Evaluation of robotic and laparoscopic partial nephrectomy for small renal tumours (T1a)

Authors


Correspondence: Kamran Ahmed, MRC Centre for Transplantation, King's College London, King's Health Partners, St Thomas Street, London SE1 9RT, UK.

e-mail: k.ahmed@imperial.ac.uk

Abstract

Objective

  • To compare laparoscopic partial nephrectomy (LPN) with robotic PN (RPN) using meta-analytical techniques, since there has been a rise in the incidence of small renal masses (SRM; <4 cm) minimally invasive approaches are becoming more popular in dealing with such pathologies.

Materials and Methods

  • A systematic review of the literature was performed to identify studies comparing LPN and RPN.
  • Comparative studies evaluating RPN and LPN that fulfilled the inclusion criteria were selected.
  • Data on preoperative, operative (operative time, estimated blood loss [EBL], and warm ischaemia time [WIT]), postoperative (length of stay [LOS]) variables and complications were collected.
  • A meta-analysis using random effect model was performed.
  • A further Bland–Altman analysis of some of the operative variables was done to compare their reproducibility and mean difference in techniques.

Results

  • Six studies matched the selection criteria. In all, 256 patients were analysed (40% RPN and 60% LPN).
  • There was no significant different in EBL (P = 0.12, 95% confidence interval [CI] –12.01 to 104.26).
  • Similarly, there was no significant different in WIT between the groups (P = 0.23, 95% CI –15.22 to 3.70).
  • Also, LOS (P = 0.22, 95% CI –0.38 to 0.09) and overall postoperative complication rates were not significantly different between the groups (P = 0.84, 95% CI –0.05 to 0.06).

Conclusions

  • Despite multiple studies reporting better perioperative variables for RPN, the present study found no significant differences between RPN and LPN. This has implications for both the surgeon and the patient.
  • Lack of randomised controlled trials in addition to a lack of long-term oncological data for RPN are current limitations.
Abbreviations
EBL

estimated blood loss

LOS

length of stay

OR

odds ratio

(L)(R)PN

(laparoscopic) (robotic) partial nephrectomy

SRM

small renal mass

WMD

weighted mean difference

Introduction

The increase in the use of new and different imaging methods over recent decades has led to an increase in incidental finding of solitary renal masses [1]. Small renal masses (SRMs; <4 cm) represent 48–66% of cases [2]. With more than half of SRMs being surgically removed, surgical excision remains the most definitive therapy for such solitary lesions. Radical nephrectomy was previously considered the ‘gold-standard’ treatment for SRMs of <4 cm. However, nephron-sparing robotic-assisted partial nephrectomy (RPN) has been gaining popularity in treating peripherally placed renal tumours, particularly in patients where preservation of renal function is important [3, 4]. Thus minimally invasive PN is increasingly replacing open radical nephrectomy and PN because of its equivalent efficacy and the advantages it potentially offers; namely lower intraoperative blood loss, a reduced length of stay (LOS) and acceptable warm ischaemia time (WIT) [3]. The evolution from open radical nephrectomy through open PN and now LPN and RPN is mostly due the advantages of robotic over laparoscopic surgery, which has been evaluated in different fields of surgery [5, 6]. In addition, RPN has the added advantage of aiding the learning curve for management of complex renal tumours [7, 8].

With a rise in incidental renal cell tumours and the ever expanding indications for robotic surgery in urology, in particular PN, an evaluation of laparoscopic and robotic approaches is required. This is warranted as most of the evidence for robotic surgery is generated from small single centre experiences or retrospective studies. In the present study we aimed to compare the RPN and LPN approaches for the treatment of SRMs. We evaluated intraoperative and postoperative variables to compare the two methods.

Materials and Methods

Study Selection

A systematic review of the literature was performed using MedLine (1950–present), EMBASE (1950–present), PubMed (1950–present), and the Cochrane database, to identify all studies published between 2000 and 2012 that compared RPN and LPN. The search terms used were ‘robotic’, ‘laparoscopic’, ‘partial nephrectomy’, ‘small renal tumour’, ‘nephron-sparing surgery’, ‘comparative’ and ‘minimally invasive surgery’. A combination of the terms mentioned was used: ‘robotic’, ‘laparoscopic’, ‘partial nephrectomy’, ‘comparative’ and ‘robotic versus laparoscopic partial nephrectomy’. The search function was used to broaden the search and all abstracts, studies, and citations were obtained and reviewed. The last search was performed on 4 June 2012.

Data Extraction

Each study was independently evaluated by two reviewers (S.F. and K.A.) for inclusion or exclusion from the review and the following data was extracted: first author, year of publication, characteristics of the study population, study design, number of patients operated with each technique, tumour size, intraoperative variables (operative time [OPT], WIT, estimated blood loss [EBL], LOS) and postoperative complications within both the RPN and LPN groups.

Inclusion Criteria

To enter our analysis, studies had to fulfil the following criteria: (i) comparison of characteristics of laparoscopic to robotic approaches in patients undergoing PN; (ii) objective evaluation of at least one of the outcome measures mentioned above; (iii) when two studies were reported by the same institution, either the study with the larger sample size or the one of higher quality was included; however, this was not applicable if the outcome measures were mutually exclusive or measured at different time intervals; (iv) studies reporting human subjects.

Exclusion Criteria

Studies that failed to fulfil the inclusion criteria were excluded. In addition, the following criteria were used to exclude studies: (i) studies in which outcomes of interest mentioned below were not reported or if it was impossible to calculate these from the published reports; (ii) studies that focused on LPN or RPN only; (iii) studies that were written in languages other than English; (iv) animal studies.

Outcomes of Interest and Definitions

RPN and LPN were compared based on several preoperative, operative (OPT, EBL, WIT) and postoperative variables (LOS, postoperative complications). To evaluate postoperative complications the Clavien grading system [9] was used.

Statistical Analysis

This meta-analysis was performed in line with recommendations from the Cochrane Collaboration and Met-analysis of Observable Studies in Epidemiological (QUORUM) guidelines [10]. The effect measures estimated were odds ratio (OR) for dichotomous data and weighted mean difference (WMD) for continuous data, both reported with 95% CIs. The OR represents the odds of an adverse event occurring in the RPN compared with the LPN group. An OR of <1 favoured the RPN group. The point estimate of the OR was considered statistically significant at the P < 0.05 level, if the 95% CI did not include the value one.

For continuous variables, the OR was calculated with the Mantle–Haenszel chi square method using a ‘random effects’ meta-analytical technique. The random effect model is particularly suitable for surgical research, as it assumes that there is natural variation between the studies, and the calculated ORs thus have a more conservative value. For continuous variables, e.g. time, statistical analysis was carried out using WMD at the summary statistic. WMD of a negative value favoured the RPN group. For studies that presented continuous data as median and/or range values, the standard deviation (sd) was calculated using statistical algorithms [11]. In reporting the results, a square is indicative of point estimates of the treatment effect (OR or MWD) with 95% CIs indicated by horizontal bars. The diamond represents the summary estimate from the pooled studies with 95% CIs.

A qualitative analysis of the studies was done using the Quality Assessment Tool for Systematic Reviews of Observational Studies [12] and the Newcastle-Ottawa Scale [13, 14] for reporting non-randomised comparative studies. This was altered to suite the analysis and studies included in this evaluation. The domains analysed for quality were patient selection, comparability of the study groups, and assessment of the outcomes. In this study, studies that achieved five or more stars were considered high quality.

All analysis was conducted using Review Manager Version 5 (The Cochrane Collaboration, Software Update, Oxford).

Comparison of Method

A further analysis of reproducibility and agreement between RPN and LPN methods is carried out using the Bland–Altman plot. Reproducibility can be described as the repeatability of measurements in time or by different observations [15]. Data gathered from different studies not included in the meta-analysis are used to compare the mean difference of the intraoperative variables in RPN and LPN. Studies used in this analysis are presented in Tables 1 [5, 7, 16-29] and 2 [30-44].

Table 1. RPN studies
ReferenceNumber of patientsMean:
Tumour size, cmOPT, minWIT, minEBL, mLLOS, days
Passerotti et al. 2012 [20]82.43216.2524213.1252.37
Dulabon et al. 2011 [21]2481.53187.419.6208.22.87
White et al. 2011 [22]673.7180192003
Mottrie et al. 2010 [19]622.891201405
Patel et al. 2010 [23]715238231002
Benway et al. 2009 [17]502.5145.317.8140.32.5
Michli and Parra 2009 [24]202.714228.12632.8
Wang and Bhayani 2009 [16]402.5141201372.5
Jeong et al. 2009 [25]313.4169.920.9198.35.2
Ho et al. 2009 [26]203.582.821.71894.8
Bhayani and Das 2008 [27]352.8142211332.5
Deane et al. 2008 [28]103.122932.11152
Rogers et al. 2008 [7]83.6192312302.6
Rogers et al. 2008 [18]113.820228.92202.6
Kaul et al. 2007 [60]102.315521923.5
Caruso et al. 2006 [29]101.9527926.42402.6
Table 2. LPN studies
AuthorNumber of patientsMean:
Tumour size, cmOPT, minWIT, minEBL, mLLOS, days
Jeon et al. 2012 [32]312.3203.531395.24.7
Olweny et al. 2012 [33]292.617426.41883.4
Olweny et al. 2012 [33]492.717832.82553.6
Haber et al. 2012 [34]483.2227.724391.34.6
Lavery et al. 2011 [30]182.3179.724.7139.72.9
Simmons et al. 2007 [35]200322834.72343.6
Häcker et al. 2007 [36]252.621128.9177.48.3
Venkatesh et al. 2006 [37]1262.6204272693
Weld et al. 2006 [38]602.417926.92262.7
Baughman et al. 2005 [39]472.119320.51882.4
Link et al. 2005 [31]2172.618627.63853.1
Johnston et al. 2005 [40]1002.519125.83582.8
Orvieto et al. 2005 [41]412.2226.529.71502
Beasley et al. 2004 [42]272.4210412502.9
Gill et al. 2007 [43]7712.720130.73003.3
Permpongkosol et al. 2006 [44]852.422529.5436.93.3

Results

Study Selection

In all, 965 potential articles published up to 2012 were identified from the literature searches. After elimination of review papers with further limiting the results to English and non-human papers 191 articles remained. After review of title and abstracts 166 articles were excluded. Of these, 25 articles were selected based on the titles and abstracts, and a full examination of the text was performed. Six studies matched the selection criteria and were suitable for meta-analysis [30, 45-49] (Fig. 1).

Figure 1.

PRISMA chart.

Study Characteristics and Results

The characteristics of these studies are summarised in Table 3 [30, 45-49]. There were no randomised controlled trials comparing the two procedures. Each study had a LPN group as control and RPN group as the experimental arm of the study. Most of the studies had a recent year of publication (2008 or later) and they contained at least 10 patients in both the RPN and LPN groups. On review of the data extraction there was 100% agreement between the two reviewers.

Table 3. Study characteristics
StudyStudy TypeCasesMatchingMean age, yearsExclusion criteriaStudy quality
RPNLPNRPNLPN
  1. Matching: 1, age; 2, BMI; 3, Preoperative eGFR; 4, ASA; 5, gender; 6, tumour size; 7, tumour type; 8, location of lesion; 10, imaging. Study type: R, retrospective; P, prospective; RP, retrospective evaluation of prospectively collected data.
Hillyer et al. 2011 [46]RP9171,2,3,4,86154NR******
Lavery et al. 2011 [30]R20181,2,3,4,6,853.655.4Conversion to laparoscopic radical nephrectomy (LRN) or hand-assisted LRN****
Williams et al. 2011 [49]P27591,2,3,4,5,6,854.655.74NR*****
Seo et al. 2011 [48]R13141,2,854.253.9NR****
Kural et al. 2009 [47]R11201,2,45159NR****
Aron et al. 2008 [45]R12121,2,4,5,7,86461NR******

In all, 256 patients were analysed, of which 101 underwent RPN (40%) and 155 LPN (60%). All patients included in studies had a mean tumour size of <4 cm, with most (>60%) of them being male patients. Tables 4 and 5 summarise the data extracted on the different variables used in meta-analysis [30, 45-49].

Table 4. Variables in the RPN arm of the studies
StudyNumber of patientsMean (range) and/or mean (sd):Complications n (%)
Tumour size, cmOPT, minWIT, minEBL, mLLOS, days
  1. CG, Clavien Grade; NR, not reported.
Hillyer et al. 2011 [46]182.85 (2.4–5)NR19 (17–31)225 (287–450)4 (3.5–5.8)Postop: 2 (11%); blood transfusion (CG II); lymph leak resulting in diet modification (CG II)
Larvery et al. 2011 [30]202.5 (1.2–4.3)189.2 (111–294)22.7 (12–40)93.3 (20–350)2.6 (1–5)Postop: 3 (15%); respiratory distress requiring intubation; pulmonary embolism; blood transfusion due to postop bleed
Williams et al. 2011 [49]272.47 (1.18)233 (43.6)18.5 (7.6)179.6 (199.6)2.51 (1.05)Postop: 5 complications: ileus, UTI, uvular oedema, rash (CG I); angina and fever (CG IIa)
Seo et al. 2011 [48]132.7 (1.2)153.2 (22.3) (110–200)35.3 (8.5) (30–47)283.6 (113.5) (126.8–945.9)6.2 (1.8) (5–10)2 blood transfusion
Kural et al. 2009 [47]113.2 (2–4.1)185 (120–270)27.3 (13–40)286.4 (50–800)3.9 (0.7)NR
Aron et al. 2008 [45]122.4 (6.9)242 (69.2)23 (6.6)329 (315)4.7 (2.4)NR
Table 5. Variables in the LPN arm of the studies
StudyNumber of patientsMean (range) and/or mean (sd):Complications, n (%)
Tumour size, cmOPT, minWIT, minEBL, mLLOS, days
  1. CG, Clavien Grade; NR, not reported.
Hillyer et al. 2011 [46]322.7 (1.9–3.6)NR37 (20–39)175 (112–262)4.5 (1–4)Postop: 4 (12.5%); blood transfusion: angioemoblisation (CG IVa) and renal failure (CG IVa) in 1 patient; prolonged urine leak requiring urethral stent insertion (CG IIIb)
Larvery et al. 2011 [30]182.3 (0.1–4.4)179.7 (132–223)24.7 (18–34)139.7 (25–300)2.9 (1–5)Postop: 2 (11%); pneumonia; renal pseudoaneurysm
Williams et al. 2011 [49]593.08 (2.17)221.4 (54.6)28 (7.6)146.3 (143.4)2.71 (0.85)Postop: 12 complications: 4 CG I: pneumonia; L median Neve Palsy, ileus, UTI; 4 CG IIa: haematuria, ileus, retroperitoneal bleed, wound infection); 4 CG IIb: angioembolisation, clot retention in 3 patients
Seo et al. 2011 [48]142.0 (1.2)117.5 (32.0) (71–190)36.4 (6.8) (27–44)264.1 (163.7) (63.4–718.6)5.3 (0.6) (5–10)0
Kural et al. 2009 [47]203.1 (1.5–7)226 (120–420)35.8 (18–50)387.5 (50–1000)4.27 (1.12)NR
Aron et al. 2008 [45]122.9 (7.1)256 (70.6)22 (10.8)300 (384)4.4 (1.1)NR

Results of Meta-analysis

All studies reported operating on SRMs measuring <4 cm [30, 45-49]. Also, all the studies reported on intraoperative variables. There was no statistical significance between the RPN and LPN for EBL (P = 0.12; MWD 46.13, 95% CI –12.01 to 104.26; Fig. 2 [30, 45-49]); OPT (P = 0.97; MWD 0.50, 95% CI –24.02 to 25.02; Fig. 3 [30, 45, 47-49]); WIT (P = 0.23; MWD –5.76, 95% CI –15.22 to 3.70; Fig. 4 [30, 45-49]) and overall LOS (P = 0.22; MWD –0.15, 95% CI –0.38 to 0.09; Fig. 5 [30, 45-49]).

Figure 2.

EBL.

Figure 3.

OPT.

Figure 4.

WIT.

Figure 5.

LOS.

Four studies reported complications including conversion to open nephrectomy, postoperative blood transfusion, pulmonary embolism, and cardio-respiratory complications. Additionally, there was no significant difference in overall complication rate (both intra and postoperative complications) between the two groups (P = 0.84; MWD 0.01, 95% CI –0.05 to 0.06; Fig. 6 [17, 30, 46, 48-50]). Data are suggestive of no significance difference with heterogeneity noted among different studies.

Figure 6.

Complications (includes both intraoperative and postoperative).

Bland–Altman Analysis

A further 16 studies were used to construct Bland–Altman plots. All 16 studies had data on intraoperative variables. Only the aforementioned variables, e.g. OPT, WIT, and EBL were used to compare correlations between the RPN and LPN groups. Results are detailed in Table 6 and presented in Figs 7-9.

Figure 7.

Bland–Altmann plot for OPT.

Figure 8.

Bland–Altman for WIT.

Figure 9.

Bland–Altman plot for EBL.

Table 6. Bland–Altman analysis for intraoperative variables
Intraoperative parameter:Value
OPT: 
Bias26.55
sd56.93
Limits of agreement–85.04, 138.13
Bias CI 
95% CI–3.98 to 57.08
Lower limits of agreement CI 
95% CI–137.92 to –32.16
Upper limits of agreement CI 
95% CI85.25 to 191.01
EBL: 
Bias95.29
sd116.11
Limits of agreement–132.29, 322.86
Bias CI 
95% CI33.02 to 157.55
Lower limits of agreement CI 
95% CI–240.14 to −24.45
Upper limits of agreement CI 
95% CI215.02 to 430.71
WIT: 
Bias5.42
sd6.01
Limits of agreement–6.37, 17.21
Bias CI 
95% CI2.19 to 8.64
Lower limits of agreement CI 
95% CI–11.95 to –0.78
Upper limits of agreement CI 
95% CI11.62 to 22.79

There was poor correlation between the RPN and LPN methods for OPT (bias 26.55, bias 95% CI –3.98 to 57.08), EBL (bias 95.28, bias 95% CI 33.02 to 157.55) and WIT (bias 5.42, bias 95% CI 2.19 to 8.64).

Discussion

In the present study we aimed to evaluate the role of RPN in treating SRMs (mean size of <4 cm) and to objectively compare RPN with LPN. The results of this meta-analysis of 256 patients showed that were no significant differences in operative variables. This is despite the fact that the individual studies report a better operative profile for RPN; namely a shorter total OPT, WIT and reduced intraoperative EBL. Because these operative factors strongly affect the likelihood of curative resection, preservation of nephrons and ultimately survival, these findings are important. Of note is also the level of laparoscopic/robotic experience of the primary surgeons which ultimately affects the OPT [31].

Similar findings have been reported by other comparative studies [16, 50, 51]. Benway et al.[17] in a multi-institutional analysis of perioperative outcomes reported better operative and similar early oncological outcomes. Despite its retrospective nature, it highlights the important role of RPN in the treatment of SRMs, where it appears to be less affected by tumour complexity [7, 17, 51]. In a recent study, Lee et al. [52], concluded the beneficial role of RPN in treating more complex and posteriorly located tumours.

One limitation to note is a lack of standardised scoring for tumour complexity. Only two of the studies included reported on a nephrometry score [30, 49]. The location of the tumour can add to the complexity of the operation and affect perioperative variables [53]. A higher nephrometry score is associated with increased EBL, WIT and LOS as was assessed by Hayn et al. [54] in a group of patients undergoing LPN. Standardising the description of anatomical features of renal tumours has resulted in development of multiple scoring systems [55].

The robotic approach has many potential advantages over the laparoscopic as the da Vinci robot provides a three-dimensional view of the surgical field with increased degrees of freedom. The robot can eliminate tremor and provides scaling of motions. Despite this, there are disadvantages associated with robotic approach. Maintaining the robot is costly with the capital expenditure being well over 1 million Euros and the annual maintenance/running costs of 100 000 Euros [56]. In a recent single centre analysis of short-term outcomes, Elsamra et al. [57] did not find any significant difference between RPN and LPN. However, RPN was associated with an increased cost. In addition to this, setting up the system is time consuming and is thought to increase the OPT. However, having said that, RPN has been suggested as an alternative to traditional LPN to improve the learning curve for trainee surgeons.

The bulk of evidence in support of RPN is gathered from a few either initial-experience, retrospective, or non-randomised studies (Table 4) [58, 59]. These reports are suggestive of RPN becoming a viable alternative to open PN and LPN in carefully selected patients [16, 60], with the advantages of the robotic system being weighed against its cost [60, 61]. In a multi-institutional study, RPN was shown to be safe and feasible for selected patients with renal hilar tumours [18]. Interestingly, most of the reports on RPN are illustrative of a hybrid method. Whereby the initial part of the surgery is performed laparoscopically and then robotic surgery is performed [58]. However, this depends on the surgeon's preference and laparoscopic experience.

One of the limitations of the present meta-analysis that must be considered is the lack of randomised control trials on the current topic. All of the studies analysed are comparative studies with no random sequencing or concealment. Most of the studies were retrospective with only one study prospective [49] in nature. Only one study [30] mentions its exclusion criteria. In addition, some of the studies had poorly defined outcome measures. Almost all of the studies lacked medium- to long-term follow-up and data on tumour margins, all of which are important when considering early and long-term success of either treatment groups. Neither the allocation nor the assessment of outcome was ‘blinded’. In addition, publication bias based on current studies, as well as variation in inclusion criteria, treatment protocols, varying operative techniques and outcome assessment between studies should be taken into account.

Additionally, the risk of positive margins may be greater in LPN and RPN than with the open technique. Thus to avoid this there is tendency to resect more normal renal parenchyma compared with open PN. This can have great impact on the long-term function of the remaining renal tissue and long-term oncological data.

The learning curve for laparoscopic surgery appears longer than robotic surgery [30, 62] and can affect outcomes [63]. Such variability in the level of experience and expertise of surgeons and their learning curve associated with the technique and its effect on postoperative outcomes could not be accounted in the present study. Only one study reported the learning curve for transition from laparoscopic to robotic surgery [30]. Although it focuses on the experience of one surgeon, it highlights the technical challenges associated with the laparoscopic technique and the quick learning curve in transition from LPN to RPN. Mottrie et al. [19] analysed the impact of learning curve on perioperative outcomes of patients with parenchymal renal tumours who underwent RPN. Interestingly, they found a short learning curve for WIT (<20 min) and console time (<100 min) but not for EBL or overall complications. However, Pierorazio et al. [64] found no difference in perioperative outcomes of RPN when comparing the early and late phase of experience. Improvement were noted in the OPT, WIT and EBL for the LPN group as their level of experience increased.

Further evaluation of surgical techniques is required, as there is variability between different centres and the particular techniques used. For example novel approaches such as zero ischaemia [65, 66] techniques can have great impact on patients, particularly where renal function is of concern.

Finally, we conducted a further Bland–Altman analysis of some of the operative variables (Table 6) to compare their reproducibility and mean difference in techniques. Comparing OPT, WIT and EBL for each group revealed poor correlation either with an absolute or proportional difference (Figs 7-9). These differences may be reflective of the variability in the actual techniques and the level of experience of surgeons in either laparoscopic or robotic surgery. However, the safety and feasibility of RPN as an alternative to LPN has been shown in other similar studies [67].

In conclusion, the present study highlights the lack of extensive evidence for using robotic surgery over laparoscopic for treating small renal tumours in selected patients. Although not significant, the results from different studies showed a trend towards use of the robotic technique in treating small and complex renal tumours. To more objectively evaluate the relation between RPN and LPN, there is a need for randomised control trials. More long-term data on surveillance and tumour recurrence is required to fully elucidate role of robotics in PN.

Conflict of Interest

None declared.

Ancillary