The Role of R.E.N.A.L. Nephrometry Score in the Era of Robot-Assisted Partial Nephrectomy

Introduction

I n determining the optimal surgical treatment for renal tumors, the anatomic features of the tumor play an important part. The R.E.N.A.L. (radius; exophytic/endophytic; nearness; anterior/posterior; location) nephrometry score (NS) was developed to quantify the anatomic characteristics of renal tumors in an objective manner. ¹ With increasing use of the da Vinci system (Intuitive Surgical Inc, Sunnyvale, CA), even complex renal tumors can now be managed with robot-assisted partial nephrectomy (RAPN). ^{2 –4} We applied the R.E.N.A.L. NS in our series of patients undergoing partial nephrectomy (PN) performed purely with robotic assistance and studied its role in predicting clinical outcomes.

Patients and Methods

All cases from a single surgeon's series (CPS) of minimally invasive PN performed at a tertiary care academic medical center from 2003 to 2011 were reviewed. Patients' demographic and clinical data were retrieved from our Institution Review Board-approved laparoscopic nephrectomy database (Table 1). Standard laparoscopic PN was performed before 2008, after which only RAPN was performed. Of 99 minimally invasive PNs, 16 standard laparoscopic cases were excluded. The remaining 83 cases performed with robotic assistance were included for this study. These included cases that were converted to standard laparoscopy during dissection because of intraoperative findings and considered as converted cases during data analysis. In the same study period, there were 27 open PN performed by the same surgeon, and the NS of these cases were compared in a subanalysis.

The R.E.N.A.L. NS was assigned in all 83 cases by a trained urologist using the preoperative cross-sectional CT scan. The nephrometry scoring methodology as described by Kutikov and colleagues ¹ can also be found online at www.nephrometry.com. For the purpose of this study, a urology resident independently reviewed the CT scans of these patients to reduce interobserver variability. When discrepancy was encountered, a consensus was made by the two investigators to arrive at the final NS.

Because of the small number of tumors with NS 10 to 12 in this study, we categorized all tumors as noncomplex (NS 4–6) or complex (NS 7–12) for the purpose of data analysis. The tumor complexity and each of the five components of the NS were analyzed for association with perioperative outcomes. The perioperative outcomes analyzed as dependent variables included warm ischemia time (WIT), operative (OR) time, estimated blood loss (EBL), conversion (to standard laparoscopy, open surgery, or radical nephrectomy), complications, and decrease in postoperative glomerular filtration rate (GFR) at last follow-up by more than 10% compared with the preoperative value. GFR was calculated by the Modification of Diet in Renal Disease formula. ⁵ Complications were classified according to the Clavien-Dindo classification system. ⁶ All statistical analyses were performed using SPSS package v 19 (SPSS Inc, Chicago, IL). Chi-square tests and Mann-Whitney or Kruskal-Wallis tests were used to analyze categorical and continuous outcomes, respectively, between nephrometry groups with type I error set at 5%. To analyze the relationship between the overall NS and outcomes, univariate linear and logistic regression analyses were performed for continuous and categorical outcomes, respectively.

Results

Of the 83 patients in the study, the CT scans of 42 patients were available for independent review by a second physician investigator. There was strong correlation between the two sets of independent reviews in these 42 cases (Spearman correlational coefficient 0.814, P<0.001). The NS and each composite score of the 83 tumors are presented in Table 2. Median pathologic tumor size was 2.4 cm (interquartile range [IQR] 1.2 cm), and malignant tumors were confirmed in 80.7% (Table 3). Median OR time was 209 minutes (IQR 60 min), WIT was 22 minutes (IQR 9.5 min), and EBL was 100 mL (IQR 100 mL).

There were four (4.8%) intraoperative and 11 (13.2%) postoperative complications. Intraoperative complications consisted of three vascular injuries and one splenic injury that were all managed laparoscopically. There were five (6.0%) Clavien grade II or higher postoperative complications. These included two postoperative hematuria (one necessitating angioembolization for pseudoaneurysm), one febrile urinary infection, one coronary event, and one cerebrovascular accident. The patient with the coronary event had chest pain 4 weeks after a normal hospital course. Cardiac angiography subsequently revealed no abnormality. The other patient who had history of transient ischemic attacks, hypertension, and diabetes had an acute right ischemic stroke on the first postoperative day. She was treated with antiplatelet agents and made a quick functional recovery. There was only one case of urine leak that resolved spontaneously after 10 days with the drain left in situ. This was in a tumor with NS of 4 for which we performed the standard two-layer renorrhaphy.

There were eight (9.6%) conversions, two of which were converted to standard laparoscopic radical nephrectomy. The first case was for a 5 cm midpolar tumor with densely adherent perirenal fat that could not be identified intraoperatively despite the use of ultrasonography. The other case involved a 2.8 cm hilar tumor that was successfully excised robotically. Frozen section results of the tumor bed tissue were positive for tumor. Because further resection of the tumor would result in compromise of the hilar vessels, the case was converted to radical nephrectomy. At median follow-up of 3.0 months, 39.8% of patients had more than 10% decrease in GFR. The postoperative outcomes are shown in Table 4.

Comparing between noncomplex (NS 4–6) and complex tumors (NS 7–12), statistical differences were found in OR time (181 min vs 215 min, P=0.028) and WIT (21 min vs 24 min, P=0.006). There were no differences in EBL, complication rates, conversion rate, and decrease in GFR. Using the overall NS in a univariate regression analysis, only WIT (P=0.007) was associated with increased NS. We then analyzed each component of the NS as an independent variable with the perioperative outcomes. Only (N)earness and (L)ocation were related to perioperative outcomes. The N score was associated with WIT (21 min vs 20 min vs 25 min, P=0.013) and intraoperative complications (0 vs 15.0% vs 3.0%, P=0.044). Both these outcomes were not significant in univariate regression analysis. The L score was associated with WIT (20 min vs 24 min vs 23 min, P=0.021), and this was significant on univariate regression analysis (P=0.031). The type I errors of these analyses are shown in Table 5. In a subanalysis comparing 27 open PNs performed in the study period by the same surgeon, the open PN group had a higher mean NS than the robotic group (8.47, standard deviation [SD] 1.77 vs 6.40, SD 1.72, P<0.001).

Discussion

Numerous indices had been developed to objectively characterize the anatomic attributes of renal tumors. The Preoperative Aspects and Dimensions Used for an Anatomic classification was introduced in 2009, and it was found to predict the risk of complications in PN. ^7,8 The Centrality Index, described in 2010, was applied to a series of 133 laparoscopic PNs and was found to be significantly associated with WIT, OR time, and complication rates. ⁹ The R.E.N.A.L NS was developed in 2009 by Kutikov and colleagues. ¹ As described by the authors, it was originally meant to communicate salient tumor anatomy in a reproducible manner. Subsequent studies did show associations between NS and certain perioperative outcomes, such as WIT and urine leak. ^10,11 These studies, however, were not conducted in a pure robot-assisted series, as in this study.

When limited to robot-assisted cases, our results suggested that the NS was limited in predicting perioperative results in RAPN. Of the outcomes analyzed, nephrometry complexity was statistically associated only with OR time and WIT. While statistically significant, the differences were not clinically significant. This was supported by our findings that postoperative and long-term morbidity, such as complication rates and worsening of GFR, were similar in the two groups. The same results were obtained when using the overall NS in a univariate regression analysis, where there were no increase in rate of complications or deterioration of GFR with increasing NS. Of the five components of the NS analyzed, only the L score was statistically related to WIT on univariate regression analysis.

Before the NS was developed, Rogers and associates ³ had reported a multi-institutional experience of 11 patients with complex hilar tumors who underwent RAPN with satisfactory outcomes. In a recent series of RAPN with reported NS, White and colleagues ¹² identified 67 patients having moderate to high complexity tumors with NS of at least 7 who underwent RAPN. Median EBL was 200 mL, and WIT was 19.0 minutes. Clavien grade II or higher complications occurred in 13 patients. Gupta and coworkers ¹³ also reported their experience of RAPN in 19 patients with tumors greater than 4 cm and median NS of 9 (range 6–11). Overall complication rate was 6%, and median WIT was 36 minutes (range 17–61 min). These studies suggested that complex tumors or those with high NS were still amenable to RAPN with satisfactory outcomes. Even in a pure laparoscopic PN series, Hayn and associates ¹⁴ had found statistical differences in WIT and EBL between nephrometry groups but no difference in preoperative and postoperative GFR, transfusion rates, and complication rates.

If a high NS does not preclude a patient from RAPN, how is it relevant in the contemporary era where the da Vinci platform is increasingly used in PN? Even in the robotic era, there continues to be a role for the open approach for specific situations, depending on the size and location of the tumor, the renal function, and the surgeon's previous experience. Previous studies had also shown that a high NS was associated with open PN compared with minimally invasive PN. In a study by Canter and colleagues, ¹⁵ a higher NS was associated with 224 open PN compared with 183 minimally invasive PN (mean NS 8.19 vs 6.62, P<0.001). From our own experience, we found that the mean NS in the open group was also significantly higher than the robotic group (8.47 vs 6.40, P<0.001). This suggested that the NS could objectively aid in the surgical treatment decision making between open or minimally invasive PN.

The NS can also be applied practically during RAPN in difficult situations. In comparing the (A)nterior/posterior groups, we found that tumors that necessitated conversion were more likely to be posteriorly located, but this did not achieve statistical significance (17.1% vs 4.2%, P=0.065). This was consistent with the experience from standard transperitoneal laparoscopic PN in which posterior lesions were technically more difficult to excise and subsequently reconstruct. ¹⁶ Therefore, in posteriorly located tumors, we would mobilize the kidney extensively to expose the tumor. Even so, the patient would be counseled on the higher possibility of conversion to open surgery or radical nephrectomy.

Our study had several limitations. Because there were only two tumors with NS 10 to 12, we had to group tumors with NS 7 to 12 as a single group of complex tumors. It is unknown if tumors with NS 10 to 12 would have inferior outcomes. To mitigate this limitation, we treated NS as a continuous variable and performed a regression analysis. In addition, in the study by White and colleagues ¹² in which 11 of 106 tumors were high complexity, they showed that these tumors displayed no difference in postoperative GFR and complication rates. Another limitation is the short follow-up period, with a median follow-up of 3.0 months. We therefore analyzed only the perioperative outcomes and postoperative GFR. Analysis of long-term oncologic outcomes and the relation to NS need longer follow-up. In terms of short-term oncologic result, our low margin positivity rate of 1.2% was comparable with most robot-assisted PN series. ^17,18

Conclusion

As the use of the da Vinci platform increases, more urologists will be performing RAPN for both complex and noncomplex renal tumors. In our study, we found that while nephrometry complexity was associated with OR and ischemia times, neither of them were clinically significant. Postoperative complications, conversion, and renal function were not affected by nephrometry. RAPN can, therefore, be safely performed even for complex tumors. In these cases, the NS may still be applied in an individualized and practical manner.