Granted Patent B2

US 12,480,877 · App. 17/557,839 · Granted Nov 25, 2025

Inventors: Jens Krauth (Lübeck, DE); Stefan Gerlach (Groß Groenau, DE); Christian Marzahl (Erlangen, DE); Christopher Krause (Büchen, DE); Jens Hocke (Luebeck, DE); Melanie Hahn (Stockelsdorf, DE); Joern Voigt (Luebeck, DE)

Assignee: EUROIMMUN Medizinische Labordiagnostika AG

G01N21/6458G01N21/6486G01N33/582G06N3/045G06T7/0012G06T7/11G06T7/62G06T2207/10056G06T2207/10064G06T2207/20084G06T2207/30024

Abstract

Claims (85)

1 . A method for detecting at least one potential presence of at least one fluorescence pattern type on an organ section by means of immunofluorescence microscopy and by means of digital image processing, the method comprising:

providing the organ section on a slide,

incubating the organ section with a liquid patient sample which potentially comprises primary antibodies, wherein said primary antibodies bind to antigens of the organ section,

incubating the organ section with secondary antibodies which have been labelled with a fluorescent dye, wherein said secondary antibodies bind to said primary antibodies,

acquiring a fluorescence image of the organ section in a color channel corresponding to the fluorescent dye,

determining, by segmentation of the fluorescence image by means of a first neural network, a sub-area of the fluorescence image that is relevant to formation of the fluorescence pattern type,

determining, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determining, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

outputting the measure of confidence of the actual presence of the fluorescence pattern type and the validity information.

2 . A method according to claim 1 , designed for detection of respective potential presences of respective fluorescence pattern types on an organ section by means of immunofluorescence microscopy and by means of digital image processing, the method further comprising:

determining, by segmentation of the fluorescence image by means of a first neural network, a sub-area of the fluorescence image that is potentially relevant to formation of the fluorescence pattern types,

determining, on the basis of the fluorescence image and on the basis of information indicating the sub-area, by means of a second neural network, respective measures of confidence that indicate respective actual presences of the respective fluorescence pattern types,

determining, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measures of confidence, and

outputting at least a subset of the respective measures of confidence of the respective actual presences of the respective fluorescence pattern types and the validity information.

3 . A method according to claim 1 , further comprising:

determining the measure of confidence on the basis of the fluorescence image and on the basis of the segmented fluorescence image by means of the second neural network.

4 . A method according to claim 1 , further comprising:

in the event of a fluorescence pattern type being determined as actually present, determining a degree of brightness of the sub-area in the fluorescence image that is potentially relevant to formation of the fluorescence pattern type.

5 . A method according to claim 4 , further comprising:

estimating a maximum degree of dilution of the patient sample at which incubation of the organ section with the patient sample still leads to a presence of a or the fluorescence pattern type.

6 . A method according to claim 1 , further comprising:

determining the measure of confidence on the basis of the fluorescence image, and on the basis of information indicating the sub-area, by means of the second neural network.

7 . An apparatus for detecting at least one potential presence of at least one fluorescence pattern type on an organ section by means of immunofluorescence microscopy and by means of digital image processing, the apparatus comprising:

a holding device for a slide containing an organ section which has been incubated with a patient sample potentially comprising primary antibodies and furthermore with secondary antibodies which have each been labelled with a fluorescent dye,

at least one image acquisition unit for acquiring a fluorescence image of the organ section in a color channel corresponding to the fluorescent dye, and

at least one computing unit configured to:

determine, by segmentation of the fluorescence image by means of a first neural network, a sub-area in the fluorescence image that is relevant to formation of the fluorescence pattern type,

determine, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determine, on the basis of the sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

output the measure of confidence of the actual presence of the fluorescent pattern type and the validity information.

8 . A computing unit which, in the course of digital image processing, is configured to:

receive a fluorescence image representing staining of an organ section due to a fluorescent dye,

determine, by segmentation of the fluorescence image by means of a first neural network, a sub-area in the fluorescence image that is relevant to formation of the fluorescence pattern type,

determine, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determine, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

output the measure of confidence of the actual presence of the fluorescence pattern type and the validity information.

9 . A data network device comprising:

at least one data interface for receiving a fluorescence image representing staining of an organ section due to a fluorescent dye, and

at least one computing unit which, in the course of digital image processing, is configured to:

determine, by segmentation of the fluorescence image by means of a first neural network, a sub-area in the fluorescence image that is relevant to formation of the fluorescence pattern type,

determine, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determine, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

output the measure of confidence of the actual presence of the fluorescence pattern type and the validity information.

10 . A method for digital image processing comprising:

receiving a fluorescence image representing staining of an organ section due to a fluorescent dye,

determining, by segmentation of the fluorescence image by means of a first neural network, a sub-area in the fluorescence image that is relevant to formation of the fluorescence pattern type,

determining, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determining, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

outputting the measure of confidence of the actual presence of the fluorescence pattern type and the validity information.

11 . A computer program product comprising commands which, upon execution of the program by a computer, prompt said computer to:

receive a fluorescence image representing staining of an organ section due to a fluorescent dye,

determine, by segmentation of the fluorescence image by means of a first neural network, a sub-area in the fluorescence image that is relevant to formation of the fluorescence pattern type,

determine, on the basis of the fluorescence image by means of a second neural network, a measure of confidence that indicates an actual presence of the fluorescence pattern type,

determine, on the basis of the previously determined sub-area, validity information that indicates a degree of a validity of the measure of confidence, wherein determining validity information comprises:

determining a first size of a total area of the fluorescence image, and a second size of the sub-area;

determining a percentage coverage of the fluorescence image due to the sub-area;

comparing the percentage coverage against a predetermined threshold value;

when the percentage coverage exceeds the predetermined threshold value, determining the validity information as valid; and

when the percentage coverage is below the predetermined threshold value, determining the validity information as invalid; and

output the measure of confidence of the actual presence of the fluorescence pattern type and the validity information.

Assignments (1)

Priority Claims (1)

Continuity (2)

References Cited (11)