Interpretable machine learning-based individual analysis of acute kidney injury in immune checkpoint inhibitor therapy.

BACKGROUND: Acute kidney injury (AKI) is a critical complication of immune checkpoint inhibitor therapy. Since the etiology of AKI in patients undergoing cancer therapy varies, clarifying underlying causes in individual cases is critical for optimal cancer treatment. Although it is essential to individually analyze immune checkpoint inhibitor-treated patients for underlying pathologies for each AKI episode, these analyses have not been realized. Herein, we aimed to individually clarify the underlying causes of AKI in immune checkpoint inhibitor-treated patients using a new clustering approach with Shapley Additive exPlanations (SHAP). METHODS: We developed a gradient-boosting decision tree-based machine learning model continuously predicting AKI within 7 days, using the medical records of 616 immune checkpoint inhibitor-treated patients. The temporal changes in individual predictive reasoning in AKI prediction models represented the key features contributing to each AKI prediction and clustered AKI patients based on the features with high predictive contribution quantified in time series by SHAP. We searched for common clinical backgrounds of AKI patients in each cluster, compared with annotation by three nephrologists. RESULTS: One hundred and twelve patients (18.2%) had at least one AKI episode. They were clustered per the key feature, and their SHAP value patterns, and the nephrologists assessed the clusters' clinical relevance. Receiver operating characteristic analysis revealed that the area under the curve was 0.880. Patients with AKI were categorized into four clusters with significant prognostic differences (p = 0.010). The leading causes of AKI for each cluster, such as hypovolemia, drug-related, and cancer cachexia, were all clinically interpretable, which conventional approaches cannot obtain. CONCLUSION: Our results suggest that the clustering method of individual predictive reasoning in machine learning models can be applied to infer clinically critical factors for developing each episode of AKI among patients with multiple AKI risk factors, such as immune checkpoint inhibitor-treated patients.

Outcomes of thoracic endovascular aortic repair using fenestrated stent grafts in patients with thoracic aortic distal arch aneurysms.

OBJECTIVES: Thoracic endovascular aortic repair (TEVAR) for aortic arch aneurysms is challenging because of anatomical restrictions and the presence of cervical branches. Revascularization of the cervical branch is required when conventional commercial stent grafts are used. TEVAR using fenestrated stent grafts (FSG) often does not require additional procedures to revascularize cervical branches. This study aimed to evaluate the features and initial and midterm outcomes of TEVAR using fenestrated stent grafts. METHODS: From April 2007 to December 2016, 101 consecutive patients underwent TEVAR using fenestrated stent grafts for distal aortic arch aneurysms at a single centre. Technical success, complications, freedom from aneurysm-related death, secondary intervention and aneurysm progression were retrospectively investigated. RESULTS: All the patients underwent TEVAR using fenestrated stent grafts. The 30-day mortality rate was zero. Cerebral infarction, access route problems and spinal cord injury occurred in 4, 3 and 2 patients, respectively. Each type of endoleak was observed in 38 of the 101 patients during the course of the study; 20/38 patients had minor type 1 endoleaks at the time of discharge. The endoleak disappeared in 2 patients and showed no significant change in 8 patients; however, the aneurysm expanded over time in 10 patients. Additional treatment was performed in 8 of the 10 patients with type 1 endoleaks and dilatation of the aneurysm. The rate of freedom from aneurysm-related death during the observation period was 98%. CONCLUSIONS: TEVAR with FSG is a simple procedure, with few complications. Additional treatment has been observed to reduce aneurysm-related deaths, even in patients with endoleaks and enlarged aneurysms. Based on this study, the outcomes of endovascular repair of aortic arch aneurysms using a fenestrated stent graft seem acceptable.

Idiopathic pulmonary fibrosis-specific Bayesian network integrating extracellular vesicle proteome and clinical information.

Idiopathic pulmonary fibrosis (IPF) is a progressive disease characterized by severe lung fibrosis and a poor prognosis. Although the biomolecules related to IPF have been extensively studied, molecular mechanisms of the pathogenesis and their association with serum biomarkers and clinical findings have not been fully elucidated. We constructed a Bayesian network using multimodal data consisting of a proteome dataset from serum extracellular vesicles, laboratory examinations, and clinical findings from 206 patients with IPF and 36 controls. Differential protein expression analysis was also performed by edgeR and incorporated into the constructed network. We have successfully visualized the relationship between biomolecules and clinical findings with this approach. The IPF-specific network included modules associated with TGF-ß signaling (TGFB1 and LRC32), fibrosis-related (A2MG and PZP), myofibroblast and inflammation (LRP1 and ITIH4), complement-related (SAA1 and SAA2), as well as serum markers, and clinical symptoms (KL-6, SP-D and fine crackles). Notably, it identified SAA2 associated with lymphocyte counts and PSPB connected with the serum markers KL-6 and SP-D, along with fine crackles as clinical manifestations. These results contribute to the elucidation of the pathogenesis of IPF and potential therapeutic targets.

Preemptive Intravenous Human Immunoglobulin G Suppresses BK Polyomavirus Replication and Spread of Infection In Vitro.

BK polyomavirus (BKPyV) infection causes various diseases in immunocompromised patients. Cells from human lung and kidney were infected with BKPyV and treated with commercially available intravenous immunoglobulin G (IVIG). Its effects on BKPyV replication and spread of infection were investigated, focusing on administration timing. IVIG treatment 3 hours after infection suppressed BKPyV replication assessed by real-time PCR and expression of the viral capsid protein 1 and large T-antigen. IVIG effectively reduced the number of BKPyV-infected cells 2 weeks after infection in an antibody titer-dependent manner. Virus release in the culture supernatants was not influenced by IVIG treatment 6-80 hours and 3-9 days after infection. Collectively, IVIG did not affect viral release from infected cells but inhibited the spread of infection by neutralizing the released virus and blocking the new infected cell formation, indicating greater efficacy in early localized infection. BKPyV replication resumed in IVIG-treated cultures at 7 days after IVIG removal. Early prophylactic administration of IVIG is expected to reduce the growth and spread of BKPyV infection, resulting in the reduction of infected cell lesions and prevention of BKPyV-associated diseases.

ggkegg: analysis and visualization of KEGG data utilizing the grammar of graphics.

SUMMARY: The Kyoto Encyclopedia of Genes and Genomes (KEGG) database serves as a valuable systems biology resource and is widely utilized in diverse research fields. However, existing software does not allow flexible visualization and network analyses of the vast and complex KEGG data. We developed ggkegg, an R package that integrates KEGG information with ggplot2 and ggraph. ggkegg enables enhanced visualization and network analyses of KEGG data. We demonstrate the utility of the package by providing examples of its application in single-cell, bulk transcriptome, and microbiome analyses. ggkegg may empower researchers to analyze complex biological networks and present their results effectively. AVAILABILITY AND IMPLEMENTATION: The package and user documentation are available at: https://github.com/noriakis/ggkegg.

Individual health-disease phase diagrams for disease prevention based on machine learning.

Early disease detection and prevention methods based on effective interventions are gaining attention worldwide. Progress in precision medicine has revealed that substantial heterogeneity exists in health data at the individual level and that complex health factors are involved in chronic disease development. Machine-learning techniques have enabled precise personal-level disease prediction by capturing individual differences in multivariate data. However, it is challenging to identify what aspects should be improved for disease prevention based on future disease-onset prediction because of the complex relationships among multiple biomarkers. Here, we present a health-disease phase diagram (HDPD) that represents an individual's health state by visualizing the future-onset boundary values of multiple biomarkers that fluctuate early in the disease progression process. In HDPDs, future-onset predictions are represented by perturbing multiple biomarker values while accounting for dependencies among variables. We constructed HDPDs for 11 diseases using longitudinal health checkup cohort data of 3,238 individuals, comprising 3,215 measurement items and genetic data. The improvement of biomarker values to the non-onset region in HDPD remarkably prevented future disease onset in 7 out of 11 diseases. HDPDs can represent individual physiological states in the onset process and be used as intervention goals for disease prevention.

Definition of a Dietary Pattern Expressing the Intake of Vegetables and Fruits and Its Association with Intestinal Microbiota.

Daily dietary habits directly or indirectly influence the intestinal microbiota, and the resulting changes in its composition and metabolic activity alter the health conditions of the host. Although many studies have analyzed the association between individual nutrients/food items and intestinal microbiota, the assessment of the diet and intestinal microbiota from a macroscopic perspective has not yet been performed in Japan. Therefore, we focused on vegetables and fruits and aimed to identify dietary patterns of high intake of these foods and to examine their relationship with the intestinal microbiota. This cross-sectional study included 1019 healthy individuals aged ≥20 years in a rural area in northern Japan. Six dietary patterns were detected by factor analysis using the brief-type self-administered diet history questionnaire (BDHQ) data to identify the "vege pattern", which was the dietary pattern rich in vegetables and fruits. Permutational multivariate analysis of variance revealed changes in ß-diversity according to dietary patterns. In multivariable-adjusted models, the adherence to the vege pattern was positively correlated with α-diversity. This is the first study to reveal a correlation between intestinal microbiota and dietary habits rich in vegetables and fruits in a rural area of Japan.

Exploring the mechanism of BK polyomavirus-associated nephropathy through consensus gene network approach.

BK polyomavirus-associated nephropathy occurs in kidney transplant recipients under immunosuppressive treatment. BK polyomavirus is implicated in cancer development and invasion, and case reports of renal cell carcinoma and urothelial carcinoma possibly associated with BK polyomavirus has been reported. Further, it has been suggested that the immune responses of KT-related diseases could play a role in the pathogenesis and progression of renal cell carcinoma. Thus, we thought to examine the relationship between BK polyomavirus-associated nephropathy and renal cell carcinoma in terms of gene expression. To identify the common and specific immune responses involved in kidney transplantation-related diseases with a specific focus on BK polyomavirus-associated nephropathy, we performed consensus weighted gene co-expression network analysis on gene profile datasets of renal biopsy samples from different institutions. After the identification of gene modules and validation of the obtained network by immunohistochemistry of the marker across kidney transplantation-related diseases, the relationship between prognosis of renal cell carcinoma and modules was assessed. We included the data from 248 patients and identified the 14 gene clusters across the datasets. We revealed that one cluster related to the translation regulating process and DNA damage response was specifically upregulated in BK polyomavirus-associated nephropathy. There was a significant association between the expression value of hub genes of the identified cluster including those related to cGAS-STING pathway and DNA damage response, and the prognosis of renal cell carcinoma. The study suggested the potential link between kidney transplantation-related diseases, especially specific transcriptomic signature of BK polyomavirus associated nephropathy and renal cell carcinoma.

Increased abundance of Ruminococcus gnavus in gut microbiota is associated with moyamoya disease and non-moyamoya intracranial large artery disease.

Moyamoya disease (MMD) is a rare cerebrovascular disease endemic in East Asia. The p.R4810K mutation in RNF213 gene confers a risk of MMD, but other factors remain largely unknown. We tested the association of gut microbiota with MMD. Fecal samples were collected from 27 patients with MMD, 7 patients with non-moyamoya intracranial large artery disease (ICAD) and 15 control individuals with other disorders, and 16S rRNA were sequenced. Although there was no difference in alpha diversity or beta diversity between patients with MMD and controls, the cladogram showed Streptococcaceae was enriched in patient samples. The relative abundance analysis demonstrated that 23 species were differentially abundant between patients with MMD and controls. Among them, increased abundance of Ruminococcus gnavus > 0.003 and decreased abundance of Roseburia inulinivorans < 0.002 were associated with higher risks of MMD (odds ratio 9.6, P = 0.0024; odds ratio 11.1, P = 0.0051). Also, Ruminococcus gnavus was more abundant and Roseburia inulinivorans was less abundant in patients with ICAD than controls (P = 0.046, P = 0.012). The relative abundance of Ruminococcus gnavus or Roseburia inulinivorans was not different between the p.R4810K mutant and wildtype. Our data demonstrated that gut microbiota was associated with both MMD and ICAD.

Use of proton pump inhibitors and macrolide antibiotics and risk of acute kidney injury: a self-controlled case series study.

BACKGROUND: Proton pump inhibitors (PPIs) are widely used for the treatment of gastrointestinal disorders such as peptic ulcer disease and dyspepsia. However, several studies have suggested that PPI use increases the risk of acute kidney injury (AKI). PPIs are often concomitantly used with antibiotics, such as macrolides and penicillins for Helicobacter pylori eradication. Although macrolide antibiotics are considered to have relatively low nephrotoxicity, they are well known to increase the risk of AKI due to drug-drug interactions. In this study, we aimed to investigate the association between PPI use and the development of AKI. We also evaluated the effect of concomitant use of PPIs and macrolide antibiotics on the risk of AKI. METHODS: This self-controlled case series study was conducted using electronic medical records at Kyoto University Hospital. We identified patients who were prescribed at least one PPI and macrolide antibiotic between January 2014 and December 2019 and underwent blood examinations at least once a year. An adjusted incident rate ratio (aIRR) of AKI with PPI use or concomitant use macrolide antibiotics with PPIs was estimated using a conditional Poisson regression model controlled for the estimated glomerular filtration rate at the beginning of observation and use of potentially nephrotoxic antibiotics. RESULTS: Of the 3,685 individuals who received PPIs and macrolide antibiotics, 766 patients with episodes of stage 1 or higher AKI were identified. Any stage of AKI was associated with PPI use (aIRR, 1.80 (95% confidence interval (CI) 1.60 to 2.04)). Stage 2 or higher AKI was observed in 279 cases, with an estimated aIRR of 2.01 (95% CI 1.57 to 2.58, for PPI use). For the period of concomitant use of macrolide antibiotics with PPIs compared with the period of PPIs alone, an aIRR of stage 1 or higher AKI was estimated as 0.82 (95% CI 0.60 to 1.13). CONCLUSIONS: Our findings added epidemiological information for the association between PPI use and an increased risk of stage 1 or higher AKI. However, we did not detect an association between the concomitant use of macrolide antibiotics and an increased risk of AKI in PPI users.

Everolimus reduces BK polyomavirus infection by suppressing its replication and spread of infection.

BK polyomavirus-associated nephropathy is one of serious complications in transplant recipients. Everolimus-a mammalian target of rapamycin inhibitor-has been shown to reduce the incidence of BK polyomavirus infection in transplant recipients. In this study, the effects of everolimus were examined on viral replication and the spread of infection in BK polyomavirus-infected cultures. BK polyomavirus replicated in renal and pulmonary cells, contrary to that in hepatocytes, and spread as diffusely scattered patterns of infected cells, unlike plaque formation through the cell-to-cell mode. BK polyomavirus is stable to heat up to 65 °C with a particle per infectivity ratio of 5000, and the replication cycle was for approximately 34 h. Everolimus administration remarkably reduced the viral replication to 20% in cells treated with 0.1-10 ng/mL, the concentration at which everolimus reached the serum of transplant recipients. In addition, it reduced the amount of viral capsid protein 1 at 5 ng/mL without reducing the ratio of viral capsid protein 1 versus ß-actin, and it also retained the pattern of viral capsid protein 1 localization in the nuclei. Everolimus suppressed the number of infected cells to 32.8% during a 14-day treatment, indicating the reduction of BK polyomavirus-infected cell mass to 18.8% of untreated cultures by modifying cellular functions. The reduction in the total number of BK polyomavirus infected cells by everolimus indicates that everolimus alleviates BK polyomavirus infection, including nephropathy in transplant recipients.

Impact of salivary and pancreatic amylase gene copy numbers on diabetes, obesity, and functional profiles of microbiome in Northern Japanese population.

Amylase genes reside in a structurally complex locus, and their copy numbers vary greatly, and several studies have reported their association with obesity. The mechanism of this effect was partially explained by changes in the oral and gut microbiome compositions; however, a detailed mechanism has been unclarified. In this study, we showed their association with diabetes in addition to obesity, and further discovered a plausible mechanism of this association based on the function of commensal bacteria. First, we confirmed that the amylase copy number in the population tends to be larger than that reported in other studies and that there is a positive association between obesity and diabetes (p = 1.89E-2 and 8.63E-3). Second, we identified that relative abundance of some genus level microbiome, Capnocytophaga, Dialister, and previously reported bacteria, were significantly associated with amylase copy numbers. Finally, through functional gene-set analysis using shotgun sequencing, we observed that the abundance of genes in the Acarbose pathway in the gut microbiome was significantly decreased with an increase in the amylase copy number (p-value = 5.80E-4). Our findings can partly explain the mechanism underlying obesity and diabetes in populations with high amylase copy numbers.

CBNplot: Bayesian network plots for enrichment analysis.

SUMMARY: When investigating gene expression profiles, determining important directed edges between genes can provide valuable insights in addition to identifying differentially expressed genes. In the subsequent functional enrichment analysis (EA), understanding how enriched pathways or genes in the pathway interact with one another can help infer the gene regulatory network (GRN), important for studying the underlying molecular mechanisms. However, packages for easy inference of the GRN based on EA are scarce. Here, we developed an R package, CBNplot, which infers the Bayesian network (BN) from gene expression data, explicitly utilizing EA results obtained from curated biological pathway databases. The core features include convenient wrapping for structure learning, visualization of the BN from EA results, comparison with reference networks, and reflection of gene-related information on the plot. As an example, we demonstrate the analysis of bladder cancer-related datasets using CBNplot, including probabilistic reasoning, which is a unique aspect of BN analysis. We display the transformability of results obtained from one dataset to another, the validity of the analysis as assessed using established knowledge and literature, and the possibility of facilitating knowledge discovery from gene expression datasets. AVAILABILITY AND IMPLEMENTATION: The library, documentation and web server are available at https://github.com/noriakis/CBNplot. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Acute aortic thrombosis in the ascending aorta after cisplatin-based chemotherapy for esophageal cancer: a case report.

BACKGROUND: The risk of thrombus development is considered to be increased by malignant tumors and chemotherapy. In addition, thrombosis of the ascending aorta is rare. We report a case of ascending aortic thrombectomy in a patient with esophageal cancer who developed ascending aortic thrombus after starting neoadjuvant chemotherapy, including operative findings and surgical treatment. CASE PRESENTATION: A 63-year-old man with esophageal cancer was administered chemotherapy comprising cisplatin plus 5-fluorouracil. A week after completing 1 cycle of chemotherapy, computed tomography angiography showed acute aortic thrombosis at the ascending aorta. The risk of embolization appeared high because the thrombosis was floating, so we performed emergency ascending aortic thrombectomy. The postoperative course was good and uncomplicated. A month after this surgery, the patient underwent surgery for esophageal cancer. As of 1 year after the cancer surgery, neither cancer nor thrombosis has recurred. CONCLUSION: We describe a case of acute aortic thrombosis in the ascending aorta after cisplatin-based chemotherapy, that was treated by aortic thrombectomy. The treatment strategy should depend on thrombus location and the condition of the patient, but surgical treatment should be considered where possible to achieve better prognosis.

Antiviral therapy for COVID-19: Derivation of optimal strategy based on past antiviral and favipiravir experiences.

Favipiravir, a broad-spectrum RNA-dependent RNA polymerase inhibitor, inhibits the replication of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) at significantly lower concentrations than the plasma trough levels achieved by the dosage adopted for influenza treatment and exhibits efficacy against coronavirus disease 2019 (COVID-19) pneumonia. Although high doses of favipiravir are required due to the molecule being a purine analog, its conversion into the active form in infected cells with active viral RNA synthesis enhances the antiviral specificity and selectivity as a chain terminator with lethal mutagenesis. Another characteristic feature is the lack of generation of favipiravir-resistant virus. COVID-19 pneumonia is caused by strong cell-mediated immunity against virus-infected cells, and the inflammatory response induced by adaptive immunity continues to peak for 3 to 5 days despite antiviral treatment. This has also been observed in herpes zoster (HZ) and cytomegalovirus (CMV) pneumonia. Inflammation due to an immune response may mask the effectiveness of favipiravir against COVID-19 pneumonia. Favipiravir significantly shortened the recovery time in patients with mild COVID-19 pneumonia by 3 days with the start of treatment by the 5th day of symptom onset. Since both CMV and COVID-19 pneumonia are caused by adaptive immunity and prevention of cytomegalovirus pneumonia is the standard treatment due to difficulties in treating refractory CMV pneumonia, COVID-19 pneumonia should be prevented with early treatment as well. In the present study, we have comprehensively reviewed the optimal antiviral therapy for COVID-19 based on clinical trials of favipiravir for the treatment of COVID-19 pneumonia and the concurrently established therapies for other viral infections, particularly HZ and CMV pneumonia. Optimally, antivirals should be administered immediately after COVID-19 diagnosis, similar to that after influenza diagnosis, to prevent COVID-19 pneumonia and complications resulting from microangiopathy.

Platypnea-orthodeoxia Syndrome Due to Right Ventricular Inflow Tract Obstruction Caused by an Elongated Ascending Aorta: Usefulness of Three-dimensional Cardiac Computed Tomography Imaging in the Sitting Position.

An 84-year-old woman presented with dyspnea in the sitting position. Platypnea-orthodeoxia syndrome (POS) was suspected based on arterial desaturation when her posture changed from the supine to the sitting position. Transesophageal echocardiography showed right-to-left shunting enhancement through a patent foramen ovale (PFO) in the sitting position. Three-dimensional (3D) cardiac CT in the sitting position revealed that the elongated ascending aorta compressed the right ventricular inflow tract, resulting in restricted blood flow to the right ventricle and increased right-to-left shunting. This case highlights the role of 3D-CT in the sitting position in the management of POS.

Maximum Standardized Uptake Value on Positron Emission Tomography is Associated With More Advanced Disease and High-risk Features in Lung Adenocarcinoma.

18F-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography (FDG-PET/CT) has been widely used for preoperative staging of lung adenocarcinomas. The aim of this study was to determine whether a high maximum standardized uptake value (SUVmax) could correlate with pathological characteristics in those patients. We retrospectively reviewed patients with clinical stage 0-IA lung adenocarcinoma who underwent preoperative 18F-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography followed by curative anatomical resection. To identify more advanced disease and high-risk features, representing visceral pleural involvement, pulmonary metastasis, lymph node involvement, and lymphovascular involvement in resected surgical specimens, univariate and multivariate logistic regression analyses were performed. The optimal cutoff point for the SUVmax was determined by receiver operating characteristic analysis. In 2 groups divided according to the cutoff point, the disease-free survivals were calculated and compared using the Kaplan-Meier method and the log-rank test. More advanced disease and high-risk features were identified in 55 (18.9%) of the 291 patients. SUVmax was significantly correlated with more advanced disease and high-risk features, as did the consolidation/tumor ratio on computed tomography. Only 2 (1.2%) of the 169 patients with a SUVmax <3.20 showed more advanced disease and high-risk features, compared with 43.4% of patients with a SUVmax ≥3.20. The disease-free survival was significantly higher in patients with a SUVmax <3.20 than in those with a SUVmax ≥3.20 (P = 0.002). A high SUVmax correlates with more advanced disease and high-risk features in patients with clinical stage 0-IA lung adenocarcinoma. The SUVmax should be considered when deciding treatment strategy in early-stage lung adenocarcinoma.

[Determination of Endograft Size in Thoracic Endovascular Aortic Repair for Blunt Thoracic Aortic Injury:Report of a Case].

Although the efficacy of thoracic endovascular aortic repair( TEVAR) for blunt thoracic aortic injury (BTAI) has been reported, the determination of the proper endograft size for TEVAR in BTAI cases is yet to be established. The management of BTAI by TEVAR may not be effective if a wrong endograft size is selected, possibly leading to fatal outcomes. In this case report, a hemodynamically unstable patient with BTAI was saved by TEVAR, and the preoperative and postoperative changes in the aortic diameter were compared. In TEVAR for BTAI, the effects of hemodynamic instability should be considered and the appropriate device size should be determined based on blood pressure, pulse, and inferior vena cava diameter.

Evaluation of Kidney Histological Images Using Unsupervised Deep Learning.

INTRODUCTION: Evaluating histopathology via machine learning has gained research and clinical interest, and the performance of supervised learning tasks has been described in various areas of medicine. Unsupervised learning of histological images has the advantage of reproducibility for labeling; however, the relationship between unsupervised evaluation and clinical information remains unclear in nephrology. METHODS: We propose an unsupervised approach combining convolutional neural networks (CNNs) and a visualization algorithm to cluster the histological images and calculate the score for patients. We applied the approach to the entire images or patched images of the glomerulus of kidney biopsy samples stained with hematoxylin and eosin obtained from 68 patients with immunoglobulin A nephropathy. We assessed the relationship between the obtained scores and clinical variables of urinary occult blood, urinary protein, serum creatinine (SCr), systolic blood pressure, and age. RESULTS: The glomeruli of the patients were classified into 12 distinct classes and 10 patches. The output of the fine-tuned CNN, which we defined as the histological scores, had significant relationships with assessed clinical variables. In addition, the clustering and visualization results suggested that the defined clusters captured important findings when evaluating renal histopathology. For the score of the patch-based cluster containing crescentic glomeruli, SCr (coefficient = 0.09, P = 0.019) had a significant relationship. CONCLUSION: The proposed approach could successfully extract features that were related to the clinical variables from the kidney biopsy images along with the visualization for interpretability. The approach could aid in the quantified evaluation of renal histopathology.