Chapter 35: Renal Cell Carcinoma

The American Cancer Society predicts that in 2015 there will be over 64,000 new cases of renal neoplasms in the United States and that 14,000 patients will die as a consequence of disease progression (¹). Renal cell carcinoma (RCC) is the most common histology found in kidney tumors, with clear cell RCC (ccRCC) being the most common histologic subtype (Fig. 35-1). Non–clear cell RCC (nccRCC) subtypes include chromophobe, papillary, oncocytoma, collecting duct carcinoma (CDC), renal medullary, translocation, and unclassified RCC.

Work by Chow et al (¹) found the worldwide incidence of RCC appearing to plateau after a steady increase over several decades. To examine the incidence in the United States, the group used the database of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program to track patients with a diagnosis of kidney cancer from 1977 to 2006. The rate of localized cancer detection has continued to increase, whereas the rates of regional, metastatic, and unstaged RCC have declined. In this work, the mortality rate associated with RCC appears to begin declining in the early 2000s across both gender and racial lines. Although the direct causal relationship for the decline in RCC mortality in this study is unclear, early detection in the era of computed tomography (CT) imaging may be contributing to this finding.

The American Joint Commission on Cancer staging schema for RCC was updated in 2010. Major staging categories are as follows: stage 1: T1 tumors that are 7 cm in maximum diameter or less and are confined to the kidney; stage 2: T2 tumors that exceed 7 cm in diameter but are confined to the kidney; stage 3: T3 tumors that demonstrate extracapsular invasion into the perinephric adipose tissue or renal sinus or extend into the renal vein or inferior vena cava (stage 3 also includes tumors with regional lymph node metastasis); and stage 4: extension of the primary tumor into the ipsilateral adrenal gland or beyond Gerota’s fascia or distant metastases (²).

The link between germline genetic mutations and the development of RCC is well established and applies to a small but biologically important subset of RCC cases (³). Although these genetic alterations certainly play an important role in the biology of RCC in both familial and sporadic cases, there are also some environmental factors that contribute to the risk of developing a renal neoplasm. Smoking has long been linked to an increase in the risk of developing RCC, in addition to its association with multiple other malignancies (⁴). As is the case with other malignancies, cessation of smoking can be associated with a diminution in the risk of developing RCC (⁵). Interestingly, this diminution in risk appears to be slower and requires a significantly longer time to reach baseline than that seen with other malignancies, such as lung cancer.

Obesity or increased body mass index (BMI) has also been linked to increased risk of developing RCC. Several studies have demonstrated a higher incidence of obesity or increased BMI in patients with RCC, suggesting an epidemiologic linkage (⁶). As is the case in most retrospective epidemiologic studies, these observations are confounded by other associated variables, including diet, occupational history, and smoking. In a report by Kamat et al., overweight and obese patients had a more favorable prognosis following surgery than did patients with a lower BMI (⁷). Additional possible risk factors identified retrospectively in epidemiologic studies include diabetes, hypertension, and treatment with diuretics; in addition, certain diets have been linked to higher or lower risk than the general population.

The natural history is variable, depending on tumor genetic factors, the general medical condition of the patient, and factors such as angiogenesis and immune response. Patients who undergo nephrectomy for localized disease remain at risk of recurrence for many years and require appropriate counseling and surveillance.

The traditional measures of performance status, tumor stage, and tumor grade each demonstrate good correlation with clinical outcome (⁸). Integration of these parameters and other clinical prognostic variables allows the classification of patients into groups with statistically significant differences in survival. There are a variety of clinical, pathologic, and molecular features that have been proposed and studied as prognostic factors (Table 35-1) (^9,10). As the genomic landscape continues to be more clearly defined in ccRCC and the genetic drivers of metastasis are identified, more refined putative biomarkers are needed to more effectively treat this heterogeneous disease (^11,12).

Clinical Prognostic Variables

In 1999, Motzer et al identified five risk factors (elevated serum corrected calcium level, anemia, lactate dehydrogenase >1.5× the upper limit of normal [ULN], Karnofsky performance score [KPS] <80, and primary tumor in place) that segregated patients into three risk groups (good, zero factors; intermediate, one to two factors; and poor, three or more factors) (¹³). In 2002, an updated version, now referred to as the Memorial Sloan-Kettering Cancer Center (MSKCC) risk criteria (also referred to as the “Motzer criteria”), was published with the original four risk factors, with the fifth being treatment of interferon within 1 year of diagnosis (¹⁴). For clinical trial purposes, the fifth factor is often considered time to initiation of systemic therapy within 1 year from initial diagnosis of RCC. With the introduction of targeted therapy, the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) was formed, and a modified prognostic scoring system was put forth in 2009 and externally validated in 2013 (^15,16). The IMDC prognostic scoring system (also referred to as the “Heng score”) has six risk factors (elevated serum corrected calcium level, anemia, KPS <80, systemic treatment for metastatic disease within 1 year, absolute neutrophil count >ULN, and platelet count >ULN). Similar to the MSKCC scoring system, good risk is defined as zero risk factors, intermediate risk as one to two risk factors, and poor risk as three or more risk factors. Although other systems have been proposed, the MSKCC and IMDC remain the most widely used (Table 35-2).

Non–clear cell histology has been associated with lower response to systemic therapy (¹⁷). In an attempt to improve on the MSKCC clinical prognostic model, we investigated the role of cytokines and angiogenic factors (CAFs) in serum of patients treated with interferon-α (IFN-α) and found elevated baseline levels of interleukin (IL)-5, IL-6, IL-12p40, and vascular endothelial growth factor A (VEGFA) to be independent risk factors associated with inferior survival. Incorporating the CAF model with the MSKCC clinical model improved the concordance index for predicting overall survival (OS). Patients with three or more of the four CAFs or with MSKCC poor-risk status had a median OS of 9 months, compared with a median OS of 32 months for patients with two or fewer CAFs (¹⁸). Similar efforts have been undertaken to develop serum and plasma-based prognostic and predictive biomarkers in individuals who received antiangiogenic therapy (^19,20).

Pathology and Molecular Markers

There are several major histologic subtypes in RCC, including clear cell, papillary, and chromophobe RCC (see Fig. 35-1). Historically, patients with nccRCC variants except chromophobe do poorly once they develop metastatic disease, mainly due to the dearth of effective systemic therapy.

The VHL gene product regulates the hypoxia-induced pathway and is commonly mutated in ccRCC (¹¹). The hypoxia-induced pathway leads to the activation of survival genes that mediate glucose transport, proliferation, angiogenesis, and pH regulation and is also implicated in the other tumor types. In a study of 187 patients undergoing nephrectomy for ccRCC in Japan, mutation or hypermethylation of the VHL gene was found in 58% of tumors, and was associated with significantly improved disease-free and cancer-specific survival in patients with organ-confined tumors (n = 134), but not in those with stage IV disease at the time of nephrectomy (n = 53) (²¹). VHL loss occurs early in the development of clear cell carcinoma (¹²), potentially identifying a subset of patients who do well after surgery.

In ccRCC, loss of the short arm of chromosome 3 was the most common genetic alteration identified in The Cancer Genome Atlas (TCGA) analysis. The most common genetic mutations related to this loss include VHL, SETD2 (a histone methyltransferase), BAP1 (histone deubiquitinase), and PBRM1 (part of chromatin remodeling complex), and all lie on chromosome 3p (¹¹). The University of Texas Southwestern group identified BAP1 mutations to be associated with poor survival, with a median OS of 4.6 years from nephrectomy in patients who present with nonmetastatic disease compared with a median OS of 10.6 years in patients whose tumors harbor PBRM1 mutations. The investigators validated these findings from tumor specimens included in the TCGA analysis (²²).

With the introduction of improved cross-sectional imaging including CT and magnetic resonance imaging (MRI), the detection of renal masses has increased substantially. The ability to characterize renal masses as benign or malignant appearing has also evolved with improved technology and the use of specific contrast-enhancing sequences. The management of small renal masses (<4 cm) will not be extensively covered in this chapter but includes options such as active surveillance, partial nephrectomy (PN), radical nephrectomy (RN), or thermal ablation depending on factors including patient comorbidities and surgical expertise (²³). Current surgical approaches to both PN and RN include open, laparoscopic, and robotic-assisted techniques (^24,25). Loss of nephrons associated with RN compared to PN increases patient risk for development of chronic kidney disease (CKD)(²⁶). In the general population, CKD is associated with increased risk of mortality and cardiovascular disease. Retrospectively, patients treated with RN were found to have higher all-cause mortality and an increased incidence of cardiovascular events (²⁷).

With the introduction of robotic partial nephrectomy, the warm ischemia and suturing times are reduced compared to laparoscopic PN even among experienced laparoscopic surgeons (²⁸). The management algorithm of potential tumors amenable to PN continues to evolve and should take into account the health of the contralateral kidney, familial cancer syndromes leading to multiple renal masses, proximity to the hilum and vascular structures, baseline kidney function, and surgeon experience and volume.

With larger, more locally advanced tumors requiring nephrectomy, the decision to perform a lymph node dissection remains an area of debate. In the European Organization for Research and Treatment of Cancer (EORTC) 30881 trial, 772 patients were randomized between lymph node resection or no resection during nephrectomy procedure. The majority of patients had ≤T2 disease (²⁹). The study found no difference in recurrence-free survival or OS. However, other series have found the presence of lymph node positivity to be directly correlated with increasing T stage. As a result, at the University of Texas MD Anderson Cancer Center (MDACC) when patients have radiographic evidence of lymph node involvement or T stage ≥T3a, a therapeutic or staging lymph node dissection is typically undertaken.

The decision to perform ipsilateral adrenalectomy as part of the standard RN procedure has evolved. In a retrospective review of the literature, O’Malley et al estimated a negative predictive value of 96% when cross-sectional imaging of the ipsilateral adrenal gland was compared to surgical pathology (³⁰). Unless there is evidence of involvement either by direct extension, radiographic evidence of metastatic spread of disease, or concern for involvement during gross inspection during surgery, the ipsilateral adrenal gland is typically spared.

Locally advanced tumors staged as T3b or higher require considerable institutional and surgeon experience in the management of these patients and will not be covered here in detail.

Role of Cytoreductive Nephrectomy in Metastatic Disease

Renal cell carcinoma remains a relatively rare malignancy where addressing the primary tumor in selected patients with metastatic disease leads to an improvement in OS (^31,32). In a pooled analysis of the results of the Southwest Oncology Group (SWOG) 8949 and EORTC 30947 studies, improvement in OS favored the nephrectomy arm, with a median survival of 13.6 months versus 7.8 months for those randomized to the surgical arm (³³). Because the initial studies compared patients treated with IFN-α with or without cytoreductive nephrectomy, the question of benefit of cytoreductive nephrectomy in the era of targeted therapy has reemerged. In a retrospective analysis, the IMDC found that patients treated in the targeted era appear to benefit from cytoreductive nephrectomy if they have three or fewer prognostic risk factors as outlined in Table 35-2, whereas those with four or more factors do not appear to benefit (³⁴). Two ongoing prospective trials, SURTIME (NCT01099423) and CARMENA (NCT0093033), are attempting to address two questions regarding targeted therapy and cytoreductive nephrectomy. The SURTIME trial is investigating whether cytoreductive surgery is better performed before or after first-line sunitinib. The CARMENA trial is an EORTC effort investigating whether cytoreductive nephrectomy is beneficial in patients treated with sunitinib. Based on the available data and our institutional experience, patients at MDACC with synchronous metastatic disease who have a good performance status, a good or intermediate prognosis based on MSKCC or IMDC score, ccRCC, and a sizable renal primary compared to metastatic disease burden are generally offered upfront cytoreductive nephrectomy.

Patients who present with metastatic disease and primary tumor in situ offer an ideal presurgical setting for clinical and translational research studies (³⁵). Retrospectively, patients at MDACC treated with targeted agents prior to cytoreductive nephrectomy compared to upfront cytoreductive nephrectomy experienced similar rates of serious adverse events (³⁶). With the ability to analyze treated tissue paired with time blood samples and clinical and radiographic parameters at defined intervals, presurgical clinical trials offer an opportunity to enhance our understanding of metastatic RCC and determinants of response and resistance to targeted and immunotherapeutic agents.

Role of Metastasectomy

In carefully selected patients, metastasectomy plays an important role in the multidisciplinary treatment of patients with metastatic RCC (mRCC). Retrospective series have found survival benefit in resecting oligometastatic disease from multiple sites. Interpretation of these results is always limited due to the highly selected and retrospective nature of these studies. However, appropriate patient selection, complete resection of disease, and metachronous single site or oligometastatic disease positively impact patient outcomes. In a pooled retrospective analysis from three centers including patients from MDACC, 22 patients who had received targeted therapy underwent consolidative metastasectomy with an acceptable postoperative complication rate. After 108 weeks of follow-up, 11 patients remained disease free, and the median time to resumption of targeted therapy was 55 weeks (³⁷). An ongoing randomized phase II study (NCT01575548) is addressing the role of pazopanib versus placebo in patients with no evidence of disease after metastasectomy. Further treatment decisions are based on the timing of recurrence, sites of recurrence, and individual patient characteristics. At MDACC, we strongly believe a multidisciplinary approach at high-volume centers is required to optimally manage these patients.

An algorithm to select candidates for metastasectomy and for the approach to postmetastasectomy systemic treatment is shown in Fig. 35-2.

Cytokine Therapy

Until the introduction of targeted therapy (discussed in more detail in the later section on targeted therapy), metastatic resection and cytokine therapy remained the only viable treatment options in mRCC. Interferon-α and IL-2 have been extensively evaluated over the past three decades. In a randomized phase II study from MDACC, treatment with low-dose IFN-α-2b compared to intermediate-dose IFN-α-2b resulted in no significant differences in progression-free survival (PFS) or OS, but patients had an improved quality of life while on low-dose therapy (³⁸). Sustained complete remissions with IFN-α are rare (1%-2%), and with the relatively low response rate of approximately 7%, IFN-α monotherapy never received US Food and Drug Administration (FDA) approval for treatment of mRCC and is no longer considered as a single agent for mRCC at MDACC.

High-dose IL-2 (HD IL-2) received FDA approval in the treatment of mRCC in 1992 based on the results of several phase II trials that found an overall response rate of 15% to 20% and a durable response in the majority of patients who achieved a complete response. A phase III trial comparing HD IL-2 with an outpatient low-dose IL-2 plus IFN-α regimen yielded a higher response rate to HD IL-2 of 23% versus 10% but no statistically significant differences in PFS and OS. However, durable responses of greater than 3 years were seen in 7% of those treated with HD IL-2 versus 0% of patients in the other arm, supporting HD IL-2 as a viable standard-of-care option for patients who are candidates for this therapy (³⁹). Although the management of toxicity related to the delivery of HD IL-2 is challenging, high-volume centers have published a toxic death rate of less than 1%, with response rates of roughly 15% to 20% and durable remissions in the range of 5% to 7% (^40,41). Given the ability to produce a sustained remission with likely cure in a small subset of patients, at MDACC, we offer frontline HD IL-2 therapy to patients with excellent performance status, previous nephrectomy, limited or no comorbidities, and low-volume metastatic disease burden, especially lung-only disease.

Targeted Therapy

Over the last decade, seven agents have been FDA approved for the treatment of mRCC. Of these agents, five are vascular endothelial growth factor (VEGF) pathway–blocking agents, with four being within the small-molecule tyrosine kinase inhibitor family and one, bevacizumab, being a monoclonal antibody directed against VEGF on the cell surface. The remaining two agents are mammalian target of rapamycin (mTOR) inhibitors. These agents target both the tumor cells and the tumor microenvironment.

Vascular Endothelial Growth Factor, or Vascular Endothelial Growth Factor Receptor–Targeted Agents

The presence of a VHL mutation in up to 80% of patients with ccRCC and the resultant increased production of angiogenic factors has made this axis the most exploited treatment target. A number of agents that target either VEGF or its receptors (VEGFR) were FDA approved in the past 10 years (Table 35-3).

Bevacizumab is a humanized recombinant anti-VEGF antibody. Two large randomized phase III studies demonstrated an improved PFS in patients who received a combination of bevacizumab plus IFN-α, when compared to IFN-α alone in patients with mRCC who had not received prior systemic therapy (^42,43). The AVOREN study compared patients treated with the combination of bevacizumab plus IFN versus placebo plus IFN and found that the bevacizumab combination resulted in a superior median PFS of 10.2 months compared to 5.4 months. The Cancer and Leukemia Group B (CALGB) 90206 study yielded a median PFS of 8.5 months for the combination arm versus 5.2 months for IFN monotherapy, a difference that was statistically significant. Of note, the OS for both studies was numerically superior in the bevacizumab-containing arm but did not reach statistical significance. Bevacizumab plus IFN was FDA approved for advanced RCC in 2009.

Sorafenib is an orally bioavailable small-molecule inhibitor of VEGFR, platelet-derived growth factor receptor (PDGFR), and Raf and was FDA approved in 2005 for the treatment of advanced RCC. In a randomized phase III trial, sorafenib yielded a significant improvement in PFS compared to placebo (median, 5.5 vs 2.8 months) in patients who had failed one prior therapy, with the majority having received prior cytokines (⁴⁴). A phase II study compared sorafenib monotherapy to sorafenib plus IFN-α and found similar overall response rates, toxicity, and median PFS, leading further exploration of this combination to be abandoned (⁴⁵).

Sunitinib is an oral inhibitor of VEGFR and PDGFR. When it was first tested in the cytokine-refractory population, a time to progression of 8.7 months was achieved (⁴⁶), earning it provisional FDA approval in early 2006. A follow-up first-line phase III trial randomizing patients between sunitinib and IFN showed a median PFS of 11.0 months for sunitinib versus 5.0 months for IFN (⁴⁷). Median OS of sunitinib-treated patients was 26.4 months versus 21.8 months for the IFN-treated group (P = .051), which although not statistically significant was likely due to crossover from IFN to sunitinib and salvage therapy after protocol (⁴⁸). When a subgroup analysis of the individuals who did not receive any subsequent therapy was performed, it was found that those who received sunitinib had a median OS of 28 months versus 14 months for the IFN group.

Pazopanib is an oral small molecule with a lower half-maximal inhibitory concentration (IC₅₀) for VEGFR when compared to sorafenib and sunitinib (⁴⁹). A randomized, phase III trial compared pazopanib to placebo in patients treated with a cytokine and previously untreated patients and demonstrated an overall PFS of 9.2 versus 4.2 months (hazard ratio, 0.46; 95% confidence interval, 0.34-0.62; P < .0000001 (⁵⁰). Pazopanib was FDA approved in 2009 for the first-line treatment of advanced RCC. Pazopanib has not been evaluated prospectively after first-line targeted therapy. At MDACC, we retrospectively studied 93 patients who received targeted therapy preceding pazopanib and observed an overall response rate of 15%, a median PFS of 6.5 months, and a median OS of 18.1 months (⁵¹). Adverse events were similar to those described in the large prospective trials with the agent.

Axitinib is a selective small-molecular inhibitor of the VEGFR-1, VEGFR-2, and VEGFR-3 signaling pathways (⁵²). In a randomized phase III trial, axitinib was compared to sorafenib as second-line therapy after sunitinib, cytokines, bevacizumab plus IFN, or temsirolimus. The overall median PFS time was 6.7 versus 4.7 months in favor of axitinib (⁵³). Median OS was not statistically different between the two groups (20.1 months for axitinib and 19.2 months for sorafenib) (⁵⁴). The largest difference between the two arms was in patients previously treated with cytokines. Patients who developed hypertension with a diastolic blood pressure of greater than 90 mm Hg on trial had a significantly improved survival in both treatment arms as compared to patients who remained normotensive. When axitinib was compared with sorafenib in the first-line setting, the median PFS was 10.1 months with axitinib and 6.5 months with sorafenib with overlapping confidence intervals. Despite the numerically longer PFS with axitinib, the difference did not reach the trial’s prespecified threshold of statistical significance (⁵⁵).

Mammalian Target of Rapamycin Inhibitors

Temsirolimus is an intravenous sirolimus ester that was tested in intermediate- and poor-risk patients after phase II data had suggested that this was the subgroup of patients most likely to benefit from temsirolimus (⁵⁶). This first-line phase III trial randomized patients to temsirolimus, temsirolimus plus IFN, or IFN monotherapy. The median OS in the temsirolimus arm was 10.9 months versus 8.4 months for the combination arm and 7.3 months for IFN monotherapy (⁵⁷). An elevated pretreatment serum lactate dehydrogenase was found to be both a predictive and prognostic biomarker in patients treated with temsirolimus compared to IFN (OS, 6.9 vs 4.2 months, P < .002) (⁵⁸).

Everolimus is an orally bioavailable sirolimus ester that was evaluated in patients who had progressed on sorafenib, sunitinib, or both. This phase III trial (RECORD-1) randomized patients between everolimus 10 mg by mouth daily versus placebo in a 2:1 fashion. Median PFS was 4.0 months with everolimus versus 1.9 months with placebo (⁵⁹).

Because everolimus and temsirolimus are mTORC1 inhibitors, agents capable of blocking both the mTORC1 and mTORC2 pathways have been developed. In a recent study comparing everolimus with GDC-0980, a pan PI3K and TORC1/2 inhibitor, the median PFS was 6.1 months with everolimus versus 3.7 months with the investigational agent, with a higher rate of adverse events in the investigational arm (⁶⁰). The reasons for this surprising result are unclear at this time.

Side Effect Profiles

The major side effects of the targeted agents are summarized in Table 35-4.

Blockade of VEGFRs or depletion of the VEGF ligand causes class effects, which include hypertension, fatigue, proteinuria, and a slightly increased risk of bleeding and thromboembolic events. All of the anti-VEGF agents will cause these side effects to varying degrees. Sorafenib, sunitinib, and pazopanib are also inhibitors of PDGF, Flt3, and a number of other receptor tyrosine kinases and thus will induce hand-foot syndrome, diarrhea, and dysgeusia. Sunitinib appears to be particularly prone to inducing hypothyroidism (^61,62) and, in some cases, cardiomyopathy with decreased ejection fraction (⁶³). The frequency of cardiac toxicity may be underappreciated. A single-institution experience from Stanford found that 33% of patients had evidence of cardiac toxicity even when excluding hypertension (⁶⁴). The rate of heart failure induced by sunitinib at MDACC in the treatment of a variety of tumor types including mRCC was 2.7% (⁶⁵). Emerging data suggest that scrupulous control of blood pressure in patients on sunitinib will decrease cardiac stress and resultant cardiac failure.

Other side effects include wound dehiscence and increased thromboembolic events. Our group reported perioperative complications in patients with mRCC and primary in situ who received antiangiogenic agents in the presurgical setting prior to cytoreductive nephrectomy. In a single-arm phase II trial (³⁵), perioperative wound healing complications in patients who received 8 weeks of bevacizumab therapy were increased when compared to a set of matched controls. The retrospective study, which looked at 44 individuals who received sorafenib, sunitinib, or bevacizumab preoperatively, did not detect a significant elevation of any perioperative events (⁶⁶).

The mTOR inhibitors demonstrate several unique side effects, including hyperglycemia, hypertriglyceridemia, and noninfectious pneumonitis. However, these agents are not as likely to induce hypertension or hand-foot skin reaction.

Agent Selection and Side Effect Management

A number of different factors are important when choosing the best initial agent for patients with metastatic ccRCC. Clearly, evidence-based criteria for selection are vital, but as seen in Fig. 35-3, there are several agents available for a particular treatment stage.

For example, if we are dealing with a patient who has undergone nephrectomy, has good-risk characteristics, and is younger than age 70 years, HD IL-2, sunitinib, bevacizumab plus IFN, and pazopanib are all reasonable options. In the COMPARZ trial, a noninferiority phase III trial with PFS as the primary end point, sunitinib and pazopanib were directly compared (⁶⁷). The trial showed noninferiority of pazopanib. Importantly, the side effect profile and quality-of-life assessment favored pazopanib. Consequently, we prefer pazopanib over the standard dosing schedule of sunitinib and over bevacizumab plus IFN given the relative ease of patient administration. Although HD IL-2 is often offered, many patients are not interested to receive it because of the low likelihood of success and the formidable toxicity.

Table 35-4 provides a summary of some of the most common side effects seen with these agents. A review of these side effects may aid in treatment selection for specific patients. Some may have occupational considerations that make hand-foot skin reaction a particular problem. Others may have cardiac comorbidities that make an agent like sunitinib, with a known effect on cardiac output in a subset of patients, less favorable. Table 35-5 summarizes the standard dose-reduction algorithms used in clinical trials of these agents.

In addition to dose interruptions and dose reductions, a change in schedule may be beneficial for some patients. Our group and others have retrospectively compared cohorts of patients treated on the standard sunitinib schedule of 4 weeks on, 2 weeks off with a schedule of 2 weeks on and 1 week off (⁶⁸). The 2-week-on, 1-week-off schedule has a better toxicity profile and retrospectively has similar treatment efficacy compared with the standard approach (⁶⁹). However, prospective comparison of the two treatment schedules has not been reported.

Symptom control is essential for patients receiving molecularly targeted agents. Table 35-6 outlines supportive care measures that can mitigate or prevent specific side effects. It is essential that the patient and health-care team maintain an ongoing dialogue during each cycle of therapy to ensure that patients proactively and appropriately manage these adverse events. Successful adverse event management will translate into higher drug compliance and a greater probability of achieving a successful outcome.

A retrospective review of patients treated with temsirolimus or everolimus at MDACC found a rate of noninfectious pneumonitis (NIP) of 6% with temsirolimus and 23% with everolimus (⁷⁰). We observed that patients who developed NIP had longer treatment duration and longer OS. Recent guidelines have been established to diagnose and treat NIP with the important caveats that patients with radiographic findings but lacking symptoms can continue therapy and patients with radiographic findings with mild to moderate cough can continue therapy with close monitoring (⁷¹). More severe symptoms require corticosteroids with a consideration of break or permanent discontinuation of mTOR inhibitor therapy depending on the severity of the event.

The decision for second-line and later therapy has become increasingly complex. Sequencing of VEGF-VEGF–directed therapy or VEGF-mTOR–directed therapy has been compared prospectively. In a randomized phase III trial of second-line therapy in patients who received first-line sunitinib, temsirolimus was compared to sorafenib. The trial’s primary end point of detecting a superior PFS with temsirolimus was not met, with a median PFS of 4.1 months with temsirolimus and 3.7 months with sorafenib. Interestingly, the secondary end point of median OS favored the sorafenib arm (16.6 vs 12.3 months, P = .01). Although median OS was not the primary end point of the trial, the authors hypothesized that VEGF-VEGF sequencing may be more beneficial compared with the so-called “sandwich approach” of VEGF-mTOR-VEGF sequence (⁷²). In the community, VEGF-mTOR-VEGF sequencing is the most common current treatment approach (⁷³). In the AXIS trial, axitinib was compared to sorafenib in the second-line setting (although first-line treatment could be cytokine-based treatment) and led to an improved PFS (⁵³). A phase II, single-arm, single-institution study found that after failure of frontline sunitinib or bevacizumab, pazopanib had activity, with a median PFS of 7.5 months (⁷⁴). As such, a reasonable treatment approach for patients with good- or intermediate-risk disease by Heng criteria who are deemed good candidates for targeted therapy and do not elect to pursue HD IL-2 therapy can receive pazopanib as first-line therapy and receive axitinib or everolimus as second-line therapy. Depending on performance status and comorbidities, the decision can be made to continue with an alternative VEGF-directed therapy with sorafenib or sunitinib or change to an mTOR-directed therapy at progression. Certainly, consideration of a clinical trial should be entertained in the frontline and later line settings.

Patients who fall into the poor prognostic group remain a considerable unmet need in the care of patients with mRCC. At this time, temsirolimus remains the only agent with category 1 evidence to support its use. Ongoing trials in this setting include the FLIPPER trial (NCT01521715), which is a phase IV trial evaluating pazopanib in poor-risk mRCC, and a randomized phase II trial at MDACC comparing temsirolimus to pazopanib in the frontline setting (NCT01392183).

Combinations of Targeted Therapy

In an attempt to improve on the success of single-agent targeted therapy, combinations of VEGF and mTOR pathway blockade have been pursued. To date, combinations are associated with significant toxicity without a positive impact on response rate, PFS, or OS, as compared with sequential single-agent targeted therapy. A phase III trial compared the standard of bevacizumab plus IFN-α versus bevacizumab plus temsirolimus (⁷⁵). The response rate, median PFS, and median OS were not found to be statistically different between the two treatment arms. As a result of early phase and later phase clinical trials to date, combinations of VEGF and mTOR inhibitors are unlikely to yield significant clinical benefit.

Vaccine Strategies

Given the rare durable responses to cytokine therapy, investigators have searched for vaccine-based strategies. At this time, no vaccine has proven successful in rigorous phase III testing. Currently, two phase III trials are ongoing. The first trial compares a multipeptide vaccine, IMA-901, paired with sunitinib versus sunitinib alone (NCT01265901). The second trial, the ADAPT phase III trial, involves a dendritic cell–based RNA vaccine, AGS-003, plus standard therapy versus standard therapy alone (NCT01582672).

Immune Checkpoint Inhibitors

A class of agents known as the immune checkpoint inhibitors has been developed allowing the targeting of molecules on both T cells and tumor cells important in downregulating the immune system, allowing for a sustained antitumor response. Agents including the cytotoxic T lymphocyte-associated antigen 4 antibody (anti-CTLA-4; ipilimumab) (⁷⁶) and the programmed death receptor-1 antibody (anti-PD-1; nivolumab and MK-3475) (⁷⁷) and its ligand antibody (anti-PDL-1; MPDL3280A) (⁷⁸) have shown significant, sustained antitumor responses in mRCC (⁷⁹).

Table 35-7 shows the available clinical results for these agents. The VEGF pathway has been found preclinically to play an important role in tumor microenvironment immunosuppression (⁸⁰). At this time, 10 ongoing studies are evaluating a variety of immune checkpoint inhibitors alone, in combination with other checkpoint inhibitors, or in combination with anti-VEGF therapy. Table 35-8 outlines the registered trials with these agents as of December 2014.

The large majority of studies performed to date excluded nccRCC. The exception was the phase III temsirolimus study, where 20% of the 626 patients had non–clear cell histology. A post hoc analysis of papillary RCC showed outcomes similar to ccRCC after treatment with temsirolimus (⁸¹). At MDACC, we conducted a phase II single-arm study of sunitinib in advanced nccRCC histologies including patients who had up to two previous lines of treatment and reported a median PFS of 2.7 months and an overall response rate of 5% (⁸²). Our group recently presented the findings of the ESPN randomized phase II clinical trial comparing everolimus to sunitinib for frontline systemic therapy in patients with metastatic non–clear cell histologies (⁸³). The primary end point was PFS after first-line treatment. The trial was stopped early due to futility of satisfying the primary end point, with a median PFS of 4.1 months with everolimus compared to 6.1 months with sunitinib. Foretinib, a c-MET, RET, and VEGFR-2 inhibitor, was tested with two different dosing schedules in a phase II trial enrolling patients with papillary RCC (⁸⁴). The study found an overall response rate of 13.5% and a median PFS of 9.5 months, and foretinib is no longer being developed in this setting. Of note, a subset of 10 patients with a germline c-MET mutation had a 50% response rate, and alternative c-MET inhibitors should be explored in this setting. An actuarial 45-patient trial evaluating erlotinib in papillary RCC showed a response rate of 11%, and an OS of 27 months (⁸⁵). Unfortunately, choosing the right agent for this relatively heterogeneous group of patients is hampered by a lack of studies and relatively small patient numbers. With the paucity of available data for clinical treatment decisions, we recommend enrolling patients with non–clear cell histologies on clinical trials when available.

Sarcomatoid Renal Cell Carcinoma

Approximately 5% of all patients with RCC will demonstrate sarcomatoid dedifferentiation in their tumors (⁸⁶). Varying percentages of sarcomatoid involvement can be seen, and in cases where there is a predominance of sarcomatoid cells, it is difficult to determine the underlying histologic background. In these extreme cases, epithelial tumor markers are still present on the tumor cells, distinguishing these tumors from sarcomas.

Presence of sarcomatoid features in the tumor portends a poor prognosis (⁸⁷). The results of chemotherapy and immunotherapy have been largely disappointing in patients with metastatic disease. Therapy targeting VEGF has been retrospectively reviewed, and in a series of 43 patients from several institutions, this approach had limited utility, with a median PFS of 5.3 months and median OS of 11.7 months (⁸⁸). At MDACC, we retrospectively reviewed 28 patients with mRCC who received the combination of gemcitabine, capecitabine, and bevacizumab, including 8 patients with sarcomatoid features, and reported a median PFS of 5.9 months and a median OS of 10.4 months. These observations formed the basis of a prospective study at MDACC using this three-drug combination in patients with sarcomatoid histology. Preliminary findings from this study in the the first 18 patients treated (out of a planned 40 patients) revealed time to treatment failure of 5.5 months and median OS of 12 months.

Collecting Duct Carcinoma

Tumors arising from the collecting duct epithelium are located in the medulla or center portion of the kidney, in contrast to RCC tumors, which arise from tubules in the cortex (⁸⁹). The diagnosis of CDC is based on both clinical and histologic features. Sarcomatoid dedifferentiation is not uncommon, and patterns of metastasis are similar to those of a high-grade, rapidly progressive RCC. There is no established effective systemic therapy for metastatic CDC, although marginal benefit is occasionally seen with chemotherapy regimens developed for transitional cell carcinoma (^90,91).

Medullary Carcinoma

Renal medullary carcinoma is a rare and virulent malignancy, afflicting patients with sickle cell hemoglobinopathies, usually sickle cell trait (⁹²). It arises from the caliceal epithelium and can be morphologically distinguished from RCC and CDC. Some investigators have suggested that it is a dedifferentiated form of transitional cell carcinoma. Other kidney disorders associated with sickle cell hemoglobinopathies include unilateral hematuria, papillary necrosis, nephrotic syndrome, renal infarction, inability to concentrate urine, and pyelonephritis. These associated conditions may contribute to the development of renal medullary carcinoma. The majority of cases reported have a highly aggressive clinical course, presenting with metastatic disease and with a survival historically of about 4 months. It is refractory to most forms of systemic therapy (⁹³), with some responses to cytotoxic chemotherapy.

Outcomes of patients with renal medullary carcinoma treated at four institutions found the median OS for all patients to be approximately 1 year from diagnosis (⁹⁴). Of 16 patients who received targeted therapy, one response was noted. An updated effort from seven institutions that evaluated clinical outcomes for 39 patients with renal medullary carcinoma was presented in abstract form at the 2015 Genitourinary Cancer Symposium (⁹⁵). Again, the median OS was 1 year, with the majority of patients presenting with advanced disease. Of seven patients who received presurgical treatment, one patient had a complete response, three had a partial response, and three had stable disease. The patient with the complete response had a near-complete pathologic response at the time of nephrectomy and remains without evidence of disease 27 months after diagnosis. Continued efforts to gain insight into the biology of this devastating entity are ongoing. In our experience, initial systemic therapy with chemotherapy with or without the addition of bevacizumab is a reasonable initial approach.

Xp11.2 Translocation Carcinoma

Xp11.2 translocation carcinoma is a rare disease entity that is proportionally more common in children but can also occur in young adults (⁹⁶). A strong female predominance is observed. The TFE3 gene located at Xp11.2 has a number of potential translocation partners leading to fusion of a novel gene. A recent analysis of 16 patients with translocation RCC found significant genomic heterogeneity, with 17q gain, 9p loss, 3p loss, and 17p loss being the most common findings (⁹⁷). Histologically, these tumors may resemble ccRCC or show mixed clear and papillary features. These histologic findings in a young adult should prompt further histologic and molecular characterization of the tumor. Staining for TFE3 permits relatively specific identification of this disease entity, but if it is highly suspected, a break apart fluorescent in situ hybridization assay should be used to increase sensitivity (^98,99). Response to standard immunotherapy or chemotherapy is poor, and patients have a relatively short OS. A retrospective review of 15 adult patients (12 females; median age, 40 years) with translocation carcinoma who were treated with targeted therapy at four centers in the United States showed an overall response rate of 20%, with a median PFS of 7.1 months and median OS of 14.3 months (¹⁰⁰). A French study reported on 21 patients with adult translocation carcinoma treated with anti-VEGF agents or mTOR inhibitors. The authors reported objective responses, PFS, and OS results in the range of those previously reported for ccRCC (¹⁰¹).

Figure 35-4 succinctly outlines the current approach at MDACC for diagnosing and treating patients with a renal mass. With rapidly improving understanding of the biology of RCC, better patient selection and surgical techniques, an enriched arsenal of targeted agents, and the promise of new immunotherapy agents in late-phase clinical testing, the future of treatment for patients with RCC appears bright. We need to continue to focus on patients in the poor-risk clinical prognostic categories with metastatic clear cell carcinoma and patients with non–clear cell histologies because these represent major areas of unmet clinical need.