Tumors in Solitary Kidneys Are Not All Equal: Outcomes of Partial Nephrectomy in High-Risk Cases

Abstract

Purpose:

While partial nephrectomy remains the preferred treatment for tumors in solitary kidneys, there is a broad range of complexity to these cases. This retrospective study refines our understanding of renal and oncologic outcomes in high-risk cases of partial nephrectomy of the solitary kidney (PNSK).

Materials and Methods:

Review of our institutional database identified patients who underwent PNSK between 1990 and 2020. Cases were classified as high-risk PNSK based on having any of the following: clinical stage ≥T3, RENAL nephrometry score ≥10, or multiple tumors suspicious for malignancy.es

Results:

Of the 124 cases included, 73 were classified as high-risk PNSK. The rate of high-grade peri-operative complications was 32% among high-risk cases, compared with 12% among low-risk cases. The high-risk group recovered less of their baseline renal function compared to the low-risk group, though this gap narrowed with time, and the respective 5-year end-stage renal disease (ESRD)-free survivals were 89% and 100%. Baseline renal function, multifocal tumors, and intraoperative blood loss were predictors of long-term renal function. The 5-year local recurrence-free survival, cancer-specific survival, and overall survival rates in the high-risk group were 83%, 80%, and 77% respectively.

Conclusions:

Outcomes of PNSK remain preferrable to ESRD in the highest-risk patients. Keys to optimization are meticulous preservation of normal parenchyma and a bloodless field. Longer ischemia time is tolerable to achieve these goals, though may benefit from the application of renal hypothermia.

INTRODUCTION

Managing renal tumors in solitary kidneys poses a unique balance between renal function outcomes and oncologic outcomes. When treatment is required, partial nephrectomy of the solitary kidney (PNSK) is generally the preferred strategy, especially over radical nephrectomy, given the known morbidity and mortality associated with end-stage renal disease (ESRD).^{1, 2} Multiple cohorts have reported adequate renal function preservation and oncologic safety with PNSK, though most represented cases have been early-stage and low surgical complexity tumors.^3–7

PNSK is often a reasonably clear-cut decision – early-stage tumors with low surgical complexity offer minimal distinction between solitary and bilateral kidney settings. Substantially more challenging are circumstances where it would be reasonable, if not standard, to perform radical nephrectomy if the patient had normal renal function and a contralateral kidney. Given the severe alternative of ESRD, PNSK remains the reluctant management strategy in these situations, though there remains a paucity of data with which to counsel patients regarding their expected clinical course.⁸

In this study we report our single-center experience enriched for high-risk PNSK. We focus on peri-operative, renal, and oncologic outcomes in the context of locally advanced, multifocal, or surgically complex tumors. Our aim is to optimize management in these challenging cases and improve prognostication of the expected clinical course.

MATERIALS AND METHODS

Cohort

We performed a retrospective review of our institutional database and identified all patients undergoing PNSK from January 1, 1990, through December 31, 2020. Solitary kidney was defined as anatomic absence of the contralateral kidney – cases with functionally solitary kidneys were not captured. Patients with heritable renal cell carcinoma (RCC) syndromes, baseline ESRD, or inadequate available preoperative imaging were excluded.

Preoperative renal ultrasounds and cross-sectional imaging were reviewed by a radiologist and assessed for clinical T stage and RENAL nephrometry score.⁹ Cases were classified as high-risk PNSK based on having any of the following: clinical stage ≥T3, RENAL score ≥10, or multiple tumors suspicious for malignancy. In cases with multiple tumors, the highest clinical stage and RENAL score tumor was used for classification. These criteria were selected on the rationale that, in patients with normal renal function and a contralateral kidney, radical rather than partial nephrectomy might be reasonably favored based on their presence.¹⁰

Renal Function Outcomes

Serum creatinine was used to calculate estimated glomerular filtration rate (eGFR) using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.¹¹ Values were captured immediately preoperative and at 4 postoperative timepoints: prior to discharge from hospital, 2–3 months, 3–12 months, and >12 months. Timepoints were selected based on clinical practice patterns as well as literature describing the course of renal function recovery following partial nephrectomy.¹² In the case of multiple available values within a window, the most recent stable value was used. Normalized eGFR values were calculated as: % baseline eGFR = postoperative value / preoperative value × 100%. ESRD-free survival was captured by identifying the timepoint when renal replacement was initiated.

Oncologic Outcomes

Local recurrence-free survival (LRFS), cancer-specific survival (CSS), and overall survival (OS) were captured for the high-risk group. Some of our high-risk criteria are known prognostic factors for oncologic outcomes in RCC. In addition, many patients will have had previous RCC, a major source of confounding. As such, comparison between high- and low-risk groups was not considered appropriate.

Additional Data

Additional data were obtained from the United States Renal Data System (USRDS) Annual Data Report, a comprehensive overview of kidney disease based on Medicare claims data. This entailed survival probabilities for hemodialysis patients with ESRD from other cause (including urologic disease), measured from the time of initiating hemodialysis, and adjusted for age, sex, race, and ethnicity.

Statistical Analysis

Categorical variables were compared using Fisher’s exact test or Pearson’s chi-squared test while continuous variables were compared using Wilcoxon rank-sum test. Multivariable linear regression was used to model long-term renal function outcomes. The Kaplan-Meier method and Cox proportional hazards regression were used to model time-to-event outcomes, with time measured starting from partial nephrectomy. All statistical tests were two-sided, with p <0.05 considered statistically significant.

Ethics Statement

This study received Memorial Sloan Kettering Cancer Center Institutional Review Board (IRB) approval (IRB # 16–172), permitting a waiver of informed consent in the context of retrospective design. The study was performed in accordance with the principles laid out in the Declaration of Helsinki.

Data Availability Statement

The data sets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

RESULTS

Cohort

A total of 145 patients who underwent PNSK were identified, of whom 124 (86%) had sufficient available preoperative imaging for inclusion. Among the included patients, 73 (59%) were classified as high-risk PNSK, with the contribution from individual criteria illustrated in Figure 1. The median follow-up after partial nephrectomy was 6.0 years across the entire cohort and 5.8 years in the high-risk group.

Figure 1.

Composition of the high-risk partial nephrectomy group (n = 73) by inclusion criteria: clinical tumor stage 3 or higher, RENAL nephrometry score 10 or higher, and/or multiple tumors.

The high-risk group was less comorbid but was otherwise similar in terms of demographic characteristics and RCC history (Table 1). While there were no major differences in technique between the groups, ischemia time, estimated blood loss, peri-operative complications, and length of stay were greater in the high-risk group (Table 1). Clinical staging correlated strongly with pathologic staging, with all but one pT3-stage tumor captured in the high-risk group (Table 1). However, the groups were similar in terms of histologies, with the majority being clear cell (cc)RCC.

Table 1.

Summary of cohort, stratified by partial nephrectomy risk group

Characteristic	Low Risk N = 51a	High Risk N = 73a	p-valueb
Demographic, Comorbidity, and RCC History
Age at Surgery (y)	62 (56 – 70)	62 (56 – 71)	0.7
Gender (Male)	43 (84)	54 (74)	0.2
BMI (kg/m2)	29 (26 – 33)	29 (26 – 33)	0.7
Charlson Comorbidity Index	2 (0 – 6)	0 (0 – 2)	0.024
Baseline CKD Stage III+	35 (69)	45 (62)	0.4
Hypertension	35 (69)	44 (60)	0.3
Diabetes	8 (16)	10 (14)	0.8
Previous Nephrectomy	51 (100)	70 (96)	0.3
Time Since Procedure (y)	7.0 (3.1 – 13.6)	6.0 (1.0 – 10.2)	0.2
Previous RCC	47 (92)	66 (90)	>0.9
Stage IV	10 (20)	10 (14)	0.4
Peri-Operative Course
Approach			0.3
Open	43 (84)	56 (77)
MIS	8 (16)	17 (23)
Ischemia Type			0.3
None	22 (43)	26 (36)
Cold	22 (43)	29 (40)
Warm	7 (14)	18 (25)
Ischemia Time (min)	18 (0 – 32)	27 (0 – 45)	0.067
EBL (mL)	400 (250 – 750)	600 (300 – 1,100)	0.024
OR Time (min)	170 (140 – 218)	206 (170 – 250)	0.01
90-Day Grade 3+ Complications	6 (12)	23 (32)	0.011
Renal Failure - HD	0 (0)	11 (15)	0.003
Urine Leak	5 (10)	12 (16)	0.3
Hemorrhage	1 (2)	2 (3)	>0.9
ICU Admission	0 (0)	5 (7)	0.077
Mortality	0 (0)	0 (0)
Post-Op LOS (d)	4 (3, 6)	6 (4, 9)	0.014
Tumor Characteristics and Pathology
Size (cm)	2.5 (2.2 – 3.5)	3.2 (2.4 – 4.7)	0.025
RENAL Scorec	7 (6 – 8)	8 (7 – 10)
pT Stagec
T1a	42 (82)	36 (49)
T1b	7 (14)	12 (16)
T2	1 (2)	2 (3)
T3	1 (2)	22 (30)
Histology			0.5
Clear Cell	35 (69)	56 (77)
Papillary	10 (20)	11 (15)
Chromophobe	3 (6)	1 (1)
Other	3 (6)	5 (7)
n Tumors Resectedc	1 (1 – 1)	1 (1 – 2)
Margin Positive	8 (16)	10 (14)	0.8

^aMedian (interquartile range [IQR]); n (%)

^bWilcoxon rank sum test; Fisher’s exact test; Pearson’s chi-squared test

^cRisk classification criterium, no statistical comparison performed

BMI, body mass index; CKD, chronic kidney disease; RCC, renal cell carcinoma; MIS, minimally invasive surgery; EBL, estimated blood loss; OR, operating room; HD, hemodialysis; ICU, intensive care unit; LOS, length of stay; pT, pathological tumor stage

Renal Function Outcomes

Eleven patients had acute renal failure requiring hemodialysis in the immediate postoperative period, all of whom were in the high-risk partial nephrectomy group. Nine patients had sufficient renal recovery to discontinue hemodialysis within a year, though two developed ESRD at 1.4 and 4.7 years postoperative, totalling 4/11 patients with long-term ESRD. These patients who required long-term hemodialysis both had baseline stage IIIb CKD. Renal function trajectories and detailed clinical courses are illustrated in Figure 2A and summarized in Supplementary Table 1.

Figure 2A.

Trajectory of estimated glomerular filtration rate (eGFR) recovery for patients with postoperative acute renal failure requiring hemodialysis (n = 11). All patients were high-risk partial nephrectomies. Annotated numbers correspond to clinical courses detailed in Supplementary Table 1. CKD, chronic kidney disease.

Patients who underwent high-risk PNSK recovered less of their baseline renal function and their renal recovery was slower when compared to the low-risk group (Figure 2B). However, the median difference in baseline eGFR recovery between groups decreased from 22% at discharge to 12% at 3–12 months. The 5-year ESRD-free survival in the high-risk group was 89% (95% confidence interval [CI], 81 – 98%), compared with 100% in the low-risk group, though there was no statistically significant difference (log-rank p = 0.09). Nonetheless, all but one patient with long-term ESRD was in the high-risk group (Supplementary Table 2). Likewise, all but one of these patients had baseline stage IIIb or higher CKD.

Figure 2B.

Trajectory of estimated glomerular filtration rate (eGFR) recovery normalized to preoperative baseline, stratified by partial nephrectomy risk group. Plots represent median, interquartile range (IQR), and range. P values represent comparison by Wilcoxon rank-sum test at each timepoint.

Multivariable regression identified baseline renal function, multifocal tumors, and intraoperative blood loss as factors predicting long-term eGFR, while ischemia time and ischemia type did not (Table 2). Ischemia type was collinear with ischemia time and was therefore removed from the model. There were 16 patients (13%), distributed between both groups, who did not have eGFR data available beyond one year and were excluded from long-term renal function analysis.

Table 2.

Multivariable linear regression for factors predicting long-term estimated glomerular filtration rate (eGFR)

Characteristic	Beta	95% CI	p-value
Age	−0.17	−0.56 to 0.22	0.4
Comorbidities
Baseline eGFR	0.72	0.50 to 0.94	<0.001
Hypertension	1.8	−6.1 to 9.6	0.7
Diabetes	1.4	−8.4 to 11	0.8
High-Risk Factors
cT3+	−0.95	−9.4 to 7.5	0.8
RENAL 10+	−0.07	−9.0 to 8.9	0.9
Multifocal	−9.3	−18.5 to 0.03	0.031
Peri-Operative Factors
Ischemia Time (/5min)	−0.60	−1.5 to 0.28	0.2
EBL (/100mL)	−0.47	−0.93 to 0.02	0.06
Grade 3+ Complication	−6.6	−16 to 2.9	0.2

CI, confidence interval; cT3+, clinical tumor stage 3 or higher; EBL, estimated blood loss.

Oncologic Outcomes

The 5-year LRFS in the high-risk group was 83% (95% CI, 74 – 93%). Multivariable Cox regression including the entire cohort (Supplementary Table 3) identified multifocal tumors as a predictor for LRFS (hazard ratio [HR] 4.8; 95% CI, 1.6 – 14.4; p = 0.005). Though not statistically significant, the confidence interval suggests cT3+ as an additional predictor (HR 3.1; 95% CI, 0.87 – 11.1; p = 0.082).

The 5-year CSS in the high-risk group was 80% (95% CI, 70– 92%). Multivariable Cox regression including the entire cohort identified previous pT3+ or ccRCC as predictors of CSS (Supplementary Table 4). Multifocal tumors, cT3+ tumors, and local recurrence were not statistically significant predictors, but once again the confidence intervals suggested potential clinical significance. There were 26 patients (21%), distributed between both groups, without sufficient previous RCC or cause of death information who were excluded from all CSS analysis.

Lastly, the 5-year OS in the high-risk partial nephrectomy group was 77% (95% CI, 68 – 88%), with the corresponding Kaplan-Meier curve shown in Figure 3. Data for patients with ESRD from urologic causes undergoing hemodialysis, adapted from USRDS 2023 Annual Data Report, are also shown.² This is for general descriptive purposes and statistical comparisons were not performed.

Figure 3.

Kaplan-Meier plot of overall survival (OS) among high-risk partial nephrectomy patients in our cohort (measured from time of surgery). Superimposed is the overall survival of patients with end-stage renal disease (ESRD) from urologic causes undergoing hemodialysis in the United States Renal Data System (USRDS) 2023 Annual Data Report (measured from the start of hemodialysis; adjusted for age sex, and race).

DISCUSSION

Partial nephrectomy for advanced-stage or complex tumors in solitary kidneys is inherently higher risk for increased peri-operative complications, decreased renal function preservation, and worse oncologic outcomes. Previous reports suggest 90-day Clavien-Dindo grade ≥3 complication rates between 9% and 49%, depending on the case mix.^{3–5, 13} In our cohort, this rate was 32% among high-risk cases, compared with 12% among low-risk cases (p = 0.011). Unsurprisingly, among the most common peri-operative complications was acute renal failure, with 12% of patients requiring transient hemodialysis, comparable to the 9% reported in previous cohorts.⁵ Importantly, all of these occurred in the context of high-risk PNSK. Temporary dialysis catheters are often prophylactically placed in patients at risk for peri-operative renal failure – our experience suggests low-risk patients can be spared this practice.

The chief motivation for PNSK is the preservation of long-term renal function. There is clear morbidity and mortality associated with severe CKD and ESRD – the mortality associated with hemodialysis is greater than many solid organ malignancies.¹ Previous cohorts have reported >80% conservation of baseline GFR and >95% 5-year dialysis-free survival among all PNSK patients.^3–5 While renal function recovery in our high-risk group followed a more gradual trajectory, recovery of baseline renal function approached that of the low-risk group by one year postoperative (Figure 2B). Though the 89% 5-year ESRD-free survival in the high-risk group was not statistically different from that of low-risk group, all but one of the seven patients with long-term ESRD underwent high-risk PNSK, suggesting that the event rate was too low to detect the difference.

Factors associated with long-term renal function following partial nephrectomy have been studied extensively in the context of bilateral kidneys and to a lesser extent in PNSK. It cannot be assumed that findings in bilateral kidneys are universally applicable to solitary kidneys, given an intraoperative anephric phase and absence of a compensatory contralateral renal unit. Consistent with previous reports, baseline renal function was a strong predictor of long-term renal function.^14–16 Indeed, all patients with long-term ESRD in our cohort had baseline stage III or greater CKD. Preservation of normal parenchyma is also widely accepted as an important factor, offering a potential rationale for multifocality as a predictor in our cohort.^16–18 The role of ischemia time and type has been more heavily debated – some have suggested no added benefit from cold ischemia, while others have suggested that ischemia in general is overemphasized.^{17, 18} In our cohort, neither ischemia time nor ischemia type was clearly predictive of long-term renal function, though these factors were colinear. Taken together, our findings suggest that in performing high-risk PNSK, meticulous and controlled resection with preservation of normal parenchyma is paramount. Longer ischemia times are tolerable to achieve these goals, though may benefit from the application of renal hypothermia.

The 83% 5-year LRFS in high-risk patients is unsurprisingly inferior to reported local recurrence rates for partial nephrectomy in general.^{3, 4} While our cohort excluded patients with heritable RCC syndromes, most patients had previous RCC and multifocality was a strong predictor on local recurrence. This suggests that local recurrences were more commonly de novo tumors attributable to an RCC predisposition rather than technical failure, though insufficient data was available to confirm this supposition. While the event rate was too low to demonstrate statistical significance, an expected association with clinical stage was also suggested. Since most patients had previous RCC, the prognosis associated with these previous tumors is an important consideration.⁵ Our cohort included patients with metastatic disease, though most had all disease sites previously treated, and all had been oncologically stable. Accounting for previous tumors, no other statistically significant predictors of CSS were identified, though once again the expected association appears to be limited by an overall low event rate, with a 5-year CSS of 80%. Cases where patients undergo radical nephrectomy are rare; while not directly comparable, USRDS overall survival data for patient patients with urologic ESRD on hemodialysis appears congruent with the generally accepted notion that patients undergoing PNSK derive a survival benefit even if compromising on oncologic rigor (Figure 2).²

Approaches to managing renal tumors have shifted with improved understanding of their biology and clinical course. Most notably, small tumors (<3–4cm) are considered appropriate for active surveillance as standard, guideline-based management.⁸ This remains true in solitary kidneys, and the best strategy for preserving renal function is to avoid intervention entirely. For tumors where treatment is indicated, percutaneous ablation remains an option, though many of the factors contributing to high-risk PNSK cases likewise preclude safe ablation.⁷ An emerging alternative, particularly in poor surgical candidates, is radiation therapy, which does not have the same size and location limitations as percutaneous ablation.¹⁹ While RCCs have historically been thought to be radioresistant, more recent evidence demonstrates promising results with SBRT.^{19, 20} Lastly, other groups have described success with neoadjuvant TKI treatment and ex-vivo partial nephrectomy with auto-transplantation as strategies for handling the most challenging cases which might otherwise be deemed unresectable.^{4, 13, 21}

As a tertiary cancer referral center, our experience is enriched for complex cases. This study represents one of the largest published cohorts of high-risk PNSK and is unique in its focus on this group of patients. Our median 6.0 years follow-up facilitated comprehensive capture of both renal function and cancer associated events. Nonetheless, these cases are overall uncommon and often uniquely nuanced, with complexity which is challenging to capture at the cohort level. Our study is limited by the usual risks of bias and confounding associated with retrospective design, though given the uncommonness and heterogeneity among these cases, prospective trials are unlikely.

CONCLUSIONS

PNSK in high-risk cases is associated with longer recovery but achieves the goals of renal function preservation while maintaining adequate oncologic control. It is the most appropriate course of management given the severe alternative. Keys to optimization are meticulous preservation of normal parenchyma and a bloodless field. While longer ischemia time is tolerable to achieve these goals, we believe that renal hypothermia should applied to help optimize functional outcomes in this setting. Surgeons should anticipate the individual risk associated with PNSK and integrate these considerations into patient counselling and operative plans accordingly. In our experience, robust surgical planning discussions among experienced colleagues are particularly helpful.