HEROHE Challenge: Predicting HER2 Status in Breast Cancer from Hematoxylin-Eosin Whole-Slide Imaging.

Breast cancer is the most common malignancy in women worldwide, and is responsible for more than half a million deaths each year. The appropriate therapy depends on the evaluation of the expression of various biomarkers, such as the human epidermal growth factor receptor 2 (HER2) transmembrane protein, through specialized techniques, such as immunohistochemistry or in situ hybridization. In this work, we present the HER2 on hematoxylin and eosin (HEROHE) challenge, a parallel event of the 16th European Congress on Digital Pathology, which aimed to predict the HER2 status in breast cancer based only on hematoxylin-eosin-stained tissue samples, thus avoiding specialized techniques. The challenge consisted of a large, annotated, whole-slide images dataset (509), specifically collected for the challenge. Models for predicting HER2 status were presented by 21 teams worldwide. The best-performing models are presented by detailing the network architectures and key parameters. Methods are compared and approaches, core methodologies, and software choices contrasted. Different evaluation metrics are discussed, as well as the performance of the presented models for each of these metrics. Potential differences in ranking that would result from different choices of evaluation metrics highlight the need for careful consideration at the time of their selection, as the results show that some metrics may misrepresent the true potential of a model to solve the problem for which it was developed. The HEROHE dataset remains publicly available to promote advances in the field of computational pathology.

Seamless Virtual Whole Slide Image Synthesis and Validation Using Perceptual Embedding Consistency.

Stain virtualization is an application with growing interest in digital pathology allowing simulation of stained tissue images thus saving lab and tissue resources. Thanks to the success of Generative Adversarial Networks (GANs) and the progress of unsupervised learning, unsupervised style transfer GANs have been successfully used to generate realistic, clinically meaningful and interpretable images. The large size of high resolution Whole Slide Images (WSIs) presents an additional computational challenge. This makes tilewise processing necessary during training and inference of deep learning networks. Instance normalization has a substantial positive effect in style transfer GAN applications but with tilewise inference, it has the tendency to cause a tiling artifact in reconstructed WSIs. In this paper we propose a novel perceptual embedding consistency (PEC) loss forcing the network to learn color, contrast and brightness invariant features in the latent space and hence substantially reducing the aforementioned tiling artifact. Our approach results in more seamless reconstruction of the virtual WSIs. We validate our method quantitatively by comparing the virtually generated images to their corresponding consecutive real stained images. We compare our results to state-of-the-art unsupervised style transfer methods and to the measures obtained from consecutive real stained tissue slide images. We demonstrate our hypothesis about the effect of the PEC loss by comparing model robustness to color, contrast and brightness perturbations and visualizing bottleneck embeddings. We validate the robustness of the bottleneck feature maps by measuring their sensitivity to the different perturbations and using them in a tumor segmentation task. Additionally, we propose a preliminary validation of the virtual staining application by comparing interpretation of 2 pathologists on real and virtual tiles and inter-pathologist agreement.

Enabling Histopathological Annotations on Immunofluorescent Images through Virtualization of Hematoxylin and Eosin.

CONTEXT: Medical diagnosis and clinical decisions rely heavily on the histopathological evaluation of tissue samples, especially in oncology. Historically, classical histopathology has been the gold standard for tissue evaluation and assessment by pathologists. The most widely and commonly used dyes in histopathology are hematoxylin and eosin (H&E) as most malignancies diagnosis is largely based on this protocol. H&E staining has been used for more than a century to identify tissue characteristics and structures morphologies that are needed for tumor diagnosis. In many cases, as tissue is scarce in clinical studies, fluorescence imaging is necessary to allow staining of the same specimen with multiple biomarkers simultaneously. Since fluorescence imaging is a relatively new technology in the pathology landscape, histopathologists are not used to or trained in annotating or interpreting these images. AIMS SETTINGS AND DESIGN: To allow pathologists to annotate these images without the need for additional training, we designed an algorithm for the conversion of fluorescence images to brightfield H&E images. SUBJECTS AND METHODS: In this algorithm, we use fluorescent nuclei staining to reproduce the hematoxylin information and natural tissue autofluorescence to reproduce the eosin information avoiding the necessity to specifically stain the proteins or intracellular structures with an additional fluorescence stain. STATISTICAL ANALYSIS USED: Our method is based on optimizing a transform function from fluorescence to H&E images using least mean square optimization. RESULTS: It results in high quality virtual H&E digital images that can easily and efficiently be analyzed by pathologists. We validated our results with pathologists by making them annotate tumor in real and virtual H&E whole slide images and we obtained promising results. CONCLUSIONS: Hence, we provide a solution that enables pathologists to assess tissue and annotate specific structures based on multiplexed fluorescence images.

Bimanual adaptation: internal representations of bimanual rhythmic movements.

From tying your shoes and clipping your tie to the claps at the end of a fine seminar, bimanual coordination plays a major role in our daily activities. An important phenomenon in bimanual coordination is the predisposition toward mirror symmetry in the performance of bimanual rhythmic movements. Although learning and adaptation in bimanual coordination are phenomena that have been observed, they have not been studied in the context of adaptive control and internal representations-approaches that were successfully employed in the arena of reaching movements and adaptation to force perturbations. In this paper we examine the dynamics of the learning mechanisms involved when subjects are trained to perform a bimanual non-harmonic polyrhythm in a bimanual index finger tapping task. Subjects are trained in this task implicitly, using altered visual feedback, while their performance is continuously monitored throughout the experiment. Our experimental results indicate the existence of significant (p<<0.01) learning curves (i.e., error plots with significantly negative slopes) during training and aftereffects with a washout period after the visual feedback ceases to be altered. These results confirm the formation of internal representations in bimanual motor control. We present a simple, physiologically plausible, neural model that combines feedback and adaptation in the control process and which is able to reproduce key phenomena of bimanual coordination and adaptation.