Introduction

Partial nephrectomy is the preferred treatment for all renal masses <7 cm and for masses >7 cm when technically feasible [1]. As surgical technique and surgeon experience have evolved, increasingly complex tumors are being managed with partial nephrectomy (PN) [2]. When compared to radical nephrectomy, PN has been shown to have at least equivalent or better oncologic outcomes with improved long-term renal function and a lower incidence of subsequent chronic kidney disease [3–5]. Historically, open PN (OPN) has been the standard approach for small renal masses, but advancements in minimally invasive techniques have led to increasing utilization of laparoscopic and robot-assisted surgical approaches [6].

Laparoscopic PN (LPN) is associated with a shorter hospital stay, decreased postoperative pain and decreased transfusion requirements while maintaining equivalent oncologic outcomes as compared to OPN [7, 8]. Prior to the introduction of robot-assisted partial nephrectomy (RAPN), OPN was preferred over LPN for more complex masses, due in part to the steep learning curve and technical challenges of LPN [3]. Since its introduction in 2004, RAPN has been shown to be feasible and safe for small renal tumors and confers many of the same benefits of LPN with equivalent outcomes [9–11]. Proponents of RAPN suggest that RAPN may be able to overcome some of the issues facing LPN while maintaining the benefits of minimally invasive surgery. Though studies have compared practice patterns of surgeons and their utilization of different modalities, none have done so with regards to the complexity of the tumors resected [12]. The objective of this study is to determine how PN tumor complexity, based on RENAL nephrometry score, has changed over time and how the introduction of robotic surgery had affected this change [13].

Materials and methods

Data

We retrospectively reviewed the data of 608 patients who underwent PN performed at Mayo Clinic Arizona (MCA) from November 1999 to July 2013 from a prospectively maintained IRB approved registry. Our laparoscopic experience has been reported previously [14, 15]. We excluded 316 patients for reasons including: patients with solitary kidneys, multiple renal masses, a follow-up time of less than 90 days or incomplete data such as inadequate cross-sectional imaging.

Study population, definitions, and statistics

RENAL nephrometry scores were calculated for the remaining 291 patients from cross-sectional imaging done <1 month prior to their surgery. The subjects were stratified into two groups: group 1 included 109 patients who underwent OPN (n = 30) or LPN (n = 79) prior to the introduction of robotic surgery to MCA in late 2007, and group 2 included 182 patients who underwent OPN (n = 37), LPN (n = 55) or RAPN (n = 90) after this time. Patient demographics (gender, age), clinical data [BMI, American Society of Anesthesiologists (ASA) score, history of diabetes or hypertension], perioperative outcomes (estimated blood loss, surgery time, units of blood transfused and length of stay) and post-surgery complications (Clavien classification score) were analyzed. Change in estimated glomerular filtration rate (eGFR) was assessed based on preoperative and most recent creatinine values using the chronic kidney disease epidemiology collaboration (CKD-EPI) equation, due to its increased accuracy over the modification of diet in renal disease equation, especially with eGFR’s ≥60 mL/min/1.73 m2 [16, 17].

Data are presented as median [interquartile range (IQR)] or frequency (percentage) for continuous and categorical variables, respectively. For group 1, continuous variables were compared using the Student t test for normally distributed data or the Mann–Whitney U test as appropriate, and categorical variables were compared using Fisher’s exact test. For group 2, continuous variables were compared across the three approaches using ANOVA for normally distributed data or Kruskal–Wallis test as appropriate, and categorical variables were compared using Pearson Chi-square test. For all data, a two-sided p value of 0.05 was considered statistically significant. Statistical analyses were performed using STATA 12®.

Results

Demographic characteristics are summarized in Table 1. No significant differences existed between approaches in either the pre-robotic era (PRE) or robotic era (RE) for any of the demographics data analyzed including gender, age, ASA score, comorbidities, BMI or pre-surgery eGFR. In the PRE, significantly more LPN patients had previous abdominal surgery, while there were no significant differences in the RE. OPN in both groups did have the highest percentage of patients with diabetes (PRE: 27.6 %, RE 27 %) and the lowest preoperative eGFR (PRE: 61.2, RE 75.9) though these differences were not statistically significant. In the PRE, OPN had significantly higher RENAL nephrometry scores compared to LPN (p = 0.009), and OPN had the only case with a RENAL nephrometry score ≥10. Table 2 summarizes RENAL nephrometry in the RE. In the RE, RAPN and OPN were significantly more complex than LPN (p = 0.007 and p = 0.002, respectively) but OPN and RAPN did not differ significantly (p = 0.14). Additionally, nine out of the ten highly complex tumors (RENAL score ≥10) resected in the RE, were performed via RAPN.

Table 1 Demographics
Full size table
Table 2 RENAL nephrometry scores in the robotic era
Full size table

Table 3 summarizes perioperative outcomes. In the PRE, estimated blood loss (EBL), transfusion rate, length of stay (LOS) and Clavien scores were significantly higher in OPN compared to LPN (p < 0.001) with no significant difference in surgery time (p = 0.096). In the RE group, significant differences existed for EBL, transfusion rate, surgery time, LOS (p < 0.001) and Clavien scores (p = 0.044). There was significantly less blood loss in LPN (median 100 mL) compared to both RAPN (median 150 mL; p = 0.046) and OPN (median 250 mL; p < 0.001), and significantly less blood loss in RAPN compared to OPN (p = 0.019). RAPN had a significantly longer median surgery time (196 min) compared to both LPN (151 min; p < 0.001) and OPN (147 min; p < 0.001). Transfusion rate and LOS were both significantly higher in OPN compared to both LPN (p < 0.001) and RAPN (p < 0.001) with no significant differences between LPN and RAPN (p = 0.53 and p = 0.17, respectively). Clavien scores were significantly higher in OPN compared to RAPN (p = 0.013). In the PRE, OPN had a significantly lower eGFR compared to LPN (p = 0.015), but change in eGFR was not significantly different between the groups. In the RE, there was a significant difference in change of eGFR between the three approaches: LPN eGFR change was significantly less than OPN (p = 0.0097) and RAPN was less than OPN though not significantly so (p = 0.056).

Table 3 Perioperative outcomes
Full size table

Discussion

RAPN has become a well-established approach for renal masses and an approach that is increasingly utilized by surgeons. One question is whether or not it provides a significant advantage over conventional laparoscopic techniques especially when cost containment is a concern. Conventional wisdom has always been that there may be certain complex tumors that are easier to perform through an open incision. Surgeon skill and experience also play a role and have been difficult to control for in attempts in comparative effectiveness analyses. Another concern has been that nephron-sparing surgery has not been adequately utilized in complex tumor cases with a default to laparoscopic radical nephrectomy [18]. In this retrospective analysis, we demonstrated that RAPN can safely and effectively be utilized for even the most complex renal masses that would have previously been reserved for OPN in the pre-robotic era.

In the PRE, the only significant differences in demographics between LPN and OPN groups were the incidence of previous abdominal surgery (greater in the LPN group) and the complexity of the tumor, as assessed by RENAL nephrometry score, with the more complex tumors being performed with OPN. The complexity of tumors resected via LPN increased from the PRE to the RE, likely reflecting increased experience and comfort with the approach. In fact, LPN included one out of the ten RENAL scores ≥10 in the robotic era. Still, the complexity of LPN in the robotic era was still significantly lower than both RAPN and OPN. This reinforces the argument that RAPN may enable surgeons to overcome both the steep learning curve of LPN and the limitations of LPN instrumentation with RAPN, allowing surgeons to undertake more complex cases with a minimally invasive approach.

In the robotic era, RAPN was utilized for tumors of similar complexity to OPN with significantly improved perioperative outcomes including lower EBL, transfusion rate, hospital stay and postoperative complication rate. Change in eGFR was also better in RAPN nearing significance (p = 0.056). While we do not suggest that RAPN allows for better preservation of renal function, it certainly does not result in worse functional outcomes. In fact, the only data point where OPN was superior to RAPN, as expected, was surgical time. Interestingly, though RENAL nephrometry scores were not significantly different between OPN and RAPN, RAPN was utilized nine out of ten times for RENAL scores ≥10, while none of the RENAL scores for OPN were ≥10. With the superior perioperative outcomes after RARP in mind, this is further evidence that RAPN can be safely utilized for nephron-sparing surgery on highly complex renal masses.

While the short-term benefits of minimally invasive surgery are demonstrated in the pre-robotic era with respect to estimated blood loss, length of stay, transfusion rate and complications, it would be easy to attribute much of this to a potential selection bias that existed with respect to tumor complexity. Clearly patients with more complex tumors and possibly more “need” for nephron-sparing due to comorbidities were being steered towards open surgery. This would explain the differences in perioperative outcomes. This has always been an issue when comparing minimally invasive approaches to open surgery for partial nephrectomy. However, when the robotic approach was introduced at MCA, the tumor complexity was essentially controlled for and still there were statistically significant short-term benefits with respect to estimated blood loss, transfusion rates, length of stay and complications. Surely, if something beyond a larger incision was influencing these factors such as complexity one would expect a difference in nephrometry scores between open and minimally invasive approaches, but that was not the case.

The limitations of the current study include its retrospective design and single center data. Although the surgeons included in this study have been performing laparoscopic surgery since 1999, differences in experience and comfort between surgeons may have influenced different utilization rates of the modalities. In addition, outcomes may have been influenced by surgeon technique and experience rather than the approach itself. Another limitation was our inability to match LPN, OPN or RAPN due to the size of the cohort, though the analysis showed no significant differences in demographic data between the three modalities. Despite these limitations, the bottom line is that the complex tumors that previously were managed with open partial nephrectomy techniques are now regularly being managed with the robotic approach. Most importantly, the robot-assisted approach may now help close the gap that has long existed between nephron-sparing surgery and radical nephrectomy, in so much as facilitating the use of partial nephrectomy.

Conclusions

Robot-assisted surgery has facilitated the minimally invasive treatment of complex renal masses in our practice. This approach appears to be safe with preservation of the benefits of minimally invasive surgery for even the most complex of tumor.