pixel intensity values in the analyte intensity images and assessing a stain quality of a slide by comparing the computed metrics to pre-determined cutoff values regarding uniformity, distribution and/or dispersion of pixel intensity, wherein the stain quality of the slide is assessed as acceptable if the computed metrics meet or exceed the pre-determined cutoff values, and wherein the stain quality of the slide is assessed as unacceptable if the computed metrics do not meet the pre-determined cutoff values. In order to establish objective criteria for assessing stain quality, in one possible embodiment, the method and system includes the step of deriving cut-off values for uniformity, distribution and/or dispersion of pixel intensity by combining the computed metrics based on the analyte intensity images with pre-established data quantifying the stain quality.

Claims20

Dependent claims

A computer device for objective stain assessment comprising one or more processors and at least one memory, the at least one memory storing non-transitory computer-readable instructions for execution by the one or more processors to cause the one or more processors to:

unmix multi-spectral image data of a tissue specimen stained in an immunohistochemical assay or an in situ hybridization assay for the presence of a particular biomarker to obtain analyte intensity images, each analyte intensity image comprising signals from a single stain;
compute metrics based on the analyte intensity images, wherein the metrics are entropy values and mean-variance ratio values;
and
assess a stain quality of a slide by comparing the computed metrics to pre-determined cutoff values, wherein the stain quality of the slide is assessed as acceptable if the computed metrics meet or exceed the pre-determined cutoff values, and wherein the stain quality of the slide is assessed as unacceptable if the computed metrics do not meet the pre-determined cutoff values.

1 more element

A computer-implemented method of stain assessment comprising:

unmixing multi-spectral image data to obtain analyte intensity images, the analyte intensity images each comprising a single image channel corresponding to signals from a particular stain;
computing entropy values for each of the analyte intensity images;
computing mean-variance ratios for each of the analyte intensity images;
and
assessing a stain quality of a slide by comparing the computed entropy and mean-variance ratio values to pre-determined entropy cutoff values and mean-variance ratio cutoff values, wherein a stain quality is assessed as acceptable if the computed entropy and mean-variance ratio values meet or exceed the pre-determined cutoff values, and wherein the stain quality is assessed as unacceptable if the computed entropy and mean-variance ratio values do not meet the pre-determined cutoff values.

2 more elements

A computer device for establishing objective criteria for stain assessment comprising one or more processors and at least one memory, the at least one memory storing non-transitory computer-readable instructions for execution by the one or more processors to cause the one or more processors to:

unmix multi-spectral image data of a tissue specimen stained in an immunohistochemical assay or an in situ hybridization assay for the presence of a particular biomarker to obtain analyte intensity images, each analyte intensity image comprising signals from a single stain, and wherein the multi-spectral image data is from a plurality of reference slides that have been identified as acceptable or unacceptable;
compute entropy values and mean-variance ratio values based on each of the obtained analyte intensity images;
and
derive cutoff values for staining quality assessment based on the computed entropy and mean-variance ratio values, wherein the cutoff values for staining quality assessment correspond to the stains used in the immunohistochemical assay or the in situ hybridization assay.

1 more element

Dependencies are inferred from the claim text.

Descriptive Measurements and Quantification of Staining Artifacts for In Situ Hybridization

Life of this patent

Descriptive Measurements and Quantification of Staining Artifacts for In Situ Hybridization

Life of this patent

Abstract

Claims20