Urological disorders in chronic kidney disease in children cohort: clinical characteristics and estimation of glomerular filtration rate

A narrative review on the application of artificial intelligence in renal ultrasound

Kidney disease is a serious public health problem and various kidney diseases could progress to end-stage renal disease. The many complications of end-stage renal disease. have a significant impact on the physical and mental health of patients. Ultrasound can be the test of choice for evaluating the kidney and perirenal tissue as it is real-time, available and non-radioactive. To overcome substantial interobserver variability in renal ultrasound interpretation, artificial intelligence (AI) has the potential to be a new method to help radiologists make clinical decisions. This review introduces the applications of AI in renal ultrasound, including automatic segmentation of the kidney, measurement of the renal volume, prediction of the kidney function, diagnosis of the kidney diseases. The advantages and disadvantages of the applications will also be presented clinicians to conduct research. Additionally, the challenges and future perspectives of AI are discussed.

Automated Society of Fetal Urology (SFU) grading of hydronephrosis on ultrasound imaging using a convolutional neural network

INTRODUCTION

Grading of hydronephrosis severity on postnatal renal ultrasound guides management decisions in antenatal hydronephrosis (ANH). Multiple systems exist to help standardize hydronephrosis grading, yet poor inter-observer reliability persists. Machine learning methods may provide tools to improve the efficiency and accuracy of hydronephrosis grading.

OBJECTIVE

To develop an automated convolutional neural network (CNN) model to classify hydronephrosis on renal ultrasound imaging according to the Society of Fetal Urology (SFU) system as potential clinical adjunct.

STUDY DESIGN

A cross-sectional, single-institution cohort of postnatal renal ultrasounds with radiologist SFU grading from pediatric patients with and without hydronephrosis of stable severity was obtained. Imaging labels were used to automatedly select sagittal and transverse grey-scale renal images from all available studies from each patient. A VGG16 pre-trained ImageNet CNN model analyzed these preprocessed images. Three-fold stratified cross-validation was used to build and evaluate the model that was used to classify renal ultrasounds on a per patient basis into five classes based on the SFU system (normal, SFU I, SFU II, SFU III, or SFU IV). These predictions were compared to radiologist grading. Confusion matrices evaluated model performance. Gradient class activation mapping demonstrated imaging features driving model predictions.

RESULTS

We identified 710 patients with 4659 postnatal renal ultrasound series. Per radiologist grading, 183 were normal, 157 were SFU I, 132 were SFU II, 100 were SFU III, and 138 were SFU IV. The machine learning model predicted hydronephrosis grade with 82.0% (95% CI: 75-83%) overall accuracy and classified 97.6% (95% CI: 95-98%) of the patients correctly or within one grade of the radiologist grade. The model classified 92.3% (95% CI: 86-95%) normal, 73.2% (95% CI: 69-76%) SFU I, 73.5% (95% CI: 67-75%) SFU II, 79.0% (95% CI: 73-82%) SFU III, and 88.4% (95% CI: 85-92%) SFU IV patients accurately. Gradient class activation mapping demonstrated that the ultrasound appearance of the renal collecting system drove the model's predictions.

DISCUSSION

The CNN-based model classified hydronephrosis on renal ultrasounds automatically and accurately based on the expected imaging features in the SFU system. Compared to prior studies, the model functioned more automatically with greater accuracy. Limitations include the retrospective, relatively small cohort, and averaging across multiple imaging studies per patient.

CONCLUSIONS

An automated CNN-based system classified hydronephrosis on renal ultrasounds according to the SFU system with promising accuracy based on appropriate imaging features. These findings suggest a possible adjunctive role for machine learning systems in the grading of ANH.

Deep learning imaging features derived from kidney ultrasounds predict chronic kidney disease progression in children with posterior urethral valves

BACKGROUND

We sought to use deep learning to extract anatomic features from postnatal kidney ultrasounds and evaluate their performance in predicting the risk and timing of chronic kidney disease (CKD) progression for boys with posterior urethral valves (PUV). We hypothesized that these features would predict CKD progression better than clinical characteristics such as nadir creatinine alone.

METHODS

We performed a retrospective cohort study of boys with PUV treated at two pediatric health systems from 1990 to 2021. Features of kidneys were extracted from initial postnatal kidney ultrasound images using a deep learning model. Three time-to-event prediction models were built using random survival forests. The Imaging Model included deep learning imaging features, the Clinical Model included clinical data, and the Ensemble Model combined imaging features and clinical data. Separate models were built to include time-dependent clinical data that were available at 6 months, 1 year, 3 years, and 5 years.

RESULTS

Two-hundred and twenty-five patients were included in the analysis. All models performed well with C-indices of 0.7 or greater. The Clinical Model outperformed the Imaging Model at all time points with nadir creatinine driving the performance of the Clinical Model. Combining the 6-month Imaging Model (C-index 0.7; 95% confidence interval [CI] 0.6, 0.79) with the 6-month Clinical Model (C-index 0.79; 95% CI 0.71, 0.86) resulted in a 6-month Ensemble Model that performed better (C-index 0.82; 95% CI 0.77, 0.88) than either model alone.

CONCLUSIONS

Deep learning imaging features extracted from initial postnatal kidney ultrasounds may improve early prediction of CKD progression among children with PUV. A higher resolution version of the Graphical abstract is available as Supplementary information.

Role of necroptosis in kidney health and disease

Cell death, particularly that of tubule epithelial cells, contributes critically to the pathophysiology of kidney disease. A body of evidence accumulated over the past 15 years has ascribed a central pathophysiological role to a particular form of regulated necrosis, termed necroptosis, to acute tubular necrosis, nephron loss and maladaptive renal fibrogenesis. Unlike apoptosis, which is a non-immunogenic process, necroptosis results in the release of cellular contents and cytokines, which triggers an inflammatory response in neighbouring tissue. This necroinflammatory environment can lead to severe organ dysfunction and cause lasting tissue injury in the kidney. Despite evidence of a link between necroptosis and various kidney diseases, there are no available therapeutic options to target this process. Greater understanding of the molecular mechanisms, triggers and regulators of necroptosis in acute and chronic kidney diseases may identify shortcomings in current approaches to therapeutically target necroptosis regulators and lead to the development of innovative therapeutic approaches.

Multi-instance Deep Learning of Ultrasound Imaging Data for Pattern Classification of Congenital Abnormalities of the Kidney and Urinary Tract in Children

OBJECTIVE

To reliably and quickly diagnose children with posterior urethral valves (PUV), we developed a multi-instance deep learning method to automate image analysis.

METHODS

We built a robust pattern classifier to distinguish 86 children with PUV from 71 children with mild unilateral hydronephrosis based on ultrasound images (3504 in sagittal view and 2558 in transverse view) obtained during routine clinical care.

RESULTS

The multi-instance deep learning classifier performed better than classifiers built on either single sagittal images or single transverse images. Particularly, the deep learning classifiers built on single images in the sagittal view and single images in the transverse view obtained area under the receiver operating characteristic curve (AUC) values of 0.796 ± 0.064 and 0.815 ± 0.071, respectively. AUC values of the multi-instance deep learning classifiers built on images in the sagittal and transverse views with mean pooling operation were 0.949 ± 0.035 and 0.954 ± 0.033, respectively. The multi-instance deep learning classifiers built on images in both the sagittal and transverse views with a mean pooling operation obtained an AUC of 0.961 ± 0.026 with a classification rate of 0.925 ± 0.060, specificity of 0.986 ± 0.032, and sensitivity of 0.873 ± 0.120, respectively. Discriminative regions of the kidney located using classification activation mapping demonstrated that the deep learning techniques could identify meaningful anatomical features from ultrasound images.

CONCLUSION

The multi-instance deep learning method provides an automatic and accurate means to extract informative features from ultrasound images and discriminate infants with PUV from male children with unilateral hydronephrosis.

Computer-Aided Diagnosis of Congenital Abnormalities of the Kidney and Urinary Tract in Children Using a Multi-Instance Deep Learning Method Based on Ultrasound Imaging Data

Ultrasound images are widely used for diagnosis of congenital abnormalities of the kidney and urinary tract (CAKUT). Since a typical clinical ultrasound image captures 2D information of a specific view plan of the kidney and images of the same kidney on different planes have varied appearances, it is challenging to develop a computer aided diagnosis tool robust to ultrasound images in different views. To overcome this problem, we develop a multi-instance deep learning method for distinguishing children with CAKUT from controls based on their clinical ultrasound images, aiming to automatic diagnose the CAKUT in children based on ultrasound imaging data. Particularly, a multi-instance deep learning method was developed to build a robust pattern classifier to distinguish children with CAKUT from controls based on their ultrasound images in sagittal and transverse views obtained during routine clinical care. The classifier was built on imaging features derived using transfer learning from a pre-trained deep learning model with a mean pooling operator for fusing instance-level classification results. Experimental results have demonstrated that the multi-instance deep learning classifier performed better than classifiers built on either individual sagittal slices or individual transverse slices.

Laboratory Findings After Urinary Tract Infection and Antimicrobial Prophylaxis in Children With Vesicoureteral Reflux

It is a common practice to monitor blood tests in patients receiving long-term trimethoprim-sulfamethoxazole (TMP-SMZ) prophylaxis for recurrent urinary tract infections. This multicenter, randomized, placebo-controlled trial enrolled 607 children aged 2 to 71 months with vesicoureteral reflux diagnosed after symptomatic urinary tract infection. Study participants received TMP-SMZ (n = 302) or placebo (n = 305) and were followed for 2 years. Serum electrolytes (n ≥ 370), creatinine (n = 310), and complete blood counts (n ≥ 206) were measured at study entry and at the 24-month study conclusion. We found no significant electrolyte, renal, or hematologic abnormalities when comparing the treatment and placebo groups. We observed changes in several laboratory parameters in both treatment and placebo groups as would normally be expected with physiologic maturation. Changes were within the normal range for age. Long-term use of TMP-SMX had no treatment effect on complete blood count, serum electrolytes, or creatinine. Our findings do not support routine monitoring of these laboratory tests in children receiving long-term TMP-SMZ prophylaxis.

Chronic kidney disease and urological disorders: systematic use of uroflowmetry in nephropathic patients

Background

Chronic kidney disease (CKD) is a highly prevalent condition. Urologic disorders are known causes of CKD, but often remain undiagnosed and underestimated also for their insidious onset and slow progression. We aimed to evaluate the prevalence of urological unrecognized diseases in CKD patients by uroflowmetry.

Methods

We enrolled consecutive stable CKD outpatients. The patients carried out two questionnaires, the International Prostate Symptom Score and Incontinence Questionnaire-Short Form, and they also underwent uroflowmetry, evaluating max flow rate (Q_max), voiding time and voided volume values.

Results

A total of 83 patients (43 males, mean age of 59.8 ± 13.3 years) were enrolled. Our study showed 28 males and 10 females with a significant reduction of Q_max (P < 0.001) while 21 females reported a significant increase of Q_max (P < 0.001) with a prevalence of 49.5% of functional urological disease. Moreover, we showed a significant association between Q_max and creatinine (P = 0.013), estimated glomerular filtration rate (P = 0.029) and voiding volume (P = 0.05). We have not shown significant associations with age (P = 0.215), body mass index (P = 0.793), systolic blood pressure (P = 0.642) or diastolic blood pressure (P = 0.305). Moreover, Pearson’s chi-squared test showed a significant association between Q_max altered with CKD (χ²  = 1.885, P = 0.170) and recurrent infection (χ² = 8.886, P = 0.012), while we have not shown an association with proteinuria (χ² = 0.484, P = 0.785), diabetes (χ² = 0.334, P = 0.563) or hypertension (χ² = 1.885, P = 0.170).

Conclusions

We showed an elevated prevalence of urological diseases in nephropathic patients; therefore, we suggest to include uroflowmetry in CKD patient assessment, considering the non-invasiveness, repeatability and low cost of examination. Uroflowmetry could be used to identify previously unrecognized urological diseases, which may prevent the onset of CKD or progression to end-stage renal disease and reduce the costs of management.

Urologic care and progression to end-stage kidney disease: a Chronic Kidney Disease in Children (CKiD) nested case-control study

INTRODUCTION

Children with chronic kidney disease (CKD) risk progressing to end-stage kidney disease (ESKD). The majority of CKD causes in children are related to congenital anomalies of the kidney and urinary tract, which may be treated by urologic care.

OBJECTIVE

To examine the association of ESKD with urologic care in children with CKD.

STUDY DESIGN

This was a nested case-control study within the Chronic Kidney Disease in Children (CKiD) prospective cohort study that included children aged 1-16 years with non-glomerular causes of CKD. The primary exposure was prior urologic referral with or without surgical intervention. Incidence density sampling matched each case of ESKD to up to three controls on duration of time from CKD onset, sex, race, age at baseline visit, and history of low birth weight. Conditional logistic regression analysis was performed to estimate rate ratios (RRs) for the incidence of ESKD.

RESULTS

Sixty-six cases of ESKD were matched to 153 controls. Median age at baseline study visit was 12 years; 67% were male, and 7% were black. Median follow-up time from CKD onset was 14.9 years. Seventy percent received urologic care, including 100% of obstructive uropathy and 96% of reflux nephropathy diagnoses. Cases had worse renal function at their baseline visit and were less likely to have received prior urologic care. After adjusting for income, education, and insurance status, urology referral with surgery was associated with 50% lower risk of ESKD (RR 0.50 [95% confidence interval [CI] 0.26-0.997), compared to no prior urologic care (Figure). After excluding obstructive uropathy and reflux nephropathy diagnoses, which were highly correlated with urologic surgery, the association was attenuated (RR 0.72, 95% CI 0.24-2.18).

DISCUSSION

In this study, urologic care was commonly but not uniformly provided to children with non-glomerular causes of CKD. Underlying specific diagnoses play an important role in both the risk of ESKD and potential benefits of urologic surgery.

CONCLUSION

Within the CKiD cohort, children with non-glomerular causes of CKD often received urologic care. Urology referral with surgery was associated with lower risk of ESKD compared to no prior urologic care but depended on specific underlying diagnoses.

Computer-aided diagnosis of congenital abnormalities of the kidney and urinary tract in children based on ultrasound imaging data by integrating texture image features and deep transfer learning image features

INTRODUCTION

Anatomic characteristics of kidneys derived from ultrasound images are potential biomarkers of children with congenital abnormalities of the kidney and urinary tract (CAKUT), but current methods are limited by the lack of automated processes that accurately classify diseased and normal kidneys.

OBJECTIVE

The objective of the study was to evaluate the diagnostic performance of deep transfer learning techniques to classify kidneys of normal children and those with CAKUT.

STUDY DESIGN

A transfer learning method was developed to extract features of kidneys from ultrasound images obtained during routine clinical care of 50 children with CAKUT and 50 controls. To classify diseased and normal kidneys, support vector machine classifiers were built on the extracted features using (1) transfer learning imaging features from a pretrained deep learning model, (2) conventional imaging features, and (3) their combination. These classifiers were compared, and their diagnosis performance was measured using area under the receiver operating characteristic curve (AUC), accuracy, specificity, and sensitivity.

RESULTS

The AUC for classifiers built on the combination features were 0.92, 0.88, and 0.92 for discriminating the left, right, and bilateral abnormal kidney scans from controls with classification rates of 84%, 81%, and 87%; specificity of 84%, 74%, and 88%; and sensitivity of 85%, 88%, and 86%, respectively. These classifiers performed better than classifiers built on either the transfer learning features or the conventional features alone (p < 0.001).

DISCUSSION

The present study validated transfer learning techniques for imaging feature extraction of ultrasound images to build classifiers for distinguishing children with CAKUT from controls. The experiments have demonstrated that the classifiers built on the transfer learning features and conventional image features could distinguish abnormal kidney images from controls with AUCs greater than 0.88, indicating that classification of ultrasound kidney scans has a great potential to aid kidney disease diagnosis. A limitation of the present study is the moderate number of patients that contributed data to the transfer learning approach.

CONCLUSIONS

The combination of transfer learning and conventional imaging features yielded the best classification performance for distinguishing children with CAKUT from controls based on ultrasound images of kidneys.

Renal Parenchymal Area Growth Curves for Children 0 to 10 Months Old

PURPOSE

Low renal parenchymal area, which is the gross area of the kidney in maximal longitudinal length minus the area of the collecting system, has been associated with increased risk of end stage renal disease during childhood in boys with posterior urethral valves. To our knowledge normal values do not exist. We aimed to increase the clinical usefulness of this measure by defining normal renal parenchymal area during infancy.

MATERIALS AND METHODS

In a cross-sectional study of children with prenatally detected mild unilateral hydronephrosis who were evaluated between 2000 and 2012 we measured the renal parenchymal area of normal kidney(s) opposite the kidney with mild hydronephrosis. Measurement was done with ultrasound from birth to post-gestational age 10 months. We used the LMS method to construct unilateral, bilateral, side and gender stratified normalized centile curves. We determined the z-score and the centile of a total renal parenchymal area of 12.4 cm(2) at post-gestational age 1 to 2 weeks, which has been associated with an increased risk of kidney failure before age 18 years in boys with posterior urethral valves.

RESULTS

A total of 975 normal kidneys of children 0 to 10 months old were used to create renal parenchymal area centile curves. At the 97th centile for unilateral and single stratified curves the estimated margin of error was 4.4% to 8.8%. For bilateral and double stratified curves the estimated margin of error at the 97th centile was 6.6% to 13.2%. Total renal parenchymal area less than 12.4 cm(2) at post-gestational age 1 to 2 weeks had a z-score of -1.96 and fell at the 3rd percentile.

CONCLUSIONS

These normal renal parenchymal area curves may be used to track kidney growth in infants and identify those at risk for chronic kidney disease progression.

Segmentation of renal parenchymal area from ultrasound images using level set evolution

This paper presents a framework for segmentation of renal parenchymal area from ultrasound images based on a 2-step level set method. We used distance regularized level set evolution method to partition the kidney boundary, followed by region-scalable fitting energy minimization method to segment the kidney collecting system, and determined renal parenchymal area by subtracting the area of the collecting system from the gross kidney area. The proposed method demonstrated excellent validity and low inter-observer variability.

Renal parenchymal area and risk of ESRD in boys with posterior urethral valves

BACKGROUND AND OBJECTIVES

Approximately 20% of boys with posterior urethral valves develop ESRD; however, few factors associated with the risk of ESRD have been identified. The objective of this study was to determine if renal parenchymal area, defined as the area of the kidney minus the area of the pelvicaliceal system on first postnatal ultrasound, is associated with the risk of ESRD in infants with posterior urethral valves.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

A retrospective cohort of boys who were diagnosed with posterior urethral valves at less than 6 months of age between 1988 and 2011 and followed for at least 1 year at a free-standing children's hospital was assembled. Cox proportional hazard regression and Kaplan-Meier analysis were used to estimate the association between renal parenchymal area and time to ESRD. Cox models were adjusted for age at presentation, minimum creatinine 1 month after bladder decompression, and vesicoureteral reflux.

RESULTS

Sixty patients were followed for 393 person-years. Eight patients developed ESRD. Median renal parenchymal area was 15.9 cm(2) (interquartile range=13.0-21.6 cm(2)). Each 1-cm(2) increase in renal parenchymal area was associated with a lower risk of ESRD (hazard ratio, 0.64; 95% confidence interval, 0.42 to 0.98). The rate of time to ESRD was 10 times higher in boys with renal parenchymal area<12.4 cm(2) than boys with renal parenchymal area≥12.4 cm(2) (P<0.001). Renal parenchymal area could best discriminate children at risk for ESRD when the minimum creatinine in the first 1 month after bladder decompression was between 0.8 and 1.1 mg/dl.

CONCLUSION

In boys with posterior urethral valves presenting during the first 6 months of life, lower renal parenchymal area is associated with an increased risk of ESRD during childhood. The predictive ability of renal parenchymal area, which is available at time of diagnosis, should be validated in a larger, prospectively-enrolled cohort.

Kidney function and tobacco smoke exposure in US adolescents

BACKGROUND AND OBJECTIVE

Active smoking and secondhand smoke (SHS) are known risk factors for kidney disease in adults. We evaluated the association between exposure to active smoking or SHS and kidney function in US adolescents.

METHODS

This is a cross-sectional study in 7516 adolescents aged 12-17 who participated in NHANES 1999-2010 and had serum creatinine and cotinine measures. Active smoking was defined as self-reported smoking or serum cotinine concentrations >10 ng/mL. SHS was defined as nonactive smokers who self-reported living with ≥1 smokers or serum cotinine concentrations ≥ 0.05 ng/mL. Kidney function was determined by using the chronic kidney disease in children estimated glomerular filtration rate (eGFR) equation.

RESULTS

Median (interquartile range) eGFR and serum cotinine concentrations were 96.8 (85.4-109.0) mL/minute per 1.73 m(2) and 0.07 (0.03-0.59) ng/mL, respectively. After multivariable adjustment, eGFR decreased 1.1 mL/minute per 1.73 m(2) (95% confidence interval [CI]: -1.8 to -0.3) per interquartile range increase in serum cotinine concentrations. The mean (95%CI) difference in eGFR for serum cotinine tertiles 1, 2, and 3 among children exposed to SHS compared to unexposed were -0.4 (-1.9 to 1.2), -0.9 (-2.7 to 0.9), and -2.2 (-4.0 to -0.4) mL/minute per 1.73 m(2), respectively (P = .03). The corresponding values among tertiles of active smokers compared to unexposed were 0.2 (-2.2 to 2.6), -1.9 (-3.8 to 0.0), and -2.6 (-4.6 to -0.6) mL/minute per 1.73 m(2) (P = .01).

CONCLUSIONS

Tobacco smoke exposure was associated with decreased eGFR in US adolescents, supporting the possibility that tobacco smoke effects on kidney function begin in childhood.