- Email: [email protected]
Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis
EURURO-7553; No. of Pages 6 E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Platinum Priority – Kidney Cancer Editorial by XXX on pp. x–y of this issue
Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis Bimal Bhindi a, Ross J. Mason a, Mustafa M. Haddad b, Stephen A. Boorjian a, Bradley C. Leibovich a, Thomas D. Atwell b, Adam J. Weisbrod b, Grant D. Schmit b, R. Houston Thompson a,* a
Department of Urology, Mayo Clinic, Rochester, MN, USA; b Department of Radiology, Mayo Clinic, Rochester, MN, USA
Article info
Abstract
Article history: Accepted September 7, 2017
Background: While partial nephrectomy (PN) is considered the standard approach for a tumor in a solitary kidney, percutaneous cryoablation (PCA) is emerging as an alternative nephronsparing option. Objective: To compare outcomes between PCA and PN for tumors in a solitary kidney. Design, setting, and participants: Patients who underwent PCA or PN between 2005 and 2015 for a single primary renal tumor in a solitary kidney were identified using Mayo Clinic Registries. Exclusion criteria were inherited tumor syndromes and salvage procedures. Intervention: PCA and PN. Outcome measurements and statistical analysis: To achieve balance in baseline characteristics, we used inverse probability of treatment weighting (IPTW) based on propensity to receive treatment. The risk of having a post-treatment complication and percent drop in estimated glomerular filtration rate (eGFR), as well as the risks of local/ipsilateral recurrence, distant metastasis, and cancer-specific mortality, were compared between groups using logistic, linear, and Fine-and-Gray competing risk regression models. Results and limitations: The cohort included 118 patients (PCA: 54; PN: 64) with a median follow-up of 47 mo (interquartile range 18, 74). In unadjusted analyses, PCA was associated with a lower risk of complications (15% vs 31%; odds ratio [OR] = 0.38; 95% confidence interval [CI] 0.15, 0.96; p = 0.04). However, upon accounting for baseline differences with IPTW adjustment, there was no longer a significant difference in the risk of complications (28% vs 29%; OR = 0.95; 95% CI 0.53, 1.69; p = 0.9). There were no significant differences between PCA and PN in percentage drop in eGFR at discharge (mean: 11% vs 16%; b = –5%; 95% CI –13, 3; p = 0.2) or at 3 mo (12% vs 9%; b = 3%; 95% CI –3, 10; p = 0.3). Likewise, no significant differences were noted in local recurrence (HR = 0.87; 95% CI 0.38, 1.98; p = 0.7), distant metastases (HR = 0.60; 95% CI 0.30, 1.20; p = 0.2), or cancer-specific mortality (HR = 1.13; 95% CI 0.32, 3.98; p = 0.8). Limitations include the sample size, given the relative rarity of renal masses in solitary kidneys. Conclusions: Our study found no significant difference in complications, renal function outcomes, and oncologic outcomes between PN and PCA for patients with a tumor in a solitary kidney. Validation in a larger multi-institutional analysis may be warranted. Patient summary: Partial nephrectomy (surgery) and percutaneous cryoablation are both options for treating a kidney tumor while preserving the normal portion of the kidney. In patients with a tumor in their only kidney, we found no difference in the risk of complications, kidney function outcomes, or cancer control outcomes between these two approaches. © 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Associate Editor: Giacomo Novara Statistical Editor: Andrew Vickers Keywords: Cryosurgery Kidney neoplasms Renal cell carcinoma Percutaneous cryoablation Partial nephrectomy
* Corresponding author. Mayo Clinic, 200 1st St. SW, Rochester, MN 55905 USA. E-mail address: [email protected] (R.H. Thompson). http://dx.doi.org/10.1016/j.eururo.2017.09.009 0302-2838/© 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 2
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
1.
Introduction
Epidemiology Collaboration formula at discharge and at 3 mo posttreatment. We chose to focus on 3-mo outcomes because most of the
A mass in a solitary kidney represents a challenge in clinical management. In addition to the need for satisfactory oncologic control, it is simultaneously necessary to preserve sufficient renal function to avoid end-stage renal disease and dialysis, which is associated with significant morbidity and 43% 5-yr survival [1]. Partial nephrectomy (PN) has long been considered the standard management strategy in patients with a mass in a solitary kidney [2], with published series supporting both open [3–5] and minimally invasive approaches [6,7]. In the past decade, percutaneous ablation, in particular percutaneous cryoablation (PCA), has emerged as an alternative nephron-sparing option [8–12]. Although not all studies are in agreement [13,14], the largest study comparing PCA and PN found comparable oncologic outcomes with these two approaches [12]. While the use of PCA for renal tumors has been reported in the setting of a solitary kidney [15], comparative data to PN are limited. In addition, these studies include large proportions of patients undergoing radiofrequency ablation (RFA) [16] or include both percutaneous and laparoscopic cryoablation (LCA) [17,18]. As such, we sought to compare the risk of having a complication, renal function outcomes, and oncologic outcomes between patients undergoing PCA and PN for a single primary renal tumor in a solitary kidney.
renal function insult after PN has been shown to occur early, with a new baseline being established at 3 mo [23]. Local treatment failure after PCA was defined as either technical failure following ablation (failure of the ablation ice ball to extend beyond the tumor margin on monitoring CT imaging during the procedure), new focal enhancement within the ablation bed on follow-up imaging, enlargement of the ablation defect on follow-up imaging, or any new mass in the ipsilateral kidney. Local recurrence following PN was defined as a new mass in the ipsilateral kidney. Metastases, cancer-specific mortality (CSM), and all-cause mortality were evaluated from the date of PN or PCA. Patients were followed until death, with the longest follow-up up to January 1, 2016.
2.4.
Statistical analyses
In order to account for baseline differences between patients who underwent PCA and PN, inverse probability of treatment weighting (IPTW)–adjusted analyses were performed [24]. The propensity to undergo PCA versus PN was estimated using a logistic regression model based on age, sex, Charlson comorbidity index, treatment year, RENAL nephrometry score (measured by one of two interventional radiologists, with complex and equivocal cases reviewed together to achieve consensus) [8,25], tumor size (maximum tumor diameter in centimeters) [20], baseline eGFR, renal cell carcinoma (RCC) tumor histology, and history of prior contralateral nephrectomy for RCC. Although tumor histology was not known at the time of treatment decision for most patients who underwent PN, it has been shown that it is important to include variables related to the outcomes of interest, even if unrelated to the exposure (choice of treatment), in order to yield the optimal propensity score model and maximally reduce bias [24,26]. Sensitivity
2.
Patients and methods
analyses to assess the impact of this a priori decision are outlined below. Propensity score weights were trimmed below and above the 1st and
2.1.
Study design and participants
Following institutional review board approval, patients with a single, clinically localized, noncystic primary renal tumor in a solitary kidney without nodal or distant metastases who underwent PN or PCA between January 2005 and October 2015 were identified using our prospectively maintained renal tumor registry [3,8–12,15,19–22], which includes over 9000 patients to date. Exclusion criteria included multiple tumors, salvage procedures (ie, prior ipsilateral treatments), inherited tumor syndromes, and a tumor in a transplanted kidney.
2.2.
Treatment
99th percentiles, respectively [24]. Patient and tumor characteristics were compared between groups pre- and postweighting using standardized differences [27]. The risk of a post-treatment complication was compared between groups using logistic regression, while percent drop in eGFR at discharge and at 3 mo was evaluated using linear regression. The probability of overall survival (OS) was compared between groups using Cox proportional hazards regression. The risk of local failure, distant metastasis, and CSM was compared between groups using Fine-and-Grey competing risk models, with death from non-RCC causes considered a competing event. A secondary analysis was performed to determine if the association between treatment group and complication risk depends on tumor complexity. As such, we first evaluated the association between tumor complexity and complication risk for each of the treatment groups
Patients were initially seen by a urologist to review management options
separately using logistic regression models. We then evaluated whether
for their renal mass and shared decision making took place [8–12]. PN
the association between treatment group and complication risk differs
was performed as previously described for patients who elected to
according to tumor complexity using an interaction term.
undergo this procedure [12,19]. If they opted for ablation, a referral was
Various sensitivity analyses were performed. First, we repeated the
then made to an interventional radiologist to evaluate feasibility and
IPTW analysis, this time truncating the nonoverlapping tails of the
suitability. PCA was performed under CT and/or ultrasound guidance as
propensity score distributions and stabilizing the weights [24]. Second,
previously described [8–12].
an analysis repeating propensity score estimation without tumor
2.3.
for tumor histology using multivariable regression models. Third, a
histology was performed, and here we instead subsequently adjusted
Follow-up and outcomes
subset analysis restricted to patients with histologic confirmation of RCC Patients were generally assessed at 3, 6, 12, 18, 24, and 36 mo, with
was performed. The propensity score was re-estimated among this
annual follow-up thereafter. Patients were recommended to undergo
subset, and the models were then fitted again. Finally, analyses were
renal function assessment and cross-sectional imaging at 3 mo and at
repeated using propensity score matching and propensity score
each visit thereafter.
regression adjustment.
Postprocedural inpatient complications were classified according to
Analyses were performed using SAS v9.4 (SAS Institute Inc., Cary, NC,
the Clavien–Dindo system for each procedure. Estimated glomerular
USA). All tests were two sided, with p < 0.05 considered statistically
filtration rate (eGFR) was calculated using the Chronic Kidney Disease
significant.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 3
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
3.
Results
The unweighted cohort included 118 patients with a single renal mass in a solitary kidney, of whom 54 underwent PCA and 64 underwent PN (Supplementary Fig. 1; Table 1). Twelve surgeons performed these PNs, and six interventional radiologists performed the PCAs. Patients who underwent PCA were older (median 67 vs 62 yr), had worse baseline renal function (median 49 vs 57 ml/min/1.73 m2), and smaller tumors (median 2.6 vs 3.8 cm). The surgical approach for PN was open in 57 (89%) patients, pure laparoscopic in five (8%) patients, and robot-assisted laparoscopic in two (3%) patients. Median clamp time was 22 min (interquartile range [IQR] 13, 26; missing n = 16). Upon IPTW adjustment, excellent balance was achieved for the majority of propensity score variables (standardized differences <0.10), while minor imbalances remained for age (median 65 vs 63 yr; standardized difference = 0.15), treatment year (median 2010 vs 2009; standardized difference = 0.12), and histologic confirmation of RCC (78% vs 74%; standardized difference = 0.11). Median follow-up among survivors was 47 mo (IQR 18, 74). 3.1.
Risk of complications
In unadjusted analysis (Table 2), the probability of having any post-treatment complication was lower after PCA versus PN (15% vs 31%; odds ratio [OR] = 0.38; 95% confidence interval [CI] 0.15, 0.96; p = 0.04). Urine leak, acute renal failure, and ileus were the most common complications after PN. Postoperative bleeding was the most common complication after PCA (Supplementary Table 1). However, upon IPTW adjustment (Table 2), there was no significant difference in the probability of
any post-treatment complication between PCA and PN (28% vs 29%; OR = 0.95; 95% CI 0.53, 1.69; p = 0.9). Meanwhile, higher nephrometry score was associated with a greater risk of complications following both PCA (OR [per one point] = 1.73; 95% CI 1.03–2.92; p = 0.04) and PN (OR [per one point] = 1.50; 95% CI 1.10–2.05; p = 0.01), although nephrometry score did not modify the effect of treatment modality on the risk of complications (p interaction = 0.1). 3.2.
Renal function outcomes
In the unweighted analysis (Table 2), there was a smaller drop in eGFR at discharge following PCA versus PN (mean: 7% vs 18%; b = –11%; 95% CI 19, 4; p = 0.004), while there was no significant difference drop in eGFR by 3 mo from baseline (mean: 11% vs 10%; b = 1%; 95% CI –6, 8; p = 0.8). However, in the IPTW analysis, there were no significant differences in drop in eGFR from baseline at discharge (mean: 11% vs 16%; b = 5%; 95% CI 13, 3; p = 0.2) or at 3 mo (12% vs 9%; b = 3%; 95% CI 3, 10; p = 0.3) between PCA and PN (Table 2). Among PN patients with clamp time data (n = 43), longer clamp time was associated with greater drop in eGFR at discharge (b = 0.9%/min; 95% CI 0.3, 1; p = 0.004) but not at 3 mo (b = 0.5%/min; 95% CI –0.03, 1; p = 0.08) on univariable analysis. 3.3.
Treatment failure and oncologic outcomes
There were 14 patients with local treatment failure and 22 patients who died in the unweighted cohort. Among those with histologic confirmation of RCC, nine were with local treatment failures, 19 developed distant metastases, and 18 died, of whom seven died of RCC. Ten of the patients
Table 1 – Cohort characteristics Unweighted cohort Parameter
Whole cohort n = 118
PN n = 64
PCA n = 54
Continuous variables
Median (IQR)
Median (IQR)
Median (IQR)
Age (yr) Charlson comorbidity index Treatment year
64 (57–72) 2 (0–3) 2010 (2007–2012) 53 (42–69) 3.2 (2.1–4.5) 8 (6–9)
62 (56–67) 2 (0–3) 2008 (2007–2011) 57 (44–76) 3.8 (2.4–6.2) 8 (6–10)
67 (57–74) 2 (0–4) 2011 (2008–2013) 49 (39–60) 2.6 (2.0–3.7) 7 (6–9)
Baseline eGFR (ml/min/1.73 m2) Tumor size (cm) RENAL nephrometry score [25]
Categorical variables Female sex Prior nephrectomy for RCC Clinical stage T1a T1b–T2 Histologically confirmed RCC
Weighted cohort p value
Std diff.
0.03 0.2 0.002
0.41 0.25 0.60
0.02 0.006 0.2
0.46 0.53 0.24
n (%)
n (%)
n (%)
32 (27) 65 (55)
17 (27) 24 (38)
15 (28) 41 (76)
0.9 <0.001
79 (67) 39 (33) 85 (72)
35 (55) 29 (45) 51 (80)
44 (81) 10 (19) 34 (63)
PN
PCA
Median (IQR)
Median (IQR)
63 (58–67) 2 (1–3) 2009 (2007–2012) 56 (47–77) 3.7 (2.0–5.0) 8 (5–9)
65 (54–75) 2 (0–4) 2010 (2007–2013) 56 (48–69) 3.5 (2.3–6.5) 7 (6–9)
Std diff.
0.15 0.07 0.12 <0.01 0.06 0.03
%
%
0.03 0.88
27 54
25 51
0.04 0.06
0.002
0.60 0.38
60 40 78
0.06
0.04
57 43 74
0.11
eGFR = estimated glomerular filtration rate; IQR = interquartile range; PCA = percutaneous cryoablation; PN = partial nephrectomy; RCC = renal cell carcinoma; Std diff. = standardized difference. The propensity to undergo PCA versus PN was estimated using a logistic regression model based on age at treatment, sex, Charlson comorbidity index, treatment year, RENAL nephrometry score (measured by two interventional radiologists), tumor size (maximum tumor diameter in centimeters), baseline eGFR, RCC tumor histology, and a history of prior contralateral nephrectomy for RCC. All patients were included in the weighted analysis.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 4
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
Table 2 – Models comparing outcomes between percutaneous renal cryoablation and partial nephrectomy for tumors in a solitary kidney Unadjusted analysis
Outcome Binary outcomes
OR (95% CI)
Any complication
0.38 (0.15, 0.96)
IPTW analysis p value
0.95 (0.53, 1.69)
0.9
beta (95% CI)
p value
beta (95% CI)
p value
% Drop in eGFR by discharge % Drop in eGFR by 3 mo
–11 (–19, –4) 1 (–6, 8)
0.004 0.8
a
Local/ipsilateral failure Distant metastasis b RCC Death b Death from any cause
p value
0.04
Continuous outcomes
Time-to-event outcomes
OR (95% CI)
–5 (–13, 3) 3 (–3, 10)
0.2 0.3
p value
HR (95% CI)
p value
HR (95% CI)
1.27 (0.45, 3.62) 0.79 (0.28, 2.19) 2.69 (0.55, 13.1) 1.54 (0.68, 3.47)
0.6 0.7 0.2 0.3
0.87 (0.38, 1.98) 0.60 (0.30, 1.20) 1.13 (0.32, 3.98) 2.23 (1.32, 3.77)
0.7 0.2 0.8 0.005
CI = confidence interval; eGFR = estimated glomerular filtration rate; HR = hazard ratio; IPTW = inverse probability of treatment weighting; OR = odds ratio; RCC = renal cell carcinoma. Odds ratios and hazard ratios >1 indicate a higher risk of an event/outcome with percutaneous renal cryoablation. Positive beta values indicate a greater decline in renal function following percutaneous renal cryoablation. a Fine-and-Grey models were used for local/ipsilateral failure, distant metastasis, and RCC death, with death from another cause considered as a competing event. Cox proportional hazards models were used for death from any cause. b Only patients or tumors that were proved on biopsy or surgical pathology to be renal cell carcinoma were eligible for analyses of distant metastasis or RCC death.
with local recurrence were managed with repeat percutaneous ablation, two patients were treated with systemic therapy due to concurrent metastatic disease, one patient was managed expectantly, and management of local recurrence was unknown in one patient. In the IPTW-adjusted analysis, there were no significant differences in the risk of local treatment failure with PCA versus PN for the whole weighted cohort (hazard ratio [HR] = 0.87; 95% CI 0.38, 1.98; p = 0.7; Table 2) or among the subset with histologic confirmation of RCC (HR = 0.50; 95% CI 0.15, 1.71; p = 0.3). Moreover, there were no significant differences in the risk of distant metastasis (HR = 0.60; 95% CI 0.30, 1.20; p = 0.2) or CSM (HR = 1.13; 95% CI = 0.32, 3.98; p = 0.8) among patients with histologic confirmation of RCC (Table 2). However, despite IPTW adjustment, PCA was associated with worse OS among the whole IPTW cohort (HR = 2.23; 95% CI 1.32, 3.77; p = 0.005) and among the subset with histologic confirmation of RCC (HR = 2.60; 95% CI 1.42, 4.76; p = 0.003; Table 2). 3.4.
Sensitivity analyses
Given that our main IPTW analysis had some imbalance for age, year of treatment, and histologic confirmation of RCC (Table 1), we used regression to adjust for each of these variables. The association between PCA and worse OS was no longer significantly different upon regression adjustment for age (HR = 1.27; 95% CI 0.72, 2.25; p = 0.4) in the main IPTW analysis. Otherwise, this sensitivity analysis did not alter conclusions (data not shown). Truncating the nonoverlapping tails of the propensity score distributions and stabilizing the weights did not alter conclusions (Supplementary Table 2). Leaving RCC histologic status out of the propensity score calculation in the IPTW analysis, and later adjusting for it using regression, did not alter conclusions, nor did the analysis recalculating the
propensity score among the subset with histologically confirmed RCC (Supplementary Table 2). Sensitivity analyses that instead employed propensity matching and propensity score regression adjustment arrived at similar conclusions, except that PCA was no longer associated with worse OS in these analyses (Supplementary Table 2). Upon further evaluation of the IPTW analysis, the observed association between PCA and worse OS was largely driven by a single patient who was assigned a large weight in the analysis. Upon excluding this patient, the association between treatment group and OS was no longer significant (HR = 1.04; 95% CI 0.52, 2.06; p = 0.9; Supplementary Table 2). 4.
Discussion
Upon accounting for differences in the propensity to receive PCA or PN, we found no significant differences in the risk of post-treatment complication between these two treatment modalities when treating a single tumor in a solitary kidney. Nephrometry score was associated with an increased risk of complications regardless of treatment approach. Additionally, there were no significant differences in renal function or oncologic outcomes between PCA and PN. Although patients who underwent PCA had worse OS in the IPTW analysis, which could reflect greater unmeasured frailty in this population, this should be interpreted with caution since this association was not consistent across all analytic approaches and was driven by an influential observation. Although thermal ablation is being increasingly utilized in the treatment of renal tumors [28], only two reports [17,18] have compared PN and cryoablation (ie, including both PCA and LCA) in patients with a solitary kidney. Meanwhile, one study [16] compared PN and ablation (predominantly RFA) for patients with an imperative indication for nephron sparing, of whom half had a solitary kidney. All three studies found that ablation was associated with a lower rate of
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
complications, which is consistent with the unadjusted analysis in our study. None of the studies found a difference in renal function outcomes between ablation and PN, which is also similar to the present study and another prior report from our group evaluating a cohort of patients with tumors in solitary kidneys who underwent percutaneous RFA or PCA [21]. In contrast to our present study, Mues et al [17] found a greater proportion of recurrence events after ablation versus PN. However, they did not perform a survival analysis or adjust for differences between treatment groups. Meanwhile, Long et al [16] also found that ablation was associated with an increased risk of recurrence, but their cohort predominantly included patients who underwent RFA, which is associated with a greater risk of local failure in tumors larger than 3 cm and central tumors [11]. PN and PCA have also been compared in studies that were not restricted to patients with solitary kidneys [12–14]. In the largest study comparing PN and PCA for cT1 renal masses, there were no differences in risks of local recurrence or distant metastasis [12]. Of note, the local recurrence rate is higher in the present study as compared with the 2% rate at 3 yr in this study of a broader population. This is because clinicians in the present analysis were tasked with maximally sparing nephrons in the setting of a solitary kidney, often with large and/or complex tumors. Caputo et al [13] evaluated cT1b patients and did not find a statistically significant difference in complication rate or eGFR preservation between PN and cryoablation (PCA or LCA). Both Caputo et al [13] and Tanagho et al [14] found that cryoablation (PCA or LCA) was associated with an increased risk of local recurrence, although they found no significant difference in cancer-specific survival. However, it is difficult to discern to what extent this is due to the inclusion of patients who underwent LCA. Although LCA allows for direct visualization of the tumor, PCA allows for real-time monitoring using cross-sectional imaging to ensure that the ice ball extends to cover the entire tumor with a sufficient margin. In addition, we perform adjunctive procedures in selected tumors, such as hydrodisplacement of adjacent bowel, preablation selective arterial embolization, and retrograde pyeloperfusion via an externalized ureteral stent, in order to reduce the risk of complications while allowing for more aggressive treatment of complex renal tumors [22,29,30]. Of note, Klatte et al [31] also reported a meta-analysis comparing LCA and PN, and found that LCA was associated with a greater risk of local recurrence. This further supports evaluating PCA and LCA separately. There are limitations worthy of discussion. First, although we accounted for several patient and tumor characteristics, including tumor complexity, observational studies cannot account for unmeasured differences between groups. For example, it is possible that patients who underwent PCA were frailer, despite controlling for Charlson comorbidity index. We employed a competing risk analysis to minimize this impact on our assessment of oncologic outcomes. It was also not possible to account for procedural learning curve. Second, we cannot rule out distant metastasis and CSM related to RCC in the contralateral kidney. We controlled for the history of contralateral RCC in order to limit this from
5
introducing a bias into our comparative analysis. Third, although this is a large study relative to existing literature, the sample size is still relatively small. Fourth, we were unable to reliably evaluate long-term renal function due to the availability of less consistent data for patients followed at outside institutions, and because PCA patients were more likely to die from competing causes and less likely to have long-term renal function data. Fifth, longer follow-ups will be needed to better assess oncologic outcomes. Finally, caution should be applied before generalizing our findings and outcomes to settings with less experience with PN and/ or PCA. Notable strengths of this study include our use of IPTWadjusted analyses and several sensitivity analyses to account for differences in factors that may have influenced patient selection as well as patient outcomes. Moreover, this is the first comparative effectiveness study in the solitary kidney population to focus specifically on PCA, which should be evaluated separately from percutaneous RFA and LCA, both of which may not have the same outcomes as PCA. In addition, our extensive experience with both treatment approaches allows us to assess and compare the true therapeutic potential of both modalities. 5.
Conclusions
In our IPTW-adjusted analysis and various sensitivity analyses, we did not find a statistically significant difference between PN and PCA in the risk of having a post-treatment complication, in renal function outcomes at 3 mo, or in oncologic outcomes. Small to moderate differences could not be ruled out, and further validation is warranted. Pending such validation, both PN and PCA can be considered for the management of a renal mass in a solitary kidney. Author contributions: Bimal Bhindi had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Thompson, Atwell, Bhindi. Acquisition of data: Haddad, Atwell, Schmit, Weisbrod, Boorjian, Leibovich, Thompson. Analysis and interpretation of data: Bhindi, Atwell, Schmit, Boorjian, Leibovich, Thompson. Drafting of the manuscript: Bhindi, Thompson. Critical revision of the manuscript for important intellectual content: Bhindi, Mason, Haddad, Boorjian, Leibovich, Atwell, Weisbrod, Schmit, Thompson. Statistical analysis: Bhindi. Obtaining funding: None. Administrative, technical, or material support: Thompson, Atwell, Schmit. Supervision: Thompson, Atwell, Schmit. Other: None. Financial disclosures: Bimal Bhindi certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 6
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
Appendix A. Supplementary data
[17] Mues AC, Korets R, Graversen JA, et al. Clinical, pathologic, and functional outcomes after nephron-sparing surgery in patients with
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. eururo.2017.09.009.
a solitary kidney: a multicenter experience. J Endourol 2012;26: 1361–6. [18] Panumatrassamee K, Kaouk JH, Autorino R, et al. Cryoablation versus minimally invasive partial nephrectomy for small renal masses in the solitary kidney: impact of approach on functional
References
outcomes. J Urol 2013;189:818–22.
[1] Mailloux LU, Bellucci AG, Napolitano B, Mossey T, Wilkes BM, Bluestone PA. Survival estimates for 683 patients starting dialysis from 1970 through 1989: identification of risk factors for survival. Clin Nephrol 1994;42:127–35. [2] Campbell SC, Novick AC, Belldegrun A, et al. Guideline for management of the clinical T1 renal mass. J Urol 2009;182:1271–9. [3] Ghavamian R, Cheville JC, Lohse CM, Weaver AL, Zincke H, Blute ML. Renal cell carcinoma in the solitary kidney: an analysis of complications and outcome after nephron sparing surgery. J Urol 2002;168:454–9. [4] Saranchuk JW, Touijer AK, Hakimian P, Snyder ME, Russo P. Partial nephrectomy for patients with a solitary kidney: the Memorial Sloan-Kettering experience. BJU Int 2004;94:1323–8. [5] Ching CB, Lane BR, Campbell SC, Li J, Fergany AF. Five to 10-year followup of open partial nephrectomy in a solitary kidney. J Urol 2013;190:470–4. [6] Gill IS, Colombo Jr JR, Moinzadeh A, et al. Laparoscopic partial nephrectomy in solitary kidney. J Urol 2006;175:454–8. [7] Hillyer SP, Bhayani SB, Allaf ME, et al. Robotic partial nephrectomy for
solitary
kidney:
a
multi-institutional
analysis.
Urology
2013;81:93–7. [8] Bhindi B, Thompson RH, Mason RJ, et al. Comprehensive assessment of renal tumor complexity in a large percutaneous cryoablation cohort. BJU Int 2017;119:905–12. [9] Atwell TD, Schmit GD, Boorjian SA, et al. Percutaneous ablation of renal masses measuring 3.0 cm and smaller: comparative local control and complications after radiofrequency ablation and cryoablation. AJR Am J Roentgenol 2013;200:461–6. [10] Schmit GD, Thompson RH, Kurup AN, et al. Percutaneous cryoablation of solitary sporadic renal cell carcinomas. BJU Int 2012;110: E526–31. [11] Schmit GD, Thompson RH, Kurup AN, et al. Usefulness of R.E.N.A.L. nephrometry scoring system for predicting outcomes and complications of percutaneous ablation of 751 renal tumors. J Urol 2013;189:30–5. [12] Thompson RH, Atwell T, Schmit G, et al. Comparison of partial nephrectomy and percutaneous ablation for cT1 renal masses. Eur Urol 2015;67:252–9. [13] Caputo PA, Zargar H, Ramirez D, et al. Cryoablation versus partial nephrectomy for clinical T1b renal tumors: a matched group comparative analysis. Eur Urol 2017;71:111–7. [14] Tanagho YS, Bhayani SB, Kim EH, Figenshau RS. Renal cryoablation versus robot-assisted partial nephrectomy: Washington University long-term experience. J Endourol 2013;27:1477–86. [15] Weisbrod AJ, Atwell TD, Frank I, et al. Percutaneous cryoablation of masses in a solitary kidney. AJR Am J Roentgenol 2010;194:1620–5.
[19] Lau WK, Blute ML, Weaver AL, Torres VE, Zincke H. Matched comparison of radical nephrectomy vs nephron-sparing surgery in patients with unilateral renal cell carcinoma and a normal contralateral kidney. Mayo Clin Proc 2000;75:1236–42. [20] Bhindi B, Lohse CM, Mason RJ, et al. Are we using the best tumor size cut-points for renal cell carcinoma staging? Urology. In press. http://dx.doi.org/10.1016/j.urology.2017.04.010. [21] Mitchell CR, Atwell TD, Weisbrod AJ, et al. Renal function outcomes in patients treated with partial nephrectomy versus percutaneous ablation for renal tumors in a solitary kidney. J Urol 2011;186:1786– 90. [22] Woodrum DA, Atwell TD, Farrell MA, Andrews JC, Charboneau JW, Callstrom MR. Role of intraarterial embolization before cryoablation of large renal tumors: a pilot study. J Vasc Interv Radiol 2010;21:930–6. [23] Mason R, Kapoor A, Liu Z, et al. The natural history of renal function after surgical management of renal cell carcinoma: results from the Canadian Kidney Cancer Information System. Urol Oncol 2016;34, 486 e1–7. [24] Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med 2015;34:3661–79. [25] Kutikov A, Uzzo RG. The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J Urol 2009;182:844–53. [26] Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Sturmer T. Variable selection for propensity score models. Am J Epidemiol 2006;163:1149–56. [27] Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009;28:3083–107. [28] Choueiri TK, Schutz FA, Hevelone ND, et al. Thermal ablation vs surgery for localized kidney cancer: a Surveillance, Epidemiology, and End Results (SEER) database analysis. Urology 2011;78:93–8. [29] Bodily KD, Atwell TD, Mandrekar JN, et al. Hydrodisplacement in the percutaneous cryoablation of 50 renal tumors. AJR Am J Roentgenol 2010;194:779–83. [30] Cantwell CP, Wah TM, Gervais DA, et al. Protecting the ureter during radiofrequency ablation of renal cell cancer: a pilot study of retrograde pyeloperfusion with cooled dextrose 5% in water. J Vasc Interv Radiol 2008;19:1034–40. [31] Klatte T, Grubmuller B, Waldert M, Weibl P, Remzi M. Laparoscopic cryoablation versus partial nephrectomy for the treatment of small renal masses: systematic review and cumulative analysis of observational studies. Eur Urol 2011;60:435–43.
[16] Long JA, Bernhard JC, Bigot P, et al. Partial nephrectomy versus ablative therapy for the treatment of renal tumors in an imperative setting. World J Urol 2017;35:649–56.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Platinum Priority – Kidney Cancer Editorial by XXX on pp. x–y of this issue
Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis Bimal Bhindi a, Ross J. Mason a, Mustafa M. Haddad b, Stephen A. Boorjian a, Bradley C. Leibovich a, Thomas D. Atwell b, Adam J. Weisbrod b, Grant D. Schmit b, R. Houston Thompson a,* a
Department of Urology, Mayo Clinic, Rochester, MN, USA; b Department of Radiology, Mayo Clinic, Rochester, MN, USA
Article info
Abstract
Article history: Accepted September 7, 2017
Background: While partial nephrectomy (PN) is considered the standard approach for a tumor in a solitary kidney, percutaneous cryoablation (PCA) is emerging as an alternative nephronsparing option. Objective: To compare outcomes between PCA and PN for tumors in a solitary kidney. Design, setting, and participants: Patients who underwent PCA or PN between 2005 and 2015 for a single primary renal tumor in a solitary kidney were identified using Mayo Clinic Registries. Exclusion criteria were inherited tumor syndromes and salvage procedures. Intervention: PCA and PN. Outcome measurements and statistical analysis: To achieve balance in baseline characteristics, we used inverse probability of treatment weighting (IPTW) based on propensity to receive treatment. The risk of having a post-treatment complication and percent drop in estimated glomerular filtration rate (eGFR), as well as the risks of local/ipsilateral recurrence, distant metastasis, and cancer-specific mortality, were compared between groups using logistic, linear, and Fine-and-Gray competing risk regression models. Results and limitations: The cohort included 118 patients (PCA: 54; PN: 64) with a median follow-up of 47 mo (interquartile range 18, 74). In unadjusted analyses, PCA was associated with a lower risk of complications (15% vs 31%; odds ratio [OR] = 0.38; 95% confidence interval [CI] 0.15, 0.96; p = 0.04). However, upon accounting for baseline differences with IPTW adjustment, there was no longer a significant difference in the risk of complications (28% vs 29%; OR = 0.95; 95% CI 0.53, 1.69; p = 0.9). There were no significant differences between PCA and PN in percentage drop in eGFR at discharge (mean: 11% vs 16%; b = –5%; 95% CI –13, 3; p = 0.2) or at 3 mo (12% vs 9%; b = 3%; 95% CI –3, 10; p = 0.3). Likewise, no significant differences were noted in local recurrence (HR = 0.87; 95% CI 0.38, 1.98; p = 0.7), distant metastases (HR = 0.60; 95% CI 0.30, 1.20; p = 0.2), or cancer-specific mortality (HR = 1.13; 95% CI 0.32, 3.98; p = 0.8). Limitations include the sample size, given the relative rarity of renal masses in solitary kidneys. Conclusions: Our study found no significant difference in complications, renal function outcomes, and oncologic outcomes between PN and PCA for patients with a tumor in a solitary kidney. Validation in a larger multi-institutional analysis may be warranted. Patient summary: Partial nephrectomy (surgery) and percutaneous cryoablation are both options for treating a kidney tumor while preserving the normal portion of the kidney. In patients with a tumor in their only kidney, we found no difference in the risk of complications, kidney function outcomes, or cancer control outcomes between these two approaches. © 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Associate Editor: Giacomo Novara Statistical Editor: Andrew Vickers Keywords: Cryosurgery Kidney neoplasms Renal cell carcinoma Percutaneous cryoablation Partial nephrectomy
* Corresponding author. Mayo Clinic, 200 1st St. SW, Rochester, MN 55905 USA. E-mail address: [email protected] (R.H. Thompson). http://dx.doi.org/10.1016/j.eururo.2017.09.009 0302-2838/© 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 2
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
1.
Introduction
Epidemiology Collaboration formula at discharge and at 3 mo posttreatment. We chose to focus on 3-mo outcomes because most of the
A mass in a solitary kidney represents a challenge in clinical management. In addition to the need for satisfactory oncologic control, it is simultaneously necessary to preserve sufficient renal function to avoid end-stage renal disease and dialysis, which is associated with significant morbidity and 43% 5-yr survival [1]. Partial nephrectomy (PN) has long been considered the standard management strategy in patients with a mass in a solitary kidney [2], with published series supporting both open [3–5] and minimally invasive approaches [6,7]. In the past decade, percutaneous ablation, in particular percutaneous cryoablation (PCA), has emerged as an alternative nephron-sparing option [8–12]. Although not all studies are in agreement [13,14], the largest study comparing PCA and PN found comparable oncologic outcomes with these two approaches [12]. While the use of PCA for renal tumors has been reported in the setting of a solitary kidney [15], comparative data to PN are limited. In addition, these studies include large proportions of patients undergoing radiofrequency ablation (RFA) [16] or include both percutaneous and laparoscopic cryoablation (LCA) [17,18]. As such, we sought to compare the risk of having a complication, renal function outcomes, and oncologic outcomes between patients undergoing PCA and PN for a single primary renal tumor in a solitary kidney.
renal function insult after PN has been shown to occur early, with a new baseline being established at 3 mo [23]. Local treatment failure after PCA was defined as either technical failure following ablation (failure of the ablation ice ball to extend beyond the tumor margin on monitoring CT imaging during the procedure), new focal enhancement within the ablation bed on follow-up imaging, enlargement of the ablation defect on follow-up imaging, or any new mass in the ipsilateral kidney. Local recurrence following PN was defined as a new mass in the ipsilateral kidney. Metastases, cancer-specific mortality (CSM), and all-cause mortality were evaluated from the date of PN or PCA. Patients were followed until death, with the longest follow-up up to January 1, 2016.
2.4.
Statistical analyses
In order to account for baseline differences between patients who underwent PCA and PN, inverse probability of treatment weighting (IPTW)–adjusted analyses were performed [24]. The propensity to undergo PCA versus PN was estimated using a logistic regression model based on age, sex, Charlson comorbidity index, treatment year, RENAL nephrometry score (measured by one of two interventional radiologists, with complex and equivocal cases reviewed together to achieve consensus) [8,25], tumor size (maximum tumor diameter in centimeters) [20], baseline eGFR, renal cell carcinoma (RCC) tumor histology, and history of prior contralateral nephrectomy for RCC. Although tumor histology was not known at the time of treatment decision for most patients who underwent PN, it has been shown that it is important to include variables related to the outcomes of interest, even if unrelated to the exposure (choice of treatment), in order to yield the optimal propensity score model and maximally reduce bias [24,26]. Sensitivity
2.
Patients and methods
analyses to assess the impact of this a priori decision are outlined below. Propensity score weights were trimmed below and above the 1st and
2.1.
Study design and participants
Following institutional review board approval, patients with a single, clinically localized, noncystic primary renal tumor in a solitary kidney without nodal or distant metastases who underwent PN or PCA between January 2005 and October 2015 were identified using our prospectively maintained renal tumor registry [3,8–12,15,19–22], which includes over 9000 patients to date. Exclusion criteria included multiple tumors, salvage procedures (ie, prior ipsilateral treatments), inherited tumor syndromes, and a tumor in a transplanted kidney.
2.2.
Treatment
99th percentiles, respectively [24]. Patient and tumor characteristics were compared between groups pre- and postweighting using standardized differences [27]. The risk of a post-treatment complication was compared between groups using logistic regression, while percent drop in eGFR at discharge and at 3 mo was evaluated using linear regression. The probability of overall survival (OS) was compared between groups using Cox proportional hazards regression. The risk of local failure, distant metastasis, and CSM was compared between groups using Fine-and-Grey competing risk models, with death from non-RCC causes considered a competing event. A secondary analysis was performed to determine if the association between treatment group and complication risk depends on tumor complexity. As such, we first evaluated the association between tumor complexity and complication risk for each of the treatment groups
Patients were initially seen by a urologist to review management options
separately using logistic regression models. We then evaluated whether
for their renal mass and shared decision making took place [8–12]. PN
the association between treatment group and complication risk differs
was performed as previously described for patients who elected to
according to tumor complexity using an interaction term.
undergo this procedure [12,19]. If they opted for ablation, a referral was
Various sensitivity analyses were performed. First, we repeated the
then made to an interventional radiologist to evaluate feasibility and
IPTW analysis, this time truncating the nonoverlapping tails of the
suitability. PCA was performed under CT and/or ultrasound guidance as
propensity score distributions and stabilizing the weights [24]. Second,
previously described [8–12].
an analysis repeating propensity score estimation without tumor
2.3.
for tumor histology using multivariable regression models. Third, a
histology was performed, and here we instead subsequently adjusted
Follow-up and outcomes
subset analysis restricted to patients with histologic confirmation of RCC Patients were generally assessed at 3, 6, 12, 18, 24, and 36 mo, with
was performed. The propensity score was re-estimated among this
annual follow-up thereafter. Patients were recommended to undergo
subset, and the models were then fitted again. Finally, analyses were
renal function assessment and cross-sectional imaging at 3 mo and at
repeated using propensity score matching and propensity score
each visit thereafter.
regression adjustment.
Postprocedural inpatient complications were classified according to
Analyses were performed using SAS v9.4 (SAS Institute Inc., Cary, NC,
the Clavien–Dindo system for each procedure. Estimated glomerular
USA). All tests were two sided, with p < 0.05 considered statistically
filtration rate (eGFR) was calculated using the Chronic Kidney Disease
significant.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 3
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
3.
Results
The unweighted cohort included 118 patients with a single renal mass in a solitary kidney, of whom 54 underwent PCA and 64 underwent PN (Supplementary Fig. 1; Table 1). Twelve surgeons performed these PNs, and six interventional radiologists performed the PCAs. Patients who underwent PCA were older (median 67 vs 62 yr), had worse baseline renal function (median 49 vs 57 ml/min/1.73 m2), and smaller tumors (median 2.6 vs 3.8 cm). The surgical approach for PN was open in 57 (89%) patients, pure laparoscopic in five (8%) patients, and robot-assisted laparoscopic in two (3%) patients. Median clamp time was 22 min (interquartile range [IQR] 13, 26; missing n = 16). Upon IPTW adjustment, excellent balance was achieved for the majority of propensity score variables (standardized differences <0.10), while minor imbalances remained for age (median 65 vs 63 yr; standardized difference = 0.15), treatment year (median 2010 vs 2009; standardized difference = 0.12), and histologic confirmation of RCC (78% vs 74%; standardized difference = 0.11). Median follow-up among survivors was 47 mo (IQR 18, 74). 3.1.
Risk of complications
In unadjusted analysis (Table 2), the probability of having any post-treatment complication was lower after PCA versus PN (15% vs 31%; odds ratio [OR] = 0.38; 95% confidence interval [CI] 0.15, 0.96; p = 0.04). Urine leak, acute renal failure, and ileus were the most common complications after PN. Postoperative bleeding was the most common complication after PCA (Supplementary Table 1). However, upon IPTW adjustment (Table 2), there was no significant difference in the probability of
any post-treatment complication between PCA and PN (28% vs 29%; OR = 0.95; 95% CI 0.53, 1.69; p = 0.9). Meanwhile, higher nephrometry score was associated with a greater risk of complications following both PCA (OR [per one point] = 1.73; 95% CI 1.03–2.92; p = 0.04) and PN (OR [per one point] = 1.50; 95% CI 1.10–2.05; p = 0.01), although nephrometry score did not modify the effect of treatment modality on the risk of complications (p interaction = 0.1). 3.2.
Renal function outcomes
In the unweighted analysis (Table 2), there was a smaller drop in eGFR at discharge following PCA versus PN (mean: 7% vs 18%; b = –11%; 95% CI 19, 4; p = 0.004), while there was no significant difference drop in eGFR by 3 mo from baseline (mean: 11% vs 10%; b = 1%; 95% CI –6, 8; p = 0.8). However, in the IPTW analysis, there were no significant differences in drop in eGFR from baseline at discharge (mean: 11% vs 16%; b = 5%; 95% CI 13, 3; p = 0.2) or at 3 mo (12% vs 9%; b = 3%; 95% CI 3, 10; p = 0.3) between PCA and PN (Table 2). Among PN patients with clamp time data (n = 43), longer clamp time was associated with greater drop in eGFR at discharge (b = 0.9%/min; 95% CI 0.3, 1; p = 0.004) but not at 3 mo (b = 0.5%/min; 95% CI –0.03, 1; p = 0.08) on univariable analysis. 3.3.
Treatment failure and oncologic outcomes
There were 14 patients with local treatment failure and 22 patients who died in the unweighted cohort. Among those with histologic confirmation of RCC, nine were with local treatment failures, 19 developed distant metastases, and 18 died, of whom seven died of RCC. Ten of the patients
Table 1 – Cohort characteristics Unweighted cohort Parameter
Whole cohort n = 118
PN n = 64
PCA n = 54
Continuous variables
Median (IQR)
Median (IQR)
Median (IQR)
Age (yr) Charlson comorbidity index Treatment year
64 (57–72) 2 (0–3) 2010 (2007–2012) 53 (42–69) 3.2 (2.1–4.5) 8 (6–9)
62 (56–67) 2 (0–3) 2008 (2007–2011) 57 (44–76) 3.8 (2.4–6.2) 8 (6–10)
67 (57–74) 2 (0–4) 2011 (2008–2013) 49 (39–60) 2.6 (2.0–3.7) 7 (6–9)
Baseline eGFR (ml/min/1.73 m2) Tumor size (cm) RENAL nephrometry score [25]
Categorical variables Female sex Prior nephrectomy for RCC Clinical stage T1a T1b–T2 Histologically confirmed RCC
Weighted cohort p value
Std diff.
0.03 0.2 0.002
0.41 0.25 0.60
0.02 0.006 0.2
0.46 0.53 0.24
n (%)
n (%)
n (%)
32 (27) 65 (55)
17 (27) 24 (38)
15 (28) 41 (76)
0.9 <0.001
79 (67) 39 (33) 85 (72)
35 (55) 29 (45) 51 (80)
44 (81) 10 (19) 34 (63)
PN
PCA
Median (IQR)
Median (IQR)
63 (58–67) 2 (1–3) 2009 (2007–2012) 56 (47–77) 3.7 (2.0–5.0) 8 (5–9)
65 (54–75) 2 (0–4) 2010 (2007–2013) 56 (48–69) 3.5 (2.3–6.5) 7 (6–9)
Std diff.
0.15 0.07 0.12 <0.01 0.06 0.03
%
%
0.03 0.88
27 54
25 51
0.04 0.06
0.002
0.60 0.38
60 40 78
0.06
0.04
57 43 74
0.11
eGFR = estimated glomerular filtration rate; IQR = interquartile range; PCA = percutaneous cryoablation; PN = partial nephrectomy; RCC = renal cell carcinoma; Std diff. = standardized difference. The propensity to undergo PCA versus PN was estimated using a logistic regression model based on age at treatment, sex, Charlson comorbidity index, treatment year, RENAL nephrometry score (measured by two interventional radiologists), tumor size (maximum tumor diameter in centimeters), baseline eGFR, RCC tumor histology, and a history of prior contralateral nephrectomy for RCC. All patients were included in the weighted analysis.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 4
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
Table 2 – Models comparing outcomes between percutaneous renal cryoablation and partial nephrectomy for tumors in a solitary kidney Unadjusted analysis
Outcome Binary outcomes
OR (95% CI)
Any complication
0.38 (0.15, 0.96)
IPTW analysis p value
0.95 (0.53, 1.69)
0.9
beta (95% CI)
p value
beta (95% CI)
p value
% Drop in eGFR by discharge % Drop in eGFR by 3 mo
–11 (–19, –4) 1 (–6, 8)
0.004 0.8
a
Local/ipsilateral failure Distant metastasis b RCC Death b Death from any cause
p value
0.04
Continuous outcomes
Time-to-event outcomes
OR (95% CI)
–5 (–13, 3) 3 (–3, 10)
0.2 0.3
p value
HR (95% CI)
p value
HR (95% CI)
1.27 (0.45, 3.62) 0.79 (0.28, 2.19) 2.69 (0.55, 13.1) 1.54 (0.68, 3.47)
0.6 0.7 0.2 0.3
0.87 (0.38, 1.98) 0.60 (0.30, 1.20) 1.13 (0.32, 3.98) 2.23 (1.32, 3.77)
0.7 0.2 0.8 0.005
CI = confidence interval; eGFR = estimated glomerular filtration rate; HR = hazard ratio; IPTW = inverse probability of treatment weighting; OR = odds ratio; RCC = renal cell carcinoma. Odds ratios and hazard ratios >1 indicate a higher risk of an event/outcome with percutaneous renal cryoablation. Positive beta values indicate a greater decline in renal function following percutaneous renal cryoablation. a Fine-and-Grey models were used for local/ipsilateral failure, distant metastasis, and RCC death, with death from another cause considered as a competing event. Cox proportional hazards models were used for death from any cause. b Only patients or tumors that were proved on biopsy or surgical pathology to be renal cell carcinoma were eligible for analyses of distant metastasis or RCC death.
with local recurrence were managed with repeat percutaneous ablation, two patients were treated with systemic therapy due to concurrent metastatic disease, one patient was managed expectantly, and management of local recurrence was unknown in one patient. In the IPTW-adjusted analysis, there were no significant differences in the risk of local treatment failure with PCA versus PN for the whole weighted cohort (hazard ratio [HR] = 0.87; 95% CI 0.38, 1.98; p = 0.7; Table 2) or among the subset with histologic confirmation of RCC (HR = 0.50; 95% CI 0.15, 1.71; p = 0.3). Moreover, there were no significant differences in the risk of distant metastasis (HR = 0.60; 95% CI 0.30, 1.20; p = 0.2) or CSM (HR = 1.13; 95% CI = 0.32, 3.98; p = 0.8) among patients with histologic confirmation of RCC (Table 2). However, despite IPTW adjustment, PCA was associated with worse OS among the whole IPTW cohort (HR = 2.23; 95% CI 1.32, 3.77; p = 0.005) and among the subset with histologic confirmation of RCC (HR = 2.60; 95% CI 1.42, 4.76; p = 0.003; Table 2). 3.4.
Sensitivity analyses
Given that our main IPTW analysis had some imbalance for age, year of treatment, and histologic confirmation of RCC (Table 1), we used regression to adjust for each of these variables. The association between PCA and worse OS was no longer significantly different upon regression adjustment for age (HR = 1.27; 95% CI 0.72, 2.25; p = 0.4) in the main IPTW analysis. Otherwise, this sensitivity analysis did not alter conclusions (data not shown). Truncating the nonoverlapping tails of the propensity score distributions and stabilizing the weights did not alter conclusions (Supplementary Table 2). Leaving RCC histologic status out of the propensity score calculation in the IPTW analysis, and later adjusting for it using regression, did not alter conclusions, nor did the analysis recalculating the
propensity score among the subset with histologically confirmed RCC (Supplementary Table 2). Sensitivity analyses that instead employed propensity matching and propensity score regression adjustment arrived at similar conclusions, except that PCA was no longer associated with worse OS in these analyses (Supplementary Table 2). Upon further evaluation of the IPTW analysis, the observed association between PCA and worse OS was largely driven by a single patient who was assigned a large weight in the analysis. Upon excluding this patient, the association between treatment group and OS was no longer significant (HR = 1.04; 95% CI 0.52, 2.06; p = 0.9; Supplementary Table 2). 4.
Discussion
Upon accounting for differences in the propensity to receive PCA or PN, we found no significant differences in the risk of post-treatment complication between these two treatment modalities when treating a single tumor in a solitary kidney. Nephrometry score was associated with an increased risk of complications regardless of treatment approach. Additionally, there were no significant differences in renal function or oncologic outcomes between PCA and PN. Although patients who underwent PCA had worse OS in the IPTW analysis, which could reflect greater unmeasured frailty in this population, this should be interpreted with caution since this association was not consistent across all analytic approaches and was driven by an influential observation. Although thermal ablation is being increasingly utilized in the treatment of renal tumors [28], only two reports [17,18] have compared PN and cryoablation (ie, including both PCA and LCA) in patients with a solitary kidney. Meanwhile, one study [16] compared PN and ablation (predominantly RFA) for patients with an imperative indication for nephron sparing, of whom half had a solitary kidney. All three studies found that ablation was associated with a lower rate of
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
complications, which is consistent with the unadjusted analysis in our study. None of the studies found a difference in renal function outcomes between ablation and PN, which is also similar to the present study and another prior report from our group evaluating a cohort of patients with tumors in solitary kidneys who underwent percutaneous RFA or PCA [21]. In contrast to our present study, Mues et al [17] found a greater proportion of recurrence events after ablation versus PN. However, they did not perform a survival analysis or adjust for differences between treatment groups. Meanwhile, Long et al [16] also found that ablation was associated with an increased risk of recurrence, but their cohort predominantly included patients who underwent RFA, which is associated with a greater risk of local failure in tumors larger than 3 cm and central tumors [11]. PN and PCA have also been compared in studies that were not restricted to patients with solitary kidneys [12–14]. In the largest study comparing PN and PCA for cT1 renal masses, there were no differences in risks of local recurrence or distant metastasis [12]. Of note, the local recurrence rate is higher in the present study as compared with the 2% rate at 3 yr in this study of a broader population. This is because clinicians in the present analysis were tasked with maximally sparing nephrons in the setting of a solitary kidney, often with large and/or complex tumors. Caputo et al [13] evaluated cT1b patients and did not find a statistically significant difference in complication rate or eGFR preservation between PN and cryoablation (PCA or LCA). Both Caputo et al [13] and Tanagho et al [14] found that cryoablation (PCA or LCA) was associated with an increased risk of local recurrence, although they found no significant difference in cancer-specific survival. However, it is difficult to discern to what extent this is due to the inclusion of patients who underwent LCA. Although LCA allows for direct visualization of the tumor, PCA allows for real-time monitoring using cross-sectional imaging to ensure that the ice ball extends to cover the entire tumor with a sufficient margin. In addition, we perform adjunctive procedures in selected tumors, such as hydrodisplacement of adjacent bowel, preablation selective arterial embolization, and retrograde pyeloperfusion via an externalized ureteral stent, in order to reduce the risk of complications while allowing for more aggressive treatment of complex renal tumors [22,29,30]. Of note, Klatte et al [31] also reported a meta-analysis comparing LCA and PN, and found that LCA was associated with a greater risk of local recurrence. This further supports evaluating PCA and LCA separately. There are limitations worthy of discussion. First, although we accounted for several patient and tumor characteristics, including tumor complexity, observational studies cannot account for unmeasured differences between groups. For example, it is possible that patients who underwent PCA were frailer, despite controlling for Charlson comorbidity index. We employed a competing risk analysis to minimize this impact on our assessment of oncologic outcomes. It was also not possible to account for procedural learning curve. Second, we cannot rule out distant metastasis and CSM related to RCC in the contralateral kidney. We controlled for the history of contralateral RCC in order to limit this from
5
introducing a bias into our comparative analysis. Third, although this is a large study relative to existing literature, the sample size is still relatively small. Fourth, we were unable to reliably evaluate long-term renal function due to the availability of less consistent data for patients followed at outside institutions, and because PCA patients were more likely to die from competing causes and less likely to have long-term renal function data. Fifth, longer follow-ups will be needed to better assess oncologic outcomes. Finally, caution should be applied before generalizing our findings and outcomes to settings with less experience with PN and/ or PCA. Notable strengths of this study include our use of IPTWadjusted analyses and several sensitivity analyses to account for differences in factors that may have influenced patient selection as well as patient outcomes. Moreover, this is the first comparative effectiveness study in the solitary kidney population to focus specifically on PCA, which should be evaluated separately from percutaneous RFA and LCA, both of which may not have the same outcomes as PCA. In addition, our extensive experience with both treatment approaches allows us to assess and compare the true therapeutic potential of both modalities. 5.
Conclusions
In our IPTW-adjusted analysis and various sensitivity analyses, we did not find a statistically significant difference between PN and PCA in the risk of having a post-treatment complication, in renal function outcomes at 3 mo, or in oncologic outcomes. Small to moderate differences could not be ruled out, and further validation is warranted. Pending such validation, both PN and PCA can be considered for the management of a renal mass in a solitary kidney. Author contributions: Bimal Bhindi had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Thompson, Atwell, Bhindi. Acquisition of data: Haddad, Atwell, Schmit, Weisbrod, Boorjian, Leibovich, Thompson. Analysis and interpretation of data: Bhindi, Atwell, Schmit, Boorjian, Leibovich, Thompson. Drafting of the manuscript: Bhindi, Thompson. Critical revision of the manuscript for important intellectual content: Bhindi, Mason, Haddad, Boorjian, Leibovich, Atwell, Weisbrod, Schmit, Thompson. Statistical analysis: Bhindi. Obtaining funding: None. Administrative, technical, or material support: Thompson, Atwell, Schmit. Supervision: Thompson, Atwell, Schmit. Other: None. Financial disclosures: Bimal Bhindi certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009
EURURO-7553; No. of Pages 6 6
E U R O P E A N U R O L O G Y X X X ( 2 0 17 ) X X X – X X X
Appendix A. Supplementary data
[17] Mues AC, Korets R, Graversen JA, et al. Clinical, pathologic, and functional outcomes after nephron-sparing surgery in patients with
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. eururo.2017.09.009.
a solitary kidney: a multicenter experience. J Endourol 2012;26: 1361–6. [18] Panumatrassamee K, Kaouk JH, Autorino R, et al. Cryoablation versus minimally invasive partial nephrectomy for small renal masses in the solitary kidney: impact of approach on functional
References
outcomes. J Urol 2013;189:818–22.
[1] Mailloux LU, Bellucci AG, Napolitano B, Mossey T, Wilkes BM, Bluestone PA. Survival estimates for 683 patients starting dialysis from 1970 through 1989: identification of risk factors for survival. Clin Nephrol 1994;42:127–35. [2] Campbell SC, Novick AC, Belldegrun A, et al. Guideline for management of the clinical T1 renal mass. J Urol 2009;182:1271–9. [3] Ghavamian R, Cheville JC, Lohse CM, Weaver AL, Zincke H, Blute ML. Renal cell carcinoma in the solitary kidney: an analysis of complications and outcome after nephron sparing surgery. J Urol 2002;168:454–9. [4] Saranchuk JW, Touijer AK, Hakimian P, Snyder ME, Russo P. Partial nephrectomy for patients with a solitary kidney: the Memorial Sloan-Kettering experience. BJU Int 2004;94:1323–8. [5] Ching CB, Lane BR, Campbell SC, Li J, Fergany AF. Five to 10-year followup of open partial nephrectomy in a solitary kidney. J Urol 2013;190:470–4. [6] Gill IS, Colombo Jr JR, Moinzadeh A, et al. Laparoscopic partial nephrectomy in solitary kidney. J Urol 2006;175:454–8. [7] Hillyer SP, Bhayani SB, Allaf ME, et al. Robotic partial nephrectomy for
solitary
kidney:
a
multi-institutional
analysis.
Urology
2013;81:93–7. [8] Bhindi B, Thompson RH, Mason RJ, et al. Comprehensive assessment of renal tumor complexity in a large percutaneous cryoablation cohort. BJU Int 2017;119:905–12. [9] Atwell TD, Schmit GD, Boorjian SA, et al. Percutaneous ablation of renal masses measuring 3.0 cm and smaller: comparative local control and complications after radiofrequency ablation and cryoablation. AJR Am J Roentgenol 2013;200:461–6. [10] Schmit GD, Thompson RH, Kurup AN, et al. Percutaneous cryoablation of solitary sporadic renal cell carcinomas. BJU Int 2012;110: E526–31. [11] Schmit GD, Thompson RH, Kurup AN, et al. Usefulness of R.E.N.A.L. nephrometry scoring system for predicting outcomes and complications of percutaneous ablation of 751 renal tumors. J Urol 2013;189:30–5. [12] Thompson RH, Atwell T, Schmit G, et al. Comparison of partial nephrectomy and percutaneous ablation for cT1 renal masses. Eur Urol 2015;67:252–9. [13] Caputo PA, Zargar H, Ramirez D, et al. Cryoablation versus partial nephrectomy for clinical T1b renal tumors: a matched group comparative analysis. Eur Urol 2017;71:111–7. [14] Tanagho YS, Bhayani SB, Kim EH, Figenshau RS. Renal cryoablation versus robot-assisted partial nephrectomy: Washington University long-term experience. J Endourol 2013;27:1477–86. [15] Weisbrod AJ, Atwell TD, Frank I, et al. Percutaneous cryoablation of masses in a solitary kidney. AJR Am J Roentgenol 2010;194:1620–5.
[19] Lau WK, Blute ML, Weaver AL, Torres VE, Zincke H. Matched comparison of radical nephrectomy vs nephron-sparing surgery in patients with unilateral renal cell carcinoma and a normal contralateral kidney. Mayo Clin Proc 2000;75:1236–42. [20] Bhindi B, Lohse CM, Mason RJ, et al. Are we using the best tumor size cut-points for renal cell carcinoma staging? Urology. In press. http://dx.doi.org/10.1016/j.urology.2017.04.010. [21] Mitchell CR, Atwell TD, Weisbrod AJ, et al. Renal function outcomes in patients treated with partial nephrectomy versus percutaneous ablation for renal tumors in a solitary kidney. J Urol 2011;186:1786– 90. [22] Woodrum DA, Atwell TD, Farrell MA, Andrews JC, Charboneau JW, Callstrom MR. Role of intraarterial embolization before cryoablation of large renal tumors: a pilot study. J Vasc Interv Radiol 2010;21:930–6. [23] Mason R, Kapoor A, Liu Z, et al. The natural history of renal function after surgical management of renal cell carcinoma: results from the Canadian Kidney Cancer Information System. Urol Oncol 2016;34, 486 e1–7. [24] Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med 2015;34:3661–79. [25] Kutikov A, Uzzo RG. The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J Urol 2009;182:844–53. [26] Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Sturmer T. Variable selection for propensity score models. Am J Epidemiol 2006;163:1149–56. [27] Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009;28:3083–107. [28] Choueiri TK, Schutz FA, Hevelone ND, et al. Thermal ablation vs surgery for localized kidney cancer: a Surveillance, Epidemiology, and End Results (SEER) database analysis. Urology 2011;78:93–8. [29] Bodily KD, Atwell TD, Mandrekar JN, et al. Hydrodisplacement in the percutaneous cryoablation of 50 renal tumors. AJR Am J Roentgenol 2010;194:779–83. [30] Cantwell CP, Wah TM, Gervais DA, et al. Protecting the ureter during radiofrequency ablation of renal cell cancer: a pilot study of retrograde pyeloperfusion with cooled dextrose 5% in water. J Vasc Interv Radiol 2008;19:1034–40. [31] Klatte T, Grubmuller B, Waldert M, Weibl P, Remzi M. Laparoscopic cryoablation versus partial nephrectomy for the treatment of small renal masses: systematic review and cumulative analysis of observational studies. Eur Urol 2011;60:435–43.
[16] Long JA, Bernhard JC, Bigot P, et al. Partial nephrectomy versus ablative therapy for the treatment of renal tumors in an imperative setting. World J Urol 2017;35:649–56.
Please cite this article in press as: Bhindi B, et al. Outcomes After Cryoablation Versus Partial Nephrectomy for Sporadic Renal Tumors in a Solitary Kidney: A Propensity Score Analysis. Eur Urol (2017), http://dx.doi.org/10.1016/j.eururo.2017.09.009