Abstract: The present disclosure relates to techniques for efficient development of initial models and efficient model update and/or adaptation to a different image domain using an adaptive learning framework. For efficient development of initial models, a two-step development strategy may be performed as follows: Phase 1: Model preconditioning, where an artificial intelligence system leverages existing annotated datasets and improves learning skills through training of these datasets; and Phase 2: Target-model training, where an artificial intelligence system utilizes the learning skills learned from Phase 1 to extend itself to a different image domain (target domain) with less number of annotations required in the target domain than conventional learning methods.

Abstract: Methods, computer-program products and systems are provided to perform actions including: receiving an image and displaying the image using a graphical user interface; receiving at least one first image annotation provided by a user via the graphical user interface; producing a first segmented image using a deep learning model, wherein the deep learning model uses the digital pathology image and the at least one first image annotation; and displaying the first segmented image using the graphical user interface; receiving at least one second image annotation provided by the user via the graphical user interface; producing a second segmented image using the deep learning model, wherein the deep learning model uses the digital pathology image, the at least one first image annotation, and the at least one second image annotation; and displaying the second segmented image using the graphical user interface.

Abstract: A computer-based specimen analyzer (10) is configured to detect a level of expression of genes in a cell sample by detecting dots that represent differently stained genes and chromosomes in a cell. The color of the stained genes and the chromosomes is enhanced and filtered to produce a dot mask that defines areas in the image that are genes, chromosomes, or non-genetic material. Metrics are determined for the dots and/or pixels in the image of the cell in areas corresponding to the dots. The metrics are fed to a classifier that separates genes from chromosomes. The results of the classifier are counted to estimate the expression level of genes in the tissue samples.

Abstract: The present disclosure relates to techniques for pre-processing training data, augmenting training data, and using synthetic training data to effectively train a machine learning model to (i) reject adversarial example images, and (ii) detect, characterize and/or classify some or all regions of images that do not include adversarial example regions. Particularly, aspects of the present disclosure are directed to receiving a training set of images for training a machine learning algorithm to detect, characterize, classify, or a combination thereof some or all regions or objects within the images, augmenting the training set of images with synthetic images generated from one or more adversarial algorithms to generate augmented batches of images, and train the machine learning algorithm using the augmented batches of images to generate a machine learning model configured to detect, characterize, classify, or a combination thereof some or all regions or objects within new images.

Abstract: A scoring functions is developed and used for identifying patients who might be responsive to a PD-1 axis directed therapy. The scoring functions are obtained by extracting features from multiplex-stained sections, selecting features that correlate with response to the therapy using a feature selection function, and fitting one or more of the selected features to a plurality of candidate scoring functions. A candidate scoring function showing the desired balance between predictive sensitivity and specificity may then selected for incorporation into a scoring system that includes at least an image analysis system.

Abstract: Method and systems for of using a machine-learning model to detect predicted artifacts at a target image resolution are provided. A machine-learning model trained to detect artifact pixels in images at a target image resolution is accessed. An image depicting at least part of the biological sample at an initial image resolution can be converted at the target image resolution. The machine-learning model is applied to the converted image to identify one or more artifact pixels from the converted image. Method and systems for training the machine-learning model to detect predicted artifacts at the target image resolution are also provided.

Abstract: Duplex immunohistochemistry (IHC) staining of tissue sections allows simultaneous detection of two biomarkers and their co-expression at the single-cell level, and does not require two IHC stains and additional registration to identify co-localization. Duplex IHC are often difficult for human including pathologists to reliably score. The methods and system herein use machine-learning models and probability maps to detect and record individual phenotype ER/PR.

Abstract: Techniques for acquiring focused images of a microscope slide are disclosed. During a calibration phase, a “base” focal plane is determined using non-synthetic and/or synthetic auto-focus techniques. Furthermore, offset planes are determined for color channels (or filter bands) and used to generate an auto-focus model. During subsequent scans, the auto-focus model can be used to quickly estimate the focal plane of interest for each color channel (or filter band) rather than re-employing the non-synthetic and/or synthetic auto-focus techniques.

Abstract: A method for analyzing an image of a tissue section may include obtaining a plurality of image locations, each corresponding to a different one of a plurality of biological structures; obtaining a plurality of locations of a first biomarker in the image; and calculating a distance transform array for at least a portion of the image that includes the plurality of seed locations. The method may include, for each of the plurality of seed locations and based on information from the first distance transform array, detecting whether the first biomarker is expressed at the seed location, and storing, to a data structure associated with the seed location, an indication of whether expression of the first biomarker at the seed location was detected. The method may include detecting, based on the stored indications, co-localization of at least two phenotypes in at least a portion of the tissue section.

Abstract: Techniques for acquiring focused images of a microscope slide are disclosed. During a calibration phase, a “base” focal plane is determined using non-synthetic and/or synthetic auto-focus techniques. Furthermore, offset planes are determined for color channels (or filter bands) and used to generate an auto-focus model. During subsequent scans, the auto-focus model can be used to quickly estimate the focal plane of interest for each color channel (or filter band) rather than re-employing the non-synthetic and/or synthetic auto-focus techniques.

Abstract: Methods and systems can include: accessing a digital pathology image; generating, using a first machine-learning model, a segmented image that identifies at least: a predicted diseased region and a background region in the digital pathology image; detecting depictions of a set of cells in the digital pathology image; generating, using a second machine-learning model, a cell classification for each cell of the set of cells, wherein the cell classification is selected from a set of potential classifications that indicate which, if any, of a set of biomarkers are expressed in the cell; detecting that a subset of the set of cells are within the background region; and updating the cell classification for each cell of at least some cells in the subset to be a background classification that was not included in the set of potential classifications.

Abstract: Methods and systems for predictive measures of anti-EGFR therapy response in wild type RAS/EGFR+ samples, e.g., histochemical staining methods for staining EGFR, AREG, and EREG, digital analysis of stained slides, and scoring algorithms that allow prediction of a response to anti-EGFR therapies. Analysis of the stained slides and scoring algorithms may include but are not limited to: a percent tumor cell positivity, computerized clustering algorithms, area density (e.g., area of tumor positive for one or more markers over total tumor area), average intensity (e.g., computerized methodology measuring average gray scale pixel intensity), average intensity broken down according to membrane, cytoplasmic, or punctate staining patterns), or any other appropriate parameter or combination of parameters. The methods of the present invention allow for resolving spatial expression patterns of the ligands and the receptor to determine what patterns are predictive for response to anti-EGFR therapies.

Abstract: The present disclosure relates to computer-implement techniques for cell localization and classification. Particularly, aspects of the present disclosure are directed to accessing an image for a biological sample, where the image depicts cells comprising a staining pattern of a biomarker; inputting the image into a machine learning model; encoding, by the machine learning model, the image into a feature representation comprising extracted discriminative features; combining, by the machine learning model, feature and spatial information of the cells and the staining pattern of the biomarker through a sequence of up-convolutions and concatenations with the extracted discriminative features from the feature representation; and generating, by the machine learning model, two or more segmentation masks for the biomarker in the image based on the combined feature and spatial information of the cells and the staining pattern of the biomarker.

Abstract: A machine learning model is accessed that is configured to use one or more parameters to process images to generate labels. The machine learning model is executed to transform at least part of each of at least one digital pathology image into a plurality of predicted labels; and generate a confidence metric for each of the plurality of predicted labels. An interface is availed that depicts the at least part of the at least one digital pathology image and that differentially represents predicted labels based on corresponding confidence metrics. In response to availing of the interface, label input is received that confirms, rejects, or replaces at least one of the plurality of predicted labels. The one or more parameters of the machine learning model are updated based on the label input.

Abstract: Techniques for acquiring focused images of a microscope slide are disclosed. During a calibration phase, a “base” focal plane is determined using non-synthetic and/or synthetic auto-focus techniques. Furthermore, offset planes are determined for color channels (or filter bands) and used to generate an auto-focus model. During subsequent scans, the auto-focus model can be used to quickly estimate the focal plane of interest for each color channel (or filter band) rather than re-employing the non-synthetic and/or synthetic auto-focus techniques.

Abstract: Efficient methods for identifying biomarkers are described. The method may include identifying a tumor area. The method may further include identifying a plurality of regions. The method may also include defining, for each region, a bounding area for the region that encompasses the region. The method may include determining, for each region of a first subset of the plurality of regions, that the region is to be ascribed to the tumor, where the bounding area is fully within the tumor area. The method may further include determining, for each region of a second subset of the plurality of regions, whether to ascribe the region to the tumor based on an intersection of the region and the tumor area. The method may also include accessing a metric characterizing a biological observation and generating a result based on the metrics. The result may be used as a biomarker.

Abstract: A method for using a federated learning classifier in digital pathology includes distributing, by a centralized server, a global model to a plurality of client devices. The client devices further train the global model using a plurality images of a specimen and corresponding annotations to generate at least one further trained model. The client devices provide further trained models to the centralized server, which aggregates the further trained models with the global model to generate an updated global model. The updated global model is then distributed to the plurality of client devices.

Abstract: Techniques for acquiring focused images of a microscope slide are disclosed. During a calibration phase, a “base” focal plane is determined using non-synthetic and/or synthetic auto-focus techniques. Furthermore, offset planes are determined for color channels (or filter bands) and used to generate an auto-focus model. During subsequent scans, the auto-focus model can be used to quickly estimate the focal plane of interest for each color channel (or filter band) rather than re-employing the non-synthetic and/or synthetic auto-focus techniques.

Abstract: Techniques relate to object classifications using bootstrapping of region-level annotations. For each of multiple images, regions within the image can be identified. For each region, a region-specific label can be identified, a set of objects within the region can be detected, and an object-specific label can be assigned to each object. The object-specific label can be the same as the region-specific label assigned to the region within which the object is located. A training data set can be defined to include, for each image of the multiple images, object-location data (indicating intra-image location data for the detected object) and label data (indicating the object-specific labels assigned to the objects). An image-processing model can be trained using the training data. Training can include learning values for a set of parameters that define calculations performed by the image-processing model.

Abstract: Systems and methods disclosed herein describe a platform that automatically creates and executes a scoring guide for use in anatomical pathology. The platform can employ a fully-automated workflow for clustering the biological objects of interest and for providing cell-by-cell read-outs of heterogeneous tumor biomarkers based on their stain appearance. The platform can include a module for automatically creating and storing a scoring guide in a training database based on training digital images (240, 250), and an object classification module that executes the scoring guide when presented with new digital images to be scored pursuant to the scoring guide (299).