Tumor Volume and Tumor Crossing of the Axial Renal Midline as Preoperative Predictors of Reduced Estimated Glomerular Filtration Rate After Robotic Partial Nephrectomy

The ability of nephrometry scoring systems, including the radius, exophytic/endophytic, nearness to collecting system or sinus, anterior/posterior, and location relative to polar lines (R.E.N.A.L.), to predict loss of renal function after robotic partial nephrectomy (RPN) is still controversial. Therefore, we veried which combination of factors from nephrometry scoring systems, including tumor volume, was the most signicant predictor of postoperative renal function. Patients who underwent RPN for cT1 renal tumors in our hospital were reviewed retrospectively (n=163). The preoperative clinical data (estimated glomerular ltration rate [eGFR], comorbidities, and nephrometry scoring systems including R.E.N.A.L.) and perioperative outcomes were evaluated. We also calculated the tumor volume using the equation applied to an ellipsoid by three-dimensional computed tomography. The primary outcome was reduced eGFR, which was dened as an eGFR reduction of ≥ 20% from baseline to 6 months after RPN. Multivariate logistic regression analyses were used to evaluate the relationships between preoperative variables and reduced eGFR. Of 163 patients, 24 (14.7%) had reduced eGFR. Multivariate analyses indicated that tumor volume (cutoff value ≥ 14.11 cm 3 , indicating a sphere with a diameter ≥ approximately 3 cm) and tumor crossing of the axial renal midline were independent factors for reduced eGFR (odds ratio [OR], 4.57; P=0.003 and OR, 3.21; P=0.034, respectively). Our classication system using these two factors showed a higher area under the receiver operating characteristic curve (AUC) than previous nephrometry scoring systems (AUC=0.786 vs. 0.653–0.719), and it may provide preoperative information for counseling patients about renal function after RPN. (all P 0.05). for of patients with 3 for the P factor of the DAP (n factors indicate the presence of a tumor crossing of the axial renal midline. These results implied that the N factor of the R.E.N.A.L. system (2 or 3 points) and the P factor of the DAP system (3 points) are signicant factors for predicting reduced eGFR.


Introduction
Partial nephrectomy (PN) is the current standard treatment for the management of small renal tumors to prevent postoperative chronic kidney disease (CKD). In particular, a warm ischemic time (WIT) within 25 minutes is considered to be associated with preventing short-and long-term loss of renal function [1]. Recent meta-analyses have shown that the WIT is signi cantly shorter in robotic partial nephrectomy (RPN) than in laparoscopic partial nephrectomy [2]; thus, RPN is generally accepted as a favorable procedure for preserving kidney function.
There are some nephrometry scoring systems such as the radius, exophytic/endophytic, nearness to collecting system or sinus, anterior/posterior, and location relative to polar lines (R.E.N.A.L.), preoperative aspects and dimensions used for anatomic (PADUA) classi cation, and diameter-axial-polar (DAP) scores for predicting surgical complexity and potential perioperative morbidity [3,4,5]. Whether these nephrometry scoring systems have the ability to predict loss of renal function after PN is still controversial [6,7]. Furthermore, some reports that used mathematical calculated scores determined from preoperative images, such as the centrality index (c-index) and tumor contact surface area, predicted that the estimated glomerular ltration rate (eGFR) decreases after PN [8,9]. However, to our knowledge, the combination of nephrometry scoring systems and mathematically calculated scores has not been evaluated for predicting renal function after RPN. Therefore, we veri ed which combination of factors from the nephrometry scoring systems, including tumor volume accurately calculated by three-dimensional computed tomography (3D-CT), was the most signi cant predictor of reduced postoperative renal function. Finally, we described the accuracy of this new classi cation system including signi cant factors for predicting eGFR reduction compared with that of R.E.N.A.L., PADUA, and DAP scores.

Results
Patient Characteristics. The clinical patient characteristics are shown in Table 1. The median percent decrease in eGFR six months after RPN was 7.55% (IQR: 1.42-15.29%). Of 163 patients, 24 (14.7%) had an eGFR reduction of ≥ 20% from baseline to 6 months after RPN. The accurately calculated median tumor volume was 6.28 cm 3 (IQR: 2.70-14.68 cm 3 ). The nephrometry scores (R.E.N.A.L. and DAP scores) are shown in Table 2. The median R.E.N.A.L. score was six, and the median DAP score was ve. No postoperative complications greater than Grade 3 in Clavien-Dindo Classi cation were observed.
Association between Reduced eGFR and Each Factor of the Nephrometry Scoring Systems including the Tumor Volume Accurately Calculated by 3D-CT. The appropriate cutoff value of tumor volume calculated by 3D-CT for predicting reduced eGFR was 14.11 cm 3 (sensitivity = 0.625 and speci city = 0.806) (Fig. 1a). The tumor volume factor (cutoff value ≥ 14.11 cm 3 , indicating a sphere with a diameter ≥ approximately 3 cm) showed a higher AUC than the size factor of the DAP and R.E.N.A.L. classi cation systems (0.715 vs 0.547-0.636; Fig. 1b). Univariate analyses of each factor of the nephrometry scoring systems are shown in Table 3. The N and L factors of the R.E.N.A.L. system, the P factor of the DAP system, and size factors were found to be signi cantly associated with the outcome of interest (all P < 0.05). The number of patients with 3 points for the L factor of the R.E.N.A.L. system was the same as the number of patients with 3 points for the P factor of the DAP system (n = 41); these factors indicate the presence of a tumor crossing of the axial renal midline. These results implied that the N factor of the R.E.N.A.L. system (2 or 3 points) and the P factor of the DAP system (3 points) are signi cant factors for predicting reduced eGFR.
Uni-and Multivariate Analyses Predicting Reduced eGFR. Those signi cant nephrometry factors including tumor volume factor (cutoff value ≥ 14.11 cm 3 ) and preoperative clinical patient characteristics were investigated on uni-and multivariate analyses. The univariate analyses showed that a comorbidity of DM, the tumor volume, the N factor of the R.E.N.A.L., and the P factor of the DAP system were signi cantly associated with reduced eGFR (all P < 0.05, Table 4). The multivariate analysis showed that 3 points for the P factor of the DAP system (OR: 3.50, P = 0.014) and tumor volume (OR: 4.57, P = 0.003) were signi cant independent factors for predicting reduced eGFR ( Table 4).
Accuracy of Our Classi cation System for Predicting Decreased Renal Function. According to the number of independent factors (3 points for the P factor of the DAP system and tumor volume), all patients were strati ed into the following three groups: low-risk group (0 factors, n = 102), intermediate-risk group (1 factor, n = 39), and high-risk group (2 factors, n = 22). The classi cation system showed a statistically signi cant trend for predicting postoperative decreases in eGFR (continuous variable) 6 months after RPN (P < 0.001; Fig. 2a) and for predicting the WIT (P < 0.001; Fig. 2b). To ascertain whether our classi cation system was useful for predicting postoperative reduced eGFR, we compared the predictive accuracy between our classi cation system and nephrometry scoring systems such as the R.E.N.A.L. score (low, intermediate, and high), PADUA score (low, intermediate, and high), and DAP sum score. Our classi cation system showed a higher AUC than these nephrometry scoring systems (0.786 vs. 0.653-0.719) in our cohort (Fig. 2c).

Discussion
The current study demonstrated that tumor volume (cutoff value ≥ 14.11 cm 3 , indicating a sphere with a diameter ≥ approximately 3 cm) and 3 points for the P factor of the DAP system, indicating a tumor crossing the axial renal midline, were signi cant factors for predicting eGFR reduction after RPN. The simple classi cation system using these two factors had the best accuracy for predicting eGFR reduction after RPN compared with existing nephrometry scoring systems such as the R.E.N.A.L., DAP, and PADUA scores. Our classi cation system can provide prognostic information for counseling patients about renal function after RPN and assist in preoperative decision making.
To assess overall survival bene ts in patients with renal cell carcinoma (RCC) after partial or radical nephrectomy, predicting both oncologic outcomes and decreased renal function to avoid chronic kidney disease is essential [12,13]. Some studies have determined that an SCr level > 2.0 mg/dl or the occurrence of stage-5 CKD is the endpoint of postoperative renal function [14,15]. However, a limitation of these endpoints is that the outcome can depend on the preoperative state. To circumvent this limitation, analysis of eGFR reduction from baseline to a point after surgery is crucial for predicting accurate postoperative renal function. General functional reduction after PN averages approximately 10% in the two-kidney and 20% in the one-kidney model [16]. Thus, we set the cutoff point for reduced eGFR to 20% with reference to a previous report [9].
Recent reports have shown that a nomogram with the sum of the R.E.N.A.L. score incorporated accurately predicts signi cant eGFR reduction after PN [17,18]. Therefore, the nephrometry scoring systems might be able to predict a decline in renal function after PN. Simmons et al. reported that the DAP score, which is a modi ed version of the R.E.N.A.L. classi cation and c-index, had simpli ed methodology and was associated with volume loss and renal function after PN. Although the cutoff value for each factor was different, similar to the DAP score including tumor diameter, axial distance from the center point, and polar distance from the midline, tumor volume, nearness to the collecting system (distance from the tumor to the collecting system), and location relative to polar lines were signi cantly associated with eGFR decline after RPN in this study. Among these three factors, nearness to the collecting system was not an independent factor associated with the outcome of interest. However, the distance from the tumor to the collecting system tended to be shorter as the P factor score of the DAP system increased, and it was negatively correlated with tumor volume (Supplementary Fig. 2, R = 0.481, p < 0.001). Thus, the two independent factors used in our classi cation system may also re ect proximity to the collecting system.
In complex cases, long ischemia times are required for complete tumor resection [19]. For renal function preservation, various techniques during PN have been described (e.g., off-clamp, selective/super-selective clamp, and early unclamp, or cooling techniques for hypothermia) [20,21]. We showed that our classi cation system was signi cantly correlated with the WIT (Fig. 2B); thus, our classi cation system might be related to the complexity of the surgery. Therefore, our classi cation system might help to select patients who need various surgical techniques to avoid renal insu ciency.
Our results should be interpreted with caution because of several limitations. First, this study was based on data from patients who were treated at a single center, and external validation is needed before applying the classi cation system to select patients. Second, the study was retrospective in design with, and the follow-up period was relatively short. Third, while perioperative variables such as the WIT and estimated blood loss were not considered, these variables are likely important in uencers of postoperative renal function. However, the purpose of this study was to determine which combination of preoperative factors such as nephrometry scoring systems were best for predicting eGFR reduction. We also calculated the tumor volume assuming that each tumor was an ellipsoid. This was not a true volume, but the calculation of tumor volume is easily obtained from preoperative 3D-CT scans.

Conclusion
Accurate calculation of tumor volume and tumor crossing of the axial renal midline were independent predictors of eGFR reduction after RPN. Our classi cation system using these two factors had the best accuracy for predicting postoperative eGFR reduction when compared with previous nephrometry scoring systems such as the R.E.N.A.L., DAP, and PADUA scores.

Methods
Patient Selection. The medical records of 165 patients who underwent RPN for localized cT1 renal tumors with warm ischemia at Kansai Medical University Hospital between August 2014 and December 2019 were retrospectively reviewed. Patients with multiple renal tumors or a solitary kidney were excluded from this study. No patient underwent presurgical treatment with tyrosine kinase inhibitors or immune checkpoint inhibitors. All procedures were performed by experienced robotic surgeons at a single institution. Among these patients, two patients who underwent conversion to nephrectomy or open partial nephrectomy were excluded from the analysis. Ultimately, 163 patients were considered for further analyses.
Data Collection. The preoperative clinical data (sex, age, body mass index (BMI), American Society of Anesthesiologists (ASA) score, comorbidities of diabetes mellitus (DM) and hypertension (HTN), previous abdominal surgery, and antiplatelet or anticoagulant therapy), perioperative outcomes, and pathological features were evaluated. Renal function was assessed by serum creatinine (SCr) and eGFR, which was calculated using the following equation established for the Japanese population [10]: eGFR (mL/min/1.73 m 2 ) = 194 × Cr − 1.094 × age − 0.287 (× 0.739 for females) (Eq. 1). The percent reduction in renal function was calculated with the preoperative and postoperative (6 months after RPN) eGFRs. All patients underwent preoperative 3D-CT with or without contrast. Based on the imaging ndings, nephrometry scoring systems including R.E.N.A.L., PADUA, and DAP scores were evaluated with several urologists at a preoperative conference. The lengths of the horizontal axis and vertical axis were measured at the transverse plane where the tumor area was the largest (x and y, respectively), and the length of maximal z axis was measured in the coronal or sagittal plane (z). Then, the tumor volume was calculated using the following equation applied to an ellipsoid ( Supplementary Fig. 1): Tumor volume (cm 3 ) = 4/3 × π(3.14)×x/2×y/2×z/2 (Eq. 2). The lengths of these three directions and the distance from the tumor to the collecting system were measured independently by two observers (HO and KA), each of whom was blinded to the clinical outcome.
Statistical Analysis. The primary outcome of this study was a reduced eGFR, which was de ned as an eGFR reduction of ≥ 20% from baseline to 6 months after RPN. All continuous data are shown as median values and interquartile ranges (IQRs). The area under the receiver operating characteristic curve (AUC) was used to decide the cutoff value for continuous variables including tumor volume. Univariate and multivariate logistic regression analyses were used to evaluate the relationship between clinical variables and reduced eGFR. The trend of our classi cation system for predicting changes in renal function was examined by performing a Jonckheere-Terpstra test. The abilities of our classi cation systems and previous nephrometry scoring systems to predict reduced eGFR were evaluated and compared using AUC analysis. A post hoc comparison was used to compare the group means. Odds ratios (ORs) estimated from the logistic regression analyses are reported as relative risks with corresponding 95% con dence intervals (CIs). All statistical analyses were performed using EZR version 1.65 (Saitama Medical Center, Jichi, Japan) [11]. A two-sided p value < 0.05 was considered as statistically signi cant.
Ethics approval. All procedures performed in the present study involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was approved by the institutional review board of the Kansai Medical University Hospital, Japan (Approval No. 2020215), and informed consent was obtained from all individual patients prior to robotic partial nephrectomy.    a Postoperative eGFR decrease (continuous variable) with our classi cation system. b Intraoperative WIT with our classi cation system. c Comparison of the AUCs of our classi cation system, the DAP sum score, the R.E.N.A.L. score, and the PADUA score for predicting eGFR reduction of ≥20%. AUC area under the receiver operating characteristic curve, DAP diameter-axial-polar, eGFR estimated glomerular ltration rate, PADUA preoperative aspects and dimensions used for anatomic, R.E.N.A.L. radius,