Joint analysis of histopathology image features and gene expression in breast cancer

Abstract

BackgroundGenomics and proteomics are nowadays the dominant techniques for novel biomarker discovery. However, histopathology images contain a wealth of information related to the tumor histology, morphology and tumor-host interactions that is not accessible through these techniques. Thus, integrating the histopathology images in the biomarker discovery workflow could potentially lead to the identification of new image-based biomarkers and the refinement or even replacement of the existing genomic and proteomic signatures. However, extracting meaningful and robust image features to be mined jointly with genomic (and clinical, etc.) data represents a real challenge due to the complexity of the images.ResultsWe developed a framework for integrating the histopathology images in the biomarker discovery workflow based on the bag-of-features approach – a method that has the advantage of being assumption-free and data-driven. The images were reduced to a set of salient patterns and additional measurements of their spatial distribution, with the resulting features being directly used in a standard biomarker discovery application. We demonstrated this framework in a search for prognostic biomarkers in breast cancer which resulted in the identification of several prognostic image features and a promising multimodal (imaging and genomic) prognostic signature. The source code for the image analysis procedures is freely available.ConclusionsThe framework proposed allows for a joint analysis of images and gene expression data. Its application to a set of breast cancer cases resulted in image-based and combined (image and genomic) prognostic scores for relapse-free survival.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1072-z) contains supplementary material, which is available to authorized users.

Computational discovery of tissue morphology biomarkers in very long-term survivors with pancreatic ductal adenocarcinoma

“…3d). Though the tiling strategy is a common strategy to reduce WSI information 26,34,35 , our study suggests that including spatial information can significantly improve disease outcome prediction and patient stratification.…”

Breast cancer outcome prediction with tumour tissue images and machine learning

“…Extracting more than 6000 predefined image features from two cohorts ( N = 248 and N = 328), authors in [35] showed that it is feasible to learn an outcome predictor for overall survival (HR = 1.78, p = 0.017) using small tumour regions. Using a subset of the whole slides from [36], an earlier study [37] proposed a joint analysis of image features and gene expression signatures for prognostic biomarker discovery. The authors used a training set of 131 patients and validated the biomarkers with H&E-stained tumour samples from 65 breast cancer patients.…”

Translational Bioinformatics for Cholangiocarcinoma: Opportunities and Challenges

“…Therefore, integrative studies are not only applied to the same omics level, such as genome, transcriptome, proteome etc. but also could be performed at cross-levels like molecular level and tissue/organ level by combining the gene expression information with the imaging features from magnetic resonance imaging (MRI), computed tomography (CT) and positron emission tomography (PET)/CT 66 - 69 . Although this kind of integrative study is not yet reported for the study of CCA, in the future the integration of gene expression or proteomics data with the clinical imaging features will provide a better way to the diagnosis of CCA and to the understanding the mechanism of CCA genesis and progression.…”