Digital pathology imaging as a novel platform for standardization and globalization of quantitative nephropathology

Time for a full digital approach in nephropathology: a systematic review of current artificial intelligence applications and future directions

INTRODUCTION

Artificial intelligence (AI) integration in nephropathology has been growing rapidly in recent years, facing several challenges including the wide range of histological techniques used, the low occurrence of certain diseases, and the need for data sharing. This narrative review retraces the history of AI in nephropathology and provides insights into potential future developments.

METHODS

Electronic searches in PubMed-MEDLINE and Embase were made to extract pertinent articles from the literature. Works about automated image analysis or the application of an AI algorithm on non-neoplastic kidney histological samples were included and analyzed to extract information such as publication year, AI task, and learning type. Prepublication servers and reviews were not included.

RESULTS

Seventy-six (76) original research articles were selected. Most of the studies were conducted in the United States in the last 7 years. To date, research has been mainly conducted on relatively easy tasks, like single-stain glomerular segmentation. However, there is a trend towards developing more complex tasks such as glomerular multi-stain classification.

CONCLUSION

Deep learning has been used to identify patterns in complex histopathology data and looks promising for the comprehensive assessment of renal biopsy, through the use of multiple stains and virtual staining techniques. Hybrid and collaborative learning approaches have also been explored to utilize large amounts of unlabeled data. A diverse team of experts, including nephropathologists, computer scientists, and clinicians, is crucial for the development of AI systems for nephropathology. Collaborative efforts among multidisciplinary experts result in clinically relevant and effective AI tools.

Automated evaluation with deep learning of total interstitial inflammation and peritubular capillaritis on kidney biopsies

BACKGROUND

Interstitial inflammation and peritubular capillaritis are observed in many diseases on native and transplant kidney biopsies. A precise and automated evaluation of these histological criteria could help stratify patients' kidney prognoses and facilitate therapeutic management.

METHODS

We used a convolutional neural network to evaluate those criteria on kidney biopsies. A total of 423 kidney samples from various diseases were included; 83 kidney samples were used for the neural network training, 106 for comparing manual annotations on limited areas to automated predictions, and 234 to compare automated and visual gradings.

RESULTS

The precision, recall and F-score for leukocyte detection were, respectively, 81%, 71% and 76%. Regarding peritubular capillaries detection the precision, recall and F-score were, respectively, 82%, 83% and 82%. There was a strong correlation between the predicted and observed grading of total inflammation, as for the grading of capillaritis (r = 0.89 and r = 0.82, respectively, all P < .0001). The areas under the receiver operating characteristics curves for the prediction of pathologists' Banff total inflammation (ti) and peritubular capillaritis (ptc) scores were respectively all above 0.94 and 0.86. The kappa coefficients between the visual and the neural networks' scores were respectively 0.74, 0.78 and 0.68 for ti ≥1, ti ≥2 and ti ≥3, and 0.62, 0.64 and 0.79 for ptc ≥1, ptc ≥2 and ptc ≥3. In a subgroup of patients with immunoglobulin A nephropathy, the inflammation severity was highly correlated to kidney function at biopsy on univariate and multivariate analyses.

CONCLUSION

We developed a tool using deep learning that scores the total inflammation and capillaritis, demonstrating the potential of artificial intelligence in kidney pathology.

Kidney fibrosis: Emerging diagnostic and therapeutic strategies

An increasing number of patients worldwide suffers from chronic kidney disease (CKD). CKD is accompanied by kidney fibrosis, which affects all compartments of the kidney, i.e., the glomeruli, tubulointerstitium, and vasculature. Fibrosis is the best predictor of progression of kidney diseases. Currently, there is no specific anti-fibrotic therapy for kidney patients and invasive renal biopsy remains the only option for specific detection and quantification of kidney fibrosis. Here we review emerging diagnostic approaches and potential therapeutic options for fibrosis. We discuss how translational research could help to establish fibrosis-specific endpoints for clinical trials, leading to improved patient stratification and potentially companion diagnostics, and facilitating and optimizing development of novel anti-fibrotic therapies for kidney patients.

Assessment of Fibrinogen-like 2 (FGL2) in Human Chronic Kidney Disease through Transcriptomics Data Analysis

Fibrinogen-like 2 (FGL2) was recently found to be associated with fibrosis in a mouse model of kidney damage and was proposed as a potential therapeutic target in chronic kidney disease (CKD). We assessed the association of renal FGL2 mRNA expression with the disease outcome in two independent CKD cohorts (NEPTUNE and Innsbruck CKD cohort) using Kaplan Meier survival analysis. The regulation of FGL2 in kidney biopsies of CKD patients as compared to healthy controls was further assessed in 13 human CKD transcriptomics datasets. The FGL2 protein expression in human renal tissue sections was determined via immunohistochemistry. The regulators of FGL2 mRNA expression in renal tissue were identified in the co-expression and upstream regulator analysis of FGL2-positive renal cells via the use of single-cell RNA sequencing data from the kidney precision medicine project (KPMP). Higher renal FGL2 mRNA expression was positively associated with kidney fibrosis and negatively associated with eGFR. Renal FGL2 mRNA expression was upregulated in CKD as compared with healthy controls and associated with CKD progression in the Innsbruck CKD cohort (p-value = 0.0036) and NEPTUNE cohort (p-value = 0.0048). The highest abundance of FGL2 protein in renal tissue was detected in the thick ascending limb of the loop of Henle and macula densa, proximal tubular cells, as well as in glomerular endothelial cells. The upstream regulator analysis identified TNF, IL1B, IFNG, NFKB1, and SP1 as factors potentially inducing FGL2-co-expressed genes, whereas factors counterbalancing FGL2-co-expressed genes included GLI1, HNF1B, or PPARGC1A. In conclusion, renal FGL2 mRNA expression is elevated in human CKD, and higher FGL2 levels are associated with fibrosis and worse outcomes.

Analysis and Estimation of Pathological Data and Findings with Deep Learning Methods

As in human diseases, rapid diagnosis of animal diseases is of great importance. In order for the disease treatments to be carried out properly, the diagnosis must be of high accuracy, as well as the rapid diagnosis. In this study, the disease types in the data set consisting of the data examined between the years 2000-2020 belonging to the Department of Pathology of the Faculty of Veterinary Medicine of Burdur Mehmet Akif Ersoy University were estimated by using the decision tree classification model and the KNN classification model. Categories such as age, type, city, and gender in the data set were analyzed in graphics. For the estimation and analysis processes to give accurate results, the data set was corrected by going through some pre-processes and the missing data in the data set was completed. It is thought that the results obtained from the estimation and analysis will allow rapid and accurate diagnosis in animal disease diagnoses.

Boundary-aware glomerulus segmentation: Toward one-to-many stain generalization

The growing availability of scanned whole-slide images (WSIs) has allowed nephropathology to open new possibilities for medical decision-making over high-resolution images. Diagnosis of renal WSIs includes locating and identifying specific structures in the tissue. Considering the glomerulus as one of the first structures analyzed by pathologists, we propose here a novel convolutional neural network for glomerulus segmentation. Our end-to-end network, named DS-FNet, combines the strengths of semantic segmentation and semantic boundary detection networks via an attention-aware mechanism. Although we trained the proposed network on periodic acid-Schiff (PAS)-stained WSIs, we found that our network was capable to segment glomeruli on WSIs stained with different techniques, such as periodic acid-methenamine silver (PAMS), hematoxylin-eosin (HE), and Masson trichrome (TRI). To assess the performance of the proposed method, we used three public data sets: HuBMAP (available in a Kaggle competition), a subset of the NEPTUNE data set, and a novel challenging data set, called WSI_Fiocruz. We compared the DS-FNet with six other deep learning networks: original U-Net, our attention version of U-Net called AU-Net, U-Net++, U-Net3Plus, ResU-Net, and DeepLabV3+. Results showed that DS-FNet achieved equivalent or superior results on all data sets: On the HuBMAP data set, it reached a dice score (DSC) of 95.05%, very close to the first place (95.15%); on the NEPTUNE and WSI_Fiocruz data sets, DS-FNet obtained the highest average DSC, whether on PAS-stained images or images stained with other techniques. To the best we know, this is the first work to show consistently high performance in a one-to-many-stain glomerulus segmentation following a thorough protocol on data sets from different medical labs.

Quantification of Glomerular Structural Lesions: Associations With Clinical Outcomes and Transcriptomic Profiles in Nephrotic Syndrome

RATIONALE & OBJECTIVE

The current classification system for focal segmental glomerulosclerosis (FSGS) and minimal change disease (MCD) does not fully capture the complex structural changes in kidney biopsies nor the clinical and molecular heterogeneity of these diseases.

STUDY DESIGN

Prospective observational cohort study.

SETTING & PARTICIPANTS

221 MCD and FSGS patients enrolled in the Nephrotic Syndrome Study Network (NEPTUNE).

EXPOSURE

The NEPTUNE Digital Pathology Scoring System (NDPSS) was applied to generate scores for 37 glomerular descriptors.

OUTCOME

Time from biopsy to complete proteinuria remission, time from biopsy to kidney disease progression (40% estimated glomerular filtration rate [eGFR] decline or kidney failure), and eGFR over time.

ANALYTICAL APPROACH

Cluster analysis was used to group patients with similar morphologic characteristics. Glomerular descriptors and patient clusters were assessed for associations with outcomes using adjusted Cox models and linear mixed models. Messenger RNA from glomerular tissue was used to assess differentially expressed genes between clusters and identify genes associated with individual descriptors driving cluster membership.

RESULTS

Three clusters were identified: X (n = 56), Y (n = 68), and Z (n = 97). Clusters Y and Z had higher probabilities of proteinuria remission (HRs of 1.95 [95% CI, 0.99-3.85] and 3.29 [95% CI, 1.52-7.13], respectively), lower hazards of disease progression (HRs of 0.22 [95% CI, 0.08-0.57] and 0.11 [95% CI, 0.03-0.45], respectively), and lower loss of eGFR over time compared with X. Cluster X had 1,920 genes that were differentially expressed compared with Y+Z; these reflected activation of pathways of immune response and inflammation. Six descriptors driving the clusters individually correlated with clinical outcomes and gene expression.

LIMITATIONS

Low prevalence of some descriptors and biopsy at a single time point.

CONCLUSIONS

The NDPSS allows for categorization of FSGS/MCD patients into clinically and biologically relevant subgroups, and uncovers histologic parameters associated with clinical outcomes and molecular signatures not included in current classification systems.

Kidney Biopsy Features Most Predictive of Clinical Outcomes in the Spectrum of Minimal Change Disease and Focal Segmental Glomerulosclerosis

BACKGROUND

Heterogeneity in disease course and treatment response among patients with MCD/FSGS necessitates a granular evaluation of kidney tissue features. This study aimed to identify histologic and ultrastructural descriptors of structural changes most predictive of clinical outcomes in the Nephrotic Syndrome Study Network (NEPTUNE).

METHODS

Forty-eight histologic (37 glomerular, 9 tubulointerstitial, 2 vascular) and 20 ultrastructural descriptors were quantified by applying the NEPTUNE Digital Pathology Scoring System to NEPTUNE kidney biopsies. Outcomes included time from biopsy to disease progression, first complete remission of proteinuria, and treatment response. Relative importance of pathology and clinical predictors was obtained from random forest models, and predictive discrimination was assessed.

RESULTS

Among 224 participants (34% Black, 24% Hispanic), model performance was excellent, with predictive discrimination of 0.9 for disease progression, 0.85 for complete remission, and 0.81 for treatment response. The most predictive descriptors of outcomes included both conventional-e.g., global sclerosis or segmental sclerosis and interstitial fibrosis/tubular atrophy-and novel features, including adhesion, interstitial foam cells, deflation, periglomerular fibrosis, mononuclear white blood cells, endothelial cell abnormalities, microvillous transformation, and acute tubular injury.

CONCLUSIONS

The most predictive descriptors of clinical outcomes among MCD/FSGS patients reflected structural changes in multiple renal compartments. Reporting these descriptors should be standardized to guide prognostication of proteinuric glomerular diseases.

Unsupervised machine learning for identifying important visual features through bag-of-words using histopathology data from chronic kidney disease

Pathologists use visual classification to assess patient kidney biopsy samples when diagnosing the underlying cause of kidney disease. However, the assessment is qualitative, or semi-quantitative at best, and reproducibility is challenging. To discover previously unknown features which predict patient outcomes and overcome substantial interobserver variability, we developed an unsupervised bag-of-words model. Our study applied to the C-PROBE cohort of patients with chronic kidney disease (CKD). 107,471 histopathology images were obtained from 161 biopsy cores and identified important morphological features in biopsy tissue that are highly predictive of the presence of CKD both at the time of biopsy and in one year. To evaluate the performance of our model, we estimated the AUC and its 95% confidence interval. We show that this method is reliable and reproducible and can achieve 0.93 AUC at predicting glomerular filtration rate at the time of biopsy as well as predicting a loss of function at one year. Additionally, with this method, we ranked the identified morphological features according to their importance as diagnostic markers for chronic kidney disease. In this study, we have demonstrated the feasibility of using an unsupervised machine learning method without human input in order to predict the level of kidney function in CKD. The results from our study indicate that the visual dictionary, or visual image pattern, obtained from unsupervised machine learning can predict outcomes using machine-derived values that correspond to both known and unknown clinically relevant features.

Impact of Consensus Definitions on Identification of Glomerular Lesions by Light and Electron Microscopy

Introduction

In 2020, a working group of 13 renal pathologists published consensus definitions for 47 individual glomerular lesions found on light microscopy (LM) and 47 glomerular lesions and 9 normal structures found on electron microscopy (EM).

Methods

To test the impact of these definitions on identification of these lesions and structures, 2 surveys were circulated to all members of the Renal Pathology Society (RPS), each having 32 images (19 LM, 13 EM) and accompanying questions with 5 multiple-choice answers, one being the consensus choice of the working group. The first survey (survey 1 [S1]), answered by 297 RPS members, was sent in September 2020, before publication of the consensus definitions. The second (survey 2 [S2]), with images of the same lesions and structures (but not the same images) and the same questions and multiple choices in different order, was sent in April 2020, 5 months after the publication of the definitions.

Results

S2 was taken by 181 RPS members; 64% also took S1 and 61% reported having read the definitions paper (def. paper). Mean agreement with the consensus answers increased modestly between the 2 surveys (65.2% vs. 72.0%, P = 0.097); the increase was greater and significant when only respondents to S2 who read the def. paper were considered (65.2% vs. 74.8%, P = 0.026). Furthermore, in S2 agreement with consensus answers was greater among respondents who read this paper versus those who did not (66.9% vs. 74.8%, P < 0.0001).

Conclusions

Publication of the consensus definitions modestly improved interobserver agreement in identification of glomerular lesions.

Semi-Supervised Segmentation of Renal Pathology: An Alternative to Manual Segmentation and Input to Deep Learning Training

Kidney biopsy interpretation is the gold standard for the diagnosis and prognosis for kidney disease. Pathognomonic diagnosis hinges on the correct assessment of different structures within a biopsy that is manually visualized and interpreted by a renal pathologist. This laborious undertaking has spurred attempts to automate the process, offloading the consumption of temporal resources. Segmentation of kidney structures, specifically, the glomeruli, tubules, and interstitium, is a precursory step for disease classification problems. Translating renal disease decision making into a deep learning model for diagnostic and prognostic classification also relies on adequate segmentation of structures within the kidney biopsy. This study showcases a semi-automated segmentation technique where the user defines starting points for glomeruli in kidney biopsy images of both healthy normal and diabetic kidney disease stained with Nile Red that are subsequently partitioned into four areas: background, glomeruli, tubules and interstitium. Five of 30 biopsies that were segmented using the semi-automated method were randomly selected and the regions of interest were compared to the manual segmentation of the same images. Dice Similarity Coefficients (DSC) between the methods showed excellent agreement; Healthy (glomeruli: 0.92, tubules: 0.86, intersititium: 0.78) and diabetic nephropathy: (glomeruli: 0.94, tubules: 0.80, intersititium: 0.80). To our knowledge this is the first semi-automated segmentation algorithm performed with human renal biopsies stained with Nile Red. Utility of this methodology includes further image processing within structures across disease states based on biological morphological structures. It can also be used as input into a deep learning network to train semantic segmentation and input into a deep learning algorithm for classification of disease states.

Modified Look-Locker Inverse-Recovery (MOLLI) Sequence of Quantitative Imaging in Dirty Magnetic Resonance Longitudinal Relaxation Time Diagnostic Value of GE Combined with Longitudinal Relaxation Time Quantitative Imaging for Myocardial Amyloidosis

The pathological changes of myocarditis include degeneration and necrosis of myocardial cells and infiltration of inflammatory cells in the myocardial interstitium, accompanied by obvious myocardial fibrosis. Myocardial fibrosis is a determinant of ventricular remodeling and an important indicator of the classification of clinical risk factors and has an important value in evaluating the prognosis of heart disease. Cardiac magnetic resonance (CMR) is the "gold standard" for evaluating the shape and function of the heart, and it can show the characteristic pathological changes of myocardial tissue. The traditional gadolinium imaging agent delays the enhanced sequence images to visually show the extent of the affected myocardial fibrosis, but it cannot effectively identify small focal fibrosis or widespread diffuse fibrosis. The CMR longitudinal relaxation time quantitative technique can directly measure the relaxation time (T1) determined by the myocardial tissue and does not depend on the signal strength of the reference tissue and can quantitatively analyze the affected myocardium. In this study, the initial and enhanced quantitative imaging techniques of CMR were used to measure the magnetic value of the myocardium in patients with myocarditis, to explore the diagnostic value of myocardial fibrosis, and to analyze the correlation between cardiac fibrosis and cardiac function.

APOL1 genotype-associated morphologic changes among patients with focal segmental glomerulosclerosis

BACKGROUND

The G1 and G2 alleles of apolipoprotein L1 (APOL1) are common in the Black population and associated with increased risk of focal segmental glomerulosclerosis (FSGS). The molecular mechanisms linking APOL1 risk variants with FSGS are not clearly understood, and APOL1's natural absence in laboratory animals makes studying its pathobiology challenging.

METHODS

In a cohort of 90 Black patients with either FSGS or minimal change disease (MCD) enrolled in the Nephrotic Syndrome Study Network (58% pediatric onset), we used kidney biopsy traits as an intermediate outcome to help illuminate tissue-based consequences of APOL1 risk variants and expression. We tested associations between APOL1 risk alleles or glomerular APOL1 mRNA expression and 83 light- or electron-microscopy traits measuring structural and cellular kidney changes.

RESULTS

Under both recessive and dominant models in the FSGS patient subgroup (61%), APOL1 risk variants were significantly correlated (defined as FDR <0.1) with decreased global mesangial hypercellularity, decreased condensation of cytoskeleton, and increased tubular microcysts. No significant correlations were detected in MCD cohort. Independent of risk alleles, glomerular APOL1 expression in FSGS patients was not correlated with morphologic features.

CONCLUSIONS

While APOL1-associated FSGS is associated with two risk alleles, both one and two risk alleles are associated with cellular/tissue changes in this study of FSGS patients. Our lack of discovery of a large group of tissue differences in FSGS and no significant difference in MCD may be due to the lack of power but also supports investigating whether machine learning methods may more sensitively detect APOL1-associated changes.

AI applications in renal pathology

The explosive growth of artificial intelligence (AI) technologies, especially deep learning methods, has been translated at revolutionary speed to efforts in AI-assisted healthcare. New applications of AI to renal pathology have recently become available, driven by the successful AI deployments in digital pathology. However, synergetic developments of renal pathology and AI require close interdisciplinary collaborations between computer scientists and renal pathologists. Computer scientists should understand that not every AI innovation is translatable to renal pathology, while renal pathologists should capture high-level principles of the relevant AI technologies. Herein, we provide an integrated review on current and possible future applications in AI-assisted renal pathology, by including perspectives from computer scientists and renal pathologists. First, the standard stages, from data collection to analysis, in full-stack AI-assisted renal pathology studies are reviewed. Second, representative renal pathology-optimized AI techniques are introduced. Last, we review current clinical AI applications, as well as promising future applications with the recent advances in AI.

Dissecting the Business Case for Adoption and Implementation of Digital Pathology: A White Paper from the Digital Pathology Association

We believe the switch to a digital pathology (DP) workflow is imminent and it is essential to understand the economic implications of conversion. Many aspects of the adoption of DP will be disruptive and have a direct financial impact, both in short term costs, such as investment in equipment and personnel, and long term revenue potential, such as improved productivity and novel tests. The focus of this whitepaper is to educate pathologists, laboratorians and other stakeholders about the business and monetary considerations of converting to a digital pathology workflow. The components of a DP business plan will be thoroughly summarized, and guidance will be provided on how to build a case for adoption and implementation as well as a roadmap for transitioning from an analog to a digital pathology workflow in various laboratory settings. It is important to clarify that this publication is not intended to list prices although some financials will be mentioned as examples. The authors encourage readers who are evaluating conversion to a DP workflow to use this paper as a foundational guide for conducting a thorough and complete assessment while incorporating in current market pricing. Contributors to this paper analyzed peer-reviewed literature and data collected from various institutions, some of which are mentioned. Digital pathology will change the way we practice through facilitating patient access to expert pathology services and enabling image analysis tools and assays to aid in diagnosis, prognosis, risk stratification and therapeutic selection. Together, they will result in the delivery of valuable information from which to make better decisions and improve the health of patients.

Towards harmony in defining and reporting glomerular diseases on kidney biopsy

PURPOSE OF REVIEW

To review recent efforts to develop uniformity and precision in defining individual glomerular histologic and ultrastructural lesions and proposals for developing greater uniformity in reporting of glomerular diseases.

RECENT FINDINGS

Over the past 2 decades, scoring systems for multiple glomerular diseases have emerged, as have several consortia for the study of glomerular diseases. However, one important limitation faced by renal pathologists and nephrologists has been a lack of uniformity and precision in defining the morphologic lesions seen by light and electron microscopy on which the scoring systems are based. In response to this, the Renal Pathology Society organized a working group that over 4 years arrived at consensus definitions for many such lesions. These definitions can be applied within the context of scoring systems for different glomerular diseases, and recently proposed reporting systems based on pathogenic categories and for defining the overall severity of chronic changes.

SUMMARY

From extensive discussions a panel of 13 renal pathologists reached consensus in defining 47 individual glomerular lesions seen on light microscopy and 56 glomerular lesions and key normal structures seen by electron microscopy. Validation of the impact of these consensus definitions on interobserver agreement in lesion identification is currently underway.

Improving data quality in observational research studies: Report of the Cure Glomerulonephropathy (CureGN) network

Background

High data quality is of crucial importance to the integrity of research projects. In the conduct of multi-center observational cohort studies with increasing types and quantities of data, maintaining data quality is challenging, with few published guidelines.

Methods

The Cure Glomerulonephropathy (CureGN) Network has established numerous quality control procedures to manage the 70 participating sites in the United States, Canada, and Europe. This effort is supported and guided by the activities of several committees, including Data Quality, Recruitment and Retention, and Central Review, that work in tandem with the Data Coordinating Center to monitor the study. We have implemented coordinator training and feedback channels, data queries of questionable or missing data, and developed performance metrics for recruitment, retention, visit completion, data entry, recording of patient-reported outcomes, collection, shipping and accessing of biological samples and pathology materials, and processing, cataloging and accessing genetic data and materials.

Results

We describe the development of data queries and site Report Cards, and their use in monitoring and encouraging excellence in site performance. We demonstrate improvements in data quality and completeness over 4 years after implementing these activities. We describe quality initiatives addressing specific challenges in collecting and cataloging whole slide images and other kidney pathology data, and novel methods of data quality assessment.

Conclusions

This paper reports the CureGN experience in optimizing data quality and underscores the importance of general and study-specific data quality initiatives to maintain excellence in the research measures of a multi-center observational study.

Artificial intelligence driven next-generation renal histomorphometry

PURPOSE OF REVIEW

Successful integration of artificial intelligence into extant clinical workflows is contingent upon a number of factors including clinician comprehension and interpretation of computer vision. This article discusses how image analysis and machine learning have enabled comprehensive characterization of kidney morphology for development of automated diagnostic and prognostic renal pathology applications.

RECENT FINDINGS

The primordial digital pathology informatics work employed classical image analysis and machine learning to prognosticate renal disease. Although this classical approach demonstrated tremendous potential, subsequent advancements in hardware technology rendered artificial neural networks '(ANNs) the method of choice for machine vision in computational pathology'. Offering rapid and reproducible detection, characterization and classification of kidney morphology, ANNs have facilitated the development of diagnostic and prognostic applications. In addition, modern machine learning with ANNs has revealed novel biomarkers in kidney disease, demonstrating the potential for machine vision to elucidate novel pathologic mechanisms beyond extant clinical knowledge.

SUMMARY

Despite the revolutionary developments potentiated by modern machine learning, several challenges remain, including data quality control and curation, image annotation and ontology, integration of multimodal data and interpretation of machine vision or 'opening the black box'. Resolution of these challenges will not only revolutionize diagnostic pathology but also pave the way for precision medicine and integration of artificial intelligence in the process of care.

Assessment of a computerized quantitative quality control tool for whole slide images of kidney biopsies

Inconsistencies in the preparation of histology slides and whole-slide images (WSIs) may lead to challenges with subsequent image analysis and machine learning approaches for interrogating the WSI. These variabilities are especially pronounced in multicenter cohorts, where batch effects (i.e. systematic technical artifacts unrelated to biological variability) may introduce biases to machine learning algorithms. To date, manual quality control (QC) has been the de facto standard for dataset curation, but remains highly subjective and is too laborious in light of the increasing scale of tissue slide digitization efforts. This study aimed to evaluate a computer-aided QC pipeline for facilitating a reproducible QC process of WSI datasets. An open source tool, HistoQC, was employed to identify image artifacts and compute quantitative metrics describing visual attributes of WSIs to the Nephrotic Syndrome Study Network (NEPTUNE) digital pathology repository. A comparison in inter-reader concordance between HistoQC aided and unaided curation was performed to quantify improvements in curation reproducibility. HistoQC metrics were additionally employed to quantify the presence of batch effects within NEPTUNE WSIs. Of the 1814 WSIs (458 H&E, 470 PAS, 438 silver, 448 trichrome) from n = 512 cases considered in this study, approximately 9% (163) were identified as unsuitable for subsequent computational analysis. The concordance in the identification of these WSIs among computational pathologists rose from moderate (Gwet's AC1 range 0.43 to 0.59 across stains) to excellent (Gwet's AC1 range 0.79 to 0.93 across stains) agreement when aided by HistoQC. Furthermore, statistically significant batch effects (p < 0.001) in the NEPTUNE WSI dataset were discovered. Taken together, our findings strongly suggest that quantitative QC is a necessary step in the curation of digital pathology cohorts. © 2020 The Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Digital pathology and computational image analysis in nephropathology

The emergence of digital pathology - an image-based environment for the acquisition, management and interpretation of pathology information supported by computational techniques for data extraction and analysis - is changing the pathology ecosystem. In particular, by virtue of our new-found ability to generate and curate digital libraries, the field of machine vision can now be effectively applied to histopathological subject matter by individuals who do not have deep expertise in machine vision techniques. Although these novel approaches have already advanced the detection, classification, and prognostication of diseases in the fields of radiology and oncology, renal pathology is just entering the digital era, with the establishment of consortia and digital pathology repositories for the collection, analysis and integration of pathology data with other domains. The development of machine-learning approaches for the extraction of information from image data, allows for tissue interrogation in a way that was not previously possible. The application of these novel tools are placing pathology centre stage in the process of defining new, integrated, biologically and clinically homogeneous disease categories, to identify patients at risk of progression, and shifting current paradigms for the treatment and prevention of kidney diseases.

Consensus definitions for glomerular lesions by light and electron microscopy: recommendations from a working group of the Renal Pathology Society

Over the past 2 decades, scoring systems for multiple glomerular diseases have emerged, as have consortia of pathologists and nephrologists for the study of glomerular diseases, including correlation of pathologic findings with clinical features and outcomes. However, one important limitation faced by members of these consortia and other renal pathologists and nephrologists in both investigative work and routine practice remains a lack of uniformity and precision in clearly defining the morphologic lesions on which the scoring systems are based. In response to this issue, the Renal Pathology Society organized a working group to identify the most frequently identified glomerular lesions observed by light microscopy and electron microscopy, review the literature to capture the published definitions most often used for each, and determine consensus terms and definitions for each lesion in a series of online and in-person meetings. The defined lesions or abnormal findings are not specific for any individual disease or subset of diseases, but rather can be applied across the full spectrum of glomerular diseases and within the context of the different scoring systems used for evaluating and reporting these diseases. In addition to facilitating glomerular disease research, standardized terms and definitions should help harmonize reporting of medical kidney diseases worldwide and lead to more-precise diagnoses and improved patient care.

Understanding the Link between Neighborhoods and Kidney Disease

Neighborhoods are where we live, learn, work, pray, and play. Growing evidence indicates that neighborhoods are an important determinant of health. The built features of our neighborhoods, such as the ways in which the streets are designed and connected and the availability of green spaces and transit stops, as well as the social features, such as the trust among neighbors and the perceptions of safety, may influence health through multiple pathways, such as access to important resources, psychosocial stress, and health behaviors. In particular, the extant literature consistently documents an association between neighborhood features and renal-associated conditions, such as cardiovascular disease, hypertension, diabetes, and obesity. There is also some evidence suggesting an association between neighborhood poverty and ESKD. The link between neighborhood and earlier stages of CKD, however, has been less clear, with most studies documenting no association. It may be that the neighborhood measures used in previous studies do not capture features of the neighborhood important for earlier stages of disease development and progression. It may also be that our current biomarkers (e.g., eGFR) and urine protein are not able to pick up very early forms of renal damage because of the kidney's overall high reserve capacity. This paper critically reviews the state of the literature on neighborhood and renal disease, with recommendations for neighborhood measures in future research. Neighborhoods are designed, built, and informed by policy, and thus, they are amenable to intervention, making them a potentially powerful way to improve renal health and reduce health inequalities at the population level.

Biobanking for glomerular diseases: a study design and protocol for KOrea Renal biobank NEtwoRk System TOward NExt-generation analysis (KORNERSTONE)

Backgrounds

Glomerular diseases, a set of debilitating and complex disease entities, are related to mortality and morbidity. To gain insight into pathophysiology and novel treatment targets of glomerular disease, various types of biospecimens linked to deep clinical phenotyping including clinical information, digital pathology, and well-defined outcomes are required. We provide the rationale and design of the KOrea Renal biobank NEtwoRk System TOward Next-generation analysis (KORNERSTONE).

Methods

The KORNERSTONE, which has been initiated by Korea Centres for Disease Control and Prevention, is designed as a multi-centre, prospective cohort study and biobank for glomerular diseases. Clinical data, questionnaires will be collected at the time of kidney biopsy and subsequently every 1 year after kidney biopsy. All of the clinical data will be extracted from the electrical health record and automatically uploaded to the web-based database. High-quality digital pathologies are obtained and connected in the database. Various types of biospecimens are collected at baseline and during follow-up: serum, urine, buffy coat, stool, glomerular complementary DNA (cDNA), tubulointerstitial cDNA. All data and biospecimens are processed and stored in a standardised manner. The primary outcomes are mortality and end-stage renal disease. The secondary outcomes will be deterioration renal function, remission of proteinuria, cardiovascular events and quality of life.

Discussion

Ethical approval has been obtained from the institutional review board of each participating centre and ethics oversight committee. The KORNERSTONE is designed to deliver pioneer insights into glomerular diseases. The study design allows comprehensive, integrated and high-quality data collection on baseline laboratory findings, clinical outcomes including administrative data and digital pathologic images. This may provide various biospecimens and information to many researchers, establish the rationale for future more individualised treatment strategies for glomerular diseases.

Trial registration

NCT03929887.

Development and evaluation of deep learning-based segmentation of histologic structures in the kidney cortex with multiple histologic stains

The application of deep learning for automated segmentation (delineation of boundaries) of histologic primitives (structures) from whole slide images can facilitate the establishment of novel protocols for kidney biopsy assessment. Here, we developed and validated deep learning networks for the segmentation of histologic structures on kidney biopsies and nephrectomies. For development, we examined 125 biopsies for Minimal Change Disease collected across 29 NEPTUNE enrolling centers along with 459 whole slide images stained with Hematoxylin & Eosin (125), Periodic Acid Schiff (125), Silver (102), and Trichrome (107) divided into training, validation and testing sets (ratio 6:1:3). Histologic structures were manually segmented (30048 total annotations) by five nephropathologists. Twenty deep learning models were trained with optimal digital magnification across the structures and stains. Periodic Acid Schiff-stained whole slide images yielded the best concordance between pathologists and deep learning segmentation across all structures (F-scores: 0.93 for glomerular tufts, 0.94 for glomerular tuft plus Bowman’s capsule, 0.91 for proximal tubules, 0.93 for distal tubular segments, 0.81 for peritubular capillaries, and 0.85 for arteries and afferent arterioles). Optimal digital magnifications were 5X for glomerular tuft/tuft plus Bowman’s capsule, 10X for proximal/distal tubule, arteries and afferent arterioles, and 40X for peritubular capillaries. Silver stained whole slide images yielded the worst deep learning performance. Thus, this largest study to date adapted deep learning for the segmentation of kidney histologic structures across multiple stains and pathology laboratories. All data used for training and testing and a detailed online tutorial will be publicly available.

Data for glomeruli characterization in histopathological images

The data presented in this article is part of the whole slide imaging (WSI) datasets generated in European project AIDPATH This data is also related to the research paper entitle "Glomerulosclerosis Identification in Whole Slide Images using Semantic Segmentation", published in Computer Methods and Programs in Biomedicine Journal [1]. In that article, different methods based on deep learning for glomeruli segmentation and their classification into normal and sclerotic glomerulous are presented and discussed. The raw data used is described and provided here. In addition, the detected glomeruli are also provided as individual image files. These data will encourage research on artificial intelligence (AI) methods, create and compare fresh algorithms, and measure their usability in quantitative nephropathology.

Ultrastructural Characterization of Proteinuric Patients Predicts Clinical Outcomes

BACKGROUND

The analysis and reporting of glomerular features ascertained by electron microscopy are limited to few parameters with minimal predictive value, despite some contributions to disease diagnoses.

METHODS

We investigated the prognostic value of 12 electron microscopy histologic and ultrastructural changes (descriptors) from the Nephrotic Syndrome Study Network (NEPTUNE) Digital Pathology Scoring System. Study pathologists scored 12 descriptors in NEPTUNE renal biopsies from 242 patients with minimal change disease or FSGS, with duplicate readings to evaluate reproducibility. We performed consensus clustering of patients to identify unique electron microscopy profiles. For both individual descriptors and clusters, we used Cox regression models to assess associations with time from biopsy to proteinuria remission and time to a composite progression outcome (≥40% decline in eGFR, with eGFR<60 ml/min per 1.73 m², or ESKD), and linear mixed models for longitudinal eGFR measures.

RESULTS

Intrarater and interrater reproducibility was >0.60 for 12 out of 12 and seven out of 12 descriptors, respectively. Individual podocyte descriptors such as effacement and microvillous transformation were associated with complete remission, whereas endothelial cell and glomerular basement membrane abnormalities were associated with progression. We identified six descriptor-based clusters with distinct electron microscopy profiles and clinical outcomes. Patients in a cluster with more prominent foot process effacement and microvillous transformation had the highest rates of complete proteinuria remission, whereas patients in clusters with extensive loss of primary processes and endothelial cell damage had the highest rates of the composite progression outcome.

CONCLUSIONS

Systematic analysis of electron microscopic findings reveals clusters of findings associated with either proteinuria remission or disease progression.

Systems Biology and Kidney Disease

The kidney is a complex organ responsible for maintaining multiple aspects of homeostasis in the human body. The combination of distinct, yet interrelated, molecular functions across different cell types make the delineation of factors associated with loss or decline in kidney function challenging. Consequently, there has been a paucity of new diagnostic markers and treatment options becoming available to clinicians and patients in managing kidney diseases. A systems biology approach to understanding the kidney leverages recent advances in computational technology and methods to integrate diverse sets of data. It has the potential to unravel the interplay of multiple genes, proteins, and molecular mechanisms that drive key functions in kidney health and disease. The emergence of large, detailed, multilevel biologic and clinical data from national databases, cohort studies, and trials now provide the critical pieces needed for meaningful application of systems biology approaches in nephrology. The purpose of this review is to provide an overview of the current state in the evolution of the field. Recent successes of systems biology to identify targeted therapies linked to mechanistic biomarkers in the kidney are described to emphasize the relevance to clinical care and the outlook for improving outcomes for patients with kidney diseases.

Histopathologic Evaluation and Scoring of Viral Lung Infection

Emergent coronaviruses such as MERS-CoV and SARS-CoV can cause significant morbidity and mortality in infected individuals. Lung infection is a common clinical feature and contributes to disease severity as well as viral transmission. Animal models are often required to study viral infections and therapies, especially during an initial outbreak. Histopathology studies allow for identification of lesions and affected cell types to better understand viral pathogenesis and clarify effective therapies. Use of immunostaining allows detection of presumed viral receptors and viral tropism for cells can be evaluated to correlate with lesions. In the lung, lesions and immunostaining can be qualitatively described to define the cell types, microanatomic location, and type of changes seen. These features are important and necessary, but this approach can have limitations when comparing treatment groups. Semiquantitative and quantitative tissue scores are more rigorous as these provide the ability to statistically compare groups and increase the reproducibility and rigor of the study. This review describes principles, approaches, and resources that can be useful to evaluate coronavirus lung infection, focusing on MER-CoV infection as the principal example.

Identification of dicarbonyl and L-xylulose reductase as a therapeutic target in human chronic kidney disease

An imbalance of nephroprotective factors and renal damaging molecules contributes to development and progression of chronic kidney disease (CKD). We investigated associations of renoprotective factor gene expression patterns with CKD severity and outcome. Gene expression profiles of 197 previously reported renoprotective factors were analyzed in a discovery cohort in renal biopsies of 63 CKD patients. Downregulation of dicarbonyl and L-xylulose reductase (DCXR) showed the strongest association with disease progression. This significant association was validated in an independent set of 225 patients with nephrotic syndrome from the multicenter NEPTUNE cohort. Reduced expression of DCXR was significantly associated with degree of histological damage as well as with lower estimated glomerular filtration rate and increased urinary protein levels. DCXR downregulation in CKD was confirmed in 3 publicly available transcriptomics data sets in the context of CKD. Expression of DCXR showed positive correlations to enzymes that are involved in dicarbonyl stress detoxification based on transcriptomics profiles. The sodium glucose cotransporter-2 (SGLT2) inhibitors canagliflozin and empagliflozin showed a beneficial effect on renal proximal tubular cells under diabetic stimuli-enhanced DCXR gene expression. In summary, lower expression of the renoprotective factor DCXR in renal tissue is associated with more severe disease and worse outcome in human CKD.

Virtual microscopy enhances the reliability and validity in histopathology curriculum: Practical guidelines

This article was migrated. The article was marked as recommended. Digital pathology innovation and application in medical education have paved the path for a significant shift in the advancement of the medical curriculum. The new technology of virtual microscopy is a proven reliable and valid pedagogy method for histopathology learning objectives, and assessments. The current transformation has brought educators around the globe nearer towards the goal of achieving competence in Curriculum Inventory in the medical curriculum. This paper emphasises the practical tips and guidelines for cost-effective implementation and the successful use of Virtual Microscope technology to enhance the histopathology curriculum in a medical school.

Extracellular Matrix in Kidney Fibrosis: More Than Just a Scaffold

Kidney fibrosis is the common histological end-point of progressive, chronic kidney diseases (CKDs) regardless of the underlying etiology. The hallmark of renal fibrosis, similar to all other organs, is pathological deposition of extracellular matrix (ECM). Renal ECM is a complex network of collagens, elastin, and several glycoproteins and proteoglycans forming basal membranes and interstitial space. Several ECM functions beyond providing a scaffold and organ stability are being increasingly recognized, for example, in inflammation. ECM composition is determined by the function of each of the histological compartments of the kidney, that is, glomeruli, tubulo-interstitium, and vessels. Renal ECM is a dynamic structure undergoing remodeling, particularly during fibrosis. From a clinical perspective, ECM proteins are directly involved in several rare renal diseases and indirectly in CKD progression during renal fibrosis. ECM proteins could serve as specific non-invasive biomarkers of fibrosis and scaffolds in regenerative medicine. The gold standard and currently only specific means to measure renal fibrosis is renal biopsy, but new diagnostic approaches are appearing. Here, we discuss the localization, function, and remodeling of major renal ECM components in healthy and diseased, fibrotic kidneys and the potential use of ECM in diagnostics of renal fibrosis and in tissue engineering.

CureGN Study Rationale, Design, and Methods: Establishing a Large Prospective Observational Study of Glomerular Disease

RATIONALE & OBJECTIVES

Glomerular diseases, including minimal change disease, focal segmental glomerulosclerosis, membranous nephropathy, and immunoglobulin A (IgA) nephropathy, share clinical presentations, yet result from multiple biological mechanisms. Challenges to identifying underlying mechanisms, biomarkers, and new therapies include the rarity of each diagnosis and slow progression, often requiring decades to measure the effectiveness of interventions to prevent end-stage kidney disease (ESKD) or death.

STUDY DESIGN

Multicenter prospective cohort study.

SETTING & PARTICIPANTS

Cure Glomerulonephropathy (CureGN) will enroll 2,400 children and adults with minimal change disease, focal segmental glomerulosclerosis, membranous nephropathy, or IgA nephropathy (including IgA vasculitis) and a first diagnostic kidney biopsy within 5 years. Patients with ESKD and those with secondary causes of glomerular disease are excluded.

EXPOSURES

Clinical data, including medical history, medications, family history, and patient-reported outcomes, are obtained, along with a digital archive of kidney biopsy images and blood and urine specimens at study visits aligned with clinical care 1 to 4 times per year.

OUTCOMES

Patients are followed up for changes in estimated glomerular filtration rate, disease activity, ESKD, and death and for nonrenal complications of disease and treatment, including infection, malignancy, cardiovascular, and thromboembolic events.

ANALYTICAL APPROACH

The study design supports multiple longitudinal analyses leveraging the diverse data domains of CureGN and its ancillary program. At 2,400 patients and an average of 2 years' initial follow-up, CureGN has 80% power to detect an HR of 1.4 to 1.9 for proteinuria remission and a mean difference of 2.1 to 3.0mL/min/1.73m² in estimated glomerular filtration rate per year.

LIMITATIONS

Current follow-up can only detect large differences in ESKD and death outcomes.

CONCLUSIONS

Study infrastructure will support a broad range of scientific approaches to identify mechanistically distinct subgroups, identify accurate biomarkers of disease activity and progression, delineate disease-specific treatment targets, and inform future therapeutic trials. CureGN is expected to be among the largest prospective studies of children and adults with glomerular disease, with a broad goal to lessen disease burden and improve outcomes.

Digital pathology in nephrology clinical trials, research, and pathology practice

PURPOSE OF REVIEW

In this review, we will discuss (i) how the recent advancements in digital technology and computational engineering are currently applied to nephropathology in the setting of clinical research, trials, and practice; (ii) the benefits of the new digital environment; (iii) how recognizing its challenges provides opportunities for transformation; and (iv) nephropathology in the upcoming era of kidney precision and predictive medicine.

RECENT FINDINGS

Recent studies highlighted how new standardized protocols facilitate the harmonization of digital pathology database infrastructure and morphologic, morphometric, and computer-aided quantitative analyses. Digital pathology enables robust protocols for clinical trials and research, with the potential to identify previously underused or unrecognized clinically useful parameters. The integration of digital pathology with molecular signatures is leading the way to establishing clinically relevant morpho-omic taxonomies of renal diseases.

SUMMARY

The introduction of digital pathology in clinical research and trials, and the progressive implementation of the modern software ecosystem, opens opportunities for the development of new predictive diagnostic paradigms and computer-aided algorithms, transforming the practice of renal disease into a modern computational science.

Reproducibility and Feasibility of Strategies for Morphologic Assessment of Renal Biopsies Using the Nephrotic Syndrome Study Network Digital Pathology Scoring System

Context Testing reproducibility is critical for the development of methodologies for morphologic assessment. Our previous study using the descriptor-based Nephrotic Syndrome Study Network Digital Pathology Scoring System (NDPSS) on glomerular images revealed variable reproducibility. Objective To test reproducibility and feasibility of alternative scoring strategies for digital morphologic assessment of glomeruli and explore use of alternative agreement statistics. Design The original NDPSS was modified (NDPSS1 and NDPSS2) to evaluate (1) independent scoring of each individual biopsy level, (2) use of continuous measures, (3) groupings of individual descriptors into classes and subclasses prior to scoring, and (4) indication of pathologists' confidence/uncertainty for any given score. Three and 5 pathologists scored 157 and 79 glomeruli using the NDPSS1 and NDPSS2, respectively. Agreement was tested using conventional (Cohen κ) and alternative (Gwet agreement coefficient 1 [AC₁]) agreement statistics and compared with previously published data (original NDPSS). Results Overall, pathologists' uncertainty was low, favoring application of the Gwet AC₁. Greater agreement was achieved using the Gwet AC₁ compared with the Cohen κ across all scoring methodologies. Mean (standard deviation) differences in agreement estimates using the NDPSS1 and NDPSS2 compared with the single-level original NDPSS were -0.09 (0.17) and -0.17 (0.17), respectively. Using the Gwet AC₁, 79% of the original NDPSS descriptors had good or excellent agreement. Pathologist feedback indicated the NDPSS1 and NDPSS2 were time-consuming. Conclusions The NDPSS1 and NDPSS2 increased pathologists' scoring burden without improving reproducibility. Use of alternative agreement statistics was strongly supported. We suggest using the original NDPSS on whole slide images for glomerular morphology assessment and for guiding future automated technologies.

Nephrin Loss Can Be Used to Predict Remission and Long-term Renal Outcome in Patients With Minimal Change Disease

Introduction

Minimal change disease is a common cause of nephrotic syndrome. In general, patients with minimal change disease respond to corticosteroids and have excellent long-term renal survival. However, some patients have less favorable outcome. These patients are often thought to have progressed to focal segmental glomerulosclerosis. We previously reported that a segmental loss of podocyte markers is present before the development of focal segmental glomerulosclerosis in a rat model. Here, we investigated whether loss of podocyte marker nephrin can serve as a biomarker for predicting poor outcome in patients with minimal change disease.

Methods

We obtained 47 kidney biopsy samples from patients diagnosed with minimal change disease and stained sections with periodic acid−Schiff and for nephrin. Nephrin loss was scored by 2 independent researchers who were blinded to clinical outcome. Clinical data were collected retrospectively, and nephrin loss was correlated with clinical follow-up data.

Results

Nephrin loss was present in 34% of the biopsy samples. During follow-up, patients with nephrin loss achieved remission less frequently (61%) compared to patients without (96%) (P = 0.002). Moreover, 5-year eGFR was lower in the patients with renal nephrin loss. The risk of eGFR decreasing to < 60 ml/min per 1.73m² increased with each percentage of glomeruli with nephrin loss (hazard ratio = 1.044, 95% confidence interval = 1.02−1.07).

Conclusion

These results indicate that nephrin loss in patients with minimal change disease can help predict both remission and long-term renal outcome.

Renal biopsy-driven molecular target identification in glomerular disease

Chronic kidney disease has severe impacts on the patient and represents a major burden to the health care systems worldwide. Despite an increased knowledge of pathophysiological processes involved in kidney diseases, the progress in defining novel treatment strategies has been limited. One reason is the descriptive disease categorization used in nephrology based on clinical findings or histopathological categories irrespective of potential different molecular disease mechanisms. To accelerate progress toward a targeted treatment, a definition of human disease extending from phenotypic disease classification to mechanism-based disease definitions is needed. In recent years, we have witnessed a major transition in biomedical research from a single gene research to an information rich and collaborative science. Tissue-based analysis in renal disease allows to link structure to molecular function. In our review, we introduce the concept of precision medicine in nephrology, describe several large cohort studies established for molecular analysis of kidney diseases, and highlight examples of renal biopsy-driven target identification by integrative systems biology approaches. Furthermore, we give an outlook on how the new disease definitions can be used for patient stratification in clinical trial design. Finally, we introduce the concept of an informational commons of renal precision medicine for joint analyses of large-scale data sets in renal failure.