The ACROBAT 2022 challenge: Automatic registration of breast cancer tissue.

The alignment of tissue between histopathological whole-slide-images (WSI) is crucial for research and clinical applications. Advances in computing, deep learning, and availability of large WSI datasets have revolutionised WSI analysis. Therefore, the current state-of-the-art in WSI registration is unclear. To address this, we conducted the ACROBAT challenge, based on the largest WSI registration dataset to date, including 4,212 WSIs from 1,152 breast cancer patients. The challenge objective was to align WSIs of tissue that was stained with routine diagnostic immunohistochemistry to its H&E-stained counterpart. We compare the performance of eight WSI registration algorithms, including an investigation of the impact of different WSI properties and clinical covariates. We find that conceptually distinct WSI registration methods can lead to highly accurate registration performances and identify covariates that impact performances across methods. These results provide a comparison of the performance of current WSI registration methods and guide researchers in selecting and developing methods.

Comparison of consecutive and restained sections for image registration in histopathology.

Significance: Although the registration of restained sections allows nucleus-level alignment that enables a direct analysis of interacting biomarkers, consecutive sections only allow the transfer of region-level annotations. The latter can be achieved at low computational cost using coarser image resolutions. Purpose: In digital histopathology, virtual multistaining is important for diagnosis and biomarker research. Additionally, it provides accurate ground truth for various deep-learning tasks. Virtual multistaining can be obtained using different stains for consecutive sections or by restaining the same section. Both approaches require image registration to compensate for tissue deformations, but little attention has been devoted to comparing their accuracy. Approach: We compared affine and deformable variational image registration of consecutive and restained sections and analyzed the effect of the image resolution that influences accuracy and required computational resources. The registration was applied to the automatic nonrigid histological image registration (ANHIR) challenge data (230 consecutive slide pairs) and the hyperparameters were determined. Then without changing the parameters, the registration was applied to a newly published hybrid dataset of restained and consecutive sections (HyReCo, 86 slide pairs, 5404 landmarks). Results: We obtain a median landmark error after registration of 6.5 µm (HyReCo) and 24.1 µm (ANHIR) between consecutive sections. Between restained sections, the median registration error is 2.2 and 0.9 µm in the two subsets of the HyReCo dataset. We observe that deformable registration leads to lower landmark errors than affine registration in both cases (p<0.001), though the effect is smaller in restained sections. Conclusion: Deformable registration of consecutive and restained sections is a valuable tool for the joint analysis of different stains.

Deep feature based cross-slide registration.

Registration of multiple sections in a tissue block is an important pre-requisite task before any cross-slide image analysis. Non-rigid registration methods are capable of finding correspondence by locally transforming a moving image. These methods often rely on an initial guess to roughly align an image pair linearly and globally. This is essential to prevent convergence to a non-optimal minimum. We explore a deep feature based registration (DFBR) method which utilises data-driven descriptors to estimate the global transformation. A multi-stage strategy is adopted for improving the quality of registration. A visualisation tool is developed to view registered pairs of WSIs at different magnifications. With the help of this tool, one can apply a transformation on the fly without the need to generate a transformed moving WSI in a pyramidal form. We compare the performance on our dataset of data-driven descriptors with that of hand-crafted descriptors. Our approach can align the images with only small registration errors. The efficacy of our proposed method is evaluated for a subsequent non-rigid registration step. To this end, the first two steps of the ANHIR winner's framework are replaced with DFBR to register image pairs provided by the challenge. The modified framework produce comparable results to those of the challenge winning team.

Stain-Independent Deep Learning-Based Analysis of Digital Kidney Histopathology.

Convolutional neural network (CNN)-based image analysis applications in digital pathology (eg, tissue segmentation) require a large amount of annotated data and are mostly trained and applicable on a single stain. Here, a novel concept based on stain augmentation is proposed to develop stain-independent CNNs requiring only one annotated stain. In this benchmark study on stain independence in digital pathology, this approach is comprehensively compared with state-of-the-art techniques including image registration and stain translation, and several modifications thereof. A previously developed CNN for segmentation of periodic acid-Schiff-stained kidney histology was used and applied to various immunohistochemical stainings. Stain augmentation showed very high performance in all evaluated stains and outperformed all other techniques in all structures and stains. Without the need for additional annotations, it enabled segmentation on immunohistochemical stainings with performance nearly comparable to that of the annotated periodic acid-Schiff stain and could further uphold performance on several held-out stains not seen during training. Herein, examples of how this framework can be applied for compartment-specific quantification of immunohistochemical stains for inflammation and fibrosis in animal models and patient biopsy specimens are presented. The results show that stain augmentation is a highly effective approach to enable stain-independent applications of deep-learning segmentation algorithms. This opens new possibilities for broad implementation in digital pathology.

Automated Detection of Portal Fields and Central Veins in Whole-Slide Images of Liver Tissue.

Many physiological processes and pathological phenomena in the liver tissue are spatially heterogeneous. At a local scale, biomarkers can be quantified along the axis of the blood flow, from portal fields (PFs) to central veins (CVs), i.e., in zonated form. This requires detecting PFs and CVs. However, manually annotating these structures in multiple whole-slide images is a tedious task. We describe and evaluate a fully automated method, based on a convolutional neural network, for simultaneously detecting PFs and CVs in a single stained section. Trained on scans of hematoxylin and eosin-stained liver tissue, the detector performed well with an F1 score of 0.81 compared to annotation by a human expert. It does, however, not generalize well to previously unseen scans of steatotic liver tissue with an F1 score of 0.59. Automated PF and CV detection eliminates the bottleneck of manual annotation for subsequent automated analyses, as illustrated by two proof-of-concept applications: We computed lobulus sizes based on the detected PF and CV positions, where results agreed with published lobulus sizes. Moreover, we demonstrate the feasibility of zonated quantification of biomarkers detected in different stainings based on lobuli and zones obtained from the detected PF and CV positions. A negative control (hematoxylin and eosin) showed the expected homogeneity, a positive control (glutamine synthetase) was quantified to be strictly pericentral, and a plausible zonation for a heterogeneous F4/80 staining was obtained. Automated detection of PFs and CVs is one building block for automatically quantifying physiologically relevant heterogeneity of liver tissue biomarkers. Perspectively, a more robust and automated assessment of zonation from whole-slide images will be valuable for parameterizing spatially resolved models of liver metabolism and to provide diagnostic information.

EMPAIA App Interface: An open and vendor-neutral interface for AI applications in pathology.

BACKGROUND AND OBJECTIVE: Artificial intelligence (AI) apps hold great potential to make pathological diagnoses more accurate and time efficient. Widespread use of AI in pathology is hampered by interface incompatibilities between pathology software. We studied the existing interfaces in order to develop the EMPAIA App Interface, an open standard for the integration of pathology AI apps. METHODS: The EMPAIA App Interface relies on widely-used web communication protocols and containerization. It consists of three parts: A standardized format to describe the semantics of an app, a mechanism to deploy and execute apps in computing environments, and a web API through which apps can exchange data with a host application. RESULTS: Five commercial AI app manufacturers successfully adapted their products to the EMPAIA App Interface and helped improve it with their feedback. Open source tools facilitate the adoption of the interface by providing reusable data access and scheduling functionality and enabling automatic validation of app compliance. CONCLUSIONS: Existing AI apps and pathology software can be adapted to the EMPAIA App Interface with little effort. It is a viable alternative to the proprietary interfaces of current software. If enough vendors join in, the EMPAIA App Interface can help to advance the use of AI in pathology.

Automated segmentation and quantification of the healthy and diseased aorta in CT angiographies using a dedicated deep learning approach.

OBJECTIVES: To develop and validate a deep learning-based algorithm for segmenting and quantifying the physiological and diseased aorta in computed tomography angiographies. METHODS: CTA exams of the aorta of 191 patients (68.1 ± 14 years, 128 male), performed between 2015 and 2018, were retrospectively identified from our imaging archive and manually segmented by two investigators. A 3D U-Net model was trained on the data, which was divided into a training, a validation, and a test group at a ratio of 7:1:2. Cases in the test group (n = 41) were evaluated to compare manual and automatic segmentations. Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff surface distance (HSD) were extracted. Maximum diameter, effective diameter, and area were quantified and compared between both segmentations at eight anatomical landmarks, and at the maximum area of an aneurysms if present (n = 14). Statistics included error calculation, intraclass correlation coefficient, and Bland-Altman analysis. RESULTS: A DSC of 0.95 [0.94; 0.95] and an MSD of 0.76 [0.06; 0.99] indicated close agreement between segmentations. HSD was 8.00 [4.47; 10.00]. The largest absolute errors were found in the ascending aorta with 0.8 ± 1.5 mm for maximum diameter and at the coeliac trunk with - 30.0 ± 81.6 mm2 for area. Results for absolute errors in aneurysms were - 0.5 ± 2.3 mm for maximum diameter, 0.3 ± 1.6 mm for effective diameter, and 64.9 ± 114.9 mm2 for area. ICC showed excellent agreement (> 0.9; p < 0.05) between quantitative measurements. CONCLUSIONS: Automated segmentation of the aorta on CTA data using a deep learning algorithm is feasible and allows for accurate quantification of the aortic lumen even if the vascular architecture is altered by disease. KEY POINTS: • A deep learning-based algorithm can automatically segment the aorta, mostly within acceptable margins of error, even if the vascular architecture is altered by disease. • Quantifications performed in the segmentations were mostly within clinically acceptable limits, even in pathologically altered segments of the aorta.

Parietal epithelial cells maintain the epithelial cell continuum forming Bowman's space in focal segmental glomerulosclerosis.

In the glomerulus, Bowman's space is formed by a continuum of glomerular epithelial cells. In focal segmental glomerulosclerosis (FSGS), glomeruli show segmental scarring, a result of activated parietal epithelial cells (PECs) invading the glomerular tuft. The segmental scars interrupt the epithelial continuum. However, non-sclerotic segments seem to be preserved even in glomeruli with advanced lesions. We studied the histology of the segmental pattern in Munich Wistar Frömter rats, a model for secondary FSGS. Our results showed that matrix layers lined with PECs cover the sclerotic lesions. These PECs formed contacts with podocytes of the uninvolved tuft segments, restoring the epithelial continuum. Formed Bowman's spaces were still connected to the tubular system. In biopsies of patients with secondary FSGS, we also detected matrix layers formed by PECs, separating the uninvolved from the sclerotic glomerular segments. PECs have a major role in the formation of glomerulosclerosis; we show here that in FSGS they also restore the glomerular epithelial cell continuum that surrounds Bowman's space. This process may be beneficial and indispensable for glomerular filtration in the uninvolved segments of sclerotic glomeruli.

Artificial Intelligence in Pathology: From Prototype to Product.

Modern image analysis techniques based on artificial intelligence (AI) have great potential to improve the quality and efficiency of diagnostic procedures in pathology and to detect novel biomarkers. Despite thousands of published research papers on applications of AI in pathology, hardly any research implementations have matured into commercial products for routine use. Bringing an AI solution for pathology to market poses significant technological, business, and regulatory challenges. In this paper, we provide a comprehensive overview and advice on how to meet these challenges. We outline how research prototypes can be turned into a product-ready state and integrated into the IT infrastructure of clinical laboratories. We also discuss business models for profitable AI solutions and reimbursement options for computer assistance in pathology. Moreover, we explain how to obtain regulatory approval so that AI solutions can be launched as in vitro diagnostic medical devices. Thus, this paper offers computer scientists, software companies, and pathologists a road map for transforming prototypes of AI solutions into commercial products.

Optimized tumour infiltrating lymphocyte assessment for triple negative breast cancer prognostics.

The tumour microenvironment has been shown to be a valuable source of prognostic information for different cancer types. This holds in particular for triple negative breast cancer (TNBC), a breast cancer subtype for which currently no prognostic biomarkers are established. Although different methods to assess tumour infiltrating lymphocytes (TILs) have been published, it remains unclear which method (marker, region) yields the most optimal prognostic information. In addition, to date, no objective TILs assessment methods are available. For this proof of concept study, a subset of our previously described TNBC cohort (n = 94) was stained for CD3, CD8 and FOXP3 using multiplex immunohistochemistry and subsequently imaged by a multispectral imaging system. Advanced whole-slide image analysis algorithms, including convolutional neural networks (CNN) were used to register unmixed multispectral images and corresponding H&E sections, to segment the different tissue compartments (tumour, stroma) and to detect all individual positive lymphocytes. Densities of positive lymphocytes were analysed in different regions within the tumour and its neighbouring environment and correlated to relapse free survival (RFS) and overall survival (OS). We found that for all TILs markers the presence of a high density of positive cells correlated with an improved survival. None of the TILs markers was superior to the others. The results of TILs assessment in the various regions did not show marked differences between each other. The negative correlation between TILs and survival in our cohort are in line with previous studies. Our results provide directions for optimizing TILs assessment methodology.

ANHIR: Automatic Non-Rigid Histological Image Registration Challenge.

Automatic Non-rigid Histological Image Registration (ANHIR) challenge was organized to compare the performance of image registration algorithms on several kinds of microscopy histology images in a fair and independent manner. We have assembled 8 datasets, containing 355 images with 18 different stains, resulting in 481 image pairs to be registered. Registration accuracy was evaluated using manually placed landmarks. In total, 256 teams registered for the challenge, 10 submitted the results, and 6 participated in the workshop. Here, we present the results of 7 well-performing methods from the challenge together with 6 well-known existing methods. The best methods used coarse but robust initial alignment, followed by non-rigid registration, used multiresolution, and were carefully tuned for the data at hand. They outperformed off-the-shelf methods, mostly by being more robust. The best methods could successfully register over 98% of all landmarks and their mean landmark registration accuracy (TRE) was 0.44% of the image diagonal. The challenge remains open to submissions and all images are available for download.

Epithelium segmentation using deep learning in H&E-stained prostate specimens with immunohistochemistry as reference standard.

Given the importance of gland morphology in grading prostate cancer (PCa), automatically differentiating between epithelium and other tissues is an important prerequisite for the development of automated methods for detecting PCa. We propose a new deep learning method to segment epithelial tissue in digitised hematoxylin and eosin (H&E) stained prostatectomy slides using immunohistochemistry (IHC) as reference standard. We used IHC to create a precise and objective ground truth compared to manual outlining on H&E slides, especially in areas with high-grade PCa. 102 tissue sections were stained with H&E and subsequently restained with P63 and CK8/18 IHC markers to highlight epithelial structures. Afterwards each pair was co-registered. First, we trained a U-Net to segment epithelial structures in IHC using a subset of the IHC slides that were preprocessed with color deconvolution. Second, this network was applied to the remaining slides to create the reference standard used to train a second U-Net on H&E. Our system accurately segmented both intact glands and individual tumour epithelial cells. The generalisation capacity of our system is shown using an independent external dataset from a different centre. We envision this segmentation as the first part of a fully automated prostate cancer grading pipeline.

Deep learning nuclei detection: A simple approach can deliver state-of-the-art results.

BACKGROUND: Deep convolutional neural networks have become a widespread tool for the detection of nuclei in histopathology images. Many implementations share a basic approach that includes generation of an intermediate map indicating the presence of a nucleus center, which we refer to as PMap. Nevertheless, these implementations often still differ in several parameters, resulting in different detection qualities. METHODS: We identified several essential parameters and configured the basic PMap approach using combinations of them. We thoroughly evaluated and compared various configurations on multiple datasets with respect to detection quality, efficiency and training effort. RESULTS: Post-processing of the PMap was found to have the largest impact on detection quality. Also, two different network architectures were identified that improve either detection quality or runtime performance. The best-performing configuration yields f1-measures of 0.816 on H&E stained images of colorectal adenocarcinomas and 0.819 on Ki-67 stained images of breast tumor tissue. On average, it was fully trained in less than 15,000 iterations and processed 4.15 megapixels per second at prediction time. CONCLUSIONS: The basic PMap approach is greatly affected by certain parameters. Our evaluation provides guidance on their impact and best settings. When configured properly, this simple and efficient approach can yield equal detection quality as more complex and time-consuming state-of-the-art approaches.

Data-Driven Discovery of Immune Contexture Biomarkers.

Background: Features characterizing the immune contexture (IC) in the tumor microenvironment can be prognostic and predictive biomarkers. Identifying novel biomarkers can be challenging due to complex interactions between immune and tumor cells and the abundance of possible features. Methods: We describe an approach for the data-driven identification of IC biomarkers. For this purpose, we provide mathematical definitions of different feature classes, based on cell densities, cell-to-cell distances, and spatial heterogeneity thereof. Candidate biomarkers are ranked according to their potential for the predictive stratification of patients. Results: We evaluated the approach on a dataset of colorectal cancer patients with variable amounts of microsatellite instability. The most promising features that can be explored as biomarkers were based on cell-to-cell distances and spatial heterogeneity. Both the tumor and non-tumor compartments yielded features that were potentially predictive for therapy response and point in direction of further exploration. Conclusion: The data-driven approach simplifies the identification of promising IC biomarker candidates. Researchers can take guidance from the described approach to accelerate their biomarker research.

Automated quantification of steatosis: agreement with stereological point counting.

BACKGROUND: Steatosis is routinely assessed histologically in clinical practice and research. Automated image analysis can reduce the effort of quantifying steatosis. Since reproducibility is essential for practical use, we have evaluated different analysis methods in terms of their agreement with stereological point counting (SPC) performed by a hepatologist. METHODS: The evaluation was based on a large and representative data set of 970 histological images from human patients with different liver diseases. Three of the evaluated methods were built on previously published approaches. One method incorporated a new approach to improve the robustness to image variability. RESULTS: The new method showed the strongest agreement with the expert. At 20× resolution, it reproduced steatosis area fractions with a mean absolute error of 0.011 for absent or mild steatosis and 0.036 for moderate or severe steatosis. At 10× resolution, it was more accurate than and twice as fast as all other methods at 20× resolution. When compared with SPC performed by two additional human observers, its error was substantially lower than one and only slightly above the other observer. CONCLUSIONS: The results suggest that the new method can be a suitable automated replacement for SPC. Before further improvements can be verified, it is necessary to thoroughly assess the variability of SPC between human observers.

Multiple sclerosis lesion segmentation using dictionary learning and sparse coding.

The segmentation of lesions in the brain during the development of Multiple Sclerosis is part of the diagnostic assessment for this disease and gives information on its current severity. This laborious process is still carried out in a manual or semiautomatic fashion by clinicians because published automatic approaches have not been universal enough to be widely employed in clinical practice. Thus Multiple Sclerosis lesion segmentation remains an open problem. In this paper we present a new unsupervised approach addressing this problem with dictionary learning and sparse coding methods. We show its general applicability to the problem of lesion segmentation by evaluating our approach on synthetic and clinical image data and comparing it to state-of-the-art methods. Furthermore the potential of using dictionary learning and sparse coding for such segmentation tasks is investigated and various possibilities for further experiments are discussed.