Metastasectomy in Advanced Renal Cell Carcinoma: A Systematic Review

Metastasectomy in solid tumors

Traditional oncology teaching holds that metastatic solid tumors represent a broader systemic disease process in which distant micrometastatic deposits will ultimately progress after resection of detectable disease. However, reports of favorable outcomes including prolonged disease-free interval have been reported across tumor types, primarily in retrospective series and case reports. These cases tended to be highly selected and enriched for patients with oligometastatic disease, slow progression, extended disease-free interval, or excellent functional status. Unfortunately, randomized data addressing whether there is a benefit from surgery are lacking. One study conducted at MD Anderson randomized patients with non-small cell lung cancer to consolidation local therapy (surgery or stereotactic radiosurgery) [1] after initial chemotherapy, versus maintenance chemotherapy alone. Local therapy was associated with improved disease-free survival suggesting that the traditional paradigm may be incomplete. A randomized phase II study in patients with oligometastatic prostate cancer found improved androgen deprivation-free survival with metastasis-directed therapy (surgery or radiation versus observation) [2]. Aside from these recently published examples, metastasectomy for solid tumors remains based on single-armed case series and physician experience.

Metastasectomy for mRCC

Renal cell carcinoma (RCC) can metastasize to many different organs and has a variable natural history. It may be rapidly progressive or indolent, requiring no immediate systemic treatment [3]. Identifiable metastases are present at diagnosis in up to 30% of cases or become apparent years after nephrectomy for clinically localized disease in nearly 40% of cases. The most common site of metastasis is the lung (45.2%), followed by bone (29.5%), lymph nodes (21.8%), and liver (20.3%) [4]. Several clinical factors have been associated with improved response to treatment of metastatic RCC (mRCC). These are important in determining suitability for surgical intervention and include performance status, length of disease-free interval, synchronous or metachronous metastasis, burden of metastatic disease and number of location and sites involved [5]. Although there have been many advances in the treatment of mRCC, including the development and approval of immunotherapy regimens, complete responses are still rare. As such, surgical management of metastatic disease remains an important aspect of therapy for long-term disease control.

Metastasectomy for RCC was first reported by Barney and Churchill in 1939 [6]. There have been no randomized controlled trials to address the role of metastasectomy in mRCC, however, there have been many retrospective studies and a few prospective series. This systematic review will summarize and discuss the available evidence.

Lung

The lung is the most common target organ for metastasis from RCC. Of the 44 studies reviewed here, 17 were comprised of patients who had metastasectomy for RCC lung metastases. These accounted for 1462 out of 4195 total patients included in this systematic review. Prognostic factors for survival with lung metastases have been studied in multiple series and meta-analyses. In a systematic review of 16 studies with a total of 1447 patients, Zhou et al. reported 1, 3, 5, and 10–year overall survival (OS) rates of 84%, 59%, 43%, and 20%, respectively following lung metastasectomy [9]. They identified lymph node involvement, incomplete resection, multifocality, size, synchronous metastases, and a short DFI as poor prognostic factors. Another report from the Mayo Clinic detailed 887 patients who underwent nephrectomy for RCC and subsequently developed metastatic disease. Of these 125 patients had complete resection of all metastatic disease. Patients who had lung-only disease had a 5–year cancer-specific survival (CSS) of 73.6% after complete resection compared with 19% for those without complete resection [10]. Similarly, in a Heidelberg series of 191 patients with pulmonary RCC metastases, 149 patients achieved complete resection [11] for a 5–year OS of 41.5% compared with 22.1% without complete resection. Notably, in this series, lymph node involvement decreased the 5–year OS to only 24.4%.

The negative impact of concomitant lymph node involvement was also emphasized by Renaud and colleagues. In their report of 122 patients who underwent pulmonary metastasectomy, 35% had lymph node involvement, decreasing OS from 107 months to 37 months (p = 0.003) [12]. Kudelin and colleagues completely resected all mediastinal lymphadenopathy at the time of pulmonary metastasectomy with overall survival of 49% and 21% at 5 and 10 years, respectively [13]. Mediastinal lymph nodes were positive in 46%. Notably, in this series the presence of intra-thoracic lymph node metastases did not result in inferior survival after metastasectomy, suggesting there may be clinical benefit to lymphadenectomy [13].

Finally, Baier and colleagues examined complete resection of multi-focal pulmonary RCC metastases (mean 13/patient). Using laser resection in 237 patients, they achieved complete resection in 208 for a median OS of 69 months compared with 19 months for those with an incomplete resection (p < 0.00001) [14]. Although a higher number of metastases was associated with a shorter OS, the advantage of complete resection was maintained for all levels of disease.

In summary, these studies demonstrate the relative safety and feasibility of resection of pulmonary metastasis and the potential for favorable outcomes. The impact of surgery on overall survival, however, is not definitively established by these non-randomized studies.

Bone

Bone is the second most common site of metastatic disease in RCC. Surgery (both resection and mechanical stabilization) and radiotherapy are often performed to mitigate local complications. Outcomes data in large series, however, are limited. A retrospective study of 295 patients with bone metastases who underwent orthopedic metastasectomy at MD Anderson showed an OS of 47% and 11% at 1 year and 5 years, respectively [15]. Another report of patients treated between 1977 and 1996 at Massachusetts General Hospital showed a 5 year OS of 55% [16]. These results are particularly notable, because they were obtained prior to the development of effective systemic therapies. Both identified a solitary site of bone metastasis as a strong predictor of overall survival. A more recent published series included 33 patients who underwent surgical resection with or without radiation, plus systemic targeted therapy [17]. Median OS was 39.1 months for surgical resection versus 7.6 months for 59 patients with bone metastases who did not undergo surgery. The authors acknowledge the potential for selection bias in these data. The impact of metastasectomy on survival in patients with RCC bone metastases remains unknown without randomized data. However, local disease control for palliation and preservation of function provides a strong clinical rationale for surgery in many cases and would make randomization of these patients to surgery versus no surgery clinically challenging.

Liver

A retrospective review from 14 centers in the Netherlands identified 33 patients who underwent resection or ablation of liver metastases [18]. The OS at 1, 3, and 5 years was 79%, 47%, and 43% respectively. Metachronous metastases and radical resection were statistically significant prognostic factors. Size <50 mm, solitary metastases and presence of extrahepatic metastases did not significantly impact survival. Another retrospective study by Staehler et al. identified 88 patients with liver metastases, including 68 who underwent metastasectomy [19]. The 5–year OS after metastasectomy was 62% compared to 29% in those who did not have a resection. A retrospective study by Thelen et al. of 31 patients who underwent surgery for liver metastases reported overall survival of 82%, 54% and 39% at 1, 2 and 5 years [20]. Incomplete resection or positive margins emerged as a statistically significant prognostic factor in multivariate analysis.

Pancreas

Renal cell carcinoma can metastasize to the pancreas, in many cases as the only site of recurrent disease. Published case series are generally small, but favorable survival outcomes have been reported. In a series of 19 patients who underwent surgery for pancreatic RCC metastases by Fikatas et al., 5–year OS was 71.4% [21]. A similar series of 20 patients with pancreatic metastasectomy showed a 72% OS at 4 years [22]. Consistent with these findings, a study of 27 patients undergoing pancreatic metastasectomy at Memorial Sloan-Kettering Cancer Center found a median OS of 8 years [23]. In contrast, a retrospective study of patients from 16 Italian centers with pancreatic metastases included 44 patients who underwent surgery with or without systemic therapy and 59 patients who had systemic therapy alone [24]. Median OS was 103 months in the patients who underwent surgery and 86 months in those who did not: a difference that was not statistically significant. MSKCC prognostic group was predictive of overall survival. Median overall survival was not reached in the good risk group; 86 months in the intermediate risk group and 42 months in the poor risk group. The authors conclude that outcomes are favorable in patients with pancreatic RCC metastases, and that pancreatic resection did not improve survival. Taken together, these studies suggest that patients with pancreatic metastases of RCC generally have an indolent course and favorable prognosis.

Brain

Brain metastases from RCC can cause devastating neurological complications. Therefore, these cases are generally treated promptly with stereotactic radiosurgery (SRS), whole brain radiation, or surgical resection. A series from MD Anderson of patients treated with SRS observed a median OS of 4–11 months after diagnosis with a 5–year OS of 12% [25]. Large surgical series were not identified in the literature. Local control of brain disease remains an objective of the highest priority regardless of impact on overall disease status, owing to the critical importance of preserving CNS function.

“Atypical” Metastases

In a study which defined “atypical” sites of metastasis as those that did not occur in the thorax, bone, liver, brain or adrenal, 37 patients who had “atypical” metastasectomy were compared to 57 patients who had lung metastasectomy. The authors reported that those who underwent atypical metastasectomy had a median overall survival of 40.8 months, whereas those who underwent lung metastasectomy had a median overall survival of 50.7 months (p = 0.372) [26].

Prognostic factors

Most of the studies included in this review attempted to identify prognostic factors associated with favorable outcomes after metastasectomy. Univariate and multivariate analysis of prognostic factors for overall survival are summarized in Table 2. Variables that were only evaluated in a minority of the studies were not included in the table. Some of these excluded variables, such as tumor size and MSKCC or IMDC risk score, were associated with survival in individual studies. In general, solitary or oligometastatic disease, metachronous metastasis with a disease-free interval of >2 years, complete resection, and absence of lymph node involvement are associated with favorable outcome after metastasectomy [27, 28].

Cytoreductive nephrectomy: A paradigm for the study of metastasectomy

While there have been no randomized trials evaluating metastasectomy for RCC, studies of cytoreductive nephrectomy demonstrate the feasibility randomizing patients to surgery in a similar population: two studies published independently in 2001 by the SWOG and EORTC cooperative groups demonstrated an overall survival benefit to cytoreductive nephrectomy followed by interferon α-2b versus interferon alone [32, 33]. A combined analysis of the studies found a 31% reduction in the risk of death associated with surgery [34] and cytoreductive nephrectomy has been incorportated into the standard of care. Since that time, however, systemic therepy for mRCC has shifted away from interferon to targeted therapies such as sunitinib, introducing uncertainty regarding the role of cytoreductive nephrectomy. To address this question, the CARMENA study randomized patients to standard of care cytoredutive nephrectomy followed by sunitinib vs sunitinib alone. The study found that sunitinib alone was non-inferior to cytoreductive nephrectomy followed by sunitinib [35]. The SWOG, EORTC and CARMENA studies of cytoreductive nephrectomy serve as a paradigm for potential randomized studies of metastasectomy in the appropriate setting.

Patient selection

Although disease-free interval, number of metastases and other factors have been associated with favorable outcomes after metastasectomy, there are no clear guidelines for selecting which patients will benefit from surgery. The MSKCC and IMDC instruments provide prognostic information for patients with mRCC based on readily available clinical data [36, 37]. In some of the studies covered by this retrospective review, MSKCC or IMDC risk group was prognostic for outcome following metastasectomy. However, these data are prognostic, and are not predictive for benefit from surgery. In recent years, the molecular characteristics of cancers have been added to these clinical features for both prognosis and prediction of benefit from particular therapies. In order to identify a molecular signature predictive of favorable outcome after metastasectomy in mRCC, Verbiest et al. performed a retrospective study of molecular tumor subtypes 43 mRCC patients who underwent complete metastasectomy, [38]. Four molecular subtypes in the metastatic setting were identified that differed in terms of mRNA expression, methylation status, mutation profile, cytogenetic abnormalities, and immune infiltration [39, 40]. The subtypes (ccrcc1–4) differed in their OS and also response rate and PFS on the angiogenesis inhibitors sunitinib and pazopanib. Patients with ccrcc1 or ccrcc4 tumors were at a higher risk of relapse after complete metastasectomy, whereas patients with a ccrcc2 or ccrcc3 tumor usually experienced a long DFS [38]. As our understanding of the molecular features of RCC improves, there is the potential for additional predictive models for metastasectomy in addition to particular systemic therapies.

Perioperative systemic therapy and metastasectomy

Adjuvant systemic therapy after nephrectomy for non-metastatic RCC has been studied in a number of randomized trials over the past 3 decades, and sunitinib was FDA approved for high risk, non-metastatic RCC based on the results of the S-TRAC study [41]. The optimal adjuvant therapy following nephrectomy remains under evaluation [42], and randomized studies of several anti-PD–1 checkpoint inhibitors are underway. An analogous question regarding the benefit of systemic therapy arises after complete metastasectomy with no evidence of disease (NED) post-operatively. A number of completed and ongoing studies have examined this question. The Cytokine Working Group conducted a randomized study of a single cycle of high dose interleukin–2 vs. observation for high risk patients after nephrectomy for RCC [43]. The study also included a separate cohort of 25 patients who were status post metastasectomy with NED. The study did not meet its endpoint of improvement in disease-free survival. The randomized phase II RESORT trial assigned patients who were NED after metastasectomy to 52 weeks of sorafenib treatment versus placebo [44]. The study was closed before planned accrual was complete, and the primary endpoint of improvement in recurrence-free survival with sorafenib was not reached. More than half the patients were free of recurrent disease at 24 months; while not randomized to no surgery, these prospectively enrolled and monitored subjects provide a unique high-quality data set regarding outcomes after metastasectomy. The integration of systemic therapy and metastasectomy was also addressed by Daliani, Jonasch et al., who published a prospective study of patients who underwent metastasectomy after systemic therapy (generally a cytokine based treatment with or with 5–fluorouracil) and were followed by post-operative systemic therapy of a similar nature [45]. Median PFS was 21.6 months, and median OS was 56.4 months.

There are additional randomized trials currently underway, one of which is studying adjuvant pazopanib versus placebo (NCT01575548) and another is a phase II trial of adjuvant sunitinib following resection of pulmonary metastases (NCT01216371). Ongoing adjuvant studies of PD–1 and PD–L1 antibody checkpoint inhibitors are including patients with completely resected small volume synchronous metastatic disease (pembrolizumab–NCT03142334, atezolizumab-NCT03024996). Pre-clinical data support the hypothesis that a minimum volume of disease is required for maximum anti–PD–1 checkpoint inhibitor activity [46]. This paradigm is being evaluated in the PROSPER-RCC randomized study of nivolumab prior to and following surgery for high risk disease (NCT03055013).