Neural network based identification of areas of interest in digital pathology images

Abstract

A CNN is applied to a histological image to identify areas of interest. The CNN classifies pixels according to relevance classes including one or more classes indicating levels of interest and at least one class indicating lack of interest. The CNN is trained on a training data set including data which has recorded how pathologists have interacted with visualizations of histological images. In the trained CNN, the interest-based pixel classification is used to generate a segmentation mask that defines areas of interest. The mask can be used to indicate where in an image clinically relevant features may be located. Further, it can be used to guide variable data compression of the histological image. Moreover, it can be used to control loading of image data in either a client-server model or within a memory cache policy. Furthermore, a histological image of a tissue sample of a tissue type that has been treated with a test compound is image processed in order to detect areas where toxic reactions to the test compound may have occurred. An autoencoder is trained with a training data set comprising histological images of tissue samples which are of the given tissue type, but which have not been treated with the test compound. The trained autoencoder is applied to detect tissue areas by their deviation from the normal variation seen in that tissue type as learnt by the training process, and so build up a toxicity map of the image. The toxicity map can then be used to direct a toxicological pathologist to examine the areas identified by the autoencoder as lying outside the normal range of heterogeneity for the tissue type. This makes the pathologist's review quicker and more reliable. The toxicity map can also be overlayed with the segmentation mask indicating areas of interest. When an area of interest and an area identified as lying outside the normal range of heterogeneity for the tissue type, and increased confidence score is applied to the overlapping area.

Description

technical field [0001] The present disclosure relates to image processing of pathology images using neural networks in order to find regions of clinical interest and regions comprising toxicological pathology. Background technique [0002] Digital pathology continues to change the way pathologists view and diagnose slides. The traditional way for pathologists to examine slides is to view them under a microscope. The pathologist will first view the slide using a low-magnification objective. When an area of ​​potential diagnostic value is observed, the pathologist will switch to a high-magnification objective to view the area in more detail. The pathologist will then switch back to low magnification to continue examining other areas on the slide. This sequence of low-high and low-magnification viewing can be repeated multiple times on the slide until a definitive and complete diagnosis can be made on the slide. [0003] Over the past two decades, the introduction of digita...

AI Technical Summary

Problems solved by technology

Despite the increased efficiency of digital pathology, detecting morphological changes caused by test compounds remains a tedious and error-prone process that requires toxicology pathologists to look at many tissue samples and determine that observed changes are due to the test compound. Caused or due to normal tissue heterogeneity or some other disease process

Images