Influence of Calcareous Deposits on Hydrogen Embrittlement Susceptibility of Q460 Steel.

Cathodic protection is widely used to protect structural steel from corrosion in marine environments. However, an inappropriate cathodic potential may lead to hydrogen embrittlement (HE). Therefore, this study investigates the relationship between cathodic protection potential, structure and composition of calcareous deposits, and hydrogen embrittlement susceptibility of Q460 steel. The slow strain rate test results and fracture analysis reveal that Q460 steel had the smallest HE susceptibility when covered with the calcareous deposits formed under -1.1 VSCE. The deposits have a relatively thin calcium-rich inner layer and a condensed magnesium-rich outer layer, which can significantly inhibit hydrogen entry. A sustained deposition reaction during slow strain rate testing (SSRT) in artificial seawater can also decrease the HE susceptibility of Q460 steel.

Tubular CPT1A deletion minimally affects aging and chronic kidney injury.

Kidney tubules use fatty acid oxidation (FAO) to support their high energetic requirements. Carnitine palmitoyltransferase 1A (CPT1A) is the rate-limiting enzyme for FAO, and it is necessary to transport long-chain fatty acids into mitochondria. To define the role of tubular CPT1A in aging and injury, we generated mice with tubule-specific deletion of Cpt1a (Cpt1aCKO mice), and the mice were either aged for 2 years or injured by aristolochic acid or unilateral ureteral obstruction. Surprisingly, Cpt1aCKO mice had no significant differences in kidney function or fibrosis compared with wild-type mice after aging or chronic injury. Primary tubule cells from aged Cpt1aCKO mice had a modest decrease in palmitate oxidation but retained the ability to metabolize long-chain fatty acids. Very-long-chain fatty acids, exclusively oxidized by peroxisomes, were reduced in kidneys lacking tubular CPT1A, consistent with increased peroxisomal activity. Single-nuclear RNA-Seq showed significantly increased expression of peroxisomal FAO enzymes in proximal tubules of mice lacking tubular CPT1A. These data suggest that peroxisomal FAO may compensate in the absence of CPT1A, and future genetic studies are needed to confirm the role of peroxisomal ß-oxidation when mitochondrial FAO is impaired.

Inhibition of transcriptional coactivator YAP Impairs the expression and function of transcription factor WT1 in diabetic podocyte injury.

Podocyte injury and loss are hallmarks of diabetic nephropathy (DN). However, the molecular mechanisms underlying these phenomena remain poorly understood. YAP (Yes-associated protein) is an important transcriptional coactivator that binds with various other transcription factors, including the TEAD family members (nuclear effectors of the Hippo pathway), that regulate cell proliferation, differentiation, and apoptosis. The present study found an increase in YAP phosphorylation at S127 of YAP and a reduction of nuclear YAP localization in podocytes of diabetic mouse and human kidneys, suggesting dysregulation of YAP may play a role in diabetic podocyte injury. Tamoxifen-inducible podocyte-specific Yap gene knockout mice (YappodKO) exhibited accelerated and worsened diabetic kidney injury. YAP inactivation decreased transcription factor WT1 expression with subsequent reduction of Tead1 and other well-known targets of WT1 in diabetic podocytes. Thus, our study not only sheds light on the pathophysiological roles of the Hippo pathway in diabetic podocyte injury but may also lead to the development of new therapeutic strategies to prevent and/or treat DN by targeting the Hippo signaling pathway.

Adverse effects of acute tubular injury on the glomerulus: contributing factors and mechanisms.

The intricate relationship between tubular injury and glomerular dysfunction in kidney diseases has been a subject of extensive research. While the impact of glomerular injury on downstream tubules has been well-studied, the reverse influence of tubular injury on the glomerulus remains less explored. This paper provides a comprehensive review of recent advances in the field, focusing on key pathways and players implicated in the pathogenesis of tubular injury on glomerular dysfunction. Anatomical and physiological evidence supports the possibility of crosstalk from the tubule to the glomerulus, whereby various mechanisms contribute to glomerular injury following tubular injury. These mechanisms include tubular backleak, dysfunctional tubuloglomerular feedback, capillary rarefaction, atubular glomeruli, and the secretion of factors from damaged tubular epithelial cells. Clinical evidence further supports the association between even mild or recovered acute kidney injury and an increased risk of chronic kidney disease, including glomerular diseases. We also discuss potential therapeutic interventions aimed at mitigating acute tubular injury, thereby reducing the detrimental effects on glomerular function. By unraveling the complex interplay from tubular injury to glomerular dysfunction, we aim to provide insights that can enhance clinical management strategies and improve outcomes for patients with kidney disease.

Effects of Dipeptidyl Peptidase-4 Inhibitor and Angiotensin-Converting Enzyme Inhibitor on Experimental Diabetic Kidney Disease.

Diabetic kidney disease (DKD) is the leading cause of end-stage kidney disease in the United States and worldwide. Proteinuria is a major marker of the severity of injury. Dipeptidyl peptidase-4 inhibitor (DPP-4I) increases incretin-related insulin production and is, therefore, used to treat diabetes. We investigated whether DPP4I could have direct effect on kidney independent of its hypoglycemic activity. We, therefore, tested the effects of DPP4I with or without angiotensin-converting enzyme inhibitor (ACEI) on the progression of diabetic nephropathy and albuminuria in a murine model of DKD. eNOS-/-db/db mice were randomized to the following groups at age 10 weeks and treated until sacrifice: baseline (sacrificed at week 10), untreated control, ACEI, DPP4I, and combination of DPP4I and ACEI (Combo, sacrificed at week 18). Systemic parameters and urine albumin-creatinine ratio were assessed at baseline, weeks 14, and 18. Kidney morphology, glomerular filtration rate (GFR), WT-1, a marker for differentiated podocytes, podoplanin, a marker of foot process integrity, glomerular collagen IV, and alpha-smooth muscle actin were assessed at the end of the study. All mice had hyperglycemia and proteinuria at study entry at week 10. Untreated control mice had increased albuminuria, progression of glomerular injury, and reduced GFR at week 18 compared with baseline. DPP4I alone reduced blood glucose and kidney DPP-4 activity but failed to protect against kidney injury compared with untreated control. ACEI alone and combination groups showed significantly reduced albuminuria and glomerular injury, and maintained GFR and WT-1+ cells. Only the combination group had significantly less glomerular collagen IV deposition and more podoplanin preservation than the untreated control. DPP-4I alone does not decrease the progression of kidney injury in the eNOS-/-db/db mouse model, suggesting that targeting only hyperglycemia is not an optimal treatment strategy for DKD. Combined DPP-4I with ACEI added more benefit to reducing the glomerular matrix.

Inhibition of retinoic acid signaling in proximal tubular epithelial cells protects against acute kidney injury.

Retinoic acid receptor (RAR) signaling is essential for mammalian kidney development but, in the adult kidney, is restricted to occasional collecting duct epithelial cells. We now show that there is widespread reactivation of RAR signaling in proximal tubular epithelial cells (PTECs) in human sepsis-associated acute kidney injury (AKI) and in mouse models of AKI. Genetic inhibition of RAR signaling in PTECs protected against experimental AKI but was unexpectedly associated with increased expression of the PTEC injury marker Kim1. However, the protective effects of inhibiting PTEC RAR signaling were associated with increased Kim1-dependent apoptotic cell clearance, or efferocytosis, and this was associated with dedifferentiation, proliferation, and metabolic reprogramming of PTECs. These data demonstrate the functional role that reactivation of RAR signaling plays in regulating PTEC differentiation and function in human and experimental AKI.

Clinical value of combined serum MMP-2, MMP-9 and TIMP-1 for the prognosis of perianal fistula patients who received minimally invasive surgery.

OBJECTIVE: This study aimed to investigate the clinical value of combined serum matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) for the prognosis of perianal fistula patients. METHODS: Patients diagnosed and treated for perianal fistula by minimally invasive surgery (MIS) were enrolled. The concentrations of serum MMP-2, MMP-9 and TIMP-1 were measured at 24 h after surgery. Different levels of wound secretion, growth of granulation tissue and wound pain were used as criteria to evaluate surgical incision healing. The receiver operating characteristic curve was used to analyze the predicted assessment value. RESULTS: The concentrations of serum MMP-2 and MMP-9 were significantly higher, while the concentrations of serum TIMP-1 at 24 h after surgery were significantly lower in the poor healing group than in the good healing group. It was further found that high levels of serum MMP-2 and MMP-9 were risk factors for poor healing, while high concentrations of serum TIMP-1 at 24 h after surgery were protective factors for poor healing. CONCLUSION: High concentrations of serum MMP-2 and MMP-9 and low concentrations of serum TIMP at 24 h after surgery are risk factors for poor healing in perianal fistula patients who receive MIS, and the combined test has a higher predictive value.

Deep multimodal fusion of image and non-image data in disease diagnosis and prognosis: a review.

The rapid development of diagnostic technologies in healthcare is leading to higher requirements for physicians to handle and integrate the heterogeneous, yet complementary data that are produced during routine practice. For instance, the personalized diagnosis and treatment planning for a single cancer patient relies on various images (e.g. radiology, pathology and camera images) and non-image data (e.g. clinical data and genomic data). However, such decision-making procedures can be subjective, qualitative, and have large inter-subject variabilities. With the recent advances in multimodal deep learning technologies, an increasingly large number of efforts have been devoted to a key question: how do we extract and aggregate multimodal information to ultimately provide more objective, quantitative computer-aided clinical decision making? This paper reviews the recent studies on dealing with such a question. Briefly, this review will include the (a) overview of current multimodal learning workflows, (b) summarization of multimodal fusion methods, (c) discussion of the performance, (d) applications in disease diagnosis and prognosis, and (e) challenges and future directions.

Omni-Seg: A Scale-Aware Dynamic Network for Renal Pathological Image Segmentation.

Comprehensive semantic segmentation on renal pathological images is challenging due to the heterogeneous scales of the objects. For example, on a whole slide image (WSI), the cross-sectional areas of glomeruli can be 64 times larger than that of the peritubular capillaries, making it impractical to segment both objects on the same patch, at the same scale. To handle this scaling issue, prior studies have typically trained multiple segmentation networks in order to match the optimal pixel resolution of heterogeneous tissue types. This multi-network solution is resource-intensive and fails to model the spatial relationship between tissue types. In this article, we propose the Omni-Seg network, a scale-aware dynamic neural network that achieves multi-object (six tissue types) and multi-scale (5× to 40× scale) pathological image segmentation via a single neural network. The contribution of this article is three-fold: (1) a novel scale-aware controller is proposed to generalize the dynamic neural network from single-scale to multi-scale; (2) semi-supervised consistency regularization of pseudo-labels is introduced to model the inter-scale correlation of unannotated tissue types into a single end-to-end learning paradigm; and (3) superior scale-aware generalization is evidenced by directly applying a model trained on human kidney images to mouse kidney images, without retraining. By learning from 150,000 human pathological image patches from six tissue types at three different resolutions, our approach achieved superior segmentation performance according to human visual assessment and evaluation of image-omics (i.e., spatial transcriptomics).

Intestinal Lymphatic Dysfunction in Kidney Disease.

Kidney disease is associated with adverse consequences in many organs beyond the kidney, including the heart, lungs, brain, and intestines. The kidney-intestinal cross talk involves intestinal epithelial damage, dysbiosis, and generation of uremic toxins. Recent studies reveal that kidney injury expands the intestinal lymphatics, increases lymphatic flow, and alters the composition of mesenteric lymph. The intestinal lymphatics, like blood vessels, are a route for transporting potentially harmful substances generated by the intestines. The lymphatic architecture and actions are uniquely suited to take up and transport large macromolecules, functions that differentiate them from blood vessels, allowing them to play a distinct role in a variety of physiological and pathological processes. Here, we focus on the mechanisms by which kidney diseases result in deleterious changes in intestinal lymphatics and consider a novel paradigm of a vicious cycle of detrimental organ cross talk. This concept involves kidney injury-induced modulation of intestinal lymphatics that promotes production and distribution of harmful factors, which in turn contributes to disease progression in distant organ systems.

Low-Dose Erythropoietin Amplifies Beneficial Effects of Angiotensin II Blockade on Glomerulosclerosis.

Exogenous erythropoietin (EPO) is used to treat anemia in patients with chronic kidney disease (CKD). Concerns about the possible adverse effect of EPO on the progression of CKD have been raised owing to nonerythroid cell effects. We investigated the effects of low-dose EPO, independent of correcting anemia, on existing glomerulosclerosis. Adult mice underwent 5/6 nephrectomy and were randomized into the following 4 groups at week 8 after surgery: vehicle (VEH), losartan (angiotensin II type 1 receptor blocker [ARB]), darbepoetin-α (DA), or combination (DA+ARB). Four weeks later, mice were euthanized, followed by evaluation of renal structure and function. Glomerular endothelial cells and podocytes were cultured to evaluate the effects of DA on cell migration, apoptosis, and Akt signaling. ARB reduced blood pressure, albuminuria, and the level of serum creatinine and increased hematocrit compared with VEH, whereas low-dose DA only reduced the level of serum creatinine. Combination treatment showed a trend to increase hematocrit and survival compared with ARB alone. Combination treatment but not ARB alone significantly reduced the progression of glomerulosclerosis compared with VEH. Low-dose DA resulted in more preserved glomerular and peritubular capillary endothelial cells with increased p-Akt and even further endothelial cell preservation in combination with ARB. In cultured glomerular endothelial cells, angiotensin II induced more apoptosis, reduced migration, and decreased p-Flk1, a receptor for the proangiogenic vascular endothelial growth factor. DA counteracted these injuries and increased p-Akt, a key factor in angiogenesis and cell survival. DA also protected cultured podocytes against transforming growth factor ß-induced apoptosis and synaptopodin loss. Low-dose EPO directly protects glomerular and peritubular endothelial cells via Akt phosphorylation. Therefore, treatment using a combination of low-dose EPO and ARB results in less progression of glomerulosclerosis in an experimental CKD model.

Democratizing Pathological Image Segmentation with Lay Annotators via Molecular-empowered Learning.

Multi-class cell segmentation in high-resolution Giga-pixel whole slide images (WSI) is critical for various clinical applications. Training such an AI model typically requires labor-intensive pixel-wise manual annotation from experienced domain experts (e.g., pathologists). Moreover, such annotation is error-prone when differentiating fine-grained cell types (e.g., podocyte and mesangial cells) via the naked human eye. In this study, we assess the feasibility of democratizing pathological AI deployment by only using lay annotators (annotators without medical domain knowledge). The contribution of this paper is threefold: (1) We proposed a molecular-empowered learning scheme for multi-class cell segmentation using partial labels from lay annotators; (2) The proposed method integrated Giga-pixel level molecular-morphology cross-modality registration, molecular-informed annotation, and molecular-oriented segmentation model, so as to achieve significantly superior performance via 3 lay annotators as compared with 2 experienced pathologists; (3) A deep corrective learning (learning with imperfect label) method is proposed to further improve the segmentation performance using partially annotated noisy data. From the experimental results, our learning method achieved F1 = 0.8496 using molecular-informed annotations from lay annotators, which is better than conventional morphology-based annotations (F1 = 0.7015) from experienced pathologists. Our method democratizes the development of a pathological segmentation deep model to the lay annotator level, which consequently scales up the learning process similar to a non-medical computer vision task. The official implementation and cell annotations are publicly available at https://github.com/hrlblab/MolecularEL.

An Accelerated Pipeline for Multi-label Renal Pathology Image Segmentation at the Whole Slide Image Level.

Deep-learning techniques have been used widely to alleviate the labour-intensive and time-consuming manual annotation required for pixel-level tissue characterization. Our previous study introduced an efficient single dynamic network - Omni-Seg - that achieved multi-class multi-scale pathological segmentation with less computational complexity. However, the patch-wise segmentation paradigm still applies to Omni-Seg, and the pipeline is time-consuming when providing segmentation for Whole Slide Images (WSIs). In this paper, we propose an enhanced version of the Omni-Seg pipeline in order to reduce the repetitive computing processes and utilize a GPU to accelerate the model's prediction for both better model performance and faster speed. Our proposed method's innovative contribution is two-fold: (1) a Docker is released for an end-to-end slide-wise multi-tissue segmentation for WSIs; and (2) the pipeline is deployed on a GPU to accelerate the prediction, achieving better segmentation quality in less time. The proposed accelerated implementation reduced the average processing time (at the testing stage) on a standard needle biopsy WSI from 2.3 hours to 22 minutes, using 35 WSIs from the Kidney Tissue Atlas (KPMP) Datasets. The source code and the Docker have been made publicly available at https://github.com/ddrrnn123/Omni-Seg.

Dicarbonyl-modified lipoproteins contribute to proteinuric kidney injury.

Lipoprotein modification by reactive dicarbonyls, including isolevuglandin (IsoLG), produces dysfunctional particles. Kidneys participate in lipoprotein metabolism, including tubular uptake. However, the process beyond the proximal tubule is unclear, as is the effect of kidney injury on this pathway. We found that patients and animals with proteinuric injury have increased urinary apolipoprotein AI (apoAI), IsoLG, and IsoLG adduct enrichment of the urinary apoAI fraction compared with other proteins. Proteinuric mice, induced by podocyte-specific injury, showed more tubular absorption of IsoLG-apoAI and increased expression of lipoprotein transporters in proximal tubular cells compared with uninjured animals. Renal lymph reflects composition of the interstitial compartment and showed increased apoAI and IsoLG in proteinuric animals, supporting a tubular cell-interstitium-lymph pathway for renal handling of lipoproteins. IsoLG-modified apoAI was not only a marker of renal injury but also directly damaged renal cells. IsoLG-apoAI increased inflammatory cytokines in cultured tubular epithelial cells (TECs), activated lymphatic endothelial cells (LECs), and caused greater contractility of renal lymphatic vessels than unmodified apoAI. In vivo, inhibition of IsoLG by a dicarbonyl scavenger reduced both albuminuria and urinary apoAI and decreased TEC and LEC injury, lymphangiogenesis, and interstitial fibrosis. Our results indicate that IsoLG-modified apoAI is, to our knowledge, a novel pathogenic mediator and therapeutic target in kidney disease.

Selonsertib Enhances Kidney Protection Beyond Standard of Care in a Hypertensive, Secondary Glomerulosclerosis CKD Model.

Background: Despite widespread use of renin-aldosterone-angiotensin system inhibitors and the benefits of lowering glomerular pressure in patients with CKD, there remains a major unmet need for therapies targeting underlying causes of CKD progression. Apoptosis signal-regulating kinase 1 (ASK1) promotes apoptosis and glomerulosclerosis, and is implicated in the progression of diabetic kidney disease (DKD), a major cause of CKD. Selonsertib is a selective ASK1 inhibitor currently in clinical development for the treatment of DKD. We examined the added benefits of selonsertib on existing glomerulosclerosis and related molecular pathways in the nondiabetic 5/6 nephrectomy (5/6 Nx) rat model in combination with the angiotensin-converting enzyme inhibitor (ACEI) enalapril. Methods: Male Sprague Dawley rats underwent 5/6 Nx with kidney biopsy 8 weeks later for assessment of glomerulosclerosis, and were randomized to four treatment groups with equal glomerulosclerosis: selonsertib, enalapril, combination (selonsertib plus enalapril), and untreated controls. Serum creatinine, systolic BP (SBP), and urinary albumin were measured at intervals. Animals were euthanized at week 12 for histologic, biochemical, and molecular analyses. Results: All rats developed hypertension, albuminuria, and glomerulosclerosis by week 8. Kidney function further declined, and glomerulosclerosis and albuminuria progressively increased in controls from week 8 to 12. Enalapril treatment alone from week 8 to 12 reduced SBP versus controls, decreased albuminuria, and resulted in numerically lower glomerulosclerosis. Selonsertib alone had no effect on SBP but preserved kidney function. Combined treatment significantly reduced glomerulosclerosis, with more regression than either monotherapy. Enalapril treatment resulted in fewer interstitial macrophages, whereas selonsertib treatment reduced apoptosis and podocyte loss. RNA-seq revealed that combined treatment influenced pathways related to extracellular matrix and wound healing. Conclusions: Selonsertib targets a novel, nonhemodynamic pathway in CKD. Our data suggest that ASK1 inhibition, when combined with ACEI, has additive effects to reduce progression of glomerulosclerosis, attenuate kidney function decline, and reduce podocyte loss.

IgG subclass deposition in diabetic nephropathy.

PURPOSE: This study aimed to analyze the distribution of IgG subclass in diabetic nephropathy (DN) and its association with clinicopathological features. METHODS: This is a single-center retrospective study enrolling 108 patients with biopsy-proven DN. Immunofluorescence and immunohistochemistry staining were applied, and clinicopathological features and renal outcomes were compared between patients with different patterns or categories of IgG subclass deposition. RESULTS: Both IgG and its subclasses colocalized with collagen IV α5 on glomerular basement membrane (GBM) and some of tubular basement membrane (TBM). IgG1 and the Mixed type were two predominant types of deposition, no matter on GBM or TBM, and IgG1 showed a much higher deposition rate on GBM than that on TBM (P = 0.004). IgG subclass deposit on multi-location was more associated with a shorter duration of nephropathy and severer tubular interstitial injury (P < 0.05). The mixed type of IgG subclass deposit on GBM was merely associated with higher levels of proteinuria, whereas the deposition on TBM was more associated with higher levels of proteinuria, lower levels of albumin, more KIM-1 positive area, and thicker TBM (P < 0.05). Survival analysis revealed that none of the pattern or the category of IgG subclass deposit was a risk factor or a renal outcome indicator. CONCLUSIONS: IgG subclass was selectively deposited along GBM and/or TBM in DN, and the mixed type of IgG subclass deposition on TBM had more clinical significance than the isotype and that on GBM. IgG subclass deposition is merely a manifestation or a consequence rather than a cause in DN.

Podocyte-Related Mechanisms Underlying Survival Benefit of Long-Term Angiotensin Receptor Blocker.

We previously found that short-term treatment (week 8 to 12 after injury) with high-dose angiotensin receptor blocker (ARB) induced the regression of existing glomerulosclerosis in 5/6 nephrectomy rats. We therefore assessed the effects of long-term intervention with ARB vs. nonspecific antihypertensives in this study. Adult rats underwent 5/6 nephrectomy and renal biopsy 8 weeks later. The rats were then divided into three groups with equivalent renal function and glomerular sclerosis and treated with high-dose losartan (ARB), nonspecific antihypertensive triple-therapy (TRX), or left untreated (Control) until week 30. We found that blood pressure, serum creatinine levels, and glomerulosclerosis were lower at sacrifice in ARB and TRX vs. Control. Only ARB reduced proteinuria and maintained the density of WT-1-positive podocytes. Glomerular tufts showed more double-positive cells for CD44, a marker of activated parietal epithelial cells, and synaptopodin after ARB vs. TRX or Control. ARB treatment reduced aldosterone levels. ARB-treated rats had significantly improved survival when compared with TRX or Control. We conclude that both long-term ARB and triple-therapy ameliorate progression, but do not sustain the regression of glomerulosclerosis. ARB resulted in the superior preservation of podocyte integrity and decreased proteinuria and aldosterone, linked to increased survival in the uremic environment.

Inflammation, Lymphatics, and Cardiovascular Disease: Amplification by Chronic Kidney Disease.

PURPOSE OF REVIEW: Kidney disease is a strong modulator of the composition and metabolism of the intestinal microbiome that produces toxins and inflammatory factors. The primary pathways for these harmful factors are blood vessels and nerves. Although lymphatic vessels are responsible for clearance of interstitial fluids, macromolecules, and cells, little is known about whether and how kidney injury impacts the intestinal lymphatic network. RECENT FINDINGS: Kidney injury stimulates intestinal lymphangiogenesis, activates lymphatic endothelial cells, and increases mesenteric lymph flow. The mesenteric lymph of kidney-injured animals contains increased levels of cytokines, immune cells, isolevuglandin (IsoLG), a highly reactive dicarbonyl, and of apolipoprotein AI (apoAI). IsoLG is increased in the ileum of kidney injured animals, and intestinal epithelial cells exposed to myeloperoxidase produce more IsoLG. IsoLG-modified apoAI directly increases lymphatic vessel contractions and activates lymphatic endothelial cells. Inhibition of IsoLG by carbonyl scavenger treatment reduces intestinal lymphangiogenesis in kidney-injured animals. Research from our group and others suggests a novel mediator (IsoLG-modified apoAI) and a new pathway (intestinal lymphatic network) in the cross talk between kidneys and intestines and heart. Kidney injury activates intestinal lymphangiogenesis and increases lymphatic flow via mechanisms involving intestinally generated IsoLG. The data identify a new pathway in the kidney gut-heart axis and present a new target for kidney disease-induced intestinal disruptions that may lessen the major adverse consequence of kidney impairment, namely cardiovascular disease.

Glo-In-One: holistic glomerular detection, segmentation, and lesion characterization with large-scale web image mining.

Purpose: The quantitative detection, segmentation, and characterization of glomeruli from high-resolution whole slide imaging (WSI) play essential roles in the computer-assisted diagnosis and scientific research in digital renal pathology. Historically, such comprehensive quantification requires extensive programming skills to be able to handle heterogeneous and customized computational tools. To bridge the gap of performing glomerular quantification for non-technical users, we develop the Glo-In-One toolkit to achieve holistic glomerular detection, segmentation, and characterization via a single line of command. Additionally, we release a large-scale collection of 30,000 unlabeled glomerular images to further facilitate the algorithmic development of self-supervised deep learning. Approach: The inputs of the Glo-In-One toolkit are WSIs, while the outputs are (1) WSI-level multi-class circle glomerular detection results (which can be directly manipulated with ImageScope), (2) glomerular image patches with segmentation masks, and (3) different lesion types. In the current version, the fine-grained global glomerulosclerosis (GGS) characterization is provided, including assessed-solidified-GSS (associated with hypertension-related injury), disappearing-GSS (a further end result of the SGGS becoming contiguous with fibrotic interstitium), and obsolescent-GSS (nonspecific GGS increasing with aging) glomeruli. To leverage the performance of the Glo-In-One toolkit, we introduce self-supervised deep learning to glomerular quantification via large-scale web image mining. Results: The GGS fine-grained classification model achieved a decent performance compared with baseline supervised methods while only using 10% of the annotated data. The glomerular detection achieved an average precision of 0.627 with circle representations, while the glomerular segmentation achieved a 0.955 patch-wise Dice dimilarity coefficient. Conclusion: We develop and release an open-source Glo-In-One toolkit, a software with holistic glomerular detection, segmentation, and lesion characterization. This toolkit is user-friendly to non-technical users via a single line of command. The toolbox and the 30,000 web mined glomerular images have been made publicly available at https://github.com/hrlblab/Glo-In-One.