Clinical and pathological outcomes of deferred nephrectomy in patients with metastatic and locally advanced RCC after immune checkpoint inhibitors

Abstract

Background and objective

Cytoreductive nephrectomy following immune checkpoint blockade (ICB) in metastatic renal cell carcinoma remains controversial, with limited data on its clinical and pathological outcomes. This study evaluated the outcomes of patients undergoing deferred cytoreductive nephrectomy (dCN) after ICB-based treatment, focusing on the radiologic and pathological responses, and postoperative clinical outcomes.

Methods

We retrospectively reviewed 24 patients with metastatic or locally advanced RCC who underwent dCN after ICB at a single institution between April 2018 and May 2024. We assessed the radiological response to ICB, pathological findings (presence and extent of necrosis) in resected primary tumors, and postoperative clinical outcomes, including the rate of patients without measurable disease and those who discontinued systemic therapy.

Results

Median ICB exposure prior to surgery was 11.3 months. Radiologically, 67% of patients had partial response, 29% had stable disease, and 4% had a complete response. Pathology showed 96% of specimens with necrosis, 21% of specimens showing no residual disease (pT0), and 21% exhibiting ≥95% necrosis. Postoperatively, 50% of patients had nonmeasurable disease of first follow-up scans, and 54% discontinued systemic therapy, with 9 patients remaining on surveillance at last follow-up. Limitations include the small sample size and retrospective design.

Conclusion

Deferred CN following ICB therapy is feasible. Extensive necrosis in the resected surgical specimen after ICB-based therapy requires further investigation as a prognostic marker for durable responses off systemic therapy.

Introduction

Two decades ago, pivotal randomized trials demonstrated a survival benefit for cytoreductive nephrectomy (CN) followed by interferon alpha in metastatic renal cell carcinoma (mRCC).^1,2 In a pooled analysis of 331 patients, there was a 4.8-month improvement in overall survival (OS) advantage for CN with cytokine therapy compared with interferon alpha alone.³

Next came the introduction of tyrosine kinase inhibitors (TKI) that demonstrated superior objective responses rates and survival outcomes with some efficacy on the primary kidney tumor. Consequently, reassessing the role of CN in during TKI treatment was a necessary.

The CARMENA trial was a noninferiority study that randomized patients to undergo CN not prior to the initiation of sunitinib in patients with mRCC. The trial included 450 participants and achieved its primary endpoint, demonstrating noninferiority in OS between the nephrectomy and TKI group versus TKI-only group (HR, 0.89; 95% CI, 0.71-1.10).⁴

Extended follow-up data revealed a rising hazard ratio and widening of the CI over time (HR, 0.97; 95% CI, 0.79-1.19), suggesting it was not noninferior.

Post-hoc analysis further indicated the potential for benefit in specific subgroups, including those with a single International Metastatic RCC Database Consortium (IMDC) risk factor (n = 126, mOS of 31.4 vs 25.2 months), those with a solitary metastatic site (n = 143, mOS 23.2 vs 14.4 months), and those who underwent delayed nephrectomy in the sunitinib-alone group (n = 40, 48.5 vs 15.7 months).⁵ These findings suggest that certain patients may derive benefit from primary tumor resection and also that delayed nephrectomy in patients who respond to systemic therapy is a reasonable strategy.

Additionally, SURTIME was a prospective clinical trial that attempted to answer the timing of surgery relative to TK administration, but it was closed prematurely for poor accrual and did not find a significant difference between immediate CN or delayed surgery after 3 cycles of sunitinib.⁶ In summary, these trials changed the standard of care from upfront nephrectomy during the cytokine era to one of prioritizing systemic therapy before considering CN.

Immune checkpoint blockade (ICB) combinations have further enhanced outcomes of mRCC patients with increased survival and, sustained responses, with median duration of response reaching impressive 76 months in CheckMate 214, for example.⁷ However, efficacy of ICB on primary tumors in the kidney has been reported from retrospective series demonstrating median tumor shrinkage of 2.3-2.5 cm and up to 15% of patients with no viable tumor (pT0) in the kidney.^8-11 As a result, consolidative CN after a systemic response to ICB-based therapy may be used to reduce tumor burden and potentially render a patient disease-free in selected cases. This study is a single-center retrospective cohort analysis aimed at characterizing pathological and clinical outcomes of patients who underwent deferred CN (dCN) following ICB-based treatment.

Methods

A retrospective analysis was conducted at Vanderbilt University Medical Center to evaluate patients with metastatic or locally advanced renal cell carcinoma who underwent dCN following prior exposure to ICB.

Eligibility criteria

The analysis included subjects with metastatic or locally advanced renal cell carcinoma who were initially deemed unsuitable for upfront surgery, regardless of tumor histology. Subjects must have received at least 1 dose of ICB before undergoing dCN. Acceptable treatments included anti-program cell death protein 1 (PD-1) or anti-programmed death-ligand 1 (PD-L1) therapies, administered either as single-agent or in combination with cytotoxic T-lymphocyte-associated protein 4 (anti-CTLA4) or VEGF-TKI. ICB could have been used as any line of treatment and must have been the last treatment received preceding CN.

Data collection

The data were collected using HIPAA-compliant data registry and were conducted under Institutional Review Board-approved protocol 160979.

Collected clinicopathological characteristics included age, gender, complete blood count, comprehensive metabolic panel before initiating systemic therapy, sites of metastasis, IMDC risk score before starting therapy, and pathologic tumor characteristics such as size, histology, nodal involvement, necrosis and its extension (reported in percentage), and sarcomatoid differentiation. Information on the first and last doses of ICB before surgery and any subsequent treatments postsurgery was also gathered.

Assessment parameters

Primary tumor size was assessed before ICB exposure and again before surgery using standard radiological imaging, and pathological size was measured after surgery. The presence and the extension of necrosis were collected in all pathology reports and were reviewed by an expert genitourinary pathologist (J.G.). Extensive necrosis was defined as ≥95% of necrosis in the pathological sample. Although there is no standard accepted definition for extensive necrosis, we adopted this threshold based on the work of Foria et al.¹²

Radiologic response was evaluated by investigators using RECIST 1.1 criteria. Tumor shrinkage was calculated based on radiological measurements (in centimeters) taken before ICB initiation and again before surgery. A complete radiological response was defined as the disappearance of the disease at all extra-renal metastatic sites and the primary tumor. Postoperatively, we described disease status based on the presence of measurable versus nonmeasurable disease per RECIST 1.1 criteria per investigator assessment. Progression of disease (PD) was defined according to RECIST 1.1 or any imaging changes that led to the reinitiation of systemic treatment per treating physician.

Post-surgical treatment data was collected during chart review and categorized into 3 groups: (1) surveillance, if systemic therapy was discontinued following surgery; (2) reinitiation of pre-surgical treatment, if patients resumed the same regimen, they were on before CN; and (3) switch to a different systemic therapy after surgery.

Time intervals and statistical analysis

The median duration of treatment before surgery was determined from the first to the last dose of ICB. Time to progression was calculated from the surgery date to either the date of PD or death, whichever occurred first.

Baseline characteristics were tabulated with categorical variables described as numbers and percents and continuous variables reported as medians and ranges. The starting date for calculation of continuous variables was set as the first day of ICB therapy.

Results

Patient and disease characteristics

Twenty-four patients were included in our analysis. Patients underwent surgery between April 2018 and April 2024. The baseline characteristics are detailed in Table 1. The median age was 62 years (range: 40 – 83), most patients were male with good performance status. 90% of patients were classified as intermediate-risk by IMDC criteria; 84% of patients had distant metastasis with a median number of metastatic sites of 2 (range: 0 – 3), representing the typical patients referred for CN. The remaining 16% (n = 4) had locally advanced disease and were deemed ineligible for upfront surgery, because of lymph node involvement (n = 3) or inferior venous cava (IVC) thrombus (n = 2). In the metastatic population, the most common sites of metastasis were the lungs (58.3%), lymph nodes (41.7%), bone (33.3%), liver (12.5%) and 29.2% of the patients presented with IVC involvement.

Treatment exposure

The majority of patients (84%) were treatment-naïve and received first-line ICB-based treatment. Among these, combination therapy of anti-PD1 plus TKI was the most common first-line treatment (67%), followed by ipilimumab-nivolumab (17%), detailed in Table 1. The remaining 16% of patients received nivolumab monotherapy as second-line treatment after progression on TKI. The median follow-up was 25.8 months (4.6– 92.6) from the first ICB infusion. The median ICB exposure prior to surgery was 11.3 months (range: 2.1-42.7) and a median time from the last ICB dose to time of surgery was 1.3 months (range: 0.1-7.9). Notably, 87.5% of patients had surgery within 3 months from the last ICB infusion date, and one patient had surgery 7.9 months after completing 2 years of ICB-TKI treatment with a persisting enhancing renal lesion on radiological assessment.

Radiological and pathological outcomes

Best systemic radiological response before surgery revealed 67% of patients with partial response (PR), 29% stable disease (SD), and one patient (4%) exhibited a complete response (CR) on radiological assessment prior to surgery. Assessing the primary tumor response, the median size before ICB was 10.0 cm (range: 2.8-30.0), the median tumor size before surgery was 7.1 cm (range: 0.0 – 20.4), and the median pathological size post-surgery was 7.2 cm (range 0.0 – 21.1). For those patients treated with anti-PD1 plus TKI (n = 16), median size before ICB was 10.5 cm (range: 3.3-17.9) and median size before surgery was 7.5 cm (range: 0 – 20.4), an absolute reduction of 3.0 cm, with 15 patients experiencing any tumor shrinkage and only one experiencing primary tumor growth. Four patients (16%) had a primary tumor increase between baseline staging and pre-surgery staging, from a median size 4.3 cm (range: 2.8. – 17.9) to 7.5 cm (range: 4.5-21.1).

All patients underwent radical nephrectomy. Histological examination revealed clear cell carcinoma in 79% of cases, 17% unclassified RCC, 4% papillary RCC. Fuhrman grade 3 or 4 was observed in 71% of cases and 12.5% exhibited sarcomatoid differentiation (Table 2).

Necrosis was reported in 96% of tumor specimens, and the median necrosis extent was 35% (range 0 – 100%). Near one fifth (21%) of the patients had no residual disease (pT0) in their primary at the time of CN, after a median time of ICB exposure of 13.1 months (range: 10.5-28.7) in these patients. All patients who had pT0 status experienced primary tumor shrinkage on radiological assessment before surgery, with a median reduction of 59.3% (range: 38.5% – 100%), from 10 cm (range: 5.4-17.2) to 3.9 cm (0 – 8.6) with 4 PRs and one CR as best systemic radiological response (Figure 1 and Table 3).

Five patients (21%) exhibited 95-99% necrosis, all of whom experienced radiological primary tumor shrinkage with a median reduction of 31.1% (range: 1% - 57%), from 10.6 cm (range: 7.6-15.7) to 7.5 cm (range: 4.5-8.7). The best overall response in radiological assessment in this subgroup included 4 PRs and one SD (Figure 1 and Table 4).

Lastly, 14 patients (58%) had less than 95% necrosis in the specimen, with a median necrosis extent of 15% (range: 0 – 80%). In this subgroup, 10 patients exhibited any radiological shrinkage in the primary tumor and the remaining 4 patients experienced primary tumor growth before surgery. Best overall radiological response assessment showed 57% PR and 43% had SD (Figure 2 and Table 4).

Postoperative follow-up

At the time of last analysis, median follow-up after surgery for the entire cohort was 10.6 months (range: 1.6-77.2). Half of the patients (n = 12) had no evidence of measurable disease (per RECIST 1.1) and the other half had residiual measurable metastatic lesions on the first post-operative scans.

Following surgery, 54% discontinued treatment and were put on surveillance at physician’s discretion, 34% resumed pre-surgery systemic treatment, 4% switched the line of treatment and 8% did not have available information. One patient, who had pT0, completed 2 years of ICB after CN and is currently on surveillance for the last 43 months. Among those who discontinued systemic treatment after surgery, after a median follow-up of 24.8 months (range: 6.7-92.6), 4 (33%) had PD.

At the time of the last analysis in August 2024, with a median follow-up of 16.2 months (range: 5.2-77.2) after surgery, 46% of the patients had PD or death event, (10 progressions and 1 death). The median time to the first event (PD or death) was 5.2 months (range: 1.0-14). Eight of these events occurred in the subgroup without extensive necrosis (n = 14) and 3 occurred in the subgroup with extensive necrosis (n = 10).

All patients who had pT0 disease were still free of recurrence after a median follow-up of 16.5 months (range: 2.8-54) after surgery (Table 3).

Discussion

The role of CN in the era of immunotherapy for RCC remains a matter of ongoing discussion, primarily due to the lack of prospective data. In our cohort, we observed primary tumor reduction after ICB that, while modest, included a notable proportion of patients with a pathologic complete response (pCR) and/or extensive necrosis and a subgroup that was able to discontinue therapy.

Historically, the efficacy of systemic therapy in the primary renal tumor is limited. A phase II trial of 30 patients with unresectable or advanced mRCC showed that TKI alone achieved a 22% decrease or absolute 1.2-cm reduction in primary tumor and none of the 13 patients who had deferred nephrectomy had a pCR.¹³ Prospective data regarding ICB efficacy in primary tumor in this setting are scarce. Two post-hoc analysis, from JAVELIN Renal 101 and from Checkmate 214 including patients without prior nephrectomy, showed objective response rates ranging from 30% to 34% in the primary tumor.^14,15

Despite the limited data regarding radiological response in the primary tumor, the observation of pT0 underscores the effectiveness of ICB-regimens in the primary tumor. Panian et al. and Gunenc et al., in retrospective cohorts, involving 52 and 38 patients, respectively, reported pCR rates ranging from 11% to 13%. These results differ slightly from our study’s, 21% pCR, which may be explained by fewer patients in those series receiving front-line therapy (25% vs 70%) and a shorter median duration of treatment duration prior to surgery, 8.1 and 4.3 months, compared with 11.3 months.

Our findings also differ from those of Panian et al., regarding the heterogeneity of patients who exhibited pT0 status in relation to IMDC-risk, tumor histology, primary tumor response to ICB, and treatment regimens. For instance, in our cohort, 4 out of 5 received ICB-TKI, whereas in the Panian et al. study, only 1 of 7 patients received this regimen. Notably, none of the 12 patients who had pT0 across both cohorts exhibited sarcomatoid differentiation in their tumor specimens.

Combination therapies involving anti-PD1 with either TKI or anti-CTLA4 have been shown to increase RECIST 1.1 complete responses when compared with TKI alone. However, around 80% of patients in these trials had previously undergone nephrectomy. As mentioned earlier, achieving complete response in the primary tumor is rare and many patients still have residual malignancy in the primary site. Therefore, dCN can remove a greater fractional percentage of the tumor volume, potentially rendering a subset of patients disease-free after surgery, which has been associated with better outcomes.^16,17

Consistent with these findings, a recent FDA-pooled exploratory analysis of JAVELIN Renal 101, KEYNOTE-426, CLEAR, and IMmotion151 revealed improved progression-free survival, 15 versus 11 months, and OS, 46 vs 28 months, for those patients who had CN compared with those who did not.¹⁸ It is important to emphasize that these patients had upfront CN, while our analysis is limited to patients who had dCN. The impact of the primary renal tumor in ICB efficacy and the timing of surgery requires further investigation.

Our study has the uniqueness of exploring extent of necrosis. It is well-established that coagulative tumor necrosis in primary tumor is a predictor of poorer outcomes in patients with clear cell RCC,¹⁹ and extensive tumor necrosis (≥95% necrosis) is uncommon in patients who undergo upfront nephrectomy, occurring in less than 2% of the cases.¹² Our findings are hypothesis generating and suggest that necrosis induced by ICB-based treatment might hold prognostic information.

The definition of extensive necrosis in RCC lacks standardization and its clinical significance after ICB exposure remains uncertain. Recently, major pathological response (defined by ≤10% of viable tumor cells in the resected primary) has been explored in phase 3 trials such as NADINA and KEYNOTE-671 trial, which tested perioperative ICB in melanoma and nonsmall cell lung cancer, respectively. In these studies, ICB increased the rate of major pathological response and demonstrated improved clinical outcomes for these patients.^20,21

Although the perioperative efficacy of ICB has been better studied in other malignancies, its role in RCC still unclear. The PROSPER EA8143 trial was the largest prospective trial of perioperative ICB in localized RCC, randomized 819 patients to nivolumab plus surgery versus surgery alone. The trial failed to improve recurrence-free survival, and the effect of ICB on the primary tumor was not reported, likely due to the administration of only 2 preoperative cycles of nivolumab.²² In contrast, the NeoAvAx trial of avelumab plus axitinib in localized RCC demonstrated 30% PR (n = 40), 20% of tumor downsizing and numerically longer disease-free survival in responders.²³

In our cohort, postoperative treatment was decided per physicians’ choice and over half of participants discontinued systemic treatment after dCN. Shapiro et al. reported 48% and Panian et al. 23% of systemic treatment discontinuation after dCN.^8,24 However, we lack information about the clinic-pathological characteristics of these patients and how they fared postdiscontinuation. This highlights the need for further investigation into the factors influencing treatment decisions and the long-term outcomes of these patients who cease therapy.

The strength of our findings lies in the fact that all kidney specimens were assessed and reviewed at a single institution by the same pathologist (J.G.), ensuring consistency in the reports. Our cohort differs from previous retrospective studies as it explores the extent of necrosis and its correlation with clinical outcomes, including the treatment received and disease status during the available postoperative follow-up.

However, our study faces several limitations, including its retrospective nature and the potential for selection bias, the small sample size, and the short follow-up for a substantial part of the patients. Additionally, pathological data were mainly collected in the primary specimen after ICB-exposure what can limit interpretation of some variables, such as Fuhrman grade which was defined in untreated RCC tumors.

Conducting prospective studies involving CN faces several challenges: slow accrual as seen in CARMENA and SURETIME (most patients have metachronous metastatic disease, a substantial number of patients are not healthy enough for surgery), the decline in CN post-CARMENA trial, and the need for a specialized multidisciplinary team. To illustrate this, our cohort analysis spans nearly 6 years at a single high-volume center. Prospective phase III trials are ongoing including PROBE [NCT04510597] and NORDIC-SUN [NCT03977571] to clarify the role of dCN in patients receiving ICB-based regimens. However, both trials will require patients to resume ICB maintenance after surgery, which will limit conclusions regarding therapy de-escalation or discontinuation.^25,26 These clinical trials will be informative for the field by assessing pathological outcomes alongside clinical outcomes.

Our cohort suggests that dCN may help reduce tumoral burden and that necrosis extension may emerge as a promising biomarker for treatment response and prognosis. Further research is needed to investigate and standardize RCC pathologic parameters associated with clinical outcomes.

Author contributions

Paulo S. do Amaral (Conceptualization, Data curation, Investigation, Methodology, Project administration, Resources, Visualization, Writing—original draft, Writing—review & editing), Kathryn E. Beckermann (Conceptualization, Resources, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Jennifer B. Gordetsky (Resources, Writing—review & editing), Sam S. Chang (Resources, Writing—review & editing), Daniel D. Joyce (Resources, Writing—review & editing), Kerry Schaffer (Writing—review & editing), Amy N. Luckenbaugh (Resources, Writing—review & editing), Morgan A. Lambrecht (Writing—review & editing), Elizabeth G. Ryan (Writing—review & editing), Daniel A. Barocas (Resources, Writing—review & editing), Kelvin A. Moses (Resources, Writing—review & editing), and Brian I. Rini (Conceptualization, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing)

Funding

Not applicable.

Conflict of interest

K.B. served as consultant to Alpine Bioscience, Aveo, Aravive, Adicet, Astrazeneca, Exelixis, BMS, Merck, Xencor, Arcus, Eisai, and Nimbus. K.A.M. received honoraria from Boston Scientific-Data Safety Monitoring Comitee. The remaining authors have no COI to declare.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References