In one of the initial studies from the National Cancer Institute, Walther et al. reported on a series of 93 patients undergoing CRN with planned postoperative interleukin-2 therapy [16]. Forty percent of patients were not able to receive postoperative systemic therapy, most commonly due to rapid progression of systemic disease. Preoperative clinical factors and laboratory values were assessed in an attempt to identify factors associated with failure to receive subsequent therapy. The only significant predictor of not receiving subsequent therapy was having a preoperative ECOG PS >1 (p = 0.047).

In an attempt to mitigate the risks of CN, Fallick et al. used strict criteria to select patients for CN [35]. Patients being considered for CN had an ECOG PS of 0 or 1; predominant clear cell histology; greater than 75 % debulking of tumor burden technically feasible; absence of central nervous system, liver, or osseous metastases; and no major comorbid medical conditions. Over a 5-year time period, 85 patients with mRCC with their primary tumor in place were evaluated for CN. Patients in whom pretreatment biopsy revealed non-clear cell predominance were not considered for surgery. Only 33 % (28/85) met the eligibility criteria for CN. By utilizing these selection criteria, the operative outcomes and the ability to receive subsequent systemic therapy (93 %) were improved over prior series. Investigators at the Cleveland Clinic performed an independent analysis of metastatic burden in 46 patients undergoing CN [37]. In this contemporary series of patients undergoing CN, fractional percentage of tumor volume (FPTV) was shown to be associated with survival. In this cohort of patients treated only with targeted therapies, FPTV removed (<90 % versus ≥90 %) and preoperative corrected calcium were independent predictors of progression-free survival (PFS) (Fig. 10.3) [37].

Published selection criteria for the two randomized trials were not as strict as those by Fallick et al. Eligibility for the SWOG 8949 and EORTC 30947 was identical [1–3]. All patients had a pre-randomization biopsy, adequate liver function (bilirubin <3× ULN), adequate renal function (Cr <3.0 mg/dl), ECOG PS of 0 or 1, and no prior malignancy within 5 years. Only 5.6 % of patients were unable to receive postoperative interferon alpha. In a later analysis of the SWOG 8949 data, Lara et al. analyzed predictive variables for OS after CN (Table 10.2) [38]. On multivariate analysis, patients with early progression (<90 days) and patients with an ECOG PS of 1 (versus 0) had significantly worse OS. Of course, early progression is not a preoperative variable, but perhaps identification of patients showing earlier signs of progression (SURTIME) would be desirable and aid in selection of patients for aggressive multimodal treatment through the use of CN.

In 2006, a multidisciplinary panel used available data and the RAND/UCLA Appropriateness Method to develop recommendations regarding optimal patient selection [39]. Patients were classified as “good risk” surgical patients if the Eastern Cooperative Oncology Group (ECOG) performance status was 0 or 1 and major comorbid conditions were absent. Metastatic burden was classified as lung metastases only, limited metastases (low-volume lung or bone disease), or extensive burden (lung and bone metastases or any liver or CNS involvement). Symptoms were defined in relation to the primary tumor. The panel recommendations were as follows: for good surgical risk patients with planned postoperative immunotherapy, nephrectomy was rated appropriate in patients who had limited metastatic burden regardless of symptoms and in symptomatic patients regardless of metastatic burden. With regard to planned targeted therapy, the panel recommended only patients with the most favorable combination of surgical risk, metastatic burden, and symptoms undergo CN. The panel highlighted the limitations in defining the role of CN in patients for whom systemic targeted therapy is planned.

In addition to selection criteria established from single center retrospective series and the two randomized trials, many authors incorporate prognostic factors for OS when deciding appropriateness of CN. Whether these overall prognostic factors can be used to predict for early progression and thus be used as selection criteria for CN is unknown. Although the MSKCC risk stratification is one of the most widely accepted and validated set of prognostic factors in mRCC, this stratification system was not intended for use as a selection criterion for performing CN, but rather was established to provide prognostication for patients undergoing systemic therapy alone or after CN [40–42]. Utilizing these prognostic variables, patients are further categorized into favorable risk (0 risk factors), intermediate risk (1–2 risk factors), or poor risk (≥3 risk factors). A poor prognostic variable in the initial report was absence of a nephrectomy (presence of the primary tumor) [42]. Due to the rapid adoption of CN after the SWOG 8949 and EORTC 30947 publications, the subsequent MSKCC criteria replaced this variable with “time from diagnosis to treatment of less than 12 months” (Table 10.3) [41]. The original as well as subsequent modified risk stratification systems have also been shown to be useful for prognostication in contemporary cohorts of mRCC patients receiving targeted therapy [43–45].

Several of these validated prognostic factors for OS after systemic therapy were either previously incorporated into patient selection criteria for CN or have been subsequently analyzed. Given the significant morbidity of CN, the indiscriminant use of CN is not advisable [22]. Kutikov et al. reported on the outcomes of 141 patients after CN treated between 1990 and 2008 [46]. Of those not receiving subsequent systemic therapy (30.5 %: 43/141), the most common reason was rapid disease progression (30.2 %). Patients not receiving systemic therapy had a trend toward lower survival although this was not statistically significant (p = 0.16). The risk of death after surgery correlated with the number of metastatic sites (p = 0.012), symptoms at presentation (p = 0.001), poor performance status (p = 0.001), high tumor grade (p = 0.006), and the presence of sarcomatoid features (p < 0.024).

Although there are significant practice variations among high-volume centers, the selection of patients for CN is based on a combination of prognostic factors for OS and predictors of surgical outcome after CN. In one of the largest series of its kind, Culp et al. attempted to identify preoperative clinical variables in a cohort of 566 patients undergoing CN and 115 receiving systemic therapy alone at the MD Anderson Cancer Center over a 15-year period (1991–2007) [31]. An extensive list of preoperative variables was analyzed which resulted in the identification of seven preoperative variables found to be significant negative predictors of overall survival (Table 10.4). The number of preoperative risk factors was correlated with OS and was inversely proportional to the median survival of patients who underwent CN. Patients who underwent CN with >3 preoperative risk factors did not appear to benefit from CN when compared to patients undergoing medical therapy alone (Fig. 10.4). Sarcomatoid dedifferentiation and Fuhrman grade 3 or 4 were also significant factors for OS, but in most cases these were not available preoperatively and thus were not included in the analysis of preoperative factors.