Accurate and reproducible invasive breast cancer detection in whole-slide images: A Deep Learning approach for quantifying tumor extent

Abstract

With the increasing ability to routinely and rapidly digitize whole slide images with slide scanners, there has been interest in developing computerized image analysis algorithms for automated detection of disease extent from digital pathology images. The manual identification of presence and extent of breast cancer by a pathologist is critical for patient management for tumor staging and assessing treatment response. However, this process is tedious and subject to inter- and intra-reader variability. For computerized methods to be useful as decision support tools, they need to be resilient to data acquired from different sources, different staining and cutting protocols and different scanners. The objective of this study was to evaluate the accuracy and robustness of a deep learning-based method to automatically identify the extent of invasive tumor on digitized images. Here, we present a new method that employs a convolutional neural network for detecting presence of invasive tumor on whole slide images. Our approach involves training the classifier on nearly 400 exemplars from multiple different sites, and scanners, and then independently validating on almost 200 cases from The Cancer Genome Atlas. Our approach yielded a Dice coefficient of 75.86%, a positive predictive value of 71.62% and a negative predictive value of 96.77% in terms of pixel-by-pixel evaluation compared to manually annotated regions of invasive ductal carcinoma.

Status and Direction of Healthcare Data in Korea for Artificial Intelligence

“…In addition, there are standards for image data, e.g., DICOM or JPEG. For the same reason, digital pathology images are good target areas for deep learning [24,25]. If we have standardized clinical data equivalent to MRI and CT images, we can apply diverse machine learning technologies as well.…”

L’intelligence artificielle à l’aide des myologues

“…Déjà en usage dans le cadre de recommandations thérapeutiques et de diagnostic, les techniques « d'apprentissage profond », outils qui permettent de transformer la masse énorme d'informations collectées en connaissance, aident des experts en imagerie médicale à mieux identifier, classer, quantifier, repérer les anomalies et interpréter des images issues de radiographies [12], de PET/Scan [13] et/ou d'IRM [14]. Ils favorisent le dépistage de certains cancers, de fractures et même d'atteintes dues à la maladie d'Alzheimer [15].…”

Context-based interpolation of coarse deep learning prediction maps for the segmentation of fine structures in immunofluorescence images

“…Classification networks as well as semantic segmentation networks such as the fully convolutional (FCNs) network 7 and the deconvolution networks (DNs) 8 previously described by Long et al and Noh et al respectively, are commonly employed for the segmentation of whole slide images. [4][5][6] As an example workflow, given an input image and a trained classification network, the corresponding segmentation map is obtained by applying the net sequentially on a rectangular grid. 6 Alternatively to this sliding window strategy, a fully convolutional network enables the direct computation of a label map.…”

“…[4][5][6] As an example workflow, given an input image and a trained classification network, the corresponding segmentation map is obtained by applying the net sequentially on a rectangular grid. 6 Alternatively to this sliding window strategy, a fully convolutional network enables the direct computation of a label map. To this end, the fully connected layers in the original classification CNN are converted by their convolutional equivalent.…”