Objective To explore artificial intelligence (AI)?based medical image analysis for constructing a renal interstitial fibrosis-diagnostic model by using ultrasound images, and provide a novel non-invasive and rapid tool in assisting the clinical identification of renal interstitial fibrosis. Methods It was a retrospective study. Patients with chronic kidney disease (CKD) who underwent renal biopsy in the Department of Nephrology, Shanghai Tongji Hospital from January 2019 to February 2025 were enrolled as study participants. According to the Banff criteria, patients were categorized into mild fibrosis and moderate-to-severe fibrosis groups. The simple random sampling method was used to divide the included patients into a training set and a test set in a 7∶3 ratio. In addition, CKD patients who underwent renal biopsy in Shanghai Ninth People's Hospital of Shanghai Jiao Tong University School of Medicine from December 2024 to December 2025 were collected as an external validation set. Python software was used to manually segment renal parenchyma, and extract radiomic features. Mann-Whitney U test and LASSO regression algorithm were used to select key radiomics biomarkers of mild and moderate fibrosis groups in the training set. Five machine learning algorithms—support vector machine, logistic regression, random forest (RF), adaptive boosting, and multi-layer perceptron—were employed to construct both radiomic-only models and radiomics-clinical characteristics combined models, respectively. The area under the receiver- operating characteristic curve (AUC), accuracy, sensitivity, and specificity were used to comprehensively evaluate the predictive performance of the models. Calibration curve was used to validate model performance, and decision curve analysis was employed to evaluate the clinical value of the models. The Shapley additive explanation method was used to quantify feature contributions of the optimal model. Results Of the 472 included patients, 148 patients (31.4%) were diagnosed with moderate-to-severe fibrosis. A total of 1 704 radiomic features were extracted from grayscale ultrasound images, 120 features demonstrated good repeatability with intraclass correlation coefficient >0.75 and 8 critical features were retained. Five machine learning models developed based on these radiomic features showed moderate discriminative ability. Among them, the support vector machine, multilayer perceptron, and adaptive boosting models achieved an AUC>0.70 in both the testing and external validation cohorts. By integrating three clinical variables, including a history of diabetes mellitus, estimated glomerular filtration rate, and hemoglobin, with the radiomic features, radiomics-clinical combined models were developed and showed improved diagnostic performance. Among these models, the RF-based combined model achieved the best diagnostic performance, with AUC of 0.89 [95% confidence interval (CI) 0.81-0.95], accuracy of 0.79, sensitivity of 0.68, and specificity of 1.00 in the testing cohort; and AUC of 0.91 (95% CI 0.82-0.96), accuracy of 0.81, sensitivity of 0.70, and specificity of 0.96 in the external validation cohort. Subgroup analysis of patients with CKD stage 2-3a revealed that the RF-based combined model remained the most effective, with AUC of 0.94 (95% CI 0.79-0.99), accuracy of 0.91, sensitivity of 0.80, and specificity of 1.00. Conclusions The AI-based ultrasound diagnosis model can identify image texture changes induced by renal interstitial fibrosis and exhibits favorable discriminative performance in patients with CKD stage 2-3a.
Objective To explore the application value of remote dielectric sensing (ReDS) in assessing volume status and predicting adverse outcome events in maintenance hemodialysis (MHD) patients. Methods This was a prospective observational study. Patients who received regular hemodialysis treatment at the Hemodialysis Center of the Second Affiliated Hospital of Naval Medical University from April to June 2024 were enrolled as research subjects. All patients underwent ReDS examination before and after dialysis to obtain ReDS values, and received body composition measurement (BCM) based on bioimpedance spectroscopy (BIS) to acquire overhydration (OH) values as the reference standard. The clinical data and composite endpoint events (defined as cardiovascular events or non-cardiovascular death) occurring during follow-up were collected. Spearman or Pearson correlation analysis was used to analyze the correlation between ReDS values and echocardiographic parameters. Receiver-operating characteristic (ROC) curve was applied to evaluate the feasibility of ReDS values in assessing severe volume overload (OH value≥2.5 L). Kaplan-Meier method was used to analyze the event-free survival rate of MHD patients, and Log-rank test was used to compare the differences of survival rates between groups. Cox regression model was adopted to analyze the value of ReDS values in predicting the risk of composite endpoint events. Results A total of 92 patients were finally included in the study, with age of (60.71±11.61) years, including 48 males (52.2%). There were statistically significant differences in body weight (F=3.768, P=0.027), dialysis vintage (H=9.927, P=0.007), hemoglobin (H=7.084, P=0.029), ultrafiltration volume (F=23.507, P<0.001), pre-dialysis ReDS values (F=4.778, P=0.011), pre-dialysis systolic blood pressure (F=3.140, P=0.049), pre-dialysis diastolic blood pressure (F=4.203, P=0.018) and post-dialysis diastolic blood pressure (F=4.199, P=0.018) among the normal volume group (-1.0 L≤OH value≤1.0 L, n=16), mild volume overload group (1.0 L<OH value<2.5 L, n=40) and severe volume overload group (OH value≥2.5 L, n=36). ReDS values (t=-8.459, P<0.001), OH values (Z=-8.286, P<0.001), systolic blood pressure (t=-6.555, P<0.001), diastolic blood pressure (t=-5.586, P<0.001) and body weight (t=-25.267, P<0.001) were significantly decreased after dialysis compared with pre-dialysis levels. Correlation analysis showed that both pre- and post-dialysis ReDS values were positively correlated with OH values (r=0.310, P=0.003; r=0.334, P=0.001), and the decreased amplitude of ReDS values was positively correlated with ultrafiltration volume (r=0.219, P=0.036). ROC curve analysis showed that the cutoff value of pre-dialysis ReDS values for distinguishing severe volume overload was 32.5%, with area under the curve of 0.812 [95% confidence interval (CI) 0.706-0.917, P<0.001], sensitivity of 0.714, and specificity of 0.808. Compared with ReDS value<32.5% group, ReDS value≥32.5% group had higher proportion of males (χ2=4.151, P=0.042), height (Z=2.545, P=0.011), body weight (t=2.835, P=0.007), body mass index (t=2.214, P=0.031), pre-dialysis OH values (Z=3.843, P<0.001) and post-dialysis OH values (Z=3.455, P<0.001), but lower age (t=-2.446, P=0.016). The mean follow-up time was 458 days. Kaplan-Meier survival analysis showed that event-free survival rate in the ReDS value≥32.5% group was significantly lower than that in ReDS value<32.5% group (Log-rank test, χ2=14.812, P<0.001), and the risk of the composite endpoint events in ReDS value≥32.5% group was significantly higher [taking ReDS value<32.5% group as the reference group, hazard ratio (HR)=5.795, 95% CI 2.220-15.127]. Multivariate Cox regression analysis showed that ReDS value≥32.5% was an independent influencing factor of composite endpoint events in MHD patients (taking ReDS value<32.5% group as the reference group, HR=7.079, 95% CI 2.288-21.902, P<0.001). Conclusion ReDS can effectively assess the volume status of MHD patients and is available for predicting adverse outcome events.
Objective To investigate the incidence and related factors of acute kidney injury (AKI) after cerebrovascular angiography intervention. Methods This study was a retrospective cohort study. It included patients hospitalized in the general ward of the Department of Neurosurgery at Xuanwu Hospital, Capital Medical University, from January 1, 2017, to December 31, 2020, who underwent interventional treatment for cerebrovascular stenosis or occlusion. Patients were divided into an AKI group and a non-AKI group according to the AKI diagnostic criteria, and intergroup differences in general data, laboratory test results, and preoperative medication were analyzed. Weighted multivariate logistic regression was used to analyze the factors associated with AKI occurrence in patients after cerebrovascular angiography intervention. Results A total of 1 017 patients undergoing cerebrovascular angiography intervention were ultimately included in this study, with an average age of (52.47±16.35) years. Among them, 573 were male (56.3%), and 44 developed AKI, accounting for 4.3%. The hemoglobin level in the AKI group was significantly lower than that in the non-AKI group [ (127.77±18.33) g/L vs. (134.49±15.95) g/L, t=2.713, P=0.007], while fasting plasma glucose level was significantly higher in the AKI group [5.89 (4.77, 8.04) mmol/L vs. 5.04 (4.51, 6.10) mmol/L, Z=3.242, P=0.001]. Regarding preoperative medication, the AKI group showed significantly higher proportions of nonsteroidal anti-inflammatory drugs (χ2=6.540, P=0.011), proton pump inhibitors (χ2=4.878, P=0.027), furosemide (χ2=14.652, P<0.001), mannitol (χ2=7.098, P=0.008), and β-blockers (χ2=9.202, P=0.002) compared to the non-AKI group. Weighted multivariate logistic regression analysis showed that fasting blood glucose (OR=1.162, 95% CI 1.055-1.280, P=0.002) and the preoperative use of beta blockers (OR=2.540, 95% CI 1.261-5.118, P=0.009) were independent factors associated with AKI after cerebrovascular angiography intervention. Conclusion The incidence of AKI after cerebrovascular angiography intervention is 4.3%. Elevated fasting blood glucose and preoperative use of β-blockers are related factors for AKI after cerebrovascular angiography intervention. However, given the limited number of AKI events and the single-center retrospective study design, the generalizability of this conclusion should be cautiously considered.
Objective To investigate the correlation of serum lactic acid and ferroptosis-related biomarkers with the occurrence and renal impairment of diabetic kidney disease (DKD). Methods Eighty-four patients diagnosed with DKD (DKD group), 67 patients with type 2 diabetes mellitus (T2DM group), and 39 healthy controls (Control group) were enrolled from the First Affiliated Hospital of Zhengzhou University between January 2021 and October 2023. Renal tissue samples were obtained from patients pathologically diagnosed with diabetic nephropathy (DN) and from patients who underwent nephrectomy in the Department of Urology of the same hospital (tissue obtained >3 cm away from the tumor margin), with 3 cases in each group. Immunohistochemistry was employed to detect the expression levels of pyruvate kinase M2 (PKM2), lactate dehydrogenase (LDH), and glutathione peroxidase 4 (GPX4) in the renal tissues of both groups. Clinical data were collected, and serum levels of lactic acid, PKM2, interleukin-6 (IL-6), and ferroptosis-related indicators, including hepcidin (Hepc), ferrous ion (Fe2+), glutathione (GSH), and malondialdehyde (MDA), were measured. Correlations between these biomarkers and renal function indicators were analyzed. Multivariable logistic regression analysis was performed to identify factors associated with DKD. Results (1) Immunohistochemistry showed increased PKM2 and LDH and decreased GPX4 expression in DN renal tissue. (2)Compared with the T2DM or control groups, the DKD group had significantly higher levels of serum lactic acid, PKM2, IL?6, Hepc, Fe2?, and MDA (all P<0.05), and significantly lower GSH level (P<0.05). (3) Stratification by estimated glomerular filtration rate (eGFR) showed that patients in stage 5 had significantly higher levels of lactic acid, PKM2, IL?6, Hepc, Fe2?, and MDA than those in stage 1-2 or stage 3-4 (all P<0.05), while GSH level was significantly lower in stage 5 than in stage 1-2 or stage 3-4 (all P<0.05). (4) Spearman correlation analysis revealed that serum lactic acid, PKM2, and IL?6 were positively correlated with serum creatinine, urinary albumin?to?creatinine ratio (UACR), and 24?hour urinary protein (all P<0.05), and negatively correlated with eGFR (all P<0.05), respectively. GSH was positively correlated with eGFR (r=0.640, P<0.001), and negatively correlated with serum creatinine, uric acid, UACR, and 24?hour urinary protein (all P<0.05), respectively. MDA was positively correlated with serum creatinine, uric acid, and 24?hour urinary protein (all P<0.05), and negatively correlated with eGFR (r=-0.794, P<0.001). Fe2? was positively correlated with serum creatinine and 24?hour urinary protein (both P<0.05), and negatively correlated with eGFR (r=-0.460, P=0.001). Hepc was positively correlated with serum creatinine, UACR, and 24?hour urinary protein (all P<0.05), and negatively correlated with eGFR (r=-0.607, P<0.001). Lactic acid was positively correlated with IL?6 (r=0.789, P<0.001), MDA (r=0.601, P<0.001), Fe2? (r=0.409, P=0.008), and Hepc (r=0.464, P=0.002), while negatively correlated with GSH (r=-0.414, P<0.001). (5) Multivariate logistic regression showed that age ≥40 years [with age <40 years as the reference, odds ratio (OR) = 5.105, 95% confidence interval (CI) 1.280-20.356, P=0.021], elevated levels of blood urea (OR=1.680，95% CI 1.250-2.259，P<0.001), serum creatinine (OR=1.057，95% CI 1.021-1.094，P=0.002), uric acid (OR=1.013，95% CI 1.005- 1.020，P<0.001), lactic acid (OR=3.229，95% CI 1.219-8.549，P=0.018), IL?6 (OR=1.178，95% CI 1.025- 1.353，P=0.020), PKM2 (OR=1.009，95% CI 1.002-1.017，P=0.015), Fe2? (OR=1.039，95% CI 1.005- 1.074，P=0.026), and Hepc (OR=1.001，95% CI 1.000-1.001，P=0.004), as well as decreased eGFR (OR=0.947，95% CI 0.923-0.973，P<0.001) and GSH (OR=0.008，95% CI 0.001-0.557，P=0.026), were independently associated with an increased risk of DKD in T2DM patients. Conclusion DKD patients exhibit pronounced lactic acid accumulation and ferroptosis imbalance, both of which are closely correlated with the severity of renal impairment. Monitoring serum lactic acid and ferroptosis-related biomarkers may facilitate the assessment of DKD development and progression.
Objective To investigate the molecular mechanism by which tripartite motif-containing protein 21 (TRIM21) regulates lipopolysaccharide (LPS)-induced inflammatory injury in peritoneal mesothelial cells (PMC) through the peroxisome proliferator?activated receptor gamma (PPARγ) signaling pathway, and to explore the downstream involvement of cluster of differentiation 14 (CD14) in this process. Methods A human PMC line stably overexpressing TRIM21 (H125-TRIM21) and its corresponding vector control cell line (H125-Vector) were established. An acute inflammatory cell model was induced by stimulation with LPS (1 μg/ml). Quantitative real?time PCR was performed to determine the mRNA expression levels of tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, IL-6, IL-8, and CD14. Western blotting was used to detect TRIM21 and PPARγ protein expression. Immunofluorescence staining was performed to assess the subcellular localization of TRIM21 and PPARγ. Co?immunoprecipitation was used to verify the interaction between the two proteins, and PPARγ ubiquitination levels were assessed via immunoblotting following ubiquitination assays. Results Compared with the vector control group, TRIM21 overexpression significantly increased the mRNA expression levels of IL-1β (t=3.891, P=0.018), IL-6 (t=12.047, P<0.001), and IL?8 (t=6.781, P<0.001). Conversely, TRIM21 knockdown significantly suppressed the expression of these inflammatory cytokines (all P<0.05), indicating that TRIM21 promotes the inflammatory response in PMC. Immunofluorescence staining showed that TRIM21 was predominantly localized in the cytoplasm and markedly colocalized with PPARγ. Co-immunoprecipitation further confirmed a protein?protein interaction between TRIM21 and PPARγ. Western blotting showed that TRIM21 overexpression-induced significantly decreased PPARγ protein expression (t=3.119, P=0.036), whereas its mRNA level was not significantly altered, suggesting that TRIM21 regulates PPARγ mainly at the post-transcriptional level. Treatment with the proteasome inhibitor MG132 reversed the TRIM21 overexpression-induced reduction in PPARγ protein levels. Moreover, ubiquitination assays demonstrated that TRIM21 overexpression markedly enhanced PPARγ ubiquitination, indicating that TRIM21 might promote PPARγ degradation through the ubiquitin proteasome pathway. Under LPS stimulation, TRIM21 overexpression further reduced the nuclear localization of PPARγ (t=6.851, P<0.01), suggesting that PPARγ nuclear translocation was impaired. In addition, TRIM21 overexpression significantly increased CD14 mRNA expression (t=15.24, P=0.004), whereas PPARγ overexpression suppressed CD14 expression (t=12.49, P=0.006), indicating that TRIM21 and PPARγ exert opposing regulatory effects on CD14 transcription. Conclusion TRIM21 interacts with PPARγ and promotes its proteasome?dependent degradation by enhancing PPARγ ubiquitination, thereby reducing PPARγ protein levels and impairing its nuclear translocation. Meanwhile, TRIM21 and PPARγ exert opposing effects on CD14 expression, ultimately enhancing the LPS-induced inflammatory response in PMC. The TRIM21- PPARγ-CD14 regulatory axis may contribute to the amplification of inflammation in Gram?negative bacterial peritonitis and may represent a potential therapeutic target.
Lipoprotein glomerulopathy (LPG) is a rare glomerular disease for which there is currently no universally accepted standard treatment regimen, and glucocorticoids and immunosuppressive agents are generally ineffective. This article reports a 14-year-old male patient with LPG, whose main clinical manifestations were foamy urine and edema. Renal pathology showed positive Oil Red O staining, and electron microscopy revealed proteinaceous substance deposition with lipid-like vacuoles within the glomerular capillary loops. Genetic testing identified a heterozygous variant in the apolipoprotein E gene. After treatment with dapagliflozin combined with finerenone, the patient's clinical symptoms and proteinuria improved significantly, renal function remained stable, and no adverse reactions, including hyperkalemia, were observed. This report suggests that dapagliflozin combined with finerenone may represent a novel therapeutic approach for LPG.
Pulmonary hypertension (PH) is a severe cardiovascular complication in patients with end-stage renal disease and is closely associated with poor prognosis. The majority of PH cases are postcapillary, related to volume overload and left ventricular dysfunction, whereas precapillary PH confirmed by right heart catheterization (RHC) is relatively rare. This article reports a 40-year-old male peritoneal dialysis patient admitted with progressively worsening dyspnea. After optimization of volume status, RHC was performed, confirming precapillary PH. Individualized targeted therapy with sildenafil combined with bosentan relieved symptoms and significantly reduced pulmonary artery pressure. This case is discussed in the context of the literature, aiming to improve clinicians' understanding of this disease and provide a reference for the diagnosis and management of such patients.
This study aimed to evaluate the efficacy of rituximab (RTX) in the treatment of relapsing minimal change disease (MCD) and to explore relapse?related factors. A retrospective analysis was conducted on 17 MCD cases who received RTX for the first time after relapse. Clinical characteristics and immunophenotypes were compared between relapse and non-relapse groups. The Kaplan-Meier method was used to analyze relapse- free survival, and univariate Cox regression was performed to analyze relapse-related factors. The follow-up duration was (30.94±8.80) months. Among the 16 patients who achieved complete remission, 7 patients relapsed during follow-up. The annual relapse rate significantly decreased from 1.90(1.06, 2.57) episodes/year before RTX treatment to 0(0, 0.41) episodes/year after treatment (Z=-3.41, P<0.001). The median relapse-free survival time was 39.23(95% CI 28.16-50.31) months. The differences in clinical characteristics, immunophenotype, and other indices between the relapse and non-relapse groups were not statistically significant.Univariate Cox regression showed that the annual relapse rate before RTX treatment was borderline significantly associated with relapse (HR=1.720, 95% CI 0.97-3.04, P=0.062). Due to the limited sample size (few events), multivariate analysis was not performed. RTX can effectively maintain remission in relapsing MCD. This study did not identify definite predictive risk factors for relapse. The annual relapse rate before treatment may be a potential warning indicator, but these results require further validation in large?scale studies.
Trogocytosis is a biological process in which membrane fragments and their associated functional molecules—including membrane proteins and lipid rafts—are transferred from donor cells to recipient cells via direct cell—cell contact. This process is broadly involved in immune recognition, antigen presentation, and intercellular communication. Emerging evidence highlights its dual biological effects: while it facilitates information transfer between immune cells, it can also compromise the clinical efficacy of biologics by mediating target antigen loss and inducing functional exhaustion of immune cells. Consequently, trogocytosis has emerged as one of the key mechanisms underlying resistance to monoclonal antibodies and CAR?T/NK cell immunotherapies. This review systematically summarizes the biological features and molecular mechanisms of trogocytosis, with a particular emphasis on its roles in promoting immune evasion and immunotherapy resistance. We also discuss targeted intervention strategies and their potential implications for kidney diseases, aiming to provide new perspectives and insights for improving the therapeutic efficacy of biologic agents.
With the widespread adoption and technical advancement of peritoneal dialysis (PD), the proportion of long-term PD patients (dialysis duration >5 years) is increasing. These patients often have virtually lost residual renal function, decreased peritoneal transport function, and are prone to malnutrition, chronic kidney disease-mineral and bone disorder (CKD-MBD), cardiovascular diseases, refractory peritonitis, and encapsulating peritoneal sclerosis (EPS), which seriously threaten their survival and quality of life. This article reviews the comprehensive management strategies for long-term PD patients based on their main clinical characteristics. For solute and fluid clearance, attention should be paid to peritoneal functional deterioration, rational selection of biocompatible PD solutions and individualized regimens, and timely transition to alternative therapies. For peritonitis, emphasis should be placed on prevention, patient re?education, early recognition of severe and refractory cases, and optimization of antibiotic therapy. For CKD-MBD, screening and monitoring of vascular calcification and calciphylaxis should be strengthened, and pharmacological and surgical interventions should be standardized. Meanwhile, malnutrition, sarcopenia, and psychological disorders should be actively managed with a multidisciplinary approach. Early identification and monitoring of high risk EPS populations are essential, with timely medical and nutritional support, and surgical referral when necessary. This article emphasizes that individualized long-term management, focusing on adequate dialysis, infection prevention, improvement of nutrition and bone metabolism, peritoneal function protection, and prevention of cardiovascular events, combined with regular follow-up and telemanagement, should be developed with the active participation of both patients and the multidisciplinary team, in order to improve the prognosis of long-term PD patients.
Kidney stones are a global health issue and have emerged as a significant contributor to chronic kidney disease. Given their strong association with metabolic abnormalities, high prevalence, and high recurrence rate, comprehensive medical management throughout the disease course is essential for correcting metabolic derangements, preventing recurrence, and reducing complications. To address this need, the Chinese Society of Nephrology organized a panel of experts to develop this guideline. This guideline formulates 28 recommendations covering the epidemiology, classification, risk factors, diagnostic evaluation, metabolic assessment,, treatment, follow-up, and special population management of kidney stones based on the Grading of Recommendations Assessment, Development and Evaluation. The aim is to provide clinicians with a systematic and standardized framework for the medical management of kidney stone disease.
