Method, system and computer program for computer-assisted evaluation of biomarkers in a tissue sample

A computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to carry out the following steps:

• automatically determining, for a plurality of locations or regions within an image of a tissue sample in which cell nuclei of the tissue as well as biomarkers are represented, a blood vessel probability that represents a probability that the image at the corresponding location or region represents a blood vessel, said automatic determining based on a combination of at least two blood vessel probability indicators,
• wherein at least one of the at least two blood vessel probability indicators is based on an automatic assessment of at least one property of a nucleus or nuclei associated with said location or region within the image, said nucleus property selected from a group consisting of the shape, the texture, the size, the intensity of the nucleus or nuclei associated with the respective location or region as represented in said image, and the absence of nuclei at said location or region, and
• wherein said automatic determining is based on an algorithm that calculates said blood vessel probability based on said at least two blood vessel probability indicators,
• automatically determining an assumed blood vessel location within said image based on a distribution of blood vessel probability with respect to said locations or regions within the image,
• automatically selecting biomarkers represented in said image at locations outside the assumed blood vessel location but within a threshold distance from said assumed blood vessel location, and
• automatically evaluating the selected biomarkers.

A system for computer-assisted evaluation of biomarkers in a tissue sample, comprising an imaging apparatus for generating an image of the tissue sample in which cell nuclei of the tissue as well as biomarkers are represented, and a processor, said processor being configured for

• automatically determining for a plurality of locations or regions within the image a blood vessel probability that represents a probability that the image at the corresponding location or region represents a blood vessel, said automatic determining based on a combination of at least two blood vessel probability indicators,
• wherein at least one of the at least two blood vessel probability indicators is based on an automatic assessment of at least one property of a nucleus or nuclei associated with said location or region within the image, said nucleus property selected from a group consisting of the shape, the texture, the size, the intensity of the nucleus or nuclei associated with the respective location or region as represented in said image, and the absence of nuclei at said location or region, and
• wherein said automatic determining is based on an algorithm that calculates said blood vessel probability based on said at least two blood vessel probability indicators,
• automatically determining an assumed blood vessel location within said image based on a distribution of blood vessel probability with respect to said locations or regions within the image,
• automatically selecting biomarkers represented in said image at locations outside the assumed vessel location but within a threshold distance from said assumed blood vessel location, and
• automatically evaluating the selected biomarkers.