Abstract
Aim: To evaluate the Mayo Adhesive Probability (MAP) score, renal pelvis score, and RENAL nephrometry score for the prediction of surgical outcome in patients with renal masses undergoing laparoscopic partial nephrectomy at a single center. Patients and Methods: A total of 280 patients who underwent laparoscopic partial nephrectomy were identified retrospectively. Thirty-eight patients were excluded because of a lack of preoperative imaging. The outcome measures included surgical technique, patient characteristics, MAP score, renal-pelvis-score, RENAL nephrometry score, and complication rates according to the Clavien-Dindo classification. Regression analysis was performed for assessment of the predictive value of the given scores. Results: Complications occurred after 32 (13%) operations. There was a significant positive association between the development of complications and RENAL nephrometry score (p=0.003). Prediction of complications was improved by the RENAL nephrometry score [area under the curve (AUC) =0.675] and the MAP score (AUC=0.655): With an increasing MAP score, there was a significantly increased operative time (p=0.033). The renal pelvis score had a minor predicitive role (AUC=0.516) and no correlation was found with postoperative urine leakage. Conclusion: The MAP score and RENAL nephrometry score seem to be able to predict a complex or complicated intra- and postoperative course, while the renal pelvis score is not suitable for predicting postoperative complications, especially urine leakage.
- Laparoscopic partial nephrectomy
- renal cell carcinoma
- complications
- Mayo Adhesive Probability score
- renal pelvis score
Radical nephrectomy in patients with small renal cell carcinoma has been supplemented by partial nephrectomy, leading to improvements in overall survival (1, 2). It has been assumed that potential negative effects of an ischemic time of between 20-30 minutes are reversible regarding renal function (3). However, it has become apparent that the detrimental effects of ischemia become more important in patients with pre-existing comorbidities, renal dysfunction and advanced age. Therefore, even in laparoscopic partial nephrectomy (LPN), the warm ischemic time (WIT) should be minimized or avoided whenever possible. In this context, tumor size and location, as well as patient predisposition may compromise feasibility and surgical outcome in LPN (3-5). It is unclear whether these factors should influence the decision of which laparoscopic technique to use in LPN, aiming to minimize surgical complication rates. Significant bleeding and urinary leakage are the most common complications after partial nephrectomy (3). Several investigations were performed to investigate whether different radiographic scores such as the Mayo Adhesive Probability score (MAP score), the renal pelvic score (RPS) or the RENAL nephrometry score (RNS) are able to predict the risk of blood loss, urinary leakage, or difficult intraoperative conditions with an increased risk of postoperative complications in LPN (6-8).
The RNS is an established score used to evaluate the complexity of the tumor and should thus also predict the complexity of the surgery. It is based on five different reproducible features: radius of the tumor, exophytic properties, proximity to the hilum, the tumor location relative to the kidney polar line and anterior or posterior location (4). The MAP score evaluates the presence of adherent perinephric fat, with a possible negative effect on the intraoperative course (7). The RPS evaluates whether the renal pelvis is endophytic or exophytic, thus, this score should be predictive of the occurrence of urinary leakage (8).
The aim of the presented study was to compare these different radiographic scores and their potential for predicting intraoperative complexity or postoperative complications in LPN.
Patients and Methods
We retrospectively identified 280 patients who underwent LPN at our Department between 03/2006 and 01/2014. All patients had a preoperative suspicious renal mass as determined by computed tomography (CT) or magnetic resonance imaging. Institutional Review Board approval was obtained before the start of data collection and analysis (123/2013 BO2). By means of the surgical technique performed, patients were divided into on-clamp and off-clamp surgical groups.
Data analysis comprised patient variables (age, gender and tumor entity), perioperative variables (on and off-clamp technique, operating time, WIT, blood loss and the use of transfusion), postoperative outcome (length of hospital stay, complications according the Clavien-Dindo classification) and pathology report (surgical margins, final histology). Parametric data are reported in a descriptive fashion using the mean and standard deviation (SD). Cohort differences were assessed by Wilcoxon–Kruskall–Wallis tests for continuous data and chi-square tests for categorical data. Statistical analysis was performed using JMP 7.2 (SAS Inc., Cary NC, USA).
Two readers (GB and KS) assessed the different radiographic scores, blinded to the available written report of the initial image analysis. Altogether, 242 patients with preoperative imaging were evaluated (and compared with different surgical techniques). In 38 patients, the imaging was either of poor quality or not digitally available, hence there was no possibility to evaluate the different radiographic scores sufficiently.
For the RNS, the readers assessed the radius of the tumors, exophytic properties, proximity to the hilum, as well as tumor location relative to the polar line as described by Kutikov et al. (4).
For the evaluation of the MAP score, the perinephretic fat was determined at the level of the renal vein. Regarding the MAP score, the posteriorly located fat of the kidney was evaluated as described by Davidiuk et al. (7) (see Figure 1).
The RPS, defining the intra- or extraparenchymal localization of the renal pelvis, was assessed as proposed by Tomaszewski et al. (4, 8) (exemplarily see Figure 2).
For the prediction of the primary endpoint of peri- and postoperative complications (occurrence of complications, yes vs. no), each radiographic score was analyzed by logistic regression analysis. Sensitivity and specificity were calculated and the predictive ability of each score was evaluated by the determination of the area under the curve (AUC) based on a receiver operator characteristics (ROC) analysis.
Results
Patient characteristics and histology. Detailed patient characteristics for each group are summarized in Table I.
The histopathological evaluation revealed the presence of renal cell carcinoma in 175 patients (72%), oncocytoma in 17 (7%), angiomyolipoma in 15 (6%) and complicated cysts in 31 (13%) patients. In four (2%) patients, a metastasis of other malignancy was found and removed. Among renal cell carcinomas, there were 112 clear cell, 47 papillary, 15 chromophobe and one cystic. Five patients had a positive surgical margin in the final pathological evaluation. Renal cell carcinoma tumors were classified as pT1a in 131, as pT1b in 27, as pT2 in one and as a pT3a in 16 patients according to the 2010 TNM classification (9).
Surgery and complications. The mean (±SD) operative time was 134 (±34) minutes and the WIT in the on-clamp group was 14.5 minutes (±11). In no case did the laparoscopic approach lead to a conversion into open partial nephrectomy or open nephrectomy. A total of 68 partial nephrectomies were performed in an off-clamp technique and 174 operations in an on-clamp technique. The mean operative time did not differ significantly between the two groups (p=0.06).
According to the Clavien-Dindo classification, overall complications were reported in 32 (13%) patients with LPN. There were no grade IV or V complications reported in the overall postoperative course. All postoperative complications were evaluated within 30 days after surgery. There was no grade I complication documented (Table II).
RNS. Using the RNS, 242 of our patients were categorized into different complexity groups, RNS 4-6: low complexity, RNS 7-9: moderate complexity, and RNS >9: high complexity. LPNs were performed in 166 patients with low complexity (69%), in 71 (29%) with moderate complexity and in five patients (2%) with high complexity. There were significantly more complications documented in the group of tumors of moderate and high complexity as compared to the group with tumors of low complexity. In the group of patients with tumors of low complexity, 14 patients (8.4% of these group) had a complication and in the group of patients with moderate and highly complex tumors, 17 patients (22.4% of the patients in these groups) had intra- or postoperative complications (p=0.003).
Logistic regression analysis for prediction of complications yielded an AUC of 0.675 for the RNS (p=0.037; r2=0.096), with a sensitivity of 65.6% and a specificity of 58.1% (cut-off value: ≥7). Regarding an RNS cut-off value of >9 for tumors with high complexity (as defined above), sensitivity and specificity were 96.9% and 8%, respectively.
RPS. The localization of the renal pelvis was classified by the RPS into intrarenal (n=160, 66.1%) or extrarenal (n=82, 43.9%). There was no significant difference between the groups concerning the incidence of urinary leakage or other complications (Table III). The sensitivity of the RPS for the prediction of the overall complication rate was 37.5% (specificity: 65.6%; AUC: 0.516; p=0.73; r2=0.0009).
MAP score. The posterior perinephretic fat thickness and the stranding of the fat were evaluated by the MAP score. The characteristics are given in Table IV.
The whole patient group was divided into those with a MAP score ≤1 and >1. In patients with a MAP score greater than 1, surgical time was significantly prolonged as compared to patients with a MAP score of 1 or less (p=0.033). No significant association between the MAP score and the overall complication rate (p=0.116) or the WIT (p=0.625) was observed.
AUC for the MAP score was 0.655 (p=0.089; r2=0.069). Using a cut-off-value of ≥3, the sensitivity for predicting peri-/postoperative complications was 87.5% (specificity 36.2%).
Discussion
Partial nephrectomy is a standard procedure in treating renal cell carcinoma and is also performed by the laparoscopic approach. The surgical technique involves a certain risk of intra- and postoperative complications, especially bleeding, urinary leakage and formation of vascular pseudoaneurysms. Currently, different radiographic scores aim to evaluate the risk of postoperative complications by means of preoperative imaging. Use of a predictive score could allow unexpected intraoperative findings, which might increase the risk of intra- or postoperative complications, to be avoided. In the present study, different radiographic scores and their reliability with regards to the respective diagnostic accuracy of predicting intra- and postoperative complications were assessed in a large cohort of patients who had undergone LPN.
The overall complication rate was 13% and was comparable to other studies of LPN or robotic partial nephrectomy, in which a complication rate of 10-21% was reported (10-13).
As expected, tumors with a moderate and high complexity were significantly associated with a higher complication rate compared to tumors with low complexity. This is in line with another study which applied the RENAL score. Liu et al. compared tumors of low, moderate and high complexity which were treated with LPN. They observed that significantly more overall complications (p=0.01) occurred in those with moderate to high complexity tumors than in those with low complexity tumors when classified by the RENAL score (14). A second study also revealed that a higher nephrometry score was significantly associated with higher complication rates (6). In the cohort presented in our study, the surgical time and WIT were significantly prolonged in the intermediate complexity group as compared to the low complexity group (p=0.001 and p=0.0001). An increased RNS was predictive for the occurrence of complications during or after surgery with an AUC of 0.675. Based on these findings, RNS is a good predictor of overall complication rates in LPN.
For the MAP score, a significantly prolonged surgical time was observed in patients with a score of >1 (p=0.033). As reported by Davidiuk et al., adherent perinephric fat is present in 31-100% of patients with a MAP score of 2-5 (7). This fact might lead to a higher grade of surgical difficulty in LPN, with the consequence of a prolonged operative time. Therefore, the MAP score might be used as a scoring system for predicting the intraoperative course in the presence of perinephritic fat. The results presented here demonstrate that the MAP score, with an AUC of 0.655, is nearly equivalent for the prediction of overall complications to the RENAL score. However, this study did not find any significant correlation between different MAP scores and the severity of complications or a longer WIT.
The RPS is an additional tool for risk stratification in patients with renal masses. Compared to the results of Tomaszewski et al. (8), the predictive value of the RPS seems to be minor compared to the RNS and MAP score. Moreover, we did not find an increased incidence of urinary leak (p=0.97) nor a higher complication rate (p=0.47) when dividing the collective by the use of the RPS. Even though 158 (66.1%) patients had an intrarenal pelvis and 81 (43.9%) patients an extrarenal pelvis, there was no significant difference with regard to the surgical approach (on-clamp vs. off-clamp), urinary leakage or complication rate. One possible explanation for this might be increasing surgical experience and technical equipment, resulting in improved surgical outcome. On the other hand, improved radiological imaging modalities (e.g. advanced magnetic resonance imaging and computed tomographic techniques) allow improved risk stratification for choosing the proper surgical approach. However, the retrospective evaluation of imaging and the low number of events possibly does not allow final conclusions regarding RPS and its value in the perioperative management of LPN to be drawn.
There are several limitations to this study, such as its retrospective design, the highly selective study population and surgery by different surgeons. However, this study reports on a comparably large cohort of patients who underwent LPN, and how these patients can be selected prior to surgery, using various radiological scores.
In conclusion, the MAP and RENAL scores are suitable for imaging-based approaches to predict complex intraoperative conditions and postoperative complications, and are superior to the RPS, even in terms of prediction of postoperative urinary leakage.
Therefore, the MAP score as well as the RENAL score are able to prepare the surgeon for intraoperative and postoperative conditions.
Footnotes
↵* These Authors contributed equally to this study.
Conflicts of Interest
None.
- Received December 21, 2016.
- Revision received February 6, 2017.
- Accepted February 8, 2017.
- Copyright© 2017, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved