A Complete Color Normalization Approach to Histopathology Images Using Color Cues Computed From Saturation-Weighted Statistics

Recurrence analysis on prostate cancer patients with Gleason score 7 using integrated histopathology whole-slide images and genomic data through deep neural networks

Prostate cancer is the most common nonskin-related cancer, affecting one in seven men in the United States. Gleason score, a sum of the primary and secondary Gleason patterns, is one of the best predictors of prostate cancer outcomes. Recently, significant progress has been made in molecular subtyping prostate cancer through the use of genomic sequencing. It has been established that prostate cancer patients presented with a Gleason score 7 show heterogeneity in both disease recurrence and survival. We built a unified system using publicly available whole-slide images and genomic data of histopathology specimens through deep neural networks to identify a set of computational biomarkers. Using a survival model, the experimental results on the public prostate dataset showed that the computational biomarkers extracted by our approach had hazard ratio as 5.73 and C -index as 0.74, which were higher than standard clinical prognostic factors and other engineered image texture features. Collectively, the results of this study highlight the important role of neural network analysis of prostate cancer and the potential of such approaches in other precision medicine applications.

A study about color normalization methods for histopathology images

Histopathology images are used for the diagnosis of the cancerous disease by the examination of tissue with the help of Whole Slide Imaging (WSI) scanner. A decision support system works well by the analysis of the histopathology images but a lot of problems arise in its decision. Color variation in the histopathology images is occurring due to use of the different scanner, use of various equipments, different stain coloring and reactivity from a different manufacturer. In this paper, detailed study and performance evaluation of color normalization methods on histopathology image datasets are presented. Color normalization of the source image by transferring the mean color of the target image in the source image and also to separate stain present in the source image. Stain separation and color normalization of the histopathology images can be helped for both pathology and computerized decision support system. Quality performances of different color normalization methods are evaluated and compared in terms of quaternion structure similarity index matrix (QSSIM), structure similarity index matrix (SSIM) and Pearson correlation coefficient (PCC) on various histopathology image datasets. Our experimental analysis suggests that structure-preserving color normalization (SPCN) provides better qualitatively and qualitatively results in comparison to the all the presented methods for breast and colorectal cancer histopathology image datasets.

Effects of different preprocessing algorithms on the prognostic value of breast tumour microscopic images

The purpose of this study was to improve the prognostic value of tumour histopathology image analysis methodology by image preprocessing. Key image qualities were modified including contrast, sharpness and brightness. The texture information was subsequently extracted from images of haematoxylin/eosin-stained tumour tissue sections by GLCM, monofractal and multifractal algorithms without any analytical limitation to predefined structures. Images were derived from patient groups with invasive breast carcinoma (BC, 93 patients) and inflammatory breast carcinoma (IBC, 51 patients). The prognostic performance was indeed significantly enhanced by preprocessing with the average AUCs of individual texture features improving from 0.68 ± 0.05 for original to 0.78 ± 0.01 for preprocessed images in the BC group and 0.75 ± 0.01 to 0.80 ± 0.02 in the IBC group. Image preprocessing also improved the prognostic independence of texture features as indicated by multivariate analysis. Surprisingly, the tonal histogram compression by the nonnormalisation preprocessing has prognostically outperformed the tested contrast normalisation algorithms. Generally, features without prognostic value showed higher susceptibility to prognostic enhancement by preprocessing whereas IDM texture feature was exceptionally susceptible. The obtained results are suggestive of the existence of distinct texture prognostic clues in the two examined types of breast cancer. The obtained enhancement of prognostic performance is essential for the anticipated clinical use of this method as a simple and cost-effective prognosticator of cancer outcome.

Spatial Statistics for Segmenting Histological Structures in H&E Stained Tissue Images

Segmenting a broad class of histological structures in transmitted light and/or fluorescence-based images is a prerequisite for determining the pathological basis of cancer, elucidating spatial interactions between histological structures in tumor microenvironments (e.g., tumor infiltrating lymphocytes), facilitating precision medicine studies with deep molecular profiling, and providing an exploratory tool for pathologists. This paper focuses on segmenting histological structures in hematoxylin- and eosin-stained images of breast tissues, e.g., invasive carcinoma, carcinoma in situ, atypical and normal ducts, adipose tissue, and lymphocytes. We propose two graph-theoretic segmentation methods based on local spatial color and nuclei neighborhood statistics. For benchmarking, we curated a data set of 232 high-power field breast tissue images together with expertly annotated ground truth. To accurately model the preference for histological structures (ducts, vessels, tumor nets, adipose, etc.) over the remaining connective tissue and non-tissue areas in ground truth annotations, we propose a new region-based score for evaluating segmentation algorithms. We demonstrate the improvement of our proposed methods over the state-of-the-art algorithms in both region- and boundary-based performance measures.

Image processing in digital pathology: an opportunity to solve inter-batch variability of immunohistochemical staining

Immunohistochemistry (IHC) is a widely used technique in pathology to evidence protein expression in tissue samples. However, this staining technique is known for presenting inter-batch variations. Whole slide imaging in digital pathology offers a possibility to overcome this problem by means of image normalisation techniques. In the present paper we propose a methodology to objectively evaluate the need of image normalisation and to identify the best way to perform it. This methodology uses tissue microarray (TMA) materials and statistical analyses to evidence the possible variations occurring at colour and intensity levels as well as to evaluate the efficiency of image normalisation methods in correcting them. We applied our methodology to test different methods of image normalisation based on blind colour deconvolution that we adapted for IHC staining. These tests were carried out for different IHC experiments on different tissue types and targeting different proteins with different subcellular localisations. Our methodology enabled us to establish and to validate inter-batch normalization transforms which correct the non-relevant IHC staining variations. The normalised image series were then processed to extract coherent quantitative features characterising the IHC staining patterns.

Stain Deconvolution Using Statistical Analysis of Multi-Resolution Stain Colour Representation

Stain colour estimation is a prominent factor of the analysis pipeline in most of histology image processing algorithms. Providing a reliable and efficient stain colour deconvolution approach is fundamental for robust algorithm. In this paper, we propose a novel method for stain colour deconvolution of histology images. This approach statistically analyses the multi-resolutional representation of the image to separate the independent observations out of the correlated ones. We then estimate the stain mixing matrix using filtered uncorrelated data. We conducted an extensive set of experiments to compare the proposed method to the recent state of the art methods and demonstrate the robustness of this approach using three different datasets of scanned slides, prepared in different labs using different scanners.