A novel nephrometry scoring system outperforms RENAL score, PADUA score, and SPARE score for feasibility prediction of nephron-sparing surgery in children with bilateral Wilms tumor

Introduction

Wilms tumor (WT) is the most common type of renal tumor in children (1). Bilateral WT only accounts for 5–8% of all WT (2). With advances in diagnosis and treatment, more than 90% of WT children are expected to have long-term survival (1,3). Radical nephrectomy is considered the gold-standard technique for unilateral WT in patients without a genetic predisposition; however, this approach would leave bilateral WT patients anephric. Therefore, nephron-sparing surgery (NSS) has been proposed and continuously improved in order to preserve normal kidney tissue as much as possible, and reduce the occurrence of end-stage renal disease and other adverse events (4).

In clinical practice, the anatomical relationship between the tumor and the kidney is critical in determining whether NSS can be performed. To date, several nephrometry scoring systems have been developed to quantify the anatomical relationships of renal masses and evaluate the feasibility of NSS (5-7). Nevertheless, these nephrometry scoring systems are based on and commonly applied to adult renal tumors. Few studies have applied these scoring systems to pediatric renal tumors, especially bilateral WT, a type more likely to require NSS (8,9). Previously, we established a radiomics nomogram to predict NSS feasibility based on the preoperative imaging data of children with bilateral WT, and the model demonstrated good predictive performance through internal validation (10).

In this study, we aimed to explore the effectiveness of our novel RSR [relation with collecting system (R), size (S), remaining renal parenchyma (R)] prediction model, compared with RENAL [radius (R), exophytic/endophytic (E), nearness to collecting system/sinus (N), anterior/posterior (A), and location relative to polar lines (L)] score, PADUA (preoperative aspects and dimensions used for an anatomical classification) score, and SPARE (Simplified PADUA REnal) scoring systems in predicting the feasibility of NSS in children with bilateral WT. We present this study in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1038/rc).

Methods

Study design

We retrospectively reviewed clinical and imaging data of bilateral WT patients who were consecutively treated at our institution from January 2008 to December 2019. Patients with complete clinical data (including baseline characteristics and follow-up information) and preoperative imaging data were included in the study, which is consistent with our previous study (10). The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the Ethics Committee of Beijing Children’s Hospital, Capital Medical University (No. [2024]-E-090-R) and the requirement for individual consent for this retrospective analysis was waived.

Study endpoints

In this study, the endpoint was the accuracy of four nephrometry scores (RSR, RENAL, PUDUA, and SPARE) as predictors of whether NSS could be successfully conducted in the tumor-bearing kidney, with the tumor-bearing kidneys were considered the study subjects. The NSS group consisted of renal units that received NSS with no macroscopic residual and pathologically confirmed negative resection margin. In contrast, the non-NSS group consisted of renal units that did not receive NSS after preoperative assessment, and those who went NSS with macroscopic residual or positive resection margin.

Nephrometry scores

The data pf preoperative abdominal computed tomography (CT) or magnetic resonance imaging (MRI) performed after chemotherapy were used to determine four nephrometry scores for each renal unit. If there were multiple lesions in one renal unit, nephrometry scores were calculated by considering all lesions to reflect the highest level of complexity. The RSR scores were assessed as previously reported (10). To validate the utility and ease of use of the RSR score, a urology resident independently re-evaluated the imaging data; for the other three nephrometry scores (RENAL, PADUA, SPARE), a senior radiologist and a senior urologist independently assessed and agreed on differences.

For the RENAL score, the complexity of renal tumors is divided into different grades, with 4–6 points being low grade, 7–9 points being intermediate, and 10–12 points being high grade (11).

For the PADUA score, tumor complexity was categorized as low, moderate, and high, with 6–7 points, 8–9 points and 10–13 points, respectively (6).

For the SPARE score, tumor complexity was stratified as low (score 0–3), moderate (score 4–7), and high (score at 10 or greater) (7).

The predictive value of RSR, RENAL, PADUA, and SPARE scores on the feasibility of NSS of bilateral WT was illustrated through the receiver operating characteristic (ROC) curve, and the area under the curve (AUC) of different nephrometry scores were compared. The net reclassification index (NRI) and integrated discriminatory index (IDI) were used to calculate the reclassification and overall discrimination ability of RENAL, PADUA, and SPARE relative to the RSR score.

Statistical analysis

Continuous variables were analyzed by the Mann-Whitney U test, and reported as median and interquartile range (IQR). Categorical variables were reported as rates and were analyzed by the Chi-squared test or Fisher’s exact test.

Inter-observer agreement of the RSR score was tested, as conducting intraclass correlation coefficient (ICC) for continuous variables [size (S)] and Fleiss’ kappa (κ) for categorical variables [relation with collecting system (R), remaining renal parenchyma (R)].

All P values were two-tailed, and statistical significance was set at P<0.05. The R programming language and environment for Windows (version 4.2.3, http://www.r-project.org) was used for analysis. Figures were created by GraphPad Prism (version 9.5.0; GraphPad Software, San Diego, CA, USA), and R. A web-based calculator for clinical application of the RSR prediction model was constructed and embellished using the “DynNom” and “shinyPredict” packages in R.

Results

Consistent with our previous study, 58 patients (115 renal units) were finally analyzed (one patient underwent a total nephrectomy on one side in another hospital and was therefore excluded) (10). The median age of the study cohorts at the time of initial surgery was 13.0 (IQR, 10.0, 22.0) months. A total of 94 (81.7%) renal units were categorized into the NSS group, and the remaining 21 (18.3%) renal units comprised the non-NSS group. Tumor characteristics are listed in Table 1.

Tumor complexity was significantly higher in the non-NSS group, regardless of whether RENAL, PADUA, or SPARE nephrometry scores were calculated. Specifically, under the RENAL scoring system, 60.6% of the renal units in the NSS group were highly complex, significantly different from the 95.2% in the non-NSS group (P=0.008). A similar circumstance was found under the SPARE scoring system, with 33% and 85.7% in the NSS and non-NSS groups, respectively (P<0.001). In the PADUA scoring system, as the original version of SPARE, more evaluation indicators make the NSS group have the highest proportion (72.3%) of high complexity among the three nephrometry scoring systems, and no statistical difference was found between the two groups.

As shown in Figure 1, the ROC curves of the RSR model and the other three nephrometry scoring systems were used to predict NSS feasibility. In particular, the AUC of the RSR score was 0.982, which showed the best predictive ability among them. Meanwhile, the AUCs of the PADUA (0.832) and SPARE (0.833) scores were similar, and the RENAL (0.779) score showed the worst predictive ability. In addition, the decision curve analysis (DCA) curve illustrated the strong clinical utility of the RSR scores; as for the other three nephrometry scores, their clinical utilities were quite the same as the AUC results (Figure 2).

Figure 1 ROC curve of the RSR model and the other three adult nephrometry scores (RENAL, PADUA, SPARE) used to predict NSS feasibility in children with BWT. RSR, relation with collecting system, size, remaining renal parenchyma; AUC, area under the curve; RENAL, radius [R], exophytic/endophytic [E], nearness to collecting system/sinus [N], anterior/posterior [A], and location relative to polar lines [L]; PADUA, preoperative aspects and dimensions used for an anatomical classification; SPARE, Simplified PADUA Renal; ROC, receiver operating characteristic; NSS, nephron-sparing surgery; BWT, bilateral Wilms tumor.

Figure 2 DCA curve of the RSR model and the other three adult nephrometry scores (RENAL, PADUA, SPARE) calculating the net benefit at different threshold probabilities. RENAL, radius [R], exophytic/endophytic [E], nearness to collecting system/sinus [N], anterior/posterior [A], and location relative to polar lines [L]; PADUA, preoperative aspects and dimensions used for an anatomical classification; SPARE, Simplified PADUA Renal; RSR, relation with collecting system, size, remaining renal parenchyma; DCA, decision curve analysis.

The NRI and IDI values were calculated to further compare the RSR model’s predictive ability and the other three nephrometry scores (Table 2). Compared to the RENAL score, the RSR model showed the best reclassification ability [0.573 (95% confidence interval: 0.161, 1.204), P=0.03], but a pretty good result was not found with RSR compared to PADUA and SPARE. Meanwhile, significant overall discrimination ability was found with RSR compared to the other three scores (P<0.001).

The RSR score showed fairly good inter-observer agreement (Table 3). The agreement was good for the assessment of relation with collecting system (kappa =0.806) and remaining renal parenchyma (kappa =0.777), almost perfect for tumor size (ICC =0.871).

According to the RSR prediction model, a web-based calculator was built to assist in clinical decision-making on NSS preoperatively (https://bchurology.shinyapps.io/NSS_prediction_for_BWT/). On the left side of the webpage are the three parameters of the RSR model. After imputing the tumor characteristics of the patients, the probability of NSS is outputted on the right side of the webpage, either graphically in the Graphical Summary window (Figure 3A) or as a table in the Numerical Summary window (Figure 3B). Take one patient for example, a 1-year-old girl with bilateral WT who received neoadjuvant chemotherapy [vincristine, adriamycin, dexamethasone (VAD) regimen] before surgery. However, the tumor legions were not sensitive to chemotherapy. The CT scans were acquired preoperatively (Figures S1-S4). After imputing the parameters into the web-based calculator, NSS probabilities of approximately 13% on the left and over 99% on the right were output. In fact, based on the surgeon’s personal experience and considering the pros and cons of NSS and radical nephrectomy, this patient underwent successful NSS bilaterally.

Figure 3 Web-based calculator according to the RSR model. The probability of NSS is outputted either graphically in the Graphical Summary window (A) or as a table in the Numerical Summary window (B). NSS, nephron-sparing surgery; WT, Wilms tumor; RSR, relation with collecting system, size, remaining renal parenchyma.

Discussion

Whether NSS can be performed mainly depends on the complexity of the tumor. In the adult field, regardless of tumor type, several complexity scoring systems have been proposed and applied clinically (5-7). In the pediatric field, there are no unified standards for evaluating the complexity of renal tumors, resulting in a lack of guidance and consensus on when NSS should be attempted. In response, a few studies have attempted to use adult scoring systems to evaluate pediatric renal tumors (8,9,12,13).

Cost et al. (8) applied the adult RENAL scoring system to children for the first time. WTs accounted for 62.7% of the total study renal units, and 88.1% of them were calculated as high complexity, demonstrating WT characteristics in children. A total of 11 renal units (16.42%) attempted NSS based on the RENAL score, and all of them completed intraoperatively as NSS, whereas only 5.9% of all high-complexity renal units conducted NSS, the specific pathology types were not given by the authors. The researchers concluded that the RENAL score helped assess tumor complexity in renal cell carcinoma and renal lesions in older children.

Other studies have also revealed the highly complex nature of WT in children. Mittal et al. (12) retrospectively analyzed 28 patients undergoing 33 NSS, and WT comprised 85.7% of the total study population. RENAL scores were calculated, and high-complexity lesions comprised about half of the cohort. Mrad et al. (13) reported that less than 15% of patients in the high-complexity group and more than half of the total patients in the low-to-medium-complexity groups underwent NSS, calculated based on the RENAL, PADUA, and Renal Tumor Invasion Index scoring systems. Additionally, only one prior study has focused on using nephrometry scoring systems within a bilateral WT pediatric cohort. Gao et al. (9) retrospectively reviewed 29 patients with 53 renal units. According to the RENAL score, about 60.4% (32/53) of renal units were high complexity. About 68.75% (22/32) of renal units received initial NSS. For the low and intermediate-complexity groups, the proportions of initial NSS were 100% and 88.89%, respectively, indicating that low-complexity tumor kidneys were fair candidates for NSS. In our study cohort, 77 of 115 (67%) renal units were classified as high complexity, according to the RENAL scores. About 74.03% (57/77) of them underwent NSS successfully, which was similar to the previous research.

Nevertheless, there are some shortcomings to applying adult nephrometry scores to children. First, the diameter of WT lesions in children is usually large, making identifying the normal renal tissue challenging. Cost et al. (14) suggested that the typical diameter of the WT lesion in children is approximately 10 cm and generally replaces much of the entire kidney. In their follow-up study, 88.2% of lesions were ≥7 cm in diameter and were classified as high complexity based on the R-score in the RENAL scoring system (8). Similar results were found in our study, as 40% of lesions were ≥7 cm in the overall cohort. This anatomical feature makes it difficult to identify the relationship between the lesion and the upper/lower poles of the kidney (L-score), as well as the anterior/posterior of the kidney (A-score). Similar confusions also exist in distinguishing whether the lesion is endophytic or exophytic (E-score). The inter-rater reliability could corroborate this issue to some extent. Cost et al. (8) reported that the rates of correlation between two independent reviewers (one pediatric urologist and one radiologist) were moderate for the A- and L-scores and extremely poor for the E-score, which contrasts with the RENAL scores’ better consistency (kappa values above 0.7 for all scores) in adult renal tumors (15). Other studies have reported varying results for inter-rater reliability of the RENAL score. Mittal et al. (12) indicated good to almost perfect agreement for the R-, E-, and L-scores, moderate for the N-score, and fair for the A-score. However, Mrad et al. (13) suggested that the inter-rater reliability was quite satisfactory for RENAL, PADUA, and Renal Tumor Invasion Index, with Pearson correlation always superior to 0.7. The varying consistency test results indicate that the application prospects of adult nephrometry scores in pediatric renal tumors are unclear. Second, the contributions of different indicators to the overall tumor complexity are unequal. Specifically, the difference between the diameter of the lesion (R-score) between 6.9 and 7.1 cm, compared with the difference between partially exophytic and completely endophytic (E-score), both add one point to the total score; however, there are great differences in the choice of surgical approaches and intraoperative difficulties.

A scoring system suitable for the pediatric population is necessary and has been called for by some studies (8,9,16). Gao et al. (9) suggested that the refined RENAL nephrometry scoring system should include tumor thrombus and multiple lesions. Previous studies have reported the incidence rate of tumor thrombus to be 2.8–9.9%, and surgical complications are more likely to occur (17). Nephrogenic rests are considered the precursor lesions of WT, which usually present as multiple lesions and are more common in patients with bilateral WT (18). In our previous study, the above two indicators were included in logistic regression together with other indicators. After univariate and multivariate analysis, a prediction model composed of RSR was finally obtained (10). We compared the RSR model with the adult nephrometry scoring systems (RENAL, PADUA, SPARE), and the results showed that the RSR model outperformed the other three scores in predicting NSS possibility (Figure 1). Meanwhile, the AUC values were relatively close for the three adult scores. The lowest was the RENAL score, with an AUC value of 0.779. In addition, through NRI and IDI analysis, the RSR model showed an improvement of ~60% in reclassification ability and ~50% in integrated discrimination ability compared to the RENAL score. The RSR model improved reclassification and integrated discrimination ability by more than 20% compared to the PADUA and SPARE scores. To verify the reproducibility of the RSR model, a urology resident re-evaluated the image data, and the interobserver variability study was performed. Good to almost perfect agreement was found for the three indicators in the RSR model. Therefore, we believe the RSR model has the following advantages: (I) the evaluation indicators are more simplified and suitable for pediatric bilateral WT patients; (II) the prediction performance is better than that of adult nephrometry scores; (III) the repeatability and reliability are satisfactory. Furthermore, the RSR model could be more practical in the clinical setting with the visual web-based calculator we developed.

There are some limitations to our study. First, the study’s retrospective nature and the rarity of bilateral WT may have caused potential biases. Second, the remaining renal parenchyma (R) was obtained by subjective assessment based on preoperative imaging data. Recently, 3D printed models and virtual reality systems have been developed, which have improved the understanding of tumor anatomy (19-21). Additional limitations might include the lack of multiple evaluators at different seniorities or specializations and data from a single institution. Future research should use multicenter data to validate the predictive efficacy of the RSR model externally.

Conclusions

The RSR model is confirmed to be a reproducible and simplified tool for predicting NSS feasibility in bilateral WT patients. The predictive performance of the RSR model outperforms the established adult nephrometry scores while offering overall fair interobserver agreement. The RSR model-based network calculator can help clinicians make individualized treatment decisions.

Acknowledgments

Funding: None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-1038/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1038/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 (as revised in 2013). The study was approved by the Ethics Committee of Beijing Children’s Hospital, Capital Medical University (No. [2024]-E-090-R) and 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. Spreafico F, Fernandez CV, Brok J, Nakata K, Vujanic G, Geller JI, Gessler M, Maschietto M, Behjati S, Polanco A, Paintsil V, Luna-Fineman S, Pritchard-Jones K. Wilms tumour. Nat Rev Dis Primers 2021;7:75. [Crossref] [PubMed]
2. Charlton J, Irtan S, Bergeron C, Pritchard-Jones K. Bilateral Wilms tumour: a review of clinical and molecular features. Expert Rev Mol Med 2017;19:e8. [Crossref] [PubMed]
3. Geller JI, Hong AL, Vallance KL, Evageliou N, Aldrink JH, Cost NG, Treece AL, Renfro LA, Mullen EACOG Renal Tumor Committee. Children's Oncology Group's 2023 blueprint for research: Renal tumors. Pediatr Blood Cancer 2023;70:e30586. [Crossref] [PubMed]
4. Murphy AJ, Davidoff AM. Bilateral Wilms Tumor: A Surgical Perspective. Children (Basel) 2018;5:134. [Crossref] [PubMed]
5. Kutikov A, Uzzo RG. The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J Urol 2009;182:844-53. [Crossref] [PubMed]
6. Ficarra V, Novara G, Secco S, Macchi V, Porzionato A, De Caro R, Artibani W. Preoperative aspects and dimensions used for an anatomical (PADUA) classification of renal tumours in patients who are candidates for nephron-sparing surgery. Eur Urol 2009;56:786-93. [Crossref] [PubMed]
7. Ficarra V, Porpiglia F, Crestani A, Minervini A, Antonelli A, Longo N, Novara G, Giannarini G, Fiori C, Simeone C, Carini M, Mirone V. The Simplified PADUA REnal (SPARE) nephrometry system: a novel classification of parenchymal renal tumours suitable for partial nephrectomy. BJU Int 2019;124:621-8. [Crossref] [PubMed]
8. Cost NG, DeFoor WR Jr, Crotty EJ, Geller JI. The initial experience with RENAL Nephrometry in children, adolescents, and young adults with renal tumors. Pediatr Blood Cancer 2014;61:1434-9. [Crossref] [PubMed]
9. Gao P, Jiang H, Wu C, Liu L, Huang M, Fu Q, Liu J, Li J, Zhang H, Wang C. RENAL nephrometry scoring system in bilateral Wilms tumor: predictive application. Pediatr Surg Int 2023;39:230. [Crossref] [PubMed]
10. Li Z, Li J, Li Z, Sun N, Zhang Q, Song H, Zhang W. A radiomics nomogram for preoperative prediction of nephron-sparing surgery in patients with bilateral Wilms tumor. Quant Imaging Med Surg 2023;13:4234-44. [Crossref] [PubMed]
11. Borgmann H, Reiss AK, Kurosch M, Filmann N, Frees S, Mager R, Tsaur I, Haferkamp A R.E.N.A.L.. Score Outperforms PADUA Score, C-Index and DAP Score for Outcome Prediction of Nephron Sparing Surgery in a Selected Cohort. J Urol 2016;196:664-71. [Crossref] [PubMed]
12. Mittal S, Long C, El-Ali A, Talwar R, Lattanzio K, Lawton B, Hamdan D, Balis FM, Mattei P, Back SJ, Kolon TF. Utilizing RENAL nephrometry in pediatric patients undergoing nephron-sparing surgery for renal tumors: A single-institutional cohort. J Pediatr Urol 2023;19:641.e1-6. [Crossref] [PubMed]
13. Mrad C, Chamouni A, Guerra X, Tordjman M, Tabone MD, Audry G, Irtan S. Nephrometry Scoring Systems for Nephron Sparing Surgery in Children. J Pediatr Hematol Oncol 2024;46:1-7. [Crossref] [PubMed]
14. Cost NG, Lubahn JD, Granberg CF, Sagalowsky AI, Wickiser JE, Gargollo PC, Baker LA, Margulis V, Rakheja D. Pathological review of Wilms tumor nephrectomy specimens and potential implications for nephron sparing surgery in Wilms tumor. J Urol 2012;188:1506-10. [Crossref] [PubMed]
15. Kolla SB, Spiess PE, Sexton WJ. Interobserver reliability of the RENAL nephrometry scoring system. Urology 2011;78:592-4. [Crossref] [PubMed]
16. Vanden Berg RN, Bierman EN, Noord MV, Rice HE, Routh JC. Nephron-sparing surgery for Wilms tumor: A systematic review. Urol Oncol 2016;34:24-32. [Crossref] [PubMed]
17. Qureshi SS, Bhagat M, Smriti V, Murli D, Baheti A, Yadav S, Prasad M, Khanna N, Laskar S, Kembhavi SA, Vora T, Chinnaswamy G, Amin N, Ramadwar M, Talole S. Intravascular extension of Wilms tumor: Characteristics of tumor thrombus and their impact on outcomes. J Pediatr Urol 2021;17:69.e1-8. [Crossref] [PubMed]
18. Vujanić GM, Apps JR, Moroz V, Ceroni F, Williams RD, Sebire NJ, Pritchard-Jones K. Nephrogenic rests in Wilms tumors treated with preoperative chemotherapy: The UK SIOP Wilms Tumor 2001 Trial experience. Pediatr Blood Cancer 2017; [Crossref]
19. Wellens LM, Meulstee J, van de Ven CP, Terwisscha van Scheltinga CEJ, Littooij AS, van den Heuvel-Eibrink MM, Fiocco M, Rios AC, Maal T, Wijnen MHWA. Comparison of 3-Dimensional and Augmented Reality Kidney Models With Conventional Imaging Data in the Preoperative Assessment of Children With Wilms Tumors. JAMA Netw Open 2019;2:e192633. [Crossref] [PubMed]
20. Girón-Vallejo Ó, García-Calderón D, Ruiz-Pruneda R, Cabello-Laureano R, Doménech-Abellán E, Fuster-Soler JL, Ruiz-Jiménez JI. Three-dimensional printed model of bilateral Wilms tumor: A useful tool for planning nephron sparing surgery. Pediatr Blood Cancer 2018; [Crossref]
21. van der Zee JM, Fitski M, Simonis FFJ, van de Ven CP, Klijn AJ, Wijnen MHWA, van der Steeg AFW. Virtual Resection: A New Tool for Preparing for Nephron-Sparing Surgery in Wilms Tumor Patients. Curr Oncol 2022;29:777-84. [Crossref] [PubMed]