Development of a prediction model for risk of breakthrough urinary tract infection in children with primary vesicoureteral reflux

Introduction

Vesicoureteral reflux (VUR) is a frequent congenital anomaly in children, and the prevalence of VUR in non-symptomatic children has been estimated at 0.4–1.8% (1). The gold standard for the diagnosis of VUR is voiding cystourethrography (VCUG), which provides precise anatomical detail and enables grading of VUR (2). Patients with VUR are at risk of developing potentially serious consequences, such as renal scarring, hypertension, and renal failure (3). For most children, the initial approach is non-surgical with continuous antibiotic prophylaxis (CAP) or close surveillance. However, breakthrough urinary tract infection (BTUTI) sometimes occurs, which leads to significant kidney injury. Surgical intervention is reserved for those who develop BTUTI during CAP therapy. Several large-scale studies have shown that the factors associated with the BTUTI include age at presentation, sex, grade of VUR, the number of past urinary tract infections (UTIs), and the presence of bladder and bowel dysfunction (BBD), among others (4,5). The VUR index (VURx) and distal ureteral diameter ratio (UDR) are predictive of BTUTI (6-10).

The purpose of this study was to analyze and summarize the clinical and imaging data of children with primary VUR in Beijing Children’s Hospital, to identify predictors for BTUTI, and to develop a prediction model and nomogram. We hypothesized that these variables would predict BTUTI better than previous indices and provide clinicians and families with guidance in management. We present this article in accordance with the TRIPOD reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-85/rc).

Methods

Patient cohort

This was a retrospective study with data collected between January 2019 and August 2021 in Beijing Children’s Hospital. Patients who had initially been diagnosed with primary VUR and underwent CAP treatment and had detailed VCUG and clinical data were identified. Trimethoprim-sulfamethoxazole, amoxicillin, and nitrofurantoin are the most commonly used CAP agents. For patients who exhibit resistance to these antibiotics, it is crucial to select alternative antibiotics based on the results of urine culture. The recommended regimen involves taking the medication orally once daily before bedtime, at a dose equivalent to one-third of the therapeutic dose. To minimize the risk of drug resistance, it is advisable to alternate between several different antibiotics every 1–3 months. All data were used to develop the prediction model. The exclusion criteria were as follows: (I) children with secondary VUR (due to posterior urethral valves or neurogenic bladder, etc.); (II) patients with previous vesicoureteral surgeries; (III) patients with incomplete clinical data or missing follow-up data; and (IV) images with significant artifacts. Patients with ureteral duplication and/or periureteral diverticula (we will refer to these two collectively as concomitant “ureteral anomalies”) associated with primary VUR were included.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Beijing Children’s Hospital, Capital Medical University, National Center for Children’s Health (No. [2023]-E-139-R). The requirement for individual consent for this retrospective analysis was waived.

Research materials

Information was obtained through a review of the medical records. Missing clinical information was obtained by follow-up visits or telephone calls to the family. Clinical variables recorded included age at presentation, sex, number of UTIs before VUR diagnosis (variable “UTI history”), and circumcision. Relevant imaging parameters including anteroposterior pelvis diameter (APD), ureteral anomalies, VUR grade at diagnosis, VUR timing at cystography, bladder morphological changes, sacral ratios (SRs), ureteropelvic junction diameter of ureter (UPJ diameter), ureterovesical junction diameter of ureter (UVJ diameter), maximum ureteral diameter (Max diameter), ureteral tortuosity, distal ureteral diameter (UD), and distal UDR. UTI was defined as positive nitrite and/or leucocyte esterase or white blood cell (WBC) >5 cells/high power field in urine, and culture showing bacterial growth of 10,000 colony forming units/mL in a catheterization sample or 100,000 colony forming units/mL in a clean-catch midstream urine sample, or any growth of a single uropathogen in a suprapubic aspiration sample (11). The primary outcome measure of the study was BTUTI, defined as a UTI that occurred while on CAP therapy.

The VCUG examination procedure was conducted as follows: patients are usually positioned supine with their legs naturally abducted. Bladder capacity is generally estimated based on the patient’s age using the formula (age + 2) × 30 mL, although this volume may be adjusted according to the patient’s comfort and tolerance. The contrast agent is introduced until the child’s bladder is filling or the urge to void is reported. During the filling phase, the primary projection angle utilized is anteroposterior (AP); however, lateral or oblique views may be additionally employed if deemed necessary for a more comprehensive assessment (12). During the voiding phase, a lateral projection angle is specifically adopted to optimize visualization. Potential variation due to patient positioning, bladder filling status, and projection angle can significantly influence the accuracy and reliability of imaging measurements. Images were taken intermittently during the VCUG examination, usually at least a few before the injection of contrast material, during filling, during voiding, and after voiding. We used averaging multiple frames during the measurement to minimize the measurement error. VUR grade was determined using the International Reflux Grading System (13). High-grade VUR was characterized as a reflux of grades IV–V. VUR timing on VCUG was designated as early-to-mid filling, late filling, and voiding only. Late filling VUR was defined as reflux onset at over 75% of estimated bladder capacity. Voiding-only reflux was defined as reflux occurring after the filling phase with or without catheter removal (14). SR was measured in the conducted VCUG and the calculation method involved using the ratio of the length of the sacroiliac joint to the coccyx to that of the length of the iliac crest to the sacroiliac joint (15). The UPJ diameter, UVJ diameter, ureteral tortuosity, and UD measurements are shown in Figure 1. UPJ diameter was defined as the diameter of the ureter before the pelvic dilatation or the most proximal aspect of the ureter fully visible for annotation. UVJ diameter was defined as the diameter of the ureter at the point where it enters the bladder if the ureter passes behind the bladder (or if the bladder is grossly distended, select the ureter at the point where the ureter is passing behind the bladder perpendicular to the midline path) (16). Max diameter was defined as the largest diameter of the ureter. UD was calculated as the largest UD within the false pelvis (defined as the area below the most superior aspect of the ilia). UDR was defined as UD divided by the distance from the L1–L3 vertebral body (17). VURx was a scoring system based on initial cystogram findings and sex, assigned a 1 to 6 point scale as follows: early-to-mid filling, 3 points; late filling, 2 points; voiding only, 1 point; female sex, 1 point; VUR grades IV/V, 1 point; and ureteral anomalies, 1 point (9). Two investigators, blinded to both predictor variables and patient outcome, extracted and measured all VCUG-related imaging information. A single investigator, blinded to the data collection forms, ascertained outcomes. To assess the reliability of our measurement methods, we conducted inter- and intra-observer reliability studies.

Figure 1 The measurement of UPJ diameter (red line), UVJ diameter (blue line), ureteral tortuosity (dotted black divided by dotted white), and UD (yellow line) based on VCUG images. UD, ureteral diameter; UPJ diameter, ureteropelvic junction diameter of ureter; UVJ diameter, ureterovesical junction diameter of ureter; VCUG, voiding cystourethrography.

Development of the prediction model

All available data on the database were used to maximize the power and generalizability of the results. Candidate variables included all demographic, clinical, and VCUG-related factors. Bivariate analysis was used to identify the correlation of variables with outcome. Variables showing association with BTUTI (P<0.05) were included in an initial multivariable logistic regression model to build a clinical prediction model. The prediction model was output as a nomogram, which calculated the total score of each patient by analyzing scores corresponding to each predictor variable, and predicted the probability of BTUTI. The model was internally validated using the bootstrap method with B=1,000. For each bootstrap sample, the model was refitted and evaluated to estimate the optimism in performance.

The performance of the model was evaluated from calibration, discrimination, and clinical benefit. The calibration curve was used to evaluate the agreement between predicted and observed probabilities. The receiver operating characteristic (ROC) curve was used to evaluate the model’s discriminative ability, and the area under the ROC curve (AUC) was calculated. We determined an optimal threshold probability based on the Youden index from the model’s ROC analysis. Decision curve analysis (DCA) was conducted to evaluate the clinical net benefit, which can determine the clinical utility of the nomograms.

Statistical analysis

All statistical analyses and graphs were performed using R software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria). All data collected from the retrospective cohort were used to develop the risk prediction model. If an outcome was missing, the patient data were excluded from the analysis. Multiple imputation was used to address missingness in our data. For categorical variables, we used the mice package of R software, and the specified method was “logreg” to extract the interpolated data. To assess the reliability of our measurement methods, we conducted inter- and intra-observer reliability studies. Categorical variables were compared using the Chi-squared test or Fisher’s exact test and were described by numbers of cases and percentages [n (%)]. Normalcy was analyzed using the Shapiro-Wilk test. Normally distributed data were analyzed using the Chi-squared test or Fisher’s exact test and presented as mean ± standard deviation (SD), whereas non-normally distributed data were analyzed using the Mann-Whitney U test and reported as median and quartile spacing [M (P25, P75)]. Bivariate analysis was used to identify the correlation of variables with BTUTI. Multicollinearity was assessed using variance inflation factor (VIF) analysis, with a VIF value > 5 indicating multicollinearity. Multivariate logistic regression was used to evaluate the correlation between the variables and the results. All tests were two-sided, and statistical significance was associated with a P<0.05 in multivariate logistic regression.

Results

Characteristics of the study cohort

A total of 205 consecutive patients were diagnosed with primary VUR and CAP therapy based on VCUG data available during the study period at Beijing Children’s Hospital. Among these, seven patients who had undergone previous vesicoureteral surgeries and five patient who had been lost to follow-up were excluded from the study. The final study population included 193 patients (123 male), 58 of whom had BTUTI. The M (P25, P75) age at diagnosis was 16.0 (10.0, 41.0) months and the M (P25, P75) follow-up duration was 35.0 (28.0, 43.0) months. A total of 96 patients had unilateral VUR, and 97 patients had bilateral VUR. Table 1 summarizes the VCUG-related parameters and clinical parameters of all patients included in the study.

Univariate and multivariate logistic regression analysis

The univariate logistic regression analysis revealed that there were statistical differences in high-grade VUR (P=0.003), UPJ diameter (P=0.045), UVJ diameter (P<0.001), ureteral tortuosity (P=0.03), Max diameter (P=0.001), UD (P<0.001), and UDR (P=0.001) between the no BTUTI and BTUTI cases. Due to sample size constraints, we limited the number of predictors in the multivariable model to ensure an adequate events-per-variable (EPV) ratio. The covariates with a P<0.05 [including sex (P=0.053)] in the univariate logistic regression analysis were included in the multivariate logistic regression analysis. Due to concerns about multicollinearity, we excluded UD, UDR, and Max diameter from the multivariate analysis. The results of the multivariate logistic regression showed that sex (female) [odds ratio (OR): 3.39; 95% confidence interval (CI): 1.57–7.33], high-grade VUR (OR: 2.27, 95% CI: 0.98–5.25), and UVJ diameter (OR: 5.85, 95% CI: 1.81–18.92) were independent predictors for BTUTI. The intraclass correlation coefficient value for the UVJ diameter measurement demonstrates a moderate level of reliability (Table S1). The results of how the univariate and multivariate logistic regression models performed to identify risk factors for BTUTI are presented in Table 2.

We determined the optimal threshold probability of 41% based on the Youden index from the ROC analysis in the revised manuscript, indicating that patients with a predicted probability above this cutoff are at higher risk of developing BTUTI. To enhance clinical applicability, we further constructed a stratified plot illustrating the predicted probability of BTUTI across high-grade VUR and UVJ diameters, with stratification by sex (Figure 2).

Figure 2 Probability of BTUTI. BTUTI, breakthrough urinary tract infection; UVJ diameter, ureterovesical junction diameter of ureter; VUR, vesicoureteral reflux.

Development of the prediction model

Based on the results of multivariate analysis, three variables of sex, high-grade VUR, and UVJ diameter were used to create a prediction model and a nomogram (Figure 3A). Internal validation of the nomogram was performed using 1,000 bootstrap sample corrections. After bootstrap validation (B=1,000), the optimism-corrected concordance index (C-index) was 0.73, with a calibration slope of 0.93. The AUCs for our model, the UDR, and the VURx in predicting the occurrence of BTUTI were 0.736, 0.680, and 0.546, respectively (Figure 3B). The calibration curve of the nomogram demonstrated agreement between the predicted and observed BTUTI probability (Figure 3C). We performed DCA for each of the single parameters (sex, high-grade VUR, and UVJ diameter) and compared them to our multivariable prediction model. In contrast, our multivariable prediction model consistently demonstrated a higher net benefit across a broader range of threshold probabilities. This indicates that the combination of these variables provides a more accurate and clinically useful prediction compared to any single parameter alone (Figure 3D).

Figure 3 Development and validation of the nomogram predicting BTUTI. (A) The nomogram for predicting BTUTI consisted of three variables, sex, high-grade VUR, and UVJ diameter. (B) ROC curves of model, UDR, and VURx for evaluating the predictive performance. (C) Calibration curve of the prognostic nomogram model. The dotted grey line represents an ideal model while the red line represents the nomogram’s prediction performance, together with a bias-corrected blue solid line. (D) DCA prediction model and each of the single parameters (sex, high-grade VUR, and UVJ diameter). UVJ diameter: the diameter of the ureter at the point where it enters the bladder if the ureter passes behind the bladder (or if the bladder is grossly distended), select the ureter at the point where the ureter is passing behind the bladder perpendicular to the midline path. UD: the largest UD within the false pelvis (defined as the area below the most superior aspect of the ilia). UDR: UD divided by the distance from the L1–L3 vertebral body. VURx: a scoring system based on initial cystogram findings and sex, assigned a 1 to 6 point scale as follows: early-to-mid filling, 3 points; late filling, 2 points; voiding only, 1 point; female sex, 1 point; VUR grades IV/V, 1 point; and ureteral anomalies, 1 point. AUC, area under the ROC curve; BTUTI, breakthrough urinary tract infection; DCA, decision curve analysis; ROC, receiver operating characteristic; UD, ureteral diameter; UDR, ureteral diameter ratio; UVJ diameter, ureterovesical junction diameter of ureter; VUR, vesicoureteral reflux; VURx, VUR index.

Discussion

In our study, the comprehensive analysis of clinical and imaging data among 193 patients has unveiled several key factors associated with BTUTI in primary VUR patients. Specifically, variables such as sex, high-grade VUR, UPJ diameter, UVJ diameter, ureteral tortuosity, and Max diameter displayed significant differences between the no BTUTI and BTUTI. In the present study, we confirmed the independent predictive value of sex (female) (OR: 3.39), high-grade VUR (OR: 2.27), and UVJ diameter (OR: 5.85) regarding the likelihood of BTUTI. Building on the independent predictors identified through multivariable analysis, we developed a prediction model and nomogram. Our prediction model exhibited good discrimination, as evidenced by an AUC of 0.736. Comparing the performance of our model to previous indices such as the VURx and the UDR, our model’s AUC demonstrates its potential to outperform existing prediction methods (Figure 3B). The calibration curve and DCA curve substantiate the clinical relevance of our model, aligning with previous studies that underscore the importance of predictive tools in shaping the management of primary VUR patients.

In this study, the prediction model consisted of sex (female), high-grade VUR, and UVJ diameter, and the prediction performance of the model. Similar to our findings, Hidas et al. (4) reported that female patients and those with higher-grade reflux were at increased risk for BTUTI. However, their study also highlighted the significance of initial presentation with a UTI as a risk factor. This suggests that the context in which VUR is diagnosed may influence the subsequent risk of BTUTI. The study by Arlen et al. (6) evaluated the effectiveness of VURx (incorporating gender, VUR grade, VUR timing, and ureteral anomalies) as a predictor of BTUTI compared to reflux grade and UDR and showed that VURx was superior to UDR and reflux grade in predicting BTUTI. These findings resonate with previous research that has highlighted the risk of female and high-grade VUR on BTUTI for primary VUR patients. Female sex is associated with higher susceptibility to UTI due to anatomical factors such as a shorter urethra. High-grade VUR increases the likelihood of kidney involvement during UTI. In our study, the measurement of the ureter diameter directly reflected anatomical changes in the ureter, which could lead to urine retention and provide an environment for bacterial growth (18). Chen et al. (7) introduced systemic immune-inflammation index (SII) as a predictor, and combined it with UDR for predicting BTUTI in VUR patients. Our study did not find the VUR timing and ureteral anomalies to be significant predictors, but instead identified UVJ diameter as a novel and significant risk factor. UVJ diameter reflects the degree of upper urinary tract dilation, which may indicate prolonged urine stasis and thereby increase the risk of renal infection. Moreover, the inclusion of UVJ diameter as an independent predictor illuminates the prospective utility of this parameter in risk stratification.

In this study, the prediction model consisted of sex (female), high-grade VUR, and UVJ diameter, and the prediction performance of the model was better than that of VURx and UDR. This suggests that our model may offer enhanced clinical utility in identifying patients at risk for BTUTI.

Our study has several strengths, including the thorough analysis of both clinical and imaging data within a well-defined cohort of primary VUR patients. The development of a prediction model and nomogram further enhances the precision of risk assessment in this patient population. The implications of our study are substantial for the management of primary VUR patients. The identification of early predictors, encompassing sex, high-grade VUR, and UVJ diameter, equips urologists to tailor patient management at an early stage. Our model assists in the identification of patients more likely to respond favorably to conservative measures, sparing them from the inherent risks and economic burdens associated with BTUTI. Conversely, it pinpoints patients at higher risk, enabling prompt interventions to avert renal damage and its associated complications. To facilitate clinical application, we identified an optimal threshold probability of 41% (based on the Youden index from the ROC curve), which may help to stratify patients by risk. Patients with predicted probabilities above this cutoff may benefit from closer follow-up or early surgical consultation. For example, a female patient presenting with high-grade VUR and significant ureteral dilation at the UVJ (i.e., a wider UVJ diameter) would yield a high total score on the nomogram, suggesting an increased risk of BTUTI. This patient may warrant early referral to a pediatric urologist. The nomogram may assist clinicians in making personalized management decisions and identifying high-risk patients who could benefit from timely intervention.

It is imperative to recognize several inherent limitations in our study. First, the retrospective nature of our investigation may have introduced selection bias and information bias, as data collection relied on existing records rather than prospective enrollment. As such, we acknowledge that the timing of infection events is not always precisely documented, and there is a lot of missing information on timing of first BTUTI events. According to previous relevant literature, a significant proportion of infections occurred within the first 2 years after diagnosis (4,19,20). Given that the median follow-up period of our study was 35 months, which was longer than the periods reported in these studies, we believe that our follow-up time was sufficient to capture most infection events. Second, the single-center design restricted the sample size and diversity, potentially diminishing the statistical power and robustness of our findings. Consequently, the generalizability of our conclusions to a broader population may be compromised. In addition, there was a lack of external validation. Although internal validation was performed, the generalizability of the model remains uncertain. Therefore, future studies are needed to externally validate this model in independent, multicenter, or prospective cohorts to ensure its robustness and applicability in broader clinical settings. Furthermore, exploring additional predictive factors and refining the model’s parameters may improve its accuracy and clinical utility. Another limitation of our study is the potential variability in the interpretation of VCUG findings. As VCUG is a partially subjective imaging modality, inter-observer differences may exist, which could introduce bias in the assessment of parameters. Continued research efforts in these directions align with the ongoing pursuit of refined predictive models and enhanced management strategies for primary VUR patients.

Conclusions

Our study presents a comprehensive analysis of clinical and VCUG imaging data in primary VUR patients. Our study offers a valuable prediction tool for distinguishing primary VUR patients at risk of BTUTI. By incorporating predictors, specifically sex, high-grade VUR grade, and UVJ diameter, into a robust prediction model, clinicians can make well-informed decisions, ultimately optimizing patient care. The clinical relevance and utility of this model provide an opportunity to improve outcomes and alleviate the burden of primary VUR on both patients and healthcare systems. These findings are consistent with prior research in this field and underscore the importance of personalized, evidence-based approaches in the effective management of primary VUR. External validation in multicenter or prospective cohorts is warranted to confirm the performance and clinical utility of the model.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-85/rc

Data Sharing Statement: Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-85/dss

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-85/prf

Funding: This work was supported by grants from the National Key R&D Program of China (No. 2016YFC 1000807).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-85/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Beijing Children’s Hospital, Capital Medical University, National Center for Children’s Health (No. [2023]-E-139-R). The requirement for individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sargent MA. What is the normal prevalence of vesicoureteral reflux? Pediatr Radiol 2000;30:587-93. [Crossref] [PubMed]
2. Darge K, Riedmiller H. Current status of vesicoureteral reflux diagnosis. World J Urol 2004;22:88-95. [Crossref] [PubMed]
3. Fanos V, Cataldi L. Antibiotics or surgery for vesicoureteric reflux in children. Lancet 2004;36:1720-2.
4. Hidas G, Billimek J, Nam A, et al. Predicting the Risk of Breakthrough Urinary Tract Infections: Primary Vesicoureteral Reflux. J Urol 2015;194:1396-401. [Crossref] [PubMed]
5. Peters CA, Skoog SJ, Arant BS Jr, et al. Summary of the AUA Guideline on Management of Primary Vesicoureteral Reflux in Children. J Urol 2010;184:1134-44. [Crossref] [PubMed]
6. Arlen AM, Leong T, Wu CQ, et al. Predicting Breakthrough Urinary Tract Infection: Comparative Analysis of Vesicoureteral Reflux Index, Reflux Grade and Ureteral Diameter Ratio. J Urol 2020;204:572-7. [Crossref] [PubMed]
7. Chen X, Zheng W, Tang X, et al. Predictive Value of Systemic Immune-Inflammation Index Combined with Distal Ureteral Diameter Ratio for Early Breakthrough Urinary Tract Infection in Children with Primary Vesicoureteral Reflux. Ann Ital Chir 2025;96:345-51. [Crossref] [PubMed]
8. Arlen AM, Leong T, Guidos PJ, et al. Distal Ureteral Diameter Ratio is Predictive of Breakthrough Febrile Urinary Tract Infection. J Urol 2017;198:1418-23. [Crossref] [PubMed]
9. Carlucci M, Damasio MB, Parodi S, et al. Does the distal ureteral diameter ratio (UDR) matter in the surgical management of vesicoureteral reflux in children? Pediatr Surg Int 2023;39:249. [Crossref] [PubMed]
10. Krishnan N, Agarwal P, Verma A, et al. Utility of ureteral diameter ratio for clinical decision-making in children with vesicoureteral reflux: a systematic review and meta analysis. Pediatr Surg Int 2024;40:296. [Crossref] [PubMed]
11. Klubdaeng A, Chaiyapak T, Sumboonnanonda A, et al. Model for predicting high-grade vesicoureteral reflux in young children presenting with febrile urinary tract infection. J Pediatr Urol 2022;18:518-24. [Crossref] [PubMed]
12. Janssen KM, Kirsch AJ, Crisostomo-Wynne TC, et al. Standardized protocol for voiding cystourethrogram: Are recommendations being followed? J Pediatr Urol 2021;17:66.e1-6. [Crossref] [PubMed]
13. Metcalfe CB, Macneily AE, Afshar K. Reliability assessment of international grading system for vesicoureteral reflux. J Urol 2012;188:1490-2. [Crossref] [PubMed]
14. Kirsch AJ, Arlen AM, Leong T, et al. Vesicoureteral reflux index (VURx): a novel tool to predict primary reflux improvement and resolution in children less than 2 years of age. J Pediatr Urol 2014;10:1249-54. [Crossref] [PubMed]
15. Metzger G, Cooper JN, Kabre RS, et al. Inter-rater Reliability of Sacral Ratio Measurements in Patients with Anorectal Malformations. J Surg Res 2020;256:272-81. [Crossref] [PubMed]
16. Khondker A, Kwong JCC, Rickard M, et al. A machine learning-based approach for quantitative grading of vesicoureteral reflux from voiding cystourethrograms: Methods and proof of concept. J Pediatr Urol 2022;18:78.e1-7. [Crossref] [PubMed]
17. Cooper CS, Birusingh KK, Austin JC, et al. Distal ureteral diameter measurement objectively predicts vesicoureteral reflux outcome. J Pediatr Urol 2013;9:99-103. [Crossref] [PubMed]
18. Karakaya D, Güngör T, Cakıcı EK, et al. Determining the effectiveness of the immature granulocyte percentage and systemic immune-inflammation index in predicting acute pyelonephritis. Postgrad Med 2023;135:155-60. [Crossref] [PubMed]
19. Arlen AM, Alexander SE, Wald M, et al. Computer model predicting breakthrough febrile urinary tract infection in children with primary vesicoureteral reflux. J Pediatr Urol 2016;12:288.e1-5. [Crossref] [PubMed]
20. Troesch VL, Wald M, Bonnett MA, et al. The additive impact of the distal ureteral diameter ratio in predicting early breakthrough urinary tract infections in children with vesicoureteral reflux. J Pediatr Urol 2021;17:208.e1-5. [Crossref] [PubMed]

(English Language Editor: J. Jones)