EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 available at www.sciencedirect.com journal homepage: www.eu-openscience.europeanurology.com Renal Disease The IRON Study: Investigation of Robot-assisted Versus Open Nephron-sparing Surgery Carlo Andrea Bravi a,b,c, Giuseppe Rosiello a, Elio Mazzone a, Andrea Minervini d, Andrea Mari d, Fabrizio Di Maida d, Karim Bensalah e, Benoit Peyronnet e, Zine-Eddine Khene e, Riccardo Schiavina f, Lorenzo Bianchi f, Alexandre Mottrie b,c, Geert De Naeyer b, Alessandro Antonelli g, Maria Furlan h, Koon Ho Rha i, Ahmad Almujalhem i, Ithaar Derweesh j, Aaronw Bradshaw j, Jihak Kaouk k, Guilherme Sawczyn k, Riccardo Bertolo k, Alberto Breda l, Francesco Montorsi a, Umberto Capitanio a, Alessandro Larcher a,*, on behalf of the Junior ERUS/Young Academic Urologist Working Group on Robot-assisted Surgery a Unit of Urology, Division of Experimental Oncology, Urological Research Institute, IRCCS San Raffaele Scientific Institute, Vita-Salute San Raffaele University, Milan, Italy; b Department of Urology, Onze Lieve Vrouw Hospital, Aalst, Belgium; c ORSI Academy, Ghent, Belgium; d Unit of Oncologic Minimally Invasive Urology and Andrology, Department of Experimental and Clinical Medicine, Careggi Hospital, University of Florence, Florence, Italy; e Department of Urology, University of Rennes, Rennes, France; f Division of Urology, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy; g Department of Urology, Azienda Ospedaliera Universitaria Integrata, University of Verona, Verona, Italy; h Department of Urology, Spedali Civili Hospital, University of Brescia, Brescia, Italy; i Urological Science Institute, Yonsei University College of Medicine, Seoul, Korea; j Department of Urology, Moores UCSD Cancer Center, University of California San Diego School of Medicine, San Diego, CA, USA; k Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, OH, USA; l Department of Urology, Fundació Puigvert, Autonomous University of Barcelona, Barcelona, Spain Article info Abstract Article history: Background: Current literature does not provide large-scale data regarding clinical Accepted December 8, 2022 outcomes of robot-assisted (RAPN) versus open (OPN) partial nephrectomy. Moreover, data assessing predictors of long-term oncologic outcomes after RAPN Associate Editor: are scarce. M. Carmen Mir Objective: To compare perioperative, functional, and oncologic outcomes of RAPN versus OPN, and to investigate the predictors of oncologic outcomes after RAPN. Keywords: Design, setting, and participants: This study included 3467 patients treated with OPN Robotic surgery (n = 1063) or RAPN (n = 2404) for a single cT1–2N0M0 renal mass from 2004 to 2018 Urological procedures at nine high-volume European, North American, and Asian institutions. Robot-assisted partial Outcome measurements and statistical analysis: The study outcomes were short-term nephrectomy postoperative, functional, and oncologic outcomes. Regression models investigated Open partial nephrectomy Perioperative outcomes the effect of surgical approach (open vs Robot assisted) on study outcomes, and inter- Complications action tests were used for subgroup analyses. Propensity score matching for demo- graphic and tumor characteristics was used in sensitivity analyses. Multivariable Cox-regression analyses identified predictors of oncologic outcomes after RAPN. * Corresponding author. Department of Urology and Division of Experimental Oncology, URI, Urological Research Institute, IRCCS San Raffaele Scientific Institute, Via Olgettina, 60, 20132 Milan, Italy. Tel. +39 02.2643.7286. E-mail address: alelarcher@gmail.com (A. Larcher). https://doi.org/10.1016/j.euros.2022.12.017 2666-1683/Ó 2023 The Authors. Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 72 EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 Results and limitations: Baseline characteristics were similar between patients receiving RAPN and OPN, with only few differences. After adjusting for confound- ing, RAPN was associated with lower odds of intraoperative (odds ratio [OR]: 0.39, 95% confidence interval [CI]: 0.22, 0.68) and Clavien-Dindo 2 postoperative (OR: 0.29, 95% CI: 0.16, 0.50) complications (both p < 0.05). This association was not affected by comorbidities, tumor dimension, PADUA score, or preoperative renal function (all p > 0.05 on interaction tests). On multivariable analyses, we found no differences between the two techniques with respect to functional and onco- logic outcomes (all p > 0.05). Overall, there were 63 and 92 local recurrences and systemic progressions, respectively, with a median follow-up after surgery of 32 mo (interquartile range: 18, 60). Among patients receiving RAPN, we assessed pre- dictors of local recurrence and systemic progression with discrimination accuracy (ie, C-index) that ranged from 0.73 to 0.81. Conclusions: While cancer control and long-term renal function did not differ between RAPN and OPN, we found that the intra- and postoperative morbidity— especially in terms of complications—was lower after RAPN than after OPN. Our pre- dictive models allow surgeons to estimate the risk of adverse oncologic outcomes after RAPN, with relevant implications for preoperative counseling and follow-up after surgery. Patient summary: In this comparative study on robotic versus open partial nephrec- tomy, functional and oncologic outcomes were similar between the two techniques, with lower morbidity—especially in terms of complications—for robot-assisted sur- gery. The assessment of prognosticators for patients receiving robot-assisted partial nephrectomy may help in preoperative counseling and provides relevant data to tai- lor postoperative follow-up. Ó 2023 The Authors. Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY-NC-ND license (http://creative- commons.org/licenses/by-nc-nd/4.0/). 1. Introduction whether RAPN might allow for similar long-term cancer control to OPN and what the meaningful prognosticators There is growing interest toward robot-assisted partial are for this patient population are often underinvestigated. nephrectomy (RAPN) for patients with renal cancer, with Based on these premises and to address these voids, we the rate of partial nephrectomies performed robotically that built a large multi-institutional database with prospective increased from 21% in 2009 to 58% in 2015 only in the USA data collection including patients treated with either OPN [1]. However, despite the increasing adoption of this surgi- or RAPN at nine high-volume centers worldwide. cal technique, large evidence comparing RAPN with other surgical approaches is limited, and long-term oncologic out- 2. Patients and methods comes of patients undergoing RAPN are, virtually, unknown. 2.1. Data source and patient selection Comparative studies are a key step in the introduction and dissemination of new surgical techniques. While The current study relied on prospectively maintained databases from prospective, comparative evidence is available for robot- nine tertiary health care institutions worldwide, including 5032 patients assisted radical prostatectomy [2] and robot-assisted radi- treated with partial nephrectomy from 2004 to 2018 (complete inclu- cal cystectomy [3], this is not the case for RAPN. Prior retro- sion/exclusion criteria are described in Supplementary Fig. 1). For the spective studies suggested that, as compared with open scope of this study, we focused on patients who underwent RAPN or partial nephrectomy (OPN), RAPN might be associated with OPN for a single cT1–2N0M0 renal mass (n = 4201). We excluded 734 more favorable perioperative [4,5] and functional outcomes patients with missing information on tumor complexity, leaving 3467 [6], with data from our group showing that this benefit (2404 RAPN and 1063 OPN) patients eligible for the analyses. The surgi- might hold true also in patients with several comorbidities cal approach was selected according to the surgeon’s choice. Systemic and/or complex tumors [7,8]. However, common limitations staging was performed using conventional imaging (computed tomogra- of these papers are the relatively small sample size, missing phy/magnetic resonance imaging scan) according to internal practice at each treating institution. All information was obtained with appropriate information on surgeon’s experience, and inclusion of ethics committee or institutional review board waivers, and data were patients treated mainly at European institutions. As a result, made anonymous before analysis. prospective, high-quality data on the comparison between RAPN and OPN are currently lacking. 2.2. Research hypotheses and outcomes of interest Among determinants of successful surgery, oncologic outcomes are the first matter of concern in cancer surgery. Our main goals were twofold. First, we wanted to compare intra-, peri-, In this regard, evidence on the long-term oncologic out- and postoperative outcomes of OPN and RAPN. In this regard, the study comes of RAPN is scarce. Research questions such as outcomes were as follows: EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 73 Intra- and post-operative outcomes: intraoperative complications; computed using a logistic regression model with the odds of receiving estimated blood loss (eBL); operative time; postoperative complica- OPN as the dependent variable and age at diagnosis, gender, CCI, clinical tions, including overall and grade-specific complications according to tumor size, tumor side, total PADUA score, any individual PADUA score the Clavien-Dindo (CD) classification [9], hemorrhagic events, and item, and year of surgery as independent variables. Moreover, the same urinary leakages; and length of stay analyses were repeated after accounting for surgeon’s experience in each Functional outcomes: warm ischemia time and postoperative esti- specific surgical approach. mated glomerular filtration rate (eGFR), defined according to the To assess long-term oncologic outcomes after RAPN, we focused on Chronic Kidney Disease Epidemiology Collaboration equation for patients who underwent RAPN with available follow-up status patients aged <70 yr and the Berlin Initiative Study formula for (n = 1687). We utilized Cox-regression analyses to predict LC and SP in patients aged 70 yr [10], and measured at the last determination pre- and postoperative settings. Given the low number of events among before discharge and 1 yr after surgery patients receiving RAPN (n = 20), predictors of cancer-specific mortality Pathologic and oncologic outcomes: positive surgical margins, local were not investigated. In the preoperative model, the adjustment for recurrence (LC; defined as evidence of disease in the resection casemix included variables available before surgery such as age at diag- bed), systemic progression (SP; defined as evidence of disease else- nosis, gender, clinical tumor size, and total PADUA score. The postoper- where than the treated kidney), and cancer-specific and all-cause ative model included age at diagnosis, gender, pathologic tumor size, mortality; vital status and cause of death were identified from death pathologic tumor grade (G3–4 vs G1–2), type of malignant histology (clear certificates and physician correspondence cell renal cell carcinoma vs other), and positive surgical margins (no vs yes). Moreover, since we found evidence of a stage migration toward Our second aim was to investigate long-term oncologic outcomes of more aggressive disease over the period of study (Supplementary patients treated with RAPN, and to assess the predictors of disease recur- Table 2), we included year of surgery in our models as a continuous vari- rence (namely, LC and SP) after RAPN. able. To allow for an adequate estimation of the risk of cancer recur- To test these research hypotheses, we performed dedicated power rences, the postoperative model included only patients with confirmed analyses that are described in Supplementary Table 1. malignancy on final pathology (n = 1333). For both models, the estima- tion accuracy was assessed using the C-index. 2.3. Covariates 3. Results Covariates consisted of age at surgery, preoperative eGFR, clinical tumor size, pathologic tumor size, year of surgery, Charlson Comorbidity Index 3.1. Descriptive characteristics (CCI), gender, tumor side, tumor grade, positive surgical margins, and specific institution. Tumor complexity was determined by the treating Descriptive characteristics of the study population are listed urologist using individual PADUA score items (ie, longitudinal location, in Table 1. A total of 2404 (69%) patients underwent RAPN. rim location, renal sinus involvement, relationship with urinary collect- As compared with patients treated with OPN, those receiv- ing system, and exophytic rate), resulting in the total PADUA score. ing RAPN were younger (median: 61 vs 65 yr) with a higher Finally, for each individual patient, surgical experience was defined as preoperative eGFR (median: 84 vs 80 ml/min/1.73 m2), had the total number of RAPNs/OPNs performed by the surgeon before the a smaller clinical tumor (median: 3.0 vs 3.2 cm), and were patient’s operation [11,12]. operated more recently (46% vs 16% from 2015 to 2018; all p < 0.001). 2.4. Statistical analyses 3.2. Comparison between RAPN and OPN Our analyses included several steps. We first compared baseline charac- teristics of patients receiving OPN versus RAPN. Subsequently, we inves- 3.2.1. Intra- and postoperative outcomes tigated the effect of surgical approach (robot assisted vs open) on study Results for intra- and postoperative outcomes are described outcomes using multivariable linear, logistic, and Cox-regression analy- in Table 2. After adjusting for confounding, patients receiv- ses. For all models, the adjustment for confounding included the follow- ing RAPN had lower eBL than those who underwent OPN ing variables: age, CCI (0 vs 1 vs 2 vs 3), gender, preoperative eGFR, (estimate [EST]: –140, 95% confidence interval [CI]: -163, clinical tumor size, tumor side, total PADUA score, year of surgery, and –115), whereas operative time was longer in the RAPN institution. group (EST: +31, 95% CI: +26, +36; both p < 0.001). Length Given the high risk of postoperative complications after partial of hospital stay was shorter after RAPN than after OPN nephrectomy [13], we focused specifically on this outcome and utilized (EST: –2, 95% CI: –2, –1; p < 0.001). regression-derived coefficients to estimate the risk of overall postopera- tive complications after RAPN and, in separate analyses, after OPN. In 3.2.2. Complications addition, a locally weighted scatter plot smoothing method was used A total of 139 (6%) and 99 (9%) patients had intraoperative to graphically explore the risk of overall complications after RAPN and complications in the RAPN versus OPN group, respectively OPN according to PADUA score, CCI, clinical tumor size, and preoperative eGFR. Finally, since the association between surgical approach and post- (Table 2). As compared with OPN, patients receiving RAPN operative complications might be different in selected subgroups had a lower rate of overall (18% vs 33%) and CD 2 (12% (namely, patients with a high PADUA score, a high CCI, large tumors, vs 20%) complications. After adjusting for confounding, and a low preoperative eGFR), we tested this hypothesis using an inter- RAPN was associated with lower odds of intraoperative action term between treatment type (robot assisted vs open) and these (odds ratio [OR]: 0.39, 95% CI: 0.22, 0.68), overall (OR: factors individually. 0.51; 95% CI: 0.33, 0.76), and CD 2 (OR: 0.29; 95% CI: Since the chances of receiving OPN or RAPN might be influenced by 0.16, 0.50) postoperative complications than OPN (all baseline characteristics of the patients [14,15], the above analyses were p < 0.05). As shown in Figures 1A–D, these findings were repeated after 1:1 propensity score matching. Propensity scores were not affected by PADUA score, CCI, clinical tumor size, or 74 EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 Table 1 – Descriptive characteristics of 3467 patients treated with preoperative eGFR (all p > 0.05 on interaction tests). Simi- robot-assisted (RAPN) or open (OPN) partial nephrectomy for a larly, our results were unaltered in sensitivity analyses after single cT1-T2 renal mass at nine institutions during 2004–2018 propensity score matching (Supplementary Table 3) and Variable RAPN OPN p after the inclusion of surgical experience (Supplementary (n = 2404; (n = 1063; value Table 4). 69%) 31%) Age (yr) 61 (51, 69) 65 (54, 72) <0.001 Gender, male 1502 (63) 690 (65) 0.1 3.2.3. Renal function CCI Despite slightly longer warm ischemia time and a lower 0 845 (35) 356 (34) 0.5 postoperative eGFR in patients treated with RAPN than in 1 416 (17) 204 (19) 2 602 (25) 270 (25) those treated with OPN, we did not find evidence of a differ- 3 541 (23) 233 (22) ence with respect to 1-yr eGFR between the two surgical eGFR (ml/min/1.73 m2) 84 (65, 97) 80 (62, 94) <0.001 approaches (adjusted estimate for RAPN vs OPN: –1; 95% Single kidney 77 (3) 52 (5) 0.02 Clinical size (cm) 3.0 (2.0, 4.2) 3.2 (2.4, 4.3) <0.001 CI: –2, +1; p = 0.5; Table 2). Clinical stage cT1a 1744 (73) 762 (72) 0.4 3.2.4. Oncologic outcomes cT1b 563 (23) 245 (23) cT2 98 (4) 56 (5) A total of 103 (4%) and 55 (5%) patients had positive surgical Tumor side, left 1155 (48) 502 (47) 0.7 margins in the RAPN versus OPN group, respectively. On PADUA score 8 (7, 10) 8 (7, 9) <0.001 multivariable analyses, the odds of positive surgical mar- PADUA risk class High 1039 (43) 426 (40) 0.2 gins were lower for patients receiving RAPN versus OPN Medium 644 (27) 307 (29) (odds ratio [OR]: 0.53; 95% CI: 0.27, 0.99; p = 0.05). Low 721 (30) 330 (31) Among 2301 patients with available follow-up data, Year of surgery 2004–2011 556 (23) 596 (56) <0.001 there were 46 cancer deaths. A total of 63 and 92 patients 2012–2014 745 (31) 295 (28) had LC and SP, respectively. The median (interquartile 2015–2018 1103 (46) 172 (16) range) follow-up for survivors was 32 (18, 60) mo. On mul- Pathologic size (cm) 3.0 (2.0, 4.1) 3.0 (2.1, 4.0) 0.1 Malignancy on final pathology 1920 (80) 834 (78) 0.7 tivariable analyses, we did not find evidence of an associa- T3-T4 on final pathology 107 (4.4) 51 (4.7) 0.8 tion between surgical approach (RAPN vs OPN) and all G3-G4 on final pathology 550 (23) 191 (18) 0.3 oncologic outcomes (all p > 0.05; Table 2). Median follow-up for 24 (14–47) 62 (30–100) 0.001 survivors (mo) 3.3. Predictors of oncologic outcomes after RAPN CCI = Charlson Comorbidity Index; eGFR = estimated glomerular filtration rate. Numbers are frequencies (proportions) and medians (interquartile range). Table 3 describes the results of our multivariable models for the prediction of oncologic outcomes after RAPN. In the Table 2 – Clinical outcomes of 3467 patients treated with robot-assisted (RAPN) and open (OPN) partial nephrectomy for a cT1–2 renal mass Outcome RAPN OPN RAPN vs OPN p value (n = 2404) (n = 1063) OR/EST/HR (95% CI) Intraoperative outcomes Estimated blood loss (ml) 150 (100, 300) 180 (100, 350) –140 (–163, –115) <0.001 Operative time (min) 150 (120, 200) 120 (100, 163) +31 (+26, +36) <0.001 Intraoperative complications 139 (6) 99 (9) 0.39 (0.22, 0.68) <0.001 Postoperative outcomes and complications Overall complications 435 (18) 355 (33) 0.51 (0.33, 0.76) 0.001 Clavien-Dindo complications 2 279 (12) 215 (20) 0.29 (0.16, 0.50) <0.001 3 97 (4) 65 (6) 0.26 (0.08, 0.65) 0.008 Hemorrhagic complications 155 (6) 96 (9) 0.22 (0.08, 0.52) 0.001 Urinary leakage 21 (1) 49 (5) 0.08 (0.02, 0.21) <0.001 Length of stay (d) 4 (3, 5) 6 (5, 7) –2 (–2, –1) <0.001 Functional outcomes Off-clamp procedure 280 (12) 200 (19) 0.32 (0.25, 0.41) 0.0001 Warm ischemia time (min) 16 (11, 22) 15 (8, 21) +4.3 (+3, +5) <0.001 Postoperative eGFR 76 (60, 89) 78 (63, 93) –6 (–8, –4) <0.001 1-yr eGFR 71 (56, 88) 68 (55, 87) –1 (–2, +1) 0.5 Oncologic outcomesa Positive surgical margins 103 (4) 55 (5) 0.53 (0.27, 0.99) 0.05 Local recurrence 97% (95%, 98%) 96% (95%, 98%) 1.02 (0.51, 2.04) 0.9 Systemic progression 94% (92%, 96%) 92% (89%, 94%) 1.18 (0.61, 2.29) 0.6 Cancer-specific mortality 97% (96%, 99%) 98% (97%, 99%) 0.99 (0.33, 2.90) 0.9 CI = confidence interval; eGFR = estimated glomerular filtration rate; EST = estimate; HR = hazard ratio; OR = odds ratio; RCC = renal cell carcinoma. All p-values refer to the association between surgical approach (robotic vs. open) and each outcome of interest, with the open approach as reference group. Data are presented as frequency (proportion) and median (interquartile range) for categorical and continuous variables, respectively. Estimated GFR is described as ml/min/1.73 m2. Multivariable models were adjusted for age, Charlson Comorbidity Index, gender, preoperative eGFR, clinical size, tumor side, PADUA score, year of surgery, and institution. a Time-dependent outcomes are shown as probability of freedom from the event at 5 yr (95% confidence interval) using Kaplan-Meier estimates, and are calculated for patients with available follow-up and confirmed RCC on final pathology (n = 994 and n = 778 in the RAPN and OPN group, respectively) EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 75 Fig. 1 – Overall risk of complications after robot-assisted (RAPN) or open (OPN) partial nephrectomy stratified according to (A) preoperative PADUA score, (B) Charlson Comorbidity Index, (C) clinical size, and (D) preoperative estimated glomerular filtration rate (ml/min/1.73 m2). Gray areas represent the distribution for the respective parameter. The risk of complications according to each individual characteristic was computed for 3467 patients treated with RAPN or OPN using an interaction term between the characteristic of interest and surgical approach (robot assisted vs open) included in a multivariable regression model adjusted for age, PADUA score, clinical size, Charlson Comorbidity Index, and preoperative estimated glomerular filtration rate. preoperative setting, women had lower hazards of LC (haz- outcomes did not differ significantly between RAPN and ards ratio [HR]: 0.41, 95% CI: 0.18, 0.95), whereas older OPN, surgical morbidity—especially in terms of complica- patients (HR: 1.03, 95% CI: 1.00, 1.06; p = 0.02) and patients tions, blood loss, and hospital stay—favored robotic surgery with a larger clinical tumor (HR: 1.25, 95% CI: 1.07, 1.47; as compared with open surgery. p = 0.004) were associated with higher hazards of LC Evidence suggesting that minimally invasive techniques (Table 3). allow for better perioperative outcomes after partial On multivariable analyses assessing the predictors of SP, nephrectomy than open surgery is increasing [16,17], and only clinical tumor size was an independent predictor of a our results are consistent with prior investigations [18]. higher risk of SP (HR: 1.20, 95% CI: 1.01, 1.43; p = 0.03). That said, it is still unclear which is the compelling indica- The discrimination accuracy—the C-index—of our preopera- tion for either a surgeon or an institution to switch from tive models was moderate, ranging from 0.73 to 0.77. open to minimally invasive surgery [19]. The debate centers In our postoperative models, pathologic tumor size (HR: on whether the transition to robotic surgery might translate 1.21, 95% CI: 1.14, 1.44), pathologic grade (HR: 3.54, 95% into better long-term functional outcomes [20]. Our results CI: 1.66, 7.56), clear cell histology (HR: 3.26, 95% CI: 1.23, add significantly to the current literature, suggesting that 8.60), and positive surgical margins (HR: 3.85, 95% CI: the functional results after RAPN are similar to those after 1.55, 9.52; all p < 0.05; Table 3) were associated with higher OPN, with the obvious advantages of reduced morbidity. hazards of LC. Still, besides these clinical factors, there are also concerns With respect to SP, patients with a larger pathologic about the costs of robotic surgery that have often contained tumor had higher hazards of SP (HR: 1.26, 95% CI: 1.08, its broad adoption. In this regard, the introduction of novel 1.47; p = 0.003). The discrimination accuracy of our postop- robotic platforms [21–23] may mitigate this limitation, erative models ranged from 0.79 to 0.81. expanding the availability of robotic surgery. Finally, another factor that might favor the transition to robotic sur- 4. Discussion gery is recent evidence on robot-assisted radical prostatec- tomy showing that robotic surgery might have the potential In this study, we analyzed one of the largest comparative to flatten the learning curve, allowing surgeons to series on OPN and RAPN. While functional and oncologic reach adequate outcome in an early phase of their career 76 EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 Table 3 – Multivariable Cox-regression models predicting local recurrence and systemic progression after RAPN relying on pre- and postoperative predictors Predictors Outcome of interest Local recurrence Systemic progression HR (95% CI) p value HR (95% CI) p value Preoperative model (n = 1687, patients treated with RAPN with available follow-up data) Age at diagnosis 1.03 (1.00, 1.06) 0.02 1.00 (0.97, 1.03) 0.8 Gender Male Ref. 0.03 – 0.07 Female 0.41 (0.18, 0.95) 0.48 (0.22, 1.07) Clinical tumor size 1.25 (1.07, 1.47) 0.004 1.20 (1.01, 1.43) 0.03 PADUA score 1.09 (0.89, 1.33) 0.4 1.17 (0.95, 1.43) 0.1 Year of surgery 0.91 (0.79, 1.05) 0.2 0.91 (0.78, 1.05) 0.2 C-index 0.73 0.77 Postoperative model (n = 1333, patients with malignancy on final pathology after RAPN with available follow-up data) Age at diagnosis 1.01 (0.99, 1.05) 0.2 0.98 (0.96, 1.01) 0.4 Gender Male Ref. 0.1 – 0.4 Female 0.53 (0.23, 1.25) 0.71 (0.33, 1.53) Pathologic tumor size 1.21 (1.14, 1.44) 0.033 1.26 (1.08, 1.47) 0.003 Pathologic tumor grade G1–2 Ref. 0.001 – 0.1 G3–4 3.54 (1.66, 7.56) 1.67 (0.83, 3.37) Type of malignant histology Other Ref. 0.01 – 0.1 Clear cell RCC 3.26 (1.23, 8.60) 2.10 (0.85, 5.15) Positive surgical margins No Ref. 0.003 – 0.7 Yes 3.85 (1.55, 9.52) 1.28 (0.30, 5.40) Year of surgery 0.88 (0.75, 1.03) 0.1 0.93 (0.79, 1.09) 0.4 C-index 0.81 0.79 CI = confidence interval; HR = hazards ratio; RCC = renal cell carcinoma; Ref. = reference. [12,24]. Whether this might apply to partial nephrectomy— Given the lack of European Association of Urology recom- a different operation, with wider variability and different mendations in favor of a specific surgical approach for par- challenges [25]—is open to discussion and should be inves- tial nephrectomy, an increasing number of comparative tigated in future studies. studies investigated whether minimally invasive surgery Oncologic control is always a concern in cancer surgery. might represent a new standard of surgical care. However, In this regard, our data showed that oncologic outcomes, since prior evidence showed different economic implica- including surgical margins rates, were not different tions according to the operating technique [28], a fair com- between patients receiving RAPN and OPN. Moreover, our parison between surgical approaches should include their results have relevant implications for clinical practice as costs. Since the multi-institutional nature of our study may we provided long-term data on predictors of oncologic out- have limited the availability of such data, further research comes after RAPN [26], allowing surgeons to adequately should address this issue. counsel patients before and after surgery. Our findings have several limitations. First, despite data being prospectively collected, the analyses were performed 5. Conclusions in a retrospective manner. Moreover, although we adjusted for clinical characteristics, we cannot completely rule out We compared the outcomes of RAPN and OPN in one of the residual confounding by differences in case mix. Such con- largest series on partial nephrectomy. While cancer control founding could have resulted from surgeons with different and long-term renal function did not differ between the levels of experience for each surgical technique. However, techniques, we found that the intra- and postoperative mor- we repeated the analyses after the inclusion of approach- bidity—especially in terms of complications—were lower specific surgical experience, with no difference in results. after RAPN than after OPN. Our predictive models allow sur- in addition, our study included patients operated over a long geons to estimate the risk of adverse oncologic outcomes period, and thus, practice changes at different institutions after RAPN with optimal predictive accuracy, with relevant might not have been captured by our data collection, and implications for preoperative counseling and follow-up it is possible that our findings might not entirely be transfer- after surgery. The global setting of the study, the relatively able to contemporary patients. We similarly did not have long follow-up, and the large population with thorough follow-up data available for the entire cohort. Moreover, clinical data collection (including contemporary patients although our models included the type of resection, this treated in the postdissemination era of RAPN) are unique information was not reported through a standardized strengths of this study and support the validity of the instrument such as the surface-intermediate-base score observed findings. [27], resulting in possible inaccuracy across centers. We also acknowledge that our study did not include a cost analysis. Author contributions: Alessandro Larcher had full access to all the data EUROPEAN UROLOGY OPEN SCIENCE 49 (2023) 71–77 77 in the study and takes responsibility for the integrity of the data and the [8] Capitanio U, Montorsi F. Renal cancer. Lancet 2016;387:894–906. accuracy of the data analysis. [9] Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications. Ann Surg 2009;250:187–96. Study concept and design: Bravi, Larcher. [10] Capitanio U, Terrone C, Antonelli A, et al. Nephron-sparing Acquisition of data: Rosiello, Mazzone, Mari, Di Maida, Bensalah, Peyron- techniques independently decrease the risk of cardiovascular net, Khene, Bianchi, Furlan, Almujalhem, Sawczyn, Bertolo. events relative to radical nephrectomy in patients with a T1a–T1b Analysis and interpretation of data: Bravi, Mazzone, Larcher. renal mass and normal preoperative renal function. Eur Urol Drafting of the manuscript: Bravi, Larcher. 2015;67:683–9. [11] Vickers A, Maschino A, Savage C, Cronin AM. Assessing the learning Critical revision of the manuscript for important intellectual content: Min- curve for prostate cancer surgery. Robotic urologic surgery, 2012. ervini, Schiavina, De Naeyer, Antonelli, Rha, Derweesh, Kaouk, Breda, [12] Bravi CA, Tin A, Vertosick E, et al. The impact of experience on the Montorsi, Capitanio, Larcher. risk of surgical margins and biochemical recurrence after robot- Statistical analysis: Bravi, Mazzone, Larcher. assisted radical prostatectomy: a learning curve study. J Urol 2019;202:108–13. Obtaining funding: None. [13] Larcher A, Muttin F, Peyronnet B, et al. The learning curve for robot- Administrative, technical, or material support: None. assisted partial nephrectomy: impact of surgical experience on Supervision: Larcher, Montorsi, Capitanio, Mottrie. perioperative outcomes. Eur Urol 2019;75:253–6. Other: None. [14] Bukavina L, Mishra K, Mahran A, et al. Gender disparity in cystectomy postoperative outcomes: propensity score analysis of the national surgical quality improvement program database. Eur Financial disclosures: Alessandro Larcher certifies that all conflicts of Urol Oncol 2021;4:84–92. [15] Austin PC. An Introduction to propensity score methods for interest, including specific financial interests and relationships and affili- reducing the effects of confounding in observational studies. ations relevant to the subject matter or materials discussed in the manu- Multivariate Behav Res 2011;46:399–424. script (eg, employment/affiliation, grants or funding, consultancies, [16] Bravi CA, Mari A, Larcher A, et al. Toward individualized approaches honoraria, stock ownership or options, expert testimony, royalties, or to partial nephrectomy: assessing the correlation between ischemia patents filed, received, or pending), are the following: None. time and patient health status (RECORD2 Project). Eur Urol Oncol 2021;4:645–50. [17] Bravi CA, Larcher A, Capitanio U, et al. Perioperative outcomes of open, laparoscopic, and robotic partial nephrectomy: a prospective Funding/Support and role of the sponsor: This work was supported by multicenter observational study (the RECORd 2 project). Eur Urol Intuitive Foundation (involved in the management of the data in the Focus 2021;7:390–6. study). [18] Garisto J, Bertolo R, Dagenais J, et al. Robotic versus open partial nephrectomy for highly complex renal masses: Comparison of perioperative, functional, and oncological outcomes. Urol Oncol Appendix A. Supplementary data 2018;36:471.e1–e9. [19] Chandrasekar T, Boorjian SA, Capitanio U, Gershman B, Mir MC, Supplementary data to this article can be found online at Kutikov A. Collaborative review: factors influencing treatment decisions for patients with a localized solid renal mass. Eur Urol https://doi.org/10.1016/j.euros.2022.12.017. 2021;80:575–88. [20] Martini A, Bravi CA. Acute kidney injury and functional outcomes References after partial nephrectomy. Int J Urol 2022;29:1243–4. [21] Bravi CA, Paciotti M, Sarchi L, et al. Robot-assisted radical [1] Mazzone E, Mistretta FA, Knipper S, et al. Contemporary national prostatectomy with the novel Hugo robotic system: initial assessment of robot-assisted surgery rates and total hospital experience and optimal surgical set-up at a tertiary referral charges for major surgical uro-oncological procedures in the robotic center. Eur Urol 2022;82:233–7. United States. J Endourol 2019;33:438–47. [22] Bravi CA, Sarchi L, Mottaran A, et al. Feasibility and optimal setting [2] Coughlin G, Yaxley J, Chambers S, et al. Robot-assisted laparoscopic of robot-assisted partial nephrectomy with the novel HUGO robotic prostatectomy versus open radical retropubic prostatectomy: 24- system: a pre-clinical study. Urol Video J 2022;15:100164. month outcomes from a randomised controlled study. Lancet Oncol [23] Sarchi L, Mottaran A, Bravi CA, et al. Robot-assisted radical 2018;19:1051–60. prostatectomy feasibility and setting with the HugoTM robot- [3] Parekh D, Reis I, Castle E, et al. Robot-assisted radical cystectomy assisted surgery system. BJU Int 2022;130:671–5. versus open radical cystectomy in patients with bladder cancer [24] Bravi CA, Dell’Oglio P, Mazzone E, et al. The surgical learning curve (RAZOR): an open-label, randomised, phase 3, non-inferiority trial. for biochemical recurrence after robot-assisted radical Lancet 2018;391:2525–36. prostatectomy. Eur Urol Oncol. In press. https://doi.org/10.1016/j. [4] Borghesi M, Schiavina R, Chessa F, et al. Retroperitoneal robot- euo.2022.06.010. assisted versus open partial nephrectomy for cT1 renal tumors: a [25] Larcher A, De Naeyer G, Turri F, et al. The ERUS Curriculum for Robot- matched-pair comparison of perioperative and early oncological assisted Partial Nephrectomy: Structure Definition and Pilot Clinical outcomes. Clin Genitourin Cancer 2018;16:e391–6. Validation. Eur Urol 2019:1–9. https://doi.org/10.1016/j.eururo. [5] Xia L, Wang X, Xu T, Guzzo TJ. Systematic review and meta-analysis 2019.02.031. of comparative studies reporting perioperative outcomes of robot- [26] Carbonara U, Simone G, Capitanio U, et al. Robot-assisted partial assisted partial nephrectomy versus open partial nephrectomy. J nephrectomy: 7-year outcomes. Minerva Urol Nephrol 2021;73: Endourol 2017;31:893–909. 540–3. [6] Chang KD, Abdel Raheem A, Kim KH, et al. Functional and oncological [27] Minervini A, Carini M, Uzzo RG, Campi R, Smaldone MC, Kutikov A. outcomes of open, laparoscopic and robot-assisted partial Standardized reporting of resection technique during nephron- nephrectomy: a multicentre comparative matched-pair analyses sparing surgery: the surface–intermediate–base margin score. Eur with a median of 5 years’ follow-up. BJU Int 2018;122:618–26. Urol 2014;66:803–5. [7] Larcher A, Capitanio U, De Naeyer G, et al. Is robot-assisted surgery [28] Alemozaffar M, Chang SL, Kacker R, Sun M, DeWolf WC, Wagner AA. contraindicated in the case of partial nephrectomy for complex Comparing costs of robotic, laparoscopic, and open partial tumours or relevant comorbidities? A comparative analysis of nephrectomy. J Endourol 2013;27:560–5. morbidity, renal function, and oncologic outcomes. Eur Urol Oncol 2018;1:61–8.