Surgical Outcomes of Cytoreductive Nephrectomy in Patients Receiving Systemic Immunotherapy for Advanced Renal Cell Carcinoma

Abstract

Purpose:

The use of systemic immune checkpoint blockade before surgery is increasing in patients with metastatic renal cell carcinoma, however, the safety and feasibility of performing consolidative cytoreductive nephrectomy after the administration of systemic therapy are not well described.

Patients and Methods:

A retrospective review of patients undergoing nephrectomy was performed using our prospectively maintained institutional database. Patients who received preoperative systemic immunotherapy were identified, and the risk of postoperative complications were compared to those who underwent surgery without upfront systemic treatment. Perioperative characteristics and surgical complications within 90 days following surgery were recorded.

Results:

Overall, we identified 220 patients who underwent cytoreductive nephrectomy from April 2015–December 2022, of which 46 patients (21%) received systemic therapy before undergoing surgery. Unadjusted rates of surgical complications included 20% (n=35) in patients who did not receive upfront systemic therapy and 20% (n=9) in those who received upfront systemic immunotherapy. In our propensity score analysis, there was no statistically significant association between receipt of upfront immunotherapy and 90-day surgical complications [odds ratio (OR): 1.82, 95% confidence interval (CI): 0.59–5.14; P=0.3]. This model, however, demonstrated an association between receipt of upfront immunotherapy and an increased odds of requiring a blood transfusion [OR: 4.53, 95% CI: 1.83–11.7; P=0.001].

Conclusion:

In our cohort, there was no significant difference in surgical complications among patients who received systemic therapy before surgery compared to those who did not receive upfront systemic therapy. Cytoreductive nephrectomy is safe and with low rates of complications following the use of systemic therapy.

1. Introduction

The advent of immunotherapy has revolutionized the treatment of advanced renal cell carcinoma (RCC), with immunotherapy alone or in combination with targeted therapy as the primary approach for patients with metastatic RCC (mRCC) [1, 2]. Nonetheless, cytoreductive nephrectomy (CN) remains a crucial aspect of RCC care, although its indications have evolved with changing systemic therapies [2–5]. Patient selection is vital for optimal outcomes after cytoreduction [6]. Patients who may benefit from CN post-systemic therapy include those initially unsuitable for surgery who have responded well to treatment or those with cancer-related symptoms [7, 8].

Preoperative therapy can have various effects, potentially altering surgical dynamics, including fibrosis and scarring of surrounding tissues [9]. Concerns have been raised about the impact of upfront immunotherapy on surgical feasabiliy and dissection. Additionally, systemic therapy adverse events may increase surgical complications and recovery issues [10]. There are differing views on the effects of treatment [11], which can, in theory, influence surgical outcomes in both positive and negative ways.

With the rise of modern treatment regimens, an increasing number of patients receive systemic therapy before surgery. However, the impact of upfront systemic therapy on surgical complications remains uncertain [9]. We retrospectively analyzed a cohort of patients who underwent CN, comparing those who received upfront systemic immunotherapy to those who did not, to assess the influence of modern treatment on perioperative outcomes.

2. Patients and Methods

2.1. Data source and patient selection

This retrospective review of all patients undergoing CN at our institution from April 2015 to December 2022 was approved by the Memorial Sloan Kettering Cancer Center Institutional Review Board. We utilized our prospectively maintained institutional database to identify consecutive patients age ≥18 years who underwent CN based on preoperative indication and had radiologically suspicious or histsologically proven mRCC. Patients noted with suspicious preoperative radiographic regional adenopathy and subsequently confirmed to have nodal involvement were also included. The decision to proceed with immediate CN or utilize upfront systemic therapy was at the discretion of the treating physician and after multidisciplinary care discussion and consensus.

Clinicodemographics, including perioperative characteristics, were recorded for each patient. Preoperative systemic therapy was recorded and categorized as either no upfront systemic therapy or upfront immunotherapy (inclusive of those treated with immunotherapy [alone or dual] or immunotherapy plus tyrosine kinase inhibitor [TKI]). See Supplementary Table 1 for systemic agents included. Patient comorbidities were captured using an adjusted Charlson-Comorbidity Index (CCI) score. Readmissions and urgent care visits within 90 days following surgery were captured. The duration and last day of therapy administered before surgery were recorded. The primary endpoint was any postoperative surgical complication occurring within 90 days of the index procedure. Complications were captured by our prospectively collected and maintained postoperative surgical complications database and confirmed by medical record review. Complications were classified using the modified Clavien-Dindo classification (CDC) [12].

As essentially all patients were considered to have metastatic disease, we removed the factor from the CCI score as to reflect a patient’s comorbidities and not be artificially elevated given the presence of metastatic disease. In addition, we compiled a list of studies for which patients received upfront systemic therapy and then underwent kidney surgery to serve as a reference (Supplementary Table 2).

2.2. Statistical analysis

Summary statistics were used to describe the clinical and demographic characteristics of our cohort, which were tabulated overall and by group. Frequency and percentages were used for categorical variables, and the median and interquartile range (IQR) were used for continuous variables. Differences between variables were assessed using the Kruskal-Wallis rank-sum test (continuous), and Fisher’s exact test, or Pearson’s Chi-square test (categorical).

Covariate adjustment using propensity scores was used to minimize confounding effects. Variables that were significantly associated with upfront immunotherapy in univariable analysis were entered into a multivariable logistic regression model, which was used to calculate propensity scores. In particular, we adjusted for procedural approach and pathologic staging, as many of the other variables did not show a difference between the two groups on univariable analysis and suggested significant equipoise between the groups. A propensity score was used to balance the two groups to minimize differences. Both endpoints, 90-day surgical complications and blood transfusion, were evaluated using the propensity score methodology. Statistical analyses were performed with R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria).

3. Results

Baseline characteristics of all 220 patients are displayed in Table 1. Overall, 46 patients (21%) received upfront systemic therapy, while 174 (79%) did not. There were few differences in demographics and clinical staging characteristics between our two groups (Table 1). A higher percentage of patients who did not receive upfront systemic therapy underwent an open procedure compared to those who received upfront immunotherapy [83% (n=145) vs 61% (n=28), respectively; P<0.001]. Differences between the two groups in both clinical and pathological T staging were observed. In terms of clinical staging, the proportion of T2 tumors varied by group and was higher in the immunotherapy group [immunotherapy: 11% (n=5) vs no immunotherapy: 1.9% (n=3); P=0.048]. For pathological T staging, the proportion of T3 tumors varied by group [immunotherapy: 59% (n=27) vs no immunotherapy: 80% (n=140); P<0.001]. Furthermore, patients in the immunotherapy group had a lower frequency of pathological M1 disease than those in the no immunotherapy group [33% (n=15) vs 57% (n=99), respectively; P=0.003].

Table 1.

Patient demographics and clinical characteristics

	Upfront systemic therapy
Characteristic	None (N = 174)a	Immunotherapy (N = 46)a	P valueb
Age (yr), median (IQR)	62 (55, 69)	61 (55, 67)	0.7
Gender			0.2
Female	45 (26)	8 (17)
Male	129 (74)	38 (83)
Race			0.2
White	149 (86)	38 (83)
Asian	13 (7.5)	2 (4.3)
Black	4 (2.3)	1 (2.2)
Hispanic	6 (3.4)	2 (4.3)
Other	2 (1.1)	V 3 (6.5)
BMI (kg/m2)			0.5
<25	44 (2 5)	8 (17)
≥30	62 (36)	18 (39)
25–29.9	68 (39)	20 (43)
Preoperative Hgb (g/L), median (IQR)	124.0 (111.3, 139.0)	130.0 (109.0, 143.8)	0.7
CCI, median (IQR)	2 (1, 3)	2 (1, 3)	0.7
Therapy duration before surgery (days), median (IQR)	-	188 (82, 319)	-
Time between last upfront therapy dose a.d surgery (days), median (IQR)	-	27 (14, 90)	-
Time between surgery and resumption of therapy (days), median (IQR)	-	39 (29, 134)	-
Clinical T stage			0.048
T1	20 (12)	6 (13)
T2	3 (1.9)	5 (11)
T3	108 (67)	29 (64)
T4	30 (19)	5 (11)
Unknown	13	1
Clinical N1 stage	100 (57)	32 (70)	0.14
Clinical M1 stage	169 (97)	43 (93)	0.4
Pathological T stage			<0.001
T1	9 (5.2)	8 (17)
T2	5 (2.9)	1 (2.2)
T3	140 (80)	27 (59)
T4	20 (11)	3 (6.5)
TX	0 (0)	7 (15)c
Pathological N stage			0.7
N0	81 (47)	18 (39)
N1	45 (26)	13 (28)
Nx	48 (28)	15 (33)
Pathological M stage			0.003
M0	75 (43)	31 (67)
M1	99 (57)	15 (33)
No. metastatic sites, median (IQR)	1 (1, 2)	2 (1, 3)	0.10
Pathologic max tumor diameter (cm), median (IQR)	9.0 (6.1, 11.5)	6.9 (5.0, 11.0)	0.076
Surgical approach			<0.001
MIS	29 (17)	18 (39)
Open	145 (83)	28 (61)
Estimated blood loss (mL)			0.042
Median (IQR)	350 (150, 788)	175 (100, 600)
Mean [range]	671 [10, 9,000]	825 [15, 13,000]
Operative duration (min)			0.11
Median (IQR)	155 (119, 182)	174 (124, 244)
Mean [range]	171 [75, 1,800]	208 [80, 643]
Blood transfusion needed	38 (22)	14 (30)	0.2
Units transfused, median (IQR)	2.0 (2.0, 3.0)	2.0 (1.3, 3.8)	0.6
Length of stay (days), median (IQR)	2 (1, 3)	2 (1, 4)	0.6
Complication (patient-level)	18 (10)	6 (13)	0.6
Death (within 90 days)	7 (4.0)	0 (0)	0.3
Reason for death
Disease progression	4 (57)	-
Other	3 (43)	-

yr = years; IQR = interquartile range; BMI = body mass index; Hgb = hemoglobin; CCI = Charlson-Comorbidity Index; no. = number; MIS = minimally invasive surgery; min = minutes.

^aReported as n (%) unless otherwise specified.

^bP values calculated using Wilcoxon rank-sum test, Pearson’s Chi-square test, or Fisher’s exact test.

^cPatients had a complete pathologic response at the primary tumor to systemic therapy.

The operative characteristics, including operative duration, need for blood transfusion, and length of stay, did not differ between patients who received upfront immunotherapy versus those who did not (Table 1). Moreover, our immunotherapy group experienced a significantly greater mean estimated blood loss (EBL) compared to the no immunotherapy group [825 mL (range: 15–13,000) vs 671 mL (range: 10–9,000), respectively; P=0.042]. Median operative duration demonstrated no significant difference between the immunotherapy and no immunotherapy groups [174 min (IQR: 124–244) vs 155 min (IQR: 119–182), respectively; P=0.11]. There was no significant difference among the need for blood transfusion in the perioperative period between the immunotherapy and no immunotherapy groups [30% (n=14) vs 22% (n=38), respectively; P=0.2]. Furthermore, there was no difference in preoperative hemoglobin between the two groups (Table 1).

Overall, 11% of patients (n=24) accounted for all 44 surgical complications, with a mean of 1.8 complications per patient, using the Clavien-Dindo classification. A smaller number of complications occurred within the first 5 days of surgery (n=13, 30%), while the majority occurred between 6 and 30 days (n=24, 55%), and 7 (16%) surgical complications occurred after 30 days (Table 2). A list of complete categorized surgical complications and median time to complication are described in Table 2. Approximately half of the recorded surgical complications were grade 3–5 (n=20, 45%; Table 2). Two patients (4.5%) were deemed to have procedural complications that occurred at the time of surgery. Three patients within the no systemic therapy group (1.7%) experienced a complication within 90 days of surgery that resulted in death. One patient experienced septic shock and another suffered hemorrhagic shock on postoperative day (POD) 37, and a third patient had cardiac arrest on POD 36. A detailed breakdown of complications after nephrectomy is displayed in Table 3.

Table 2.

Complications after nephrectomy, stratified by no upfront systemic therapy and upfront immunotherapy

Characteristica	None (n = 35)b	Immunotherapy (n = 9)b
Category
Pulmonary	8 (23)	1 (11)
Gastrointestinal	6 (17)	1 (11)
Renal and urinary	3 (8.6)	0 (0)
Vascular	8 (23)	4 (44)
Infectious	6 (17)	1 (11)
Other	3 (8.6)	1 (11)
Procedural	1 (2.9)	1 (11)
Complication grade
1	0 (0)	1 (11)
2	17 (49)	4 (44)
3	9 (26)	2 (22)
4	5 (14)	1 (11)
5	3 (8.6)	0 (0)
Procedural	1 (2.9)	1 (11)
Time to complication (categorical)
Within 5 days	9 (26)	4 (44)
6 to 30 days	22 (63)	2 (22)
>30 days	4 (11)	3 (33)
Time to complication (days), median (IQR)	8 (6, 18)	13 (0, 50)

IQR = interquartile range

^aComplication-level.

^bReported as n (%) unless otherwise specified.

Table 3.

Detailed characterization of postoperative surgical complications by system

Characteristic	n (%)
Pulmonary (N=9, 20%)
Pleural effusion	3 (33)
Pneumonia	1 (11)
Pulmonary embolus	3 (33)
Respiratory failure	1 (11)
Respiratory failure requiring re-intubation	1 (11)
Gastrointestinal (N=7, 16%)
Bowel perforation	2 (29)
Enterocutaneous fistula	1 (14)
Ileus	2 (29)
Small bowel obstruction	2 (29)
Renal and urinary (N=3, 6.8%)
Acute renal failure	2 (67)
Renal insufficiency requiring dialysis	1 (33)
Vascular (N=12, 27%)
Acute blood loss anemia	3.25
Chylothorax	1 (8.3)
Chylous ascites	3 (25)
Deep venous thrombosis	1 (8.3)
Hematoma	1 (8.3)
Hemorrhagic shock	3 (25)
Infectious (N=7, 16%)
Intra-abdominal infection	5 (71)
Septic shock	1 (14)
Superficial wound infection	1 (14)
Other (N=4, 9.1%)
Cardiac arrest	1 (25)
Cardiac event with cardiopulmonary resuscitation	1 (25)
Hyponatremia	1 (25)
Wound breakdown	1 (25)
Procedural (N=2, 4.5%)
Intraoperative gastrointestinal injury	1 (50)
Splenic injury	1 (50)

Of all patients, 20% had a hospital readmission or urgent care visit within 90 days following surgery, with no differences observed among therapy group (P=0.7; Table 4). Median and mean number of readmissions and urgent care visits are lisited in Table 4 and were similar across the groups.

Table 4.

Readmission and urgent care visits within 90 days of surgery

Characteristic	None (N = 174)a	Immunotherapy (N = 46)a	P valueb
Readmission (any)	36 (21)	8 (17)	0.7
Urgent care visit (any)	37 (21)	8 (17)	0.7
Readmission c			0.2
Median (IQR)	1.0 (1.0, 1.0)	1.0 (1.0, 1.0)
Mean [range]	1.3 [1.0, 3.0]	1.0 [1.0, 1.0]
Urgent care d			0.8
Median (IQR)	1.0 (1.0, 1.0)	1.0 (1.0, 1.0)
Mean [range]	1.2 [1.0, 2.0]	1.5 [1.0, 5.0]

^aReported as n (%) unless otherwise specified.

^bCalculated using Fisher’s exact test or Wilcoxon rank-sum test.

^cOf those who were readmitted.

^dOf those who presented to urgent care.

Using a propensity score analysis, we estimated the odds of 90-day surgical complications and receiving perioperative blood transfusion. With respect to 90-day surgical complications, upfront systemic therapy was not associated with a higher odds of surgical complications when compared to that of patients who received no upfront therapy while controlling for propensity score [Immunotherapy: odds ratio (OR): 1.82 (95% CI: 0.59–5.14); P=0.3; Table 5A]. For perioperative blood transfusion, perioperative immunotherapy was significantly associated with an increased odds of receiving blood transfusion compared to no upfront systemic therapy after adjusting for differences [Immunotherapy: OR: 4.53 (95% CI: 1.83–11.7); P=0.001; Table 5B].

Table 5.

Propensity weighted analysis of (A) 90-day surgical complications and (B) receipt of blood transfusion

Characteristic	(A) 90-day complications
	OR	95% CI	P value
Neoadjuvant therapy
None	—	—
Immunotherapy	1.82	0.59, 5.14	0.3
Propensity score	0.59	0.30, 1.08	0.11
Characteristic	(B) Blood transfusion
	OR	95% CI	P value
Neoadjuvant therapy
None	—	—
Immunotherapy	4.53	1.83, 11.7	0.001
Propensity score	0.21	0.11, 0.39	<0.001

OR = odds ratio; CI = confidence interval.

4. Discussion

We evaluated the impact of modern upfront systemic regimens on surgical outcomes in a large, prospectively maintained single-institutional database. In our cohort, no difference between groups, either based on operative characteristics or postoperative surgical complications, were observed.

As more effective systemic therapy agents become available, there has been an increased emphasis on patient selection in the role of CN for patients with mRCC [13]. The CARMENA trial did not demonstrate a benefit to CN, however, the study population from the trial is generally considered not to be reflective of modern surgical patient selection and concerns about the high percentage of International Metastatic RCC Database Consortium (IMDC) poor risk patients, poor patient accrual, and the low-surgical volume of participating sites have led many to question the generalizability of these findings [14]. Multiple studies, albeit retrospective, have demonstrated a potential benefit to CN, including in patients receiving immunotherapy, and this evidence forms the basis for current practice [3, 8, 15]. A management dilemma that is currently debated in the appropriately selected patient is the choice of proceeding with upfront nephrectomy versus a delayed nephrectomy after upfront systemic therapy [7]. Recently, the use of upfront systemic therapy has been increasingly utilized due to data from the SURTIME trial demonstrating delayed CN being associated with improved survival [7], the potential benefit of tumor downstaging before undergoing resection, and/or the possibility that some patients deemed unfit for surgery and/or unresectable at time of presentation may demonstrate an extraordinary response to systemic therapy, leading to a window where consolidative surgery becomes possible. Additionally, there remains ongoing interest in the role of upfront systemic therapy to downsize the tumor before surgery.

We present one of the largest cohort studies comparing treatment-naïve patients and patients receiving upfront modern-era systemic therapy undergoing CN. Our study demonstrates similar surgical complication rates that have been previously reported in the literature. For example, in an older series that predominantly examined patients undergoing CN without upfront systemic therapy, CDC grade ≥II complications occured in 27% of patients within 8 weeks of surgery [16]; in our more contemporary cohort, postoperative surgical complication rates were approximately 20%, occurring in 11% of patients. Studies that examined upfront use of TKIs on postsurgical outcomes have found a small to negligible association between upfront use and postoperative complications, despite potential theoretical reasons why targeted therapy agents may alter postoperative surgical complications, including the effect on angiogenesis and wound healing [17–19]. In the SURTIME trial, the rate of CDC grade ≥III complications was 41.2% [7], similar to the rate of high-grade complications in our cohort. In a large retrospective study, researchers compared patients who received systemic therapy before CN with those who underwent upfront surgery and found no difference in surgical complications between the groups. However, they identified upfront systemic therapy as an independent risk factor for wound complications, which was mainly attributed to the frequent use of bevacizumab and its impact on angiogenesis [19].

Currently, few studies examine how upfront immunotherapy may affect the surgical outcomes of patients undergoing resection of the primary renal tumor [10, 20, 21]. One small series of 11 patients reported a 30-day postoperative complication rate of 54.6% [20], and another series of 21 patients described the finding of intraoperative major fibrosis in 38% of cases [21]. These results must be interpreted with caution in regard of their limited sample size. Theoretical considerations suggest that posttreatment effects might result in increased scarring and fibrosis, leading to the obliteration of tissue planes, which can make surgical dissection more challenging. Although we did not record these operative characteristics and complications, to an extent, our findings can only be regarded as an indirect indicator of the potential scarring and fibrosis associated with systemic therapy. Moreover, we did not find that upfront immunotherapy significantly altered operative characteristics between the groups in our cohort. Like our study, literature from other surgical disciplines did not find an association with upfront immunotherapy and postoperative surgical complications [22, 23]. In general, our results support that oncologic surgery after immunotherapy is not associated with a significant difference in rates of surgical complications compared to patients who have not received upfront systemic therapy.

Potential reasons why upfront systemic therapy may not lead to increased surgical risk is through downstaging of the renal mass, thus making the surgical dissection more favorable. In a study of patients who recieved upfront immunotherapy, tumor size was smaller compared to immediate CN, with 8% (2/24) of patients who were downstaged from cT3/cT2b to pT0 [24]. Another study evaluating the shrinkage of the primary tumor after immunotherapy/immunotherapy or TKI therapy found that ten weeks after treatment, the median primary tumor volume reduced from 523.6 cm³ to 258.3 cm³ [25]. In our study, we did not quantify changes in size of the tumor mass. However, it’s worth mentioning that while minor disparities were observed in clinical staging, a notable difference emerged in pathological staging. Specifically, we observed a reduced number of patients with pT3 and M1 stages who received CN after immunotherapy, as compared to those who underwent CN without prior systemic therapy. This discrepancy suggests the potential influence of upfront immunotherapy in downstaging the disease. Additionally, use of systemic therapy may lead to some patients becoming candidates for a minimally invasive approach, as we noted a statistically significant difference in rates of utilization amongst patients who had received upfront immunotherapy. However, it remains unclear how this could effect treatment-related morbidity and optimize postoperative recovery. These may be potential reasons why we did not note differences between patients receiving upfront immunotherapy or targeted therapy compared to those who received no therapy on metrics like operative duration, despite earlier reports suggesting a more difficult surgical dissection.

In our propensity weighted model when controlling for differences between groups, it remains unclear why we observed an increased association with systemic immunotherapy and odds of receiving a blood transfusion. Possible explanations may reflect variation in practices over time and/or that blood loss and transfusion occurred in a delayed fashion, thus decoupling operative blood loss with transfusion.

Our study is not without limitation. The retrospective nature of this study may not capture unmeasured confounders in patient selection, nor do we offer clear guidance as to whom may benefit most from surgical resection of the primary tumor or which characteristics are associated with improved postoperative complications. There is also uncertainty in evaluating individual agents and extrapolating inferences to a class of therapeutics. We also acknowledge that the outcomes of interest are relatively rare and that our study may be underpowered to detect small differences between groups. The results of our study at a high-volume oncology center may not necessarily be extrapolated to other settings where patients with kidney cancer receive care. Although efforts were made to capture and categorize postoperative complications, it is possible we were unable to capture CDC grade I complications and thus our results are skewed towards higher grade complications. Lastly, although our institution has standardized means to report communication with outside hospitals, it may be possible (although we have no reason to suspect) that we were unable to capture some patients who presented to an outside institution, leading to underreporting of surgical complications.

5. Conclusion

In sum, we were unable to establish a difference in intraoperative characteristics or postoperative surgical complications among patients receiving upfront modern systemic therapy regimens compared to those not receiving any upfront therapy. Further, the rate of postoperative complications among patients undergoing CN in a modern series is low. This suggests that use of these agents do not significantly contribute to additional postoperative surgical risk. Questions remain about how and when surgical therapy should be incorporated into the multidisciplinary treatment for patients with mRCC and include understanding the optimal duration of preoperative immunotherapy, should that be indicated. Future studies should examine survival endpoints in patients with advance RCC who are treated with systemic immunotherapy followed by surgery.

Highlights.

• Rates of surgical complications are low after cytoreductive nephrectomy

• Upfront immunotherapy not associated with increased intraoperative complications

• Upfront immunotherapy not associated with increased 90d postoperative complications

Abbreviations:

CDC

Clavien-Dindo classification

CN

Cytoreductive nephrectomy

EBL

Estimated blood loss

IQR

Interquartile range

mRCC

Metastatic renal cell carcinoma

POD

Postoperative day

RCC

Renal cell carcinoma

TKI

Tyrosine kinase inhibitor