EP0896661A1 - Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques - Google Patents

Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques

Info

Publication number
EP0896661A1
EP0896661A1 EP97920751A EP97920751A EP0896661A1 EP 0896661 A1 EP0896661 A1 EP 0896661A1 EP 97920751 A EP97920751 A EP 97920751A EP 97920751 A EP97920751 A EP 97920751A EP 0896661 A1 EP0896661 A1 EP 0896661A1
Authority
EP
European Patent Office
Prior art keywords
sample
cell
image
segments
evaluation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP97920751A
Other languages
German (de)
English (en)
Inventor
Bela Molnar
Rainer Schäfer
Winfried Albert
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
3DHistech KFT
Original Assignee
Roche Diagnostics GmbH
Boehringer Mannheim GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Roche Diagnostics GmbH, Boehringer Mannheim GmbH filed Critical Roche Diagnostics GmbH
Publication of EP0896661A1 publication Critical patent/EP0896661A1/fr
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Optical investigation techniques, e.g. flow cytometry
    • G01N15/1429—Signal processing
    • G01N15/1433—Signal processing using image recognition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Optical investigation techniques, e.g. flow cytometry
    • G01N2015/1488—Methods for deciding

Definitions

  • the present invention relates to a method for examining tissue samples or body fluids on a slide with an automatic microscope as well as a video system and an evaluation computer with neural networks.
  • the details of the method result from claims 1 and 8.
  • the above object is achieved by a method according to claim 1 and claim 8.
  • the method according to the invention proposes a procedure in which neither the intervention of an operator nor the use of derived variables, such as, for example, morphometric or densitometric variables, are necessary.
  • derived variables such as, for example, morphometric or densitometric variables
  • the present invention falls in the field of medical image diagnosis, which includes tomography, ultrasound and microscopy. Of these methods, microscopic diagnosis is the most important means to rule out or confirm malignant diseases. For gynecological, urological or endoscopic examinations and operations, samples for oncological diagnostics are routinely sent to the pathological, cytological or hematological laboratories.
  • the sample material consists of cell swabs, cell suspensions or cell and structure-forming parts which are obtained from biopsies or tissue sections
  • the doctor In today's usual examination, the doctor first decides at the cellular level whether all cell parts are present, are in a normal relationship to each other and whether the cell nucleus has a normal chromatin distribution. This level also analyzes whether unusual cell types are present. At the structural level, the doctor examines whether the cell groups found correspond to the organ from which they were taken and whether there are cell types that are atypical for the tissue examined. It also analyzes whether the boundaries between different types of tissue are normal or disturbed. The most common cause of such changes is the malignant conversion of cells. Because the sought structural changes in the samples are usually not visible without staining, special stains are used. The preparations are stained by hand or with automatic stainers, with special immunological markers such as e.g. B.
  • tumor marker antibodies can be used in so-called in-situ hybridization, special DNA probes that bind to the oncogene part of the cellular DNA are used to find tumor genes.
  • a special embodiment of the detection of original tumor cells is the PCR method, in which a sufficient amount of a specific DNA or RNA is formed by an intermediate amplification step for different staining methods. The tissue components or cells marked in this way are then accessible to the microscopic evaluation described above
  • the present invention addresses these problems and proposes an automated method for examining tissue samples or body fluid samples.
  • Samples that can be examined with the method according to the invention are, for example, tissue sections, smears, cell smears and body fluids, such as blood, urine, punctate fluid and the like.
  • tissue sections tissue sections
  • smears cell smears and body fluids
  • body fluids such as blood, urine, punctate fluid and the like.
  • the sample is placed on a slide, so that a thin layer of the sample is obtained, which can be examined.
  • neural networks are used to examine cells, cell complexes, histological structures, etc. These units are collectively referred to below as cell objects. These are also to be understood to mean those objects which lie on the edge of the recorded image and are only partially detected.
  • the sample is evaluated with an automatic microscope.
  • Microscopes known in the prior art for example WO 91/20048 have devices for adjusting the focus and for adjusting desired image sections.
  • the most common example of such a device for the lateral displacement of the sample consists in a cross table, through which the slide is moved within a plane.
  • a microscope for carrying out the method according to the invention can have devices for automatically changing objectives and filters.
  • the sensor of a video system is arranged in the image plane of the microscope.
  • Current video cameras have a CCD array of 1/2 inches with 440,000 pixels.
  • the frame grabbers used for image acquisition deliver arrays of 1024 x 1024 or 512 x 512 pixels. Both CCD arrays are known which provide gray values as the output value and those which provide color values (red-green-blue values).
  • a system according to the invention for the automatic examination of samples furthermore has an evaluation computer which is connected to the video system and also to the microscope.
  • Evaluation computers for such an application have a so-called "frame grabber", which is used to read the signals provided by the video system into the evaluation computer.
  • the evaluation computer has devices for controlling the microscope in order to control focusing and lateral displacement of the sample relative to the objective of the microscope.
  • the evaluation computer must have a high computing power to solve the present problem.
  • several neural networks are implemented in the evaluation computer.
  • Neural networks are used as mathematical models of biological brain functions. By selecting suitable network topologies and processing functions, an attempt is made to simulate complex thinking and decision-making processes with the aid of computers.
  • the main findings on artificial neural networks are presented in the book “Parallel distributed processing: Explorations in the microstructure of cognition” by Rummelhart and McCulloch.
  • the principle of operation of a neural network is shown in FIG. 1.
  • input neurons On the input side of the neural network there are so-called input neurons which are connected to "hidden neurons".
  • Each neuron has one or more weighted inputs, either in the form of external signals or as output from other neurons. Both positive and negative weightings are possible.
  • the sum of the weighted inputs is transferred to the output value (s) via a transfer function, which in turn control other neurons or serve as output values.
  • the hidden neurons shown in FIG. 1 are connected to the output neurons. Of course, the area of the hidden neurons can also be constructed much more complex and consist of several interconnected levels.
  • the entirety of all neurons of a certain functionality is called a layer, such as B. Input layer, of the organizational structures of neural networks known in the prior art, only the types Perceptron, Hopfield, Kohen and Grossberg model should be mentioned at this point.
  • the most important parameters of a network are the basic neuron voltages and the Strength of the connections between the neurons. To set the parameters, a representative training set is evaluated several times by the network.
  • the weightings and the basic voltages are changed and reset. This iteration is repeated until the average error rate falls below a predetermined minimum or a previously defined task-related termination criterion is reached.
  • the test set is also evaluated iteratively for control purposes.
  • the present invention is particularly characterized in that several Neural networks are used with different tasks.
  • the digitized image is divided into contiguous segments, preferably of the same size, and the segments obtained are evaluated separately from one another by the networks.
  • the examined sample is first classified according to its type. This can either be done by an operator or the digitized image of the sample is examined by a neural network (NN ⁇ ).
  • an image of the sample is first recorded and the digitized signals of the image are assigned to individual pixels.
  • the pixels obtained are assigned to contiguous segments, the evaluation of which is carried out separately.
  • the segments are examined by neural networks NN P j, in that the digitized image signals of the segment are fed to the input neurons.
  • a first type of neural network NNpi the neural network examines whether there is a cell, a cell complex, a tissue part or artifact structures in the respective segment.
  • a further evaluation can be carried out by a neural network NNp 2 , which examines whether the tissue of the sample is epithelium, endothelium, connective tissue or muscles or whether there is a body fluid.
  • the evaluation by a third neural network NNp 3 is possible, which examines whether a cell or a cell complex is present in the segment.
  • an examination can also be carried out by an fourth neural network NN p4 , which detects when a cell object protrudes beyond the edge of the respective segment If so, the entire cell object is evaluated.
  • the cell object protrudes into other segments for which an image has already been taken, the cell object can be examined by forming new segments so that the cell object lies completely in one of these segments However, if the cell object protrudes from the image recorded with the video system, it is necessary to record another image that completely captures the cell object
  • the segments are further evaluated with one or more neural networks NN ⁇ and classified as pathological if there are cell types that do not belong to the type of the sample or if there are structural cell or tissue changes to detect whether foreign cell types or structural ones If changes are present, it is important to include the type of sample in the evaluation.
  • a further examination is preferably carried out by taking a picture of the relevant one or more Segments in which peculiarities were found are recorded and examined with a higher magnification.
  • the new recording is again divided into segments as already described and neural networks NN P j are used, as has already been done at low resolution.
  • the neural networks NN ⁇ j are also used for the recording with higher magnification, and again it is examined whether there are cell types that do not belong to the type of the sample or whether there are structural cell / tissue changes. If this is the case, the diagnosis made on the image taken with a smaller magnification is verified. It can also be provided that the image is output directly at a higher magnification on a monitor so that an experienced surgeon can carry out a visual evaluation.
  • a first neural network is used to evaluate the digitized signals of the video system to determine whether there is a cell object in the segment that was assigned to the input layer of the neural network.
  • This first neural network (NN A ) preferably has a structure in which a 16 x 16 pixel image in the input field is assigned to each output neuron. For a CCD camera with 512 x 512 pixels, this results in a 32 x 32 output field.
  • a digitized pixel is assigned to an input neuron.
  • a neural network according to the invention preferably has a single output neuron for each segment.
  • NN can also be used with several output neurons per segment if the desired information can be easily read from these output neurons.
  • the task of this first network is to determine in which of the segments a cell object is located. For this analysis, it is necessary to train the neural network used accordingly. It has proven particularly advantageous to train the neural network with samples of the same type that are later to be analyzed by the network. The training of neural networks to achieve the desired behavior is well known in the prior art and is therefore not explained in more detail here.
  • the information is obtained in which segments cell objects are present. This is indicated by the activity of the output neuron (s) which are assigned to each segment.
  • the segments in which a cell object has been detected are evaluated by a second neural network (NN B ) by supplying the digitized signals of the respective segment to the input neurons of the NN B and preferably assigning an output neuron to each of the segments whose activity indicates whether a single cell or a cell complex is present in the respective segment.
  • Suitable neural networks can also be obtained for this application by training on samples of known composition. Analogous to the neural network A, it is also advantageous here to carry out the training on samples of the same type.
  • a procedure in which a segment is only evaluated with the neural network B if a cell object was previously detected with the neural network A in this segment has the advantage that the number of segments that have to be evaluated is greatly reduced can, which leads to a saving in computing time.
  • a third neural network (NNc) is used to evaluate the segments that lie at the image borders and in which a cell object was found.
  • the neural network C provides the information as to whether a cell object is located on the image boundary in a specific segment. If this is the case, a further image of the sample is taken, in which the cell objects are completely contained.
  • the automatic microscope is controlled by the evaluation computer in such a way that the sample and objective are laterally displaced relative to one another. This can be achieved, for example, by means of a motor-driven cross table.
  • the segments in which cell objects were detected are taken up again with a higher magnification.
  • the images obtained in this way can, for example, be visually evaluated by an operator.
  • the evaluation of these recordings is advantageously carried out by a further neural network (NN D ). This network carries out an analysis as to whether the sample is to be assessed as pathological.
  • a special feature of the method according to the invention is that the microscope does not focus on individual cells or structures in any of the steps. It is only necessary to take pictures of suitable layers of the sample.
  • the methods known in the prior art for automated image analysis for cellular diagnosis carry out a localization of cells or cell comglomerates and evaluate these with regard to morphometric and densitometric parameters. It was therefore surprising that with the method according to the invention the detection of pathological structural changes is possible with great certainty even without such a focus on individual objects.
  • a classification of the type of sample has already been described as advantageous, since suitable neural networks can be selected on the basis of this information.
  • the evaluation computer therefore advantageously has a large number of different neural networks which are adapted to different types of samples. If the type of sample is known, suitable neural networks can then automatically be used for the evaluation steps of the method according to the invention.
  • the classification of the sample itself can either be carried out by the user, who makes the classification based on the origin, or a further neural network can be used which specializes in differentiating different types of samples. For routine use, it is preferred to mark the slide on which the sample is located with respect to the type of sample (for example with a barcode).
  • the information can be automatically read in before the start of the examination procedure (for example with a bar code reader) and this information can be transmitted to the evaluation computer so that suitable neural networks are selected.
  • the magnifications with which the microscope works are for the steps with the neural networks NN A , NN B and NNc in the range from 30 to 200, preferably around 100.
  • the image with higher magnification for the analysis of segments in which cell objects are detected were carried out with a magnification of 200 to 600, preferably about 400
  • neural networks NN A or NNp in which a 16 x 16 input field is assigned to each output neuron.
  • the methods described above can also be modified such that they can be used to carry out a three-dimensional, microscopic examination of tissue samples.
  • sections are made from a tissue sample, each of which is examined using the described methods.
  • the cuts are made in a defined manner, so that after evaluation has been carried out, it is possible to assign the evaluation results to the original locations in the sample.
  • Tissue cuts are preferably used, the cutting planes of which are arranged essentially parallel to one another.
  • Such an examination of tissue samples was not possible in the prior art with the manual or partially automated methods previously available.
  • the sections are marked with two or more different preparation methods. For example, a first layer with hematoxilin, a subsequent layer with an immunohistochemical label and a next layer with an in situ hybridization are marked. The next cut can then be marked again with hematoxilin and so on.
  • the data of several sections and preparations are logically combined, an evaluation of these three-dimensional preparations is possible, according to a 4-D slide map.
  • the same effect is also obtained if the sections of a biopsy sample are prepared alternately using several labeling methods, e.g. Antibodies with different epitope specificity and fluorescent labeling can be used.
  • Antibodies with different epitope specificity and fluorescent labeling can be used.
  • individual RGB image components By selecting individual RGB image components, several three-dimensional, microscopic images of the sample can be obtained. Each fluorescence wavelength corresponds to a different sample characterization.
  • the logical combination of the data creates a 4-D slide map in which the doctor can browse both by sample topology and by cell characteristics (type of marking). Different views of a sample can be selected by showing or hiding individual image data sets. This supports the user of the system considerably in the verification or falsification of diagnostic hypotheses.
  • slide maps Another possible application of slide maps is to generate various views from them using suitable software modules. It is therefore possible for the viewer to view the sample from different spatial directions. This is particularly important because different cell types, particularly in the case of pathological samples, occur more locally accumulated than homogeneously distributed and are therefore easier to find with the described procedure.
  • a blood smear or a sternum puncture smear is prepared according to prescribed methods and evaluated according to the method according to the invention.
  • the selected segments have to contain an entire cell, but an evaluation is also possible if the segments contain only parts of the cell.
  • the color information of the pixels is preferably evaluated according to the RGB (red-green-blue) or HSI (Hue Saturation Intensity) system.
  • a second type of sample is a biopsy.
  • a sample is cut out with a biopsy needle or cut out with scissors. The evaluation of such a sample generally takes about 1 week.
  • the third type of sample is taken in one operation but is not evaluated immediately.
  • the fourth type are samples with difficult differential diagnosis that are sent to diagnostic centers.
  • DNA stains and immunohistochemical stains can also be used. Since the samples are very different, it is particularly important in this application to select a suitable neural network for the particular type of sample.
  • the network that makes the final evaluation differentiates between structures that are normally present in this type of tissue and structures that are based on an abnormal change or are based on structures that are normally not present in this type of sample. Since the case configurations that occur here are very different, an evaluation process is preferred in which the neural network locates structures that have been recognized as unusual and shows these to the operator on a display, so that a separate evaluation takes place.
  • differential diagnostic neural networks can also be used for this, which are adapted to the possible case designs.
  • the immunohistochemical reactions serve to facilitate the differential diagnosis and to contribute to the prognosis.
  • a system based on neural networks divides the image of the preparation stained with special antibodies into image segments and checks these segments for positive reactions.
  • the training of the neural network is preferably carried out at the manufacturer, with special training having to be carried out for each test.
  • several, differently structured neural networks e.g. B. for different test procedures. If the analysis system receives the current test request via bar code or manual programming, the associated network type and the valid network parameters are selected automatically. The evaluation can run very quickly since the network only has to make a yes / no decision for the image segment.
  • neural networks an immunohistochemical evaluation can be carried out after the sample has been stained with special antibodies.
  • the image of the sample is divided into different image segments and these are checked for a positive reaction with the antibodies.
  • several differently structured neural networks can also be used for different test procedures.
  • In-situ hybridization is used to detect special DNA segments.
  • the cellular DNA is first unfolded, then DNA probes provided with a labeling system are added. After completion of the reaction, it is determined whether and how many probes have taken up a cell.
  • Example 1 Evaluation with a neural network A
  • Figure 2 shows the image of an object to be analyzed, which is located on a slide. The image is read into an electronic memory for further processing.
  • Figure 3 shows the optical intensity distribution for the same object on the Z axis for the object located in the XY plane. Instead of the monochromatic intensity measurement shown here, the intensities of the colors red, yellow and blue are advantageously measured and used as the basis for a subsequent evaluation.
  • the evaluation of the intensity distribution in Figure 3 is shown schematically in Figure 4.
  • the grid drawn in above the intensity distribution represents the division of the image into segments.
  • a neural network evaluates the individual segments and checks whether a tissue sample is present in the respective segment. In further examinations, only these image segments are taken into account. For the example of a parenchymal tissue, the intensity information is further examined for abnormalities with a neural network. Image segments in which such abnormalities were found are classified with the aid of a further neural network, which can differentiate between vascular objects, neuron complexes and immediately pathological changes.
  • a neural network analyzes whether the object to be analyzed contains connective tissue muscles, epithelium or endothelium.
  • Another Neuronlaes network analyzes the boundaries of the epithelium and connective tissue in order to discover invasive processes.
  • Yet another neural network can be used to investigate whether the connective tissue contains epi / endothelial cell islands.
  • Another neural network can search for vascular or neuronal complexes in the segments.
  • Example 2 Differentiation of tissue types with the help of neural networks
  • a sample from a gastric biopsy is examined, the sample size is 3 mm x 8 mm, and hematoxilin eosin is used for staining.
  • a Zeiss Axioplan 2 MOT microscope is used for the automatic image acquisition, for the manual recording of the image an Olympus BH2 microscope is used.
  • a Grundig Vi "CCD camera is used, which can also supply black and white images.
  • a camera from Stemmer technik can be used.
  • the data volume is approximately 1 Mbyte for an image section of 40 ⁇ m x 40 ⁇ m
  • the memory requirement can be reduced to approximately 10-20% without significant loss of the image information, and the individual images are stored in such a way that a cross-segment, electronic search for image content is possible.
  • a neural network A checks all partial images for existing biological material, all “empty” segments are discarded / deleted. The result of this analysis is a segmented “map” of the slide in which the position of the tissue sample is drawn. This slide map is stored in the computer and serves as a reference for further examinations and classifications.
  • a neural network B is used to check whether the segments contain an ephitelium. If the evaluation is positive, the decision tree branches into the neural networks B1 and B2. Network B1 checks whether other types of tissue are also present and whether there is an edge region of the epithelium. Network B2 differentiates between the upper and lower epithelial area.
  • the neural network C checks whether the segment contains connective tissue or muscles.
  • the further investigation of the image segments is carried out by the neural networks D and E.
  • Network D searches for epithelium with pathological cells or cell groups.
  • Network E checks the connective tissue segments for possible inclusions of epithelial islands.
  • the results of the individual analyzes are entered step by step in the slide map and are then available to the doctor as a basis for decision-making.
  • the different types of tissue are represented by different colors, for example green for epithelium, yellow for connective tissue, orange for muscles, pathological areas are marked in red.
  • An advantage of this display is that, in addition to differentiating the different types of tissue, their relative size and position with respect to one another is also immediate is recognizable. With the help of an appropriate evaluation tool, it is also possible to call up an image segment directly by clicking on the slide map.
  • Example 3 Detection of proliferating cells using neural networks
  • a hematoxilin-stained preparation from a gastric biopsy is examined with the PCNA antibody system (the antibodies are directed against the proliferation cellular nuclear antigen and labeled with the avidin-biotin peroxidase system).
  • Nuclei with a positive reaction turn light brown, non-reactive nuclei take on a light blue color.
  • the image segments are recorded automatically as described in Example 2.
  • the Grundig camera 8301 is used, which provides both a black / white image and the three RGB color components.
  • a neural network A checks which segments come from the epithelium. These are then examined with the neural network B for centers with positive staining.
  • results of the analyzes are entered in the slide map, e.g. the positive epithelial areas are shown in brown, the negative areas in green and the areas without epithelium in yellow. Metastases can be recognized quickly and safely in the slide map.
  • Example 4 spatial distribution of tissue types in a preparation

Landscapes

  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Molecular Biology (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Biomedical Technology (AREA)
  • Dispersion Chemistry (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un procédé d'analyse automatisée assistée par ordinateur d'échantillons tissulaires ou d'échantillons de liquides organiques à l'aide de réseaux neuronaux. Selon un premier procédé d'analyse, l'échantillon est d'abord classé en fonction de son type, puis une image numérisée est répartie en segments qui sont explorés par un ou plusieurs réseaux neuronaux. L'échantillon est qualifié de pahtologique s'il réveèle la présence de types de cellules n'appartenant pas au type de l'échantillon ou en cas de modifications cellulaires ou tissulaires. Selon un second procédé d'analyse, il est également prévu une segmentation de l'image numérisée et les segments sont ensuite observés, cet examen permettant de voir s'ils contiennent un objet cellulaire. On observe ensuite si l'objet cellulaire est une cellule individuelle ou un complexe cellulaire. Dans une troisième étape d'évaluation, on détermine si l'objet cellulaire identifié se trouve sur une des limites de l'image. Dans l'affirmative, on prend à nouveau une image dans laquelle les objets cellulaires trouvés sont compris en totalité. Finalement, les segments dans lesquels des objets cellulaires ont été détectés sont enregistrés avec un agrandissement plus important, à des fins d'évaluation.
EP97920751A 1996-04-27 1997-04-25 Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques Withdrawn EP0896661A1 (fr)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
DE19616997 1996-04-27
DE19616997A DE19616997A1 (de) 1996-04-27 1996-04-27 Verfahren zur automatisierten mikroskopunterstützten Untersuchung von Gewebeproben oder Körperflüssigkeitsproben
PCT/EP1997/002130 WO1997041416A1 (fr) 1996-04-27 1997-04-25 Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques

Publications (1)

Publication Number Publication Date
EP0896661A1 true EP0896661A1 (fr) 1999-02-17

Family

ID=7792708

Family Applications (1)

Application Number Title Priority Date Filing Date
EP97920751A Withdrawn EP0896661A1 (fr) 1996-04-27 1997-04-25 Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques

Country Status (5)

Country Link
US (1) US6246785B1 (fr)
EP (1) EP0896661A1 (fr)
JP (2) JP2000513465A (fr)
DE (1) DE19616997A1 (fr)
WO (1) WO1997041416A1 (fr)

Families Citing this family (57)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6404906B2 (en) * 1997-03-03 2002-06-11 Bacus Research Laboratories,Inc. Method and apparatus for acquiring and reconstructing magnified specimen images from a computer-controlled microscope
US6272235B1 (en) * 1997-03-03 2001-08-07 Bacus Research Laboratories, Inc. Method and apparatus for creating a virtual microscope slide
DE19801400C2 (de) * 1998-01-16 2001-10-18 Petra Perner Verfahren zur automatischen Erkennung, Eigenschaftsbeschreibung und Interpretation von Hep-2-Zellmustern
US6959981B2 (en) 1998-06-09 2005-11-01 Silverbrook Research Pty Ltd Inkjet printhead nozzle having wall actuator
US6611833B1 (en) * 1999-06-23 2003-08-26 Tissueinformatics, Inc. Methods for profiling and classifying tissue using a database that includes indices representative of a tissue population
US6993170B2 (en) * 1999-06-23 2006-01-31 Icoria, Inc. Method for quantitative analysis of blood vessel structure
US6581011B1 (en) * 1999-06-23 2003-06-17 Tissueinformatics, Inc. Online database that includes indices representative of a tissue population
US6979425B1 (en) * 1999-10-04 2005-12-27 Robodesign International, Inc. High capacity microarray dispensing
US6558623B1 (en) * 2000-07-06 2003-05-06 Robodesign International, Inc. Microarray dispensing with real-time verification and inspection
CN1214340C (zh) * 2000-04-24 2005-08-10 国际遥距成象系统公司 多个神经网络的成像设备和方法
US7236623B2 (en) * 2000-04-24 2007-06-26 International Remote Imaging Systems, Inc. Analyte recognition for urinalysis diagnostic system
US7025933B2 (en) * 2000-07-06 2006-04-11 Robodesign International, Inc. Microarray dispensing with real-time verification and inspection
US6985616B2 (en) 2001-10-18 2006-01-10 Robodesign International, Inc. Automated verification and inspection device for sequentially inspecting microscopic crystals
WO2002076282A2 (fr) * 2001-01-05 2002-10-03 Tissueinformatics, Inc. Procede d'analyse de la structure de vaisseaux sanguins
WO2002097714A1 (fr) * 2001-04-09 2002-12-05 Lifespan Biosciences, Inc. Procedes informatiques de reconnaissance de motifs d'images dans de la matiere organique
US20020186875A1 (en) * 2001-04-09 2002-12-12 Burmer Glenna C. Computer methods for image pattern recognition in organic material
DE10124340A1 (de) 2001-05-18 2002-12-05 Fraunhofer Ges Forschung Verfahren zur Analyse einer biologischen Probe
CA2468861C (fr) * 2001-12-05 2012-06-12 The Regents Of The University Of California Systemes de microscopie robotique
JP3516144B1 (ja) * 2002-06-18 2004-04-05 オムロン株式会社 光学情報コードの読取方法および光学情報コード読取装置
US20050033713A1 (en) 2003-01-07 2005-02-10 Microsoft Corporation Automatic text generation
KR101131297B1 (ko) 2003-11-10 2012-03-30 알토 바이오사이언스 코포레이션 가용성 티씨알 분자 및 이용 방법
EP1574988B1 (fr) * 2004-03-08 2014-06-18 Siemens Product Lifecycle Management Software Inc. Détermination et utilisation de données de caractéristiques géométriques
DE102004018174B4 (de) * 2004-04-08 2006-12-21 Perner, Petra, Dr.-Ing. Verfahren zur Akquisition von Formen aus Bildern mit Fällen und zum fallbasierten Erkennen von Objekten in digitalen Bildern, Computer-Programm-Produkt und digitales Speichermedium zur Ausführung dieses Verfahrens
DE102004018171B4 (de) * 2004-04-08 2006-12-21 Perner, Petra, Dr.-Ing. Verfahren zur Akquisition von Formen aus Bildern luftgetragener Keime und/oder Pollen und zum fallbasierten Erkennen dieser Objekte in digitalen Bildern, Computer-Programm-Produkt und digitales Speichermedium zur Ausführung dieses Verfahrens
DE502004005734D1 (de) 2004-04-08 2008-01-24 Petra Perner Verfahren zur akquisition von formen aus hep-2-zellschnitten und zum fallbasierten erkennen von hep-2-zellen
RU2303812C2 (ru) * 2004-12-29 2007-07-27 Общество с ограниченной ответственностью "НПФ РЕНАМ" Способ распознавания и подсчета клеток в биологических средах человека и животных и устройство для его осуществления
US20060189893A1 (en) * 2005-01-06 2006-08-24 Diamics, Inc. Systems and methods for detecting abnormal cells
US20060161076A1 (en) * 2005-01-06 2006-07-20 Diamics, Inc. Systems and methods for collection of cell clusters
DE102007003448A1 (de) 2007-01-19 2008-07-24 Perner, Petra, Dr.-Ing. Verfahren und Datenverarbeitungssystem zur automatischen Erkennung, Verarbeitung, Interpretation und Schlussfolgerung von als digitale Datenmengen vorliegenden Objekten
DE102007009485A1 (de) * 2007-02-22 2008-08-28 Perner, Petra, Dr.-Ing. Verfahren und Datenverarbeitungssystem zur Modellbildung für die Bildsegmentierung
US8065166B2 (en) 2007-10-30 2011-11-22 Onemednet Corporation Methods, systems, and devices for managing medical images and records
US9171344B2 (en) 2007-10-30 2015-10-27 Onemednet Corporation Methods, systems, and devices for managing medical images and records
US8059136B2 (en) * 2007-12-11 2011-11-15 Honeywell International Inc. Hierarchichal rapid serial visual presentation for robust target identification
WO2010126797A1 (fr) 2009-04-29 2010-11-04 Onemednet Corporation Procédés, systèmes et dispositifs de gestion d'images et de dossiers médicaux
JP5387147B2 (ja) * 2009-06-03 2014-01-15 日本電気株式会社 病理画像診断システム、病理画像処理方法、病理画像診断プログラム
WO2010151761A2 (fr) * 2009-06-26 2010-12-29 Cim Software Corporation Procédé pour identifier et suivre des échantillons de tissu et des préparations histologiques
DE202009011850U1 (de) 2009-08-28 2011-01-05 Perner, Petra, Dr.-Ing. Einrichtung zur automatischen Erfassung der dynamischen Prozesse von Zellen von Zellproben
DE102010035908A1 (de) 2009-08-28 2011-03-10 Perner, Petra, Dr.-Ing. Einrichtung und Verfahren zur automatischen Erfassung der dynamischen Prozesse von Zellen von Zellproben
HUP0900741A2 (en) 2009-11-27 2011-07-28 3Dhistech Kft Assembly and procedure for digitization of moving slide
CA2893590C (fr) 2012-12-14 2023-10-10 The J. David Gladstone Institutes Systemes de microscopie robotique automatisee
US9904842B2 (en) 2012-12-19 2018-02-27 Koninklijke Philips N.V. System and method for classification of particles in a fluid sample
CN105096294B (zh) * 2014-04-30 2019-01-18 西门子医疗保健诊断公司 用于对尿液沉渣图像的待处理区块进行区块检索的方法和装置
US10755078B2 (en) 2015-04-10 2020-08-25 President And Fellows Of Harvard College Methods and devices for live cell imaging analysis
JP6959218B2 (ja) * 2015-08-12 2021-11-02 モレキュラー デバイシーズ， エルエルシー 細胞の表現型応答を自動的に分析するためのシステムおよび方法
JP2018061440A (ja) * 2016-10-11 2018-04-19 憲隆 福永 受精判定システム
CN110431637A (zh) 2017-02-09 2019-11-08 莱维特医疗公司 用于组织样本处理的系统和方法
US10540535B2 (en) 2017-03-13 2020-01-21 Carl Zeiss Microscopy Gmbh Automatically identifying regions of interest on images of biological cells
TWI699816B (zh) * 2017-12-26 2020-07-21 雲象科技股份有限公司 自動化顯微鏡系統之控制方法、顯微鏡系統及電腦可讀取記錄媒體
DE102018217901A1 (de) 2018-10-18 2020-04-23 Leica Microsystems Cms Gmbh Optimierung von Arbeitsabläufen von Mikroskopen
WO2020092550A1 (fr) * 2018-11-01 2020-05-07 Leavitt Medical, Inc. Lames d'anatomopathologie ainsi que systèmes et méthodes associés
PT3884463T (pt) * 2018-11-30 2024-03-14 Amgen Inc Sistemas e métodos para deteção do efeito citopático nas células
BR112021010404A2 (pt) 2018-11-30 2021-08-24 Amgen Inc. Sistemas e métodos para facilitar a seleção de clones
JP2020005661A (ja) * 2019-10-16 2020-01-16 憲隆 福永 受精判定システム
CN111339904B (zh) * 2020-02-21 2023-11-03 腾讯科技（深圳）有限公司 动物精子图像识别方法和装置
WO2022081863A2 (fr) 2020-10-16 2022-04-21 Life Technologies Corporation Systèmes et procédés de mise au point automatique et de comptage de cellules automatisé à l'aide d'une intelligence artificielle
CN112396583B (zh) * 2020-11-18 2024-01-26 深思考人工智能机器人科技（北京）有限公司 一种基于巴氏染色方式的dna倍体定量分析的方法及系统
WO2022107435A1 (fr) * 2020-11-20 2022-05-27 コニカミノルタ株式会社 Procédé d'analyse d'image, système d'analyse d'image et programme

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4965725B1 (en) * 1988-04-08 1996-05-07 Neuromedical Systems Inc Neural network based automated cytological specimen classification system and method
US5313532A (en) * 1990-01-23 1994-05-17 Massachusetts Institute Of Technology Recognition of patterns in images
WO1991020048A1 (fr) 1990-06-21 1991-12-26 Applied Electronic Vision, Inc. Analyse cellulaire utilisant un traitement video et un reseau neuronal
US5857030A (en) * 1995-08-18 1999-01-05 Eastman Kodak Company Automated method and system for digital image processing of radiologic images utilizing artificial neural networks
DE69630935T2 (de) * 1995-09-29 2004-11-04 Koninklijke Philips Electronics N.V. Bilverarbeitungsverfahren und -vorrichtung zur automatischen Erfassung von Bereichen eines vorbestimmten Krebstyps in einem Intensitätsbild
US5741213A (en) * 1995-10-25 1998-04-21 Toa Medical Electronics Co., Ltd. Apparatus for analyzing blood

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See references of WO9741416A1 *

Also Published As

Publication number Publication date
DE19616997A1 (de) 1997-10-30
US6246785B1 (en) 2001-06-12
JP2007133859A (ja) 2007-05-31
JP2000513465A (ja) 2000-10-10
WO1997041416A1 (fr) 1997-11-06

Similar Documents

Publication Publication Date Title
EP0896661A1 (fr) Procede d'analyse automatisee assistee par ordinateur d'echantillons tissulaires ou d'echantillons de liquides organiques
DE69232413T2 (de) Verfahren und Vorrichtung zum automatischen Test biologischer Proben
DE2823490C2 (fr)
DE69322498T2 (de) Verfahren zur identifizierung von normalen biomedizinischen proben
DE19747415C2 (de) Verfahren zur Unterstützung eines Betrachters bei der Durchmusterung einer Probe und zytologisches Probenanalysiersystem
EP1181525B1 (fr) Procede pour l'analyse automatique d'images microscopiques
DE69422128T2 (de) Verfahren zur interaktiven automatischen Zytologie
DE3689856T3 (de) Analyseverfahren und vorrichtung für biologische specimen.
AT411066B (de) Verfahren und anordnung zur untersuchung von zellen
DE69131102T2 (de) Untersuchungskontrollverfahren für bilder die auf einer anzeige gezeigt werden
DE102008059788B4 (de) Analyse und Klassifizierung insbesondere biologischer oder biochemischer Objekte auf Basis von Zeitreihen-Bildern, anwendbar bei der zytometrischen Time-Lapse-Zellanalyse in der bildbasierten Zytometrie
DE69532276T2 (de) Automatisiertes cytologisches probenklassifizierungsverfahren
DE69429145T2 (de) Klassifikation und Prüfvorrichtung für Teilchen in einer Flüssigkeit
EP3575848A1 (fr) Analyseur destiné à l'analyse tridimensionnelle d'un échantillon médical au moyen d'un appareil photographique plénoptique
DE4211904C2 (de) Automatisches Verfahren zum Erstellen einer Liste unterschiedlicher Arten für eine flüssige Probe
DE102007018034A1 (de) Automatische Bildanalyse und Quantifizierung für Fluoreszenz-In Situ-Hybridisierung
DE1498824A1 (de) Vorrichtung zur maschinellen Krebsdiagnose
DE69229036T2 (de) Methode und Vorrichtung zur automatischen Zellanalyse
DE112007001907T5 (de) Verfahren für die Zellanalyse
EP3828617A1 (fr) Procédé de colorisation numérique de cellules
DE19801400C2 (de) Verfahren zur automatischen Erkennung, Eigenschaftsbeschreibung und Interpretation von Hep-2-Zellmustern
DE19637741B4 (de) Automatisches Zytologisches Probenanalysesystem unter Verwendung von individualisierten Patientendaten
DE10124340A1 (de) Verfahren zur Analyse einer biologischen Probe
DE69831573T2 (de) Inspektionssystem mit probenvorschau
EP3540631A1 (fr) Procédé in vitro destiné à la détermination sans marquage d'un type de cellule d'une cellule sanguine blanche

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19981127

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): BE CH DE ES FR GB IT LI NL SE

RAP3 Party data changed (applicant data changed or rights of an application transferred)

Owner name: ROCHE DIAGNOSTICS GMBH

17Q First examination report despatched

Effective date: 20031107

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: 3DHISTECH KFT

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20101103