Incidence

RCC, which accounts for 2% to 3% of all adult malignant neoplasms, is the most lethal of the common urologic cancers. Traditionally, 30% to 40% of patients with RCC have died of their cancer, in contrast to the 20% mortality rates associated with prostate and bladder carcinomas (Landis et al, 1999; Pantuck et al, 2001b). Approximately 54,000 new diagnoses of RCC are made each year in the United States, and 13,000 patients die of disease (Russo et al, 2008, Carrizosa and Godley, 2009; Jemal et al, 2009). Overall, approximately 12 new cases are diagnosed per 100,000 population per year, with a male-to-female predominance of 3 : 2 (Landis et al, 1999; Wallen et al, 2007; DeCastro and McKiernon, 2008; Woldrich et al, 2008; Carrizosa and Godley, 2009). This is primarily a disease of the elderly patient, with typical presentation in the sixth and seventh decades of life (Pantuck et al, 2001b; Wallen et al, 2007). Incidence rates are 10% to 20% higher in African Americans for unknown reasons (Chow et al, 1999, Lipworth et al, 2006; Stafford et al, 2008). The majority of cases of RCC are believed to be sporadic; only 2% to 3% are familial (Lipworth et al, 2006).

The incidence of RCC has increased since the 1970s by an average of 3% per year for whites and 4% per year for African-Americans, largely related to the more prevalent use of ultrasonography and CT for the evaluation of a variety of abdominal complaints (Chow et al, 1999; DeCastro and McKiernon, 2008; Kummerlin et al, 2008b). This trend has correlated with an increased proportion of incidentally discovered and localized tumors and with improved 5-year survival rates for patients with this stage of disease (Konnak and Grossman, 1985; Thompson and Peek, 1988; Kessler et al, 1994; Pantuck et al, 2001b; Parsons et al, 2001, Kane et al, 2008). However, other factors must also be at play because Chow and colleagues (1999) have documented a steadily increasing mortality rate from RCC per unit population since the 1980s, and this was observed in all ethnic and both sex groups. They reported that the incidence of advanced tumors per unit population has also increased; and although the proportion of advanced tumors has decreased, the mortality rate per unit population has still been negatively affected (Chow et al, 1999; Hock et al, 2002; Wallen et al, 2007; DeCastro and McKiernon, 2008). This suggests that a deleterious change in tumor biology may have occurred during the past several decades, perhaps related to tobacco use, dietary factors, or exposure to other carcinogens (Chow et al, 1999; Pantuck et al, 2001b; Hock et al, 2002; Parsons et al, 2002; Kane et al, 2008).

RCC in childhood is uncommon, representing only 2.3% to 6.6% of all renal tumors in children (Castellanos et al, 1974; Chan et al, 1983; Freedman et al, 1996; Asanuma et al, 1999; Broecker, 2000). Mean age at presentation in children is 8 to 9 years, and the incidence is similar in boys and in girls. Although Wilms tumor is much more common in younger children, RCC is as common as Wilms tumor during the second decade of life. RCC in children and young adults is more likely to be symptomatic and to exhibit papillary histology, and a predilection for locally advanced, high-grade disease, and unfavorable histologic subtypes has also been reported (Freedman et al, 1996; Renshaw et al, 1999; Sánchez-Ortiz et al, 2004b; Cook et al, 2006; Estrada et al, 2005). TFE3 protein overexpression, which correlates with the presence of ASPL-TFE3 and PRCC-TFE3 gene translocation events involving the X and first chromosomes, is relatively common in children and young adults with RCC and is unique to this population (Heimann et al, 2001; Geller et al, 2008). The clinical significance of TFE3 protein overexpression is not well defined, although preliminary data suggest that these tumors may show differential sensitivity to certain chemotherapeutic agents (Argani et al, 2002; Heimann et al, 2001; Perot et al, 2003; Bruder et al, 2004). Most studies suggest that stage for stage, children and young adults with RCC may respond better to surgical therapy, and a number of long-term survivors have been reported after radical nephrectomy and lymphadenectomy for lymph node–positive disease (Freedman et al, 1996; Asanuma et al, 1999; Abou El Fettouh et al, 2002; Geller et al, 2008; Sánchez-Ortiz et al, 2004b; Geller et al, 2008). An aggressive surgical approach with formal lymphadenectomy has thus been recommended at the time of radical nephrectomy when RCC is suspected in children or young adults (Freedman et al, 1996; Asanuma et al, 1999; Geller et al, 2008; Selle et al, 2006; Bosquet et al, 2008).

Etiology

RCCs were traditionally thought to arise primarily from the proximal convoluted tubules, and this is probably true for the clear cell and papillary variants. However, it is now established that other histologic subtypes of RCC, such as chromophobe and collecting duct RCC, are derived from the more distal components of the nephron (Störkel, 1996; Oyasu, 1998; Pantuck et al, 2001a).

The most generally accepted environmental risk factor for RCC is tobacco exposure, although the relative associated risks have been modest, ranging from 1.4 to 2.5 compared with controls (Table 49–5). All forms of tobacco use have been implicated, and risk increases with cumulative dose or pack-years (Kantor, 1977; La Vecchia et al, 1990; McLaughlin et al, 1995; McLaughlin and Lipworth, 2000; Moyad, 2001; Dhote et al, 2004; Lindblad, 2004; Lipworth et al, 2006; Carrizosa and Godley, 2009). Relative risk is directly related to duration of smoking and begins to fall after cessation, further supporting a cause-and-effect relationship (La Vecchia et al, 1990; McLaughlin et al, 1995; Parker et al, 2003b). Tobacco use accounts for 20% to 30% of cases of RCC in men and 10% to 20% in women (McLaughlin et al, 1995; McLaughlin and Lipworth, 2000).

Table 49–5 Etiology and Environmental Factors of Malignant Renal Tumors

Established

Tobacco exposure Obesity Hypertension

Putative

Lead compounds Various chemicals (e.g., aromatic hydrocarbons) Trichloroethylene exposure Occupational exposure (metal, chemical, rubber, and printing industries) Asbestos or cadmium exposure Radiation therapy Dietary (high fat/protein and low fruits/vegetables)

Obesity is now accepted as another major risk factor for RCC with an increased relative risk of 1.07 for each unit of rising body mass index (Chow et al, 2000; Bergstrom et al, 2001; Bjorge et al, 2004; Calle and Kaaks, 2004; Reeves et al, 2007). The increased prevalence of obesity likely contributes to the increased incidence of RCC in Western countries, and it has been estimated that more than 40% of cases of RCC in the United States may be causally linked to obesity (Calle and Kaaks, 2004). Potential mechanisms linking obesity to RCC include lipid peroxidation leading to DNA adducts, increased insulin-like growth factor-1 expression, increased circulating estrogen levels, and increased arterionephrosclerosis and local inflammation (Kasiske et al, 1992; Huang et al, 1998; Gago-Dominguez et al, 2002; Calle and Kaaks, 2004).

Hypertension appears to be the third major etiologic factor for RCC. Diuretics and other antihypertensive medications have also been implicated, but the weight of the epidemiologic evidence suggests that it is the underlying disorder, hypertension, rather than the treatment, that increases the risk of RCC (McLaughlin et al, 1995; Yuan et al, 1998; McLaughlin and Lipworth, 2000; Lipworth et al, 2006). The proposed mechanisms are hypertension-induced renal injury and inflammation or metabolic or functional changes in the renal tubules that may increase susceptibility to carcinogens (Lipworth et al, 2006).

Although a number of other potential etiologic factors have been identified in animal models, including viruses, lead compounds, and more than 100 chemicals such as aromatic hydrocarbons, no specific agent has been definitively established as causative in human RCC (Bennington and Beckwith, 1947; Kantor, 1977). The potential role of trichloroethylene exposure has been actively investigated; some studies showed relative risks ranging from twofold to sixfold, but others have argued that inherent biases likely account for these results (Vamvakas et al, 2000; Mandel, 2001; Moyad, 2001; Bruning et al, 2003, Lipworth et al, 2006; Lock and Reed, 2006). However, Brauch and colleagues (1999) reported an increased incidence and unique pattern of von Hippel-Lindau gene (VHL) mutations in this population, which would argue in favor of a potential causative role for this compound. Slightly increased relative risks for RCC have been reported for workers in the metal, chemical, rubber, and printing industries and those exposed to asbestos or cadmium, but the data are not particularly convincing (Kolonel, 1976; Pesch et al, 2000; Moyad, 2001; Hu et al, 2002; Dhote et al, 2004; Lindblad, 2004; Lipworth et al, 2006; Carrizosa and Godley, 2009).

Case-control studies have shown that RCC is more common among individuals with low socioeconomic status and urban background, although the causative factors have not been defined (Kantor, 1977; Goodman et al, 1986; Muscat et al, 1995; Yuan et al, 1998). The typical modern Western diet (high in fat and protein and low in fruits and vegetables), increased intake of dairy products, and increased consumption of coffee or tea have been associated with RCC, but the relative risks have been modest, and conflicting data are available in most instances (Yu et al, 1986; Lindblad et al, 1997; Moyad, 2001; Handa and Kreiger, 2002; Dhote et al, 2004; Lindblad, 2004; Murai and Oya, 2004, Lipworth et al, 2006; Wolk et al, 2006; Carrizosa and Godley, 2009). A family history of RCC may also be a factor; one study showed a relative risk of 2.9 for individuals with a first- or second-degree relative with RCC (Gago-Dominguez et al, 2001).

Other potential iatrogenic causes include Thorotrast (which was used as a contrast agent in the past), and radiation therapy, but, again, the relative risks are low (Wenz, 1967; Romanenko et al, 2000). Vogelzang and colleagues (1998) reported four cases of RCC developing in a previously irradiated field, and a slightly increased incidence of RCC has been reported in men who received retroperitoneal irradiation for the treatment of testicular cancer. Survivors of childhood Wilms tumor also appear to be at increased risk for RCC, possibly related to prior radiation therapy or chemotherapy (Cherullo et al, 2001). An increased incidence of RCC is also observed in patients with end-stage renal failure and certain familial syndromes such as tuberous sclerosis, as discussed later (Ishikawa et al, 1990; Bjornsson et al, 1996; Neumann and Zbar, 1997).

Familial Renal Cell Carcinoma and Molecular Genetics

Since the early 1990s, significant advances have been made in our understanding of the molecular genetics of RCC. Novel familial syndromes of RCC have been identified, and the tumor suppressor genes and oncogenes contributing to the development of both sporadic and familial forms of this malignancy have been characterized (Table 49–6) (Linehan et al, 1995; Zbar et al, 1995; Schmidt et al, 1997; Weirich et al, 1998; Choyke et al, 2003; Linehan et al, 2003; Pavlovich et al, 2003; Zimmer and Iliopoulos, 2003; Pavlovich and Schmidt, 2004; Klatte and Pantuck, 2008; Nathanson and Stephenson, 2009). The impact of this new information should not be underestimated because it has fundamentally changed our perceptions about RCC. We now, more than ever, recognize the distinct nature of the various histologic subtypes of RCC, and advances in molecular genetics have contributed to a major revision in the histologic classification of this malignant neoplasm (Oyasu, 1998; Linehan et al, 2003; Young et al, 2008; Roma and Zhou, 2007; Pfaffenroth and Linehan, 2008; Zhou, 2009). A direct and beneficial impact on management of patients has also been achieved, with molecular targeted agents now extending survival for patients with advanced RCC (Linehan, 2002; Vira et al, 2007; Hudes et al, 2007, Motzer et al, 2007; Lane et al, 2007c; Kroog and Motzer, 2008; Clark and Cookson, 2008).

Table 49–6 Familial Renal Cell Carcinoma (RCC) Syndromes

SYNDROME	GENETIC ELEMENT	MAJOR CLINICAL MANIFESTATIONS
von Hippel-Lindau	VHL gene (chromosome 3p25-26)	Clear cell RCC Hemangioblastomas of the central nervous system Retinal angiomas Pheochromocytoma
Hereditary papillary RCC	c-MET proto-oncogene (chromosome 7q31)	Type 1 papillary RCC
Familial leiomyomatosis and RCC	Fumarate hydratase (chromosome 1q42)	Type 2 papillary RCC Cutaneous leiomyomas Uterine leiomyomas
Birt-Hogg-Dubé	BHD1 gene (chromosome 17p12q11)	Chromophobe RCC Oncocytoma Transitional tumors* Occasional clear cell RCC Cutaneous fibrofolliculomas Lung cysts Spontaneous pneumothorax

* Also known as hybrid oncocytic tumors and containing features of both chromophobe RCC and oncocytoma.

Data from Linehan, 2002; Choyke et al, 2003; Linehan et al, 2003; Maranchie and Linehan, 2003; Pavlovich et al, 2003; Pavlovich and Schmidt, 2004; Sudarshan and Linehan, 2006; Coleman, 2008; Pfaffenroth and Linehan, 2008; Hansel and Rini, 2008; and Nathanson and Stephenson, 2009.

Knudson and Strong recognized that familial forms of cancer might hold the key to the identification of important regulatory elements known as tumor suppressor genes (Knudson, 1971; Knudson and Strong, 1972). Their observations about the childhood tumor retinoblastoma, in which familial cases tend to be multifocal and early onset, led them to propose a two-hit theory of carcinogenesis. They hypothesized that a gene product that could suppress tumor development must be involved and that both alleles of this “tumor suppressor gene” must be mutated or inactivated for tumorigenesis to occur. Furthermore, Knudson postulated that patients with the familial form of the cancer are born with one mutant allele and that all cells in that organ or tissue are at risk, accounting for the early onset and multifocal nature of the disease. In contrast, sporadic tumors develop only if a mutation occurs in both alleles within the same cell; and because each event occurs with low frequency, most tumors develop late in life and in a unifocal manner (Knudson, 1971; Knudson and Strong, 1972). Knudson’s hypothesis has proved true for retinoblastoma and a number of other tumor types, including RCC (Choyke et al, 2003; Linehan et al, 2003; Pavlovich et al, 2003; Zimmer and Iliopoulos, 2003; Pavlovich and Schmidt, 2004; Sudarshan and Linehan, 2006). Identification of familial cases of RCC was particularly important because it allowed linkage analysis between affected family members.

Tumor Biology and Clinical Implications

Pathology

Most RCCs are round to ovoid and circumscribed by a pseudocapsule of compressed parenchyma and fibrous tissue rather than a true histologic capsule. Unlike upper tract transitional cell carcinomas, most RCCs are not grossly infiltrative, with the notable exception of collecting duct RCC and some sarcomatoid variants (Farrow, 1997). Tumor size has averaged between 4 and 8 cm in most series but can vary from a few millimeters to large enough to fill the entire abdomen. Tumors smaller than 3 cm were previously classified as benign adenomas, but some small tumors have been associated with metastases (Nguyen and Gill, 2009), and most pathologists agree that with the exception of oncocytomas and some small (<1 cm) low-grade papillary adenomas there are no reliable histologic or ultrastructural criteria to differentiate benign from malignant renal epithelial tumors (Farrow, 1997; Eble et al, 2004) (see Chapter 51). When they are bivalved, RCCs consist of yellow, tan, or brown tumor interspersed with fibrotic, necrotic, or hemorrhagic areas; few are uniform in gross appearance. Cystic degeneration is found in 10% to 25% of RCCs and appears to be associated with a better prognosis compared with purely solid RCC (Corica et al, 1999; Koga et al, 2000; Onishi et al, 2001; Nassir et al, 2002; Imura et al, 2004; Webster et al, 2007). Calcification can be stippled or plaquelike and is found in 10% to 20% of RCCs.

Nuclear features can be highly variable. Grading has been based primarily on nuclear size and shape and the presence or absence of prominent nucleoli. Fuhrman’s system (Table 49–10) has been most generally adopted and is now recognized as an independent prognostic factor for RCC generally and for clear cell RCC in particular (Fuhrman et al, 1982; Pantuck et al, 2001a; Lang et al, 2005; Lohse et al, 2005; Zhou, 2009).

Aggressive local behavior is not uncommon with RCC and can be expressed in a variety of ways. Frank invasion and perforation of the renal capsule, renal sinus, or collecting system are found in approximately 20% of cases, although displacement of these structures is a more common finding. Further spread to involve adjacent organs or the abdominal wall is often precluded by the Gerota fascia, although some high-grade RCCs are able to overcome this natural barrier. One unique feature of RCC is its predilection for involvement of the venous system, which is found in 10% of RCCs, more often than in any other tumor type (Skinner et al, 1972; Schefft et al, 1978). This is most commonly manifested in the form of a contiguous tumor thrombus that can extend into the inferior vena cava (IVC) as high as the right atrium (Blute et al, 2007). Many such tumor thrombi are highly vascularized by arterial blood flow (Novick et al, 1990), and some directly invade the wall of the renal vein or vena cava, which correlates with compromised prognosis (Zini et al, 2008).

Most sporadic RCCs are unilateral and unifocal. Bilateral involvement can be synchronous or asynchronous and is found in 2% to 4% of sporadic RCCs, although it is considerably more common in patients with familial forms of RCC, such as von Hippel-Lindau disease (Farrow, 1997; Linehan et al, 2003). Multicentricity, which is found in 10% to 20% of cases, is more common in association with papillary histology and familial RCC (Mukamel et al, 1988; Cheng et al, 1991; Whang et al, 1995; Campbell et al, 1997a; Richstone et al, 2004; Krambeck et al, 2008). Satellite lesions are often small and difficult to identify by preoperative imaging, intraoperative ultrasonography, or visual inspection; they appear to be the main factor contributing to local recurrence after partial nephrectomy (Campbell et al, 1996a; Richstone et al, 2004). Microsatellite analysis suggests a clonal origin for most multifocal RCC within the same kidney (Junker et al, 2002), but tumor in the contralateral kidney is likely to be an independent growth if it is synchronous or a metastasis if it is asynchronous (Kito et al, 2002). Molecular analyses, such as gene expression profiling, may help to determine whether an asynchronous tumor is a second primary tumor or a metastasis (Lane et al, 2009).

All RCCs are, by definition, adenocarcinomas, derived from renal tubular epithelial cells (Zambrano et al, 1999; Pantuck et al, 2001a; Renshaw, 2002; Zhou, 2009). Most RCCs share ultrastructural features, such as surface microvilli and complex intracellular junctions, with normal proximal tubular cells, and are believed to be derived from this region of the nephron (reviewed in Farrow, 1997). Similarly, immunohistochemistry for lectins and other cell surface antigens has supported derivation from the proximal convoluted tubule (Kim and Kim, 2002). However, more recent data suggest that this information applies primarily to the more common clear cell and papillary variants of RCC (Table 49–11), whereas most other histologic subtypes of RCC appear to be derived from more distal elements of the nephron (Störkel et al, 1997; Zambrano et al, 1999; Pantuck et al, 2001a; Renshaw, 2002; Linehan et al, 2003; Pavlovich and Schmidt, 2004).

Since the early 1990s the histologic classification of RCC has undergone several major revisions (Zambrano et al, 1999; Renshaw, 2002; Linehan et al, 2003; Eble et al, 2004; Zhou, 2009). Traditionally, RCC was divided into four histologic subtypes: clear cell, granular cell, tubulopapillary, and sarcomatoid. On the basis of advances in the molecular genetics of RCC and a more discerning interpretation of histologic and ultrastructural features, a newer classification scheme was proposed by Kovacs (1993). This classification system was approved by an international consensus workshop of clinicians and researchers in the field (Weiss et al, 1995; Störkel et al, 1997; Zambrano et al, 1999; Pantuck et al, 2001a; Renshaw, 2002; Linehan et al, 2003; Pavlovich and Schmidt, 2004). In this system, granular cell tumors were reclassified into other categories based on distinct histopathologic features, chromophobe RCC was recognized as a new RCC subtype, and sarcomatoid features were categorized as variants of other histologic subtypes rather than a distinct tumor type (Weiss et al, 1995; Störkel et al, 1997; Oyasu, 1998; Zambrano et al, 1999; Renshaw, 2002; Linehan et al, 2003; Pavlovich and Schmidt, 2004). Current practice is to identify the primary histologic subtype and comment on the presence and extent of sarcomatoid differentiation rather than to separate these tumors into a distinct category, although the prognostic implications have not changed (Mian et al, 2002; Cheville et al, 2004; Eble et al, 2004; Zhou, 2009). Depending on well-defined histologic and ultrastructural criteria, granular cell tumors were reclassified as papillary RCC, eosinophilic variants of chromophobe RCC, or combined with clear cell RCC (Weiss et al, 1995; Kovacs et al, 1997; Störkel et al, 1997; Oyasu, 1998; Zambrano et al, 1999; Renshaw, 2002; Zhou, 2009). Another important development was the identification of renal medullary carcinoma that is common in young African-Americans with sickle cell trait (Davis et al, 1995b; Störkel et al, 1997; Hakimi et al, 2007). With additional advances in ancillary pathologic studies, including electron microscopy, immunohistochemistry, molecular genetics and cytogenetics, several additional unique subtypes of RCC have been identified since the implementation of the 1993 classification system. Based on these findings, an updated classification of malignant epithelial tumors of the kidney was presented by the World Health Organization in 2004 (see Table 49–11; Eble et al, 2004).

The 2004 World Health Organization classification reflects current understanding of RCC not as a single malignant neoplasm but rather a group comprising several different tumor subtypes, each with a distinct genetic basis and unique clinical features (see Table 49–11). Important changes include the addition of several RCC subtypes with distinct pathologic and clinical features that were previously grouped within “conventional” or unclassified RCC, such as RCC associated with XP11.2 translocations/TFE3 gene fusions, which has microscopic features of both clear cell and papillary RCC and occurs primarily in children and young adults (Argani et al, 2001; Camparo et al, 2008; Geller et al, 2008), and the indolent mucinous tubular and spindle cell carcinoma (Hes et al, 2002; Ferlicot et al 2005; Fine et al, 2006). Sophisticated gene expression profiling and proteomic analyses support the individuality of each of these tumor subtypes and hold great promise for differentiating additional subtypes in the future (Takahashi et al, 2003, 2004; Amy-Bazille et al, 2004; Furge et al, 2004; Sugimura et al, 2004; Yang et al, 2004, 2006; Young et al, 2008). This is clearly a field in evolution, with changes stimulated by basic science advances and astute clinical observation.

Clinical Presentation

Because of the sequestered location of the kidney within the retroperitoneum, many renal masses remain asymptomatic and nonpalpable until they are advanced. With the more pervasive use of noninvasive imaging for the evaluation of a variety of nonspecific symptom complexes, more than 50% of RCCs are now detected incidentally (Pantuck et al, 2000). Several studies have shown that such tumors are more likely to be confined to the kidney, and a positive impact on survival has been reported, although the potential contributions of lead and length time biases have not been defined (Konnak and Grossman, 1985; Thompson and Peek, 1988; Kessler et al, 1994; Tsui et al, 2000b; Parsons et al, 2001; Lee et al, 2002; Leslie et al, 2003; Nguyen et al, 2006; DeCastro and McKiernan, 2008; Kane et al, 2008).

Symptoms associated with RCC can be due to local tumor growth, hemorrhage, paraneoplastic syndromes, or metastatic disease (Table 49–12). Flank pain is usually due to hemorrhage and clot obstruction, although it can also occur with locally advanced or invasive disease. The classic triad of flank pain, gross hematuria, and palpable abdominal mass is now rarely found (Jayson and Sanders, 1994). This is fortunate because this constellation of findings almost always denotes advanced disease, and some refer to it as the “too late triad.” Before the advent of ultrasonography and CT, most patients with RCC presented with one or more of these signs or symptoms, and many were incurable. Other indicators of advanced disease include constitutional symptoms, such as weight loss, fever, and night sweats, and physical examination findings such as palpable cervical lymphadenopathy, nonreducing varicocele, and bilateral lower extremity edema due to venous involvement. A minority of patients present with symptoms directly related to metastatic disease, such as bone pain or persistent cough. A less common but important presentation of RCC is that of spontaneous perirenal hemorrhage, although the underlying mass is often obscured by the blood. Zhang and colleagues (2002) have shown that more than 50% of patients with perirenal hematoma of unclear etiology have an occult renal tumor, most often AML or RCC. Repeat CT a few months later will often provide a definitive diagnosis.

Table 49–12 Clinical Presentation of Renal Cell Carcinoma

Incidental Local Tumor Growth Hematuria Flank pain Abdominal mass Perirenal hematoma Metastases Persistent cough Bone pain Cervical lymphadenopathy Constitutional symptoms Weight loss/fever/malaise Obstruction of the Inferior Vena Cava Bilateral lower extremity edema Nonreducing or right-sided varicocele Paraneoplastic Syndromes Hypercalcemia Hypertension Polycythemia Stauffer’s syndrome

Paraneoplastic syndromes are found in 20% of patients with RCC, and few tumors are associated with the diversity of such syndromes (Table 49–13). In fact, RCC was previously referred to as the internist’s tumor because of the predominance of systemic rather than local manifestations (Sufrin et al, 1989; Gold et al, 1996; Moein and Dehghani, 2000; DeLuca et al, 2002; Kamra et al, 2002; Hagel et al, 2005; Costanzi et al, 2008). Now, a more appropriate name would be the radiologist’s tumor, given the frequency of incidental detection (Parsons et al, 2001; DeCastro and McKiernan, 2008; Kane et al, 2008). Nevertheless, it is still important to evaluate for paraneoplastic phenomena because they can be a source of major morbidity and can affect clinical decision making. Under normal circumstances, the kidney produces 1,25-dihydroxycholecalciferol, renin, erythropoietin, and various prostaglandins, all of which are tightly regulated to maintain homeostasis. RCC may produce these substances in pathologic amounts, and it may also elaborate a variety of other physiologically important factors, such as parathyroid hormone–like peptides, lupus-type anticoagulant, human chorionic gonadotropin, insulin, and various cytokines and inflammatory mediators (Sufrin et al, 1989; Gold et al, 1996; Ather et al, 2002; Elias, 2005). These substances are believed to be responsible for the development of constitutional symptoms such as weight loss, fever, and anemia as well as some of the distinct paraneoplastic syndromes observed with this malignancy (Altundag et al, 2005; Elias, 2005).

Table 49–13 Incidence of Systemic Syndromes Associated with Renal Cell Carcinoma

SYNDROME	%
Elevated erythrocyte sedimentation rate	55.6
Hypertension	37.5
Anemia	36.3
Cachexia, weight loss	34.5
Pyrexia	17.2
Abnormal liver function	14.4
Hypercalcemia	4.9
Polycythemia	3.5
Neuromyopathy	3.2
Amyloidosis	2.0

From Gold PJ, Fefer A, Thompson JA. Paraneoplastic manifestations of renal cell carcinoma. Semin Urol Oncol 1996;14:216–22.

Hypercalcemia has been reported in up to 13% of patients with RCC and can be due to either paraneoplastic phenomena or osteolytic metastatic involvement of the bone (Sufrin et al, 1989; Gold et al, 1996; Magera et al, 2004; Klatte et al, 2007d; Pepper et al, 2007; Schwarzberg and Michaelson, 2009). The production of parathyroid hormone–like peptides is the most common paraneoplastic etiology, although tumor-derived 1,25-dihydroxycholecalciferol and prostaglandins may contribute in a minority of cases (Goldberg et al, 1964; Mangin et al, 1988; Sufrin et al, 1989; Gold et al, 1996; Walther et al, 1997b; Magera et al, 2004; Massfelder et al, 2004; Klatte et al, 2007d; Pepper et al, 2007). Recent data suggest that the expression of parathyroid hormone–like peptides is suppressed by the wild-type VHL protein, and these peptides may act as potent growth factors for RCC (Massfelder et al, 2004). This may account in part for the observation that patients with RCC who present with hypercalcemia have a compromised prognosis, with a relative risk of death from cancer progression of 1.78 compared with patients with normal serum calcium levels (Magera et al, 2004). The signs and symptoms of hypercalcemia are often nonspecific and include nausea, anorexia, fatigue, and decreased deep tendon reflexes. Medical management predominates and includes vigorous hydration followed by diuresis with furosemide and the selective use of bisphosphonates, corticosteroids, or calcitonin (reviewed in Gold et al, 1996; Coleman, 2004; Pepper et al, 2007; Schwarzberg and Michaelson, 2009). Bisphosphonate therapy is now established as standard of care for patients with hypercalcemia of malignancy, as long as renal function is adequate (Schwarzberg and Michaelson, 2009). Zoledronic acid, 4 mg intravenously every 4 weeks, appears to be particularly effective in patients with RCC but must be withheld in the presence of renal insufficiency (Lipton et al, 2003, 2004; Coleman et al, 2004; Schwarzberg and Michaelson, 2009). Indomethacin has also proved useful in a minority of cases (Gold et al, 1996; Walther et al, 1997b). More definite management includes nephrectomy and occasional metastasectomy, depending on the clinical circumstances. Systemic therapy to reduce the burden of disease is also desirable but difficult to achieve in patients with RCC (Goldberg et al, 1980). Hypercalcemia related to extensive osteolytic metastases is much more difficult to palliate because it is not amenable to surgical approaches, but many such patients may respond to bisphosphonate therapy (Lipton et al, 2003, 2004; Coleman et al, 2004). Some patients with hypercalcemia related to osteolytic metastases may also benefit from focused radiation therapy if limited sites of involvement can be identified (Gold et al, 1996).

Hypertension and polycythemia are other important paraneoplastic syndromes commonly found in patients with RCC (Moein and Dehghani, 2000). Hypertension associated with RCC can be secondary to increased production of renin directly by the tumor; compression or encasement of the renal artery or its branches, effectively leading to renal artery stenosis; or arteriovenous fistula within the tumor (Robertson et al, 1967; Sufrin et al, 1989). Less common causes include polycythemia, hypercalcemia, ureteral obstruction, and increased intracranial pressure associated with cerebral metastases (Sufrin et al, 1989). Polycythemia associated with RCC can be due to increased production of erythropoietin, either directly by the tumor or by the adjacent parenchyma in response to hypoxia induced by tumor growth (Gross et al, 1994; Wiesener et al, 2007).

One of the more fascinating paraneoplastic syndromes associated with RCC is nonmetastatic hepatic dysfunction, or Stauffer syndrome, which has been reported in 3% to 20% of cases (Stauffer, 1961; Rosenblum, 1987; Giannakos et al, 2005; Moria et al, 2006; Young et al, 2008). Almost all patients with Stauffer syndrome have an elevated serum alkaline phosphatase level, 67% have elevated prothrombin time or hypoalbuminemia, and 20% to 30% have elevated serum bilirubin or transaminase levels (Sufrin et al, 1989). Other common findings include thrombocytopenia and neutropenia, and typical symptoms include fever and weight loss, which is not surprising given that many patients are found to harbor discrete regions of hepatic necrosis (Sufrin et al, 1989; Gold et al, 1996). Hepatic metastases must be excluded. Biopsy, when indicated, often demonstrates nonspecific hepatitis associated with a prominent lymphocytic infiltrate (Hanash, 1982). Elevated serum levels of IL-6 have been found in patients with Stauffer syndrome, and it is believed that this and other cytokines may play a pathogenic role (Blay et al, 1997). Hepatic function normalizes after nephrectomy in 60% to 70% of cases. Persistence or recurrence of hepatic dysfunction is almost always indicative of the presence of viable tumor and thus represents a poor prognostic finding (reviewed in Sufrin, 1989).

A variety of other less common but distinct paraneoplastic syndromes associated with RCC have been reviewed by Sufrin and colleagues (1989), including Cushing syndrome, hyperglycemia, galactorrhea, neuromyopathy, clotting disorders, and cerebellar ataxia (Ather et al, 2002; Kamra et al, 2002; Hagel et al, 2005; Mohammed et al, 2006; Ammar et al, 2008).

In general, treatment of paraneoplastic syndromes associated with RCC has required surgical excision or systemic therapy and, except for hypercalcemia, medical therapies have not proved helpful.

Screening and Clinical Associations

A number of factors make screening for RCC appealing (Cohn and Campbell, 2000; Herts, 2005; Carrizosa and Godley, 2009). Most important, RCC remains primarily a surgical disease requiring early diagnosis to optimize the opportunity for cure. Unfortunately, our ability to salvage patients with advanced disease remains limited. Consistent with these observations, several studies have demonstrated an apparent advantage to early or incidental diagnosis of RCC (Konnak and Grossman, 1985; Thompson and Peek, 1988; Kessler et al, 1994; Cohn and Campbell, 2000; Tsui et al, 2000b; Parsons et al, 2001; Lee et al, 2002; Leslie et al, 2003).

The primary factor that limits the widespread implementation of screening for RCC is the relatively low incidence of RCC in the general population (approximately 12 cases per 100,000 population/year) (Wallen et al, 2007; DeCastro and McKiernon, 2008; Woldrich et al, 2008; Carrizosa and Godley, 2009). In this setting a screening test must be almost 100% specific to avoid an unacceptably high false-positive rate, which would lead to unnecessary, expensive, and potentially harmful diagnostic or therapeutic procedures. In addition, even if the test were 100% sensitive and specific, the yield from screening would be so low that it would not be considered cost effective (Cohn and Campbell, 2000; Carrizosa and Godley, 2009). Even when one considers populations with established risk factors for RCC, such as male sex, increased age, and heavy tobacco use, generalized screening would be difficult to justify because the increase in relative risk associated with each of these factors is at best twofold to threefold (Paganini-Hill et al, 1988; Cohn and Campbell, 2000; Carrizosa and Godley, 2009). Another confounding factor is the prevalence of clinically insignificant tumors such as renal adenomas, which are found at autopsy in 10% to 20% of individuals, and other benign or slow-growing tumors (Xipell, 1971; Bosniak et al, 1995; Cohn and Campbell, 2000; Pantuck et al, 2000; Parsons et al, 2001). There is clearly a risk that such clinically insignificant lesions could be detected, leading to unnecessary evaluation and treatment (Pantuck et al, 2000; Parsons et al, 2001). All these factors recommend against generalized screening efforts for the detection of RCC.

Review of the literature describing the use of dipstick analysis for hematuria and ultrasonography or CT for screening for RCC supports these conclusions (Cohn and Campbell, 2000; Herst, 2005; Carrizosa and Godley, 2009). Urinalysis is simple and inexpensive, but the yield of RCC in several screening studies has been exceedingly low (Mohr et al, 1986; Thompson, 1987; Mariani et al, 1989; Murakami et al, 1990). In part, this may be because small RCCs are often not associated with hematuria (gross or microscopic) because this is a parenchymal rather than an urothelium-based malignant neoplasm (Tosaka et al, 1990). The incidence of RCC in ultrasound or CT screening studies has ranged from 23 to 300 per 100,000 population, and an apparent advantage of screening has been debated because an increased proportion of organ-confined tumors has been found in screened populations compared with historical controls (Sohma et al, 1989; Tosaka et al, 1990; Spouge et al, 1996; Tsuboi et al, 2000; Filipas et al, 2003; Fenton and Weiss, 2004; Herts, 2005; Malaeb et al, 2005; Turney et al, 2006; Carrizosa and Godley, 2009). However, although the yield of RCC has been higher than expected, it is still relatively low; and it is unlikely that such efforts would be considered cost effective (Cohn and Campbell, 2000; Herts, 2005). Overall, the yield of RCC in such studies is an order of magnitude lower than the yield from prostate-specific antigen–based screening for prostate cancer, and many of the same controversies about lead and length time biases that have plagued the debate about screening for prostate cancer also apply to RCC (Cohn and Campbell, 2000; Pantuck et al, 2000; Parsons et al, 2001; Fenton and Weiss, 2004; Herts, 2005; Carrizosa and Godley, 2009). Because of these considerations, it is difficult to justify generalized screening efforts for RCC given the currently available technology.

Several investigators are now reporting novel molecular assays to detect RCC-related biomarkers in the urine or serum that may substantially alter our perspective about screening for RCC. These assays can detect microsatellite alterations in the DNA, VHL gene mutations or hypermethylation, expression of RCC-specific proteins such as CA-9, or upregulation of angiogenic factors, including VEGF (Eisenberger et al, 1999; De la Taille et al, 2000; Chang et al, 2001; Ashida et al, 2003; Battagli et al, 2003; Chen and Getzenberg, 2004; Hoque et al, 2004; Uzzo et al, 2004). Proteomic profiling of the urine to detect RCC-specific markers also holds much promise for the future (Rogers et al, 2003).

For now, however, the focus of screening for RCC must be on well-defined target populations, such as patients with end-stage renal disease and acquired renal cystic disease, tuberous sclerosis, and familial RCC (Table 49–14). Eighty percent of patients with end-stage renal disease eventually develop acquired renal cystic disease, and 1% to 2% of this subgroup develops RCC (Ishikawa et al, 1980, 1990; Matson and Cohen, 1990; Levine et al, 1991; Cheuk et al, 2002; Brown, 2004). Overall, the relative risk of RCC in patients with end-stage renal disease has been estimated to be 5- to 20-fold higher than that in the general population (Ishikawa et al, 1990, 1991, 2004; Matson and Cohen, 1990; Levine et al, 1991; Levine, 1992; Cowie, 2002; Denton et al, 2002; Brown, 2004; Farivar-Mohseni et al, 2006). Fifteen percent of patients with RCC in the setting of end-stage renal disease have metastases at the time of presentation, and many such patients die of malignant progression (Levine et al, 1991; Brown, 2004; Ishikawa, 2004). Given these considerations, screening for RCC is recommended in this population, which is substantial, representing almost 300,000 patients in the United States alone. There are, however, a number of problems associated with screening this population of patients. These include concerns about short life expectancy, increased incidence of adenomas (20% to 40% vs. 10% to 20% in the general population), complexity of imaging given the altered architecture associated with acquired renal cystic disease, and inevitable cost-related issues (Mindell, 1989; Levine et al, 1991, 1997; Sarasin et al, 1995; Cowie, 2002). A reasonable compromise for patients with end-stage renal disease is to target those without other major comorbidities, to delay screening until the third year on dialysis, and to take into account the sex and type of renal replacement therapy, although data about the last factors are admittedly controversial (Brown, 2004; Carrizosa and Godley, 2009). Ishizuka and colleagues (2000) have demonstrated elevated serum levels of VEGF in dialysis patients with RCC, suggesting a potential role for biomarkers for screening this population in the future. Interestingly, recent data suggest that renal transplant recipients remain at high risk for RCC in the native kidneys, and Neuzillet and colleagues (2005) have recommended continued periodic radiologic screening even after transplantation (Ianhez et al, 2007).

Table 49–14 Screening for Renal Cell Carcinoma: Target Populations

Patients with End-Stage Renal Disease

Screen only patients with long life expectancy and minimal major comorbidities. Perform periodic ultrasound examination or CT scan beginning during third year on dialysis.

Patients with Known von Hippel-Lindau Syndrome

Obtain biannual abdominal CT or ultrasound study beginning at age 15 to 20 years. Perform periodic clinical and radiographic screening for nonrenal manifestations.

Relatives of Patients with von Hippel-Lindau Syndrome

Obtain genetic analysis. If positive, follow screening recommendations for patients with known von Hippel-Lindau syndrome. If negative, less stringent follow-up is required.

Relatives of Patients with Other Familial Forms of Renal Cell Carcinoma

Obtain periodic ultrasound or CT study and consider genetic analysis.

Patients with Tuberous Sclerosis

Perform periodic screening with ultrasound examination or CT scan.

Patients with Autosomal-Dominant Polycystic Kidney Disease

Routine screening is not justified.

An increased incidence of RCC has also been debated in tuberous sclerosis, an autosomal dominant disorder in which patients can develop adenoma sebaceum (a distinctive skin lesion), epilepsy, mental retardation, and renal cysts and AMLs (Shapiro et al, 1984; Bernstein et al, 1986; Washecka and Hanna, 1991; Aoyama et al, 1996; Bjornsson et al, 1996; Robertson et al, 1996; Sampson, 1996; Tello et al, 1998; Choyke et al, 2003; Lendvay and Marshall, 2003; Narayanan, 2003; Cohen and Zhou, 2005; Radkowski et al, 2006). Many cases of RCC in this syndrome have been characterized by early onset and multifocality, suggesting a genetic predisposition (Washecka and Hanna, 1991; Lendvay and Marshall, 2003). In addition, the Eker rat, which is mutant for the rodent homologue of the TSC2 gene responsible for the development of tuberous sclerosis in humans, develops RCC at high frequency, as do TSC2-deficient knockout mice (Yeung et al, 1994; Kobayashi et al, 1997, 1999; Hino et al, 1999; Lendvay and Marshall, 2003; McDorman and Wolf, 2002). Furthermore, an increased incidence of TSC2 mutations has been found in human RCC (Duffy et al, 2002; Lendvay and Marshall, 2003), and Liu and colleagues (2003) have shown that loss of this tumor suppressor gene leads to upregulated expression of VEGF through an mTOR and HIF-2α–mediated mechanism, analogous to the role of the VHL protein. Such biologic and clinical observations argue in favor of an increased predisposition for RCC in this syndrome, which is consistent with most, although admittedly not all, relevant demographic data (Shapiro et al, 1984; Bernstein et al, 1986; Washecka and Hanna, 1991; Aoyama et al, 1996; Bjornsson et al, 1996; Robertson et al, 1996; Sampson, 1996; Tello et al, 1998; Lendvay and Marshall, 2003; Martignoni et al, 2003; Cohen and Zhou, 2005; Rakowski et al, 2006; Nathanson and Stephenson, 2009). A reasonable conclusion is that periodic renal imaging should be pursued in patients with tuberous sclerosis; such a policy will also facilitate follow-up for the development and progression of AML.

Screening for RCC in autosomal dominant polycystic kidney disease (ADPKD) was previously recommended, but several factors have prompted a reassessment of this policy. More recent studies suggest no significantly increased risk of RCC in ADPKD, and imaging is extremely difficult in this population related to the altered intrarenal architecture (Torres et al, 1985; Gregoire et al, 1987; Keith et al, 1994; Sessa et al, 1997; Soderdahl et al, 1997; Cohn and Campbell, 2000; Gupta et al, 2000; Mosetti et al, 2003). The increased incidence of adenomas in ADPKD would also mitigate against a potential benefit of screening (Torres et al, 1985; Gregoire et al, 1987). Taken together, these considerations suggest that screening for RCC in patients with ADPKD should not be pursued.

Special consideration should also be given to von Hippel-Lindau disease in any discussion of the value of screening for RCC. This syndrome should be considered in any patient with early-onset or multifocal RCC or RCC in combination with any of the following: a history of visual or neurologic disorders; a family history of blindness, central nervous system tumors, or renal cancer; or coexistent pancreatic cysts, epididymal lesions, or inner ear tumors (Neumann and Zbar, 1997; Choyke et al, 2003; Linehan et al, 2003; Pavlovich and Schmidt, 2004; Vira et al, 2007; Pfaffenroth and Linehan, 2008; Coleman, 2008; Hansel and Rini, 2008; Nathanson and Stephenson, 2009). Patients suspected of having von Hippel-Lindau disease, or the appropriate relatives of those with documented disease, should strongly consider genetic evaluation. The entire VHL gene has now been sequenced and sophisticated molecular analysis is readily available and offers a number of important advantages (Zbar et al, 1999; Linehan et al, 2003; Vira et al, 2007; Pfaffenroth and Linehan, 2008). Patients with germline mutations can be identified and offered clinical and radiographic screening that can identify the major manifestations of von Hippel-Lindau disease at a presymptomatic phase, allowing potential amelioration of the considerable morbidity associated with this syndrome (Glenn et al, 1992; Maranchie and Linehan, 2003; Vira et al, 2007; Pfaffenroth and Linehan, 2008). Investigators at the National Institutes of Health have recommended that such patients be evaluated with (1) annual physical examination and ophthalmologic evaluation beginning in infancy; (2) estimation of urinary catecholamines at the age of 2 years and every 1 to 2 years thereafter; (3) MRI of the central nervous system biannually beginning at the age of 11 years; (4) ultrasound examination of the abdomen and pelvis annually beginning at the age of 11 years, followed by CT every 6 months if cysts or tumors develop; and (5) periodic auditory examinations (Choyke et al, 1995; Friedrich, 1999; Maranchie and Linehan, 2003). Less intensive protocols have also been advocated, although all relevant organ systems should be addressed (Fraser et al, 2007; Foulkes and Hodgson, 1998). Individuals who are found to be wild type for both alleles of VHL also benefit because they can be spared much of the expense and anxiety associated with such intensive surveillance protocols.

Molecular screening is also available for patients suspected of having hereditary papillary RCC and other familial forms of RCC and should be discussed with appropriate family members (Schmidt et al, 1997; Linehan et al, 2003; Zbar et al, 2007; Vira et al, 2007; Pfaffenroth and Linehan, 2008; Coleman, 2008; Hansel and Rini, 2009, Nathanson and Stephenson, 2009). Again, individuals at risk, as defined by the presence of mutations of the c-MET proto-oncogene or other relevant genetic alterations, and those with suggestive clinical or family histories should be evaluated with abdominal ultrasonography or CT at periodic intervals. Further testing may be indicated according to the syndrome involved.

Staging

Until the 1990s the most commonly used staging system for RCC was Robson’s modification of the system of Flocks and Kadesky, and this schema is still embedded in the mindset of many urologists (Fig. 49–15) (Robson, 1963; Robson et al, 1969). In retrospect, the limitations of this classification scheme are readily evident. The primary problem can be found in stage III, where tumors with lymphatic metastases, a very poor prognostic finding, were combined with those with venous involvement, many of which can be treated and potentially cured with an aggressive surgical approach (Gettman and Blute, 2002; Leibovich et al, 2003c; Nguyen and Campbell, 2006). Further imprecision resulted from the fact that the extent of venous involvement was not delineated in this system, and tumor size, an important prognostic parameter, was not incorporated. The net effect was that the prognostic significance of the various stages was blunted, with some studies reporting equivalent survival for patients with stage II and stage III tumors (Skinner et al, 1971).

The tumor, nodes, and metastasis (TNM) system proposed by the Union International Contre le Cancer (UICC) represented a major improvement because it separated tumors with venous involvement from those with lymphatic invasion and defined the anatomic extent of disease more explicitly (Beahrs et al, 1988; Leibovich et al, 2003b; Nguyen and Campbell, 2006; DeCastro and McKiernon, 2008). Another advantage of the TNM system is that it has facilitated comparison of clinical and pathologic data from various centers across the globe (Leung and Ghavamian, 2002; Leibovich et al, 2003b; Nguyen and Campbell, 2006).

In 2009 the American Joint Committee on Cancer (AJCC) proposed a revision of the TNM system that is now the recommended staging system for RCC (Table 49–15) (Edge et al, 2010). The TNM classification for RCC has undergone several modifications in the past 3 decades in an effort to more accurately reflect tumor biology and prognosis and to guide clinical management. It is important to be cognizant of these changes when comparing studies from different eras (Nguyen and Campbell, 2006). In 1997 the previous division of stages T1 and T2 at tumor size of 2.5 cm was abandoned because several studies showed no prognostic significance at this level. Analysis of the Surveillance, Epidemiology, and End Results (SEER) program database demonstrated survival differences associated with 5-, 7.5-, and 10-cm cut points, and the 7-cm cut point between stages T1 and T2 was adopted because it reflected the mean tumor size in the database (Guinan et al, 1995a; Gettman et al, 2001b; Ficcarra et al, 2006; Salama et al, 2005). In the 2002 version stage T1 was subdivided: T1a represents tumor size of 4 cm or less, and T1b represents tumor size between 4 and 7 cm, reflecting data in the literature demonstrating excellent outcomes for patients with small (≤4 cm), unilateral, confined tumors managed by either partial or radical nephrectomy (Butler et al, 1994; Lerner et al, 1996; Hafez et al, 1999; Igarashi et al, 2001; Frank et al, 2004a, 2005b, Nguyen and Campbell, 2006). The most recent change for organ-confined tumors is a subdivision of T2 tumors: T2a tumor represents tumors between 7 and 10 cm, and T2b represents tumors greater than 10 cm (see Table 49–15), supported by a number of studies demonstrating prognostic relevance at this breakpoint (Frank et al, 2005b; Klatte et al, 2007c).

Other major revisions in 2009 included a reclassification of tumors with adrenal metastasis, venous thrombi, and lymphatic involvement, representing a substantial departure from previous staging paradigms for RCC (Table 49–16) (Edge et al, 2010). Contiguous extension of tumor into the ipsilateral adrenal gland is now classified as T4 and metastatic involvement of either adrenal as M1, reflecting likely patterns of dissemination. The poor prognosis of adrenal involvement from RCC is well documented and supported this important change (Sagalowsky et al, 1994; Sandock et al, 1997; Paul et al, 2001a; Von Knobloch et al, 2004; Lam et al, 2005a; Shvarts et al, 2005a; Siemer et al, 2005; Thompson et al, 2005a; Nguyen and Campbell, 2006; Kirkali et al, 2007). The favorable prognosis of isolated renal venous thrombi prompted a downgrading from stage T3b to stage T3a in the 2009 version (Moinzadeh and Libertino, 2004; DeKernion, 2005; Leibovich et al, 2005; Shvarts et al, 2005a; Thompson et al, 2005a; Ficarra and Artibani, 2006; Gofrit et al, 2007; Ficarra et al, 2007b; Margulis et al, 2007c). Finally, lymphatic extension, which previously was subdivided based on the number of involved nodes, has now been compressed to simplify this aspect of the staging process, because prognostic relevance of the previous version was not observed (Edge et al, 2010).

Table 49–16 2010 Revision of AJCC Staging for Kidney Cancer (7th Edition): Summary of Substantial Changes

Division of T2: organ confined T2a: 7-10 cm T2b: >10 cm Upgrading of adrenal involvement T4 if contiguous extension into ipsilateral adrenal M1 otherwise Downgrading of isolated renal venous thrombus Previously T3b, now T3a Compression of nodal involvement Any nodal involvement now N1 Recommendation that all potentially adverse features be captured at least parenthetically For example, T3b (tumor with inferior vena cava thrombus below level of diaphragm, also exhibiting extension into the perinephric fat)

Data from Edge SB, Byrd DR, Compton CC, et al, editors. AJCC cancer staging manual. 7th ed. New York: Springer; 2010. p. 479–89.

TNM staging classically is defined by the most advanced feature demonstrated by the tumor, yet important prognostic information can be lost in the process. Many tumors exhibit multiple adverse findings, such as high-level tumor thrombus along with ipsilateral adrenal involvement. Ideally all of the relevant anatomic staging information would be captured, at least parenthetically (e.g., pT4 [ipsilateral adrenal involvement; also exhibiting IVC thrombus above the diaphragm]). Future staging systems will ideally capture all of this information, because a number of investigators have demonstrated strong prognostic relevance to such combinations of parameters (Liebovich et al, 2005; Shvarts et al, 2005a; Thompson et al, 2005a; Fujita et al, 2007; Ficarra et al, 2007a, 2007b; Klatte et al, 2007c; Terrone et al, 2008).

The clinical staging of renal malignant disease begins with a thorough history, physical examination, and judicious use of laboratory tests (Nguyen and Campbell, 2006; DeCastro and McKiernon, 2008). Presentation with systemic symptoms such as significant weight loss (>10% of body weight), cachexia, or poor performance status all suggest advanced disease, as do physical examination findings of a palpable mass or lymphadenopathy. A nonreducing varicocele and lower extremity edema suggest venous involvement. Abnormal liver function test results, elevated serum alkaline phosphatase or lactate dehydrogenase level or sedimentation rate, hypercalcemia, and significant anemia point to the probability of advanced disease (Gelb, 1997; Srigley et al, 1997; Nguyen and Campbell, 2006; Lane and Kattan, 2008).

The radiographic staging of RCC can be accomplished in most cases with a high-quality abdominal CT scan and a routine chest radiograph, with selective use of MRI and other studies as indicated (reviewed in Bechtold and Zagoria, 1997; Choyke et al, 2001; Griffin et al, 2007, Zhang et al, 2007a; Ng et al, 2008; Herts, 2009). MRI can be reserved primarily for patients with locally advanced malignant disease, equivocal venous involvement, or allergy to intravenous contrast material (Choyke, 1997; Pretorius et al, 2000; Choyke et al, 2001; Zhang et al, 2007a; Herts, 2009). CT findings suggestive of extension into the perinephric fat include perinephric stranding (Fig. 49–16), which is a nonspecific finding, or a distinct soft tissue density within the perinephric space, which is a more definitive but uncommon finding (Bechtold and Zagoria, 1997; Herts, 2009). Overall, the accuracy of CT or MRI for detection of involvement of the perinephric fat is low, reflecting the fact that extracapsular spread often occurs microscopically (Pretorius et al, 2000; Choyke et al, 2001; Leung and Ghavamian, 2002; Kamel et al, 2004; Zhang et al, 2007a). Many of these potentially locally advanced cases are managed with radical nephrectomy, so the clinical relevance of this imprecision in staging is blunted. Ipsilateral adrenal involvement can be assessed with reasonable accuracy through a combination of preoperative CT and intraoperative inspection. Patients with an enlarged or indistinct adrenal gland on CT, extensive malignant replacement of the kidney, or palpably abnormal adrenal gland are at risk for ipsilateral adrenal involvement and should be managed accordingly (Sagalowsky et al, 1994; Sandock et al, 1997; Paul et al, 2001a; Sawai et al, 2002; Kobayashi et al, 2003; Zhang et al, 2007a; Ng et al, 2008; Lane et al, 2009). Enlarged hilar or retroperitoneal lymph nodes 2 cm or more in diameter on CT almost always harbor malignant change, but this should be confirmed by surgical exploration or percutaneous biopsy if the patient is not a surgical candidate. Many smaller nodes prove to be inflammatory rather than neoplastic and should not preclude surgical therapy (Studer et al, 1990; Choyke et al, 2001; Israel and Bosniak, 2003a; Zhang et al, 2007a; Bach and Zhang, 2008; Ng et al, 2008; Herts, 2009). MRI can add specificity to the evaluation of retroperitoneal nodes by distinguishing vascular structures from lymphatic ones (Bechtold and Zagoria, 1997; Bassignani, 2006). MRI is still the premier study for evaluation of invasion of tumor into adjacent structures and for surgical planning in these challenging cases (Bechtold and Zagoria, 1997; Choyke, 1997; Pretorius et al, 2000; Choyke et al, 2001; Bassignani, 2006; Bach and Zhang, 2008; Herts, 2009). Obliteration of the fat plane between the tumor and adjacent organs (e.g., the liver) can be a misleading finding on CT and should prompt further imaging with MRI. In reality, surgical exploration is often required to make an absolute differentiation.

The sensitivities of CT for detection of renal venous tumor thrombus and IVC involvement are 78% and 96%, respectively (Bechtold and Zagoria, 1997; Zhang et al, 2007a; Ng et al, 2008; Herts, 2009). CT findings suggestive of venous involvement include venous enlargement, abrupt change in the caliber of the vein, and intraluminal areas of decreased density or filling defects. The diagnosis is strengthened by the demonstration of collateral vessels. Most false-negative findings occur in patients with right-sided tumors in whom the short length of the vein and the mass effect from the tumor combine to make detection of the tumor thrombus difficult (Bechtold and Zagoria, 1997; Herts, 2009). Fortunately, most such cases are readily identified and dealt with intraoperatively. MRI is well established as the premier study for the evaluation and staging of IVC tumor thrombus, although recent data suggest that multiplanar CT is likely equivalent (Choyke et al, 1987; Goldfarb et al, 1990; Kallman et al, 1992; Bechtold and Zagoria, 1997; Choyke, 1997; Oto et al, 1998; Sun et al, 1999; Pretorius et al, 2000; Sohaib et al, 2002; Israel and Bosniak, 2003a; Ergen et al, 2004; Hallscheidt et al, 2005; Zhang et al, 2007a; Ng et al, 2008). MRI and multiplanar CT are noninvasive methods that provide reliable information about both the cephalad and caudal extent of the thrombus and can often distinguish bland from tumor thrombus (Pretorius et al, 2000; Choyke et al, 2001, Bassignani, 2006; Bach and Zhang, 2008; Herts, 2009). Venacavography is now best reserved for patients with equivocal MRI or CT findings or for patients who cannot tolerate or have other contraindications to cross-sectional imaging. Transesophageal echocardiography also appears to be accurate for establishing the cephalad extent of the tumor thrombus, but it is invasive and provides no distinct advantages over MRI or CT in the preoperative setting (Glazer and Novick, 1997). Doppler ultrasonography is operator dependent and does not provide the anatomic resolution available with MRI or multiplanar CT (Habboub et al, 1997).

Metastatic evaluation in all cases should include a routine chest radiograph, careful and systematic review of the abdominal and pelvic CT or MRI findings, and liver function tests (Choyke et al, 1987; Griffin et al, 2007; Zhang et al, 2007a; Ng et al, 2008; Herts, 2009). Most investigators agree that a bone scintiscan can be reserved for patients with elevated serum alkaline phosphatase or bone pain and that a chest CT scan can be reserved for patients with pulmonary symptoms or an abnormal chest radiograph (Lim and Carter, 1993; Seaman et al, 1996; Choyke et al, 2001). However, patients with locally advanced disease, enlarged retroperitoneal lymph nodes, or significant comorbid disease may mandate more thorough imaging to rule out metastatic disease and to aid in treatment planning (Choyke et al, 2001; Griffin et al, 2007). As always, evaluation and management must be individualized on the basis of the clinical circumstances. Shvarts and colleagues (2004) have shown that performance status is a powerful predictor of bone metastasis. In their analysis, patients with good performance status (Eastern Cooperative Oncology Group performance status score of 0), no evidence of extraosseous metastases, and no bone pain were at extremely low risk and did not benefit from bone scintigraphy. They recommended a bone scintiscan for all other patients, and the incidence of bone metastasis in this group was above 15%.

Positron emission tomography (PET) has also been investigated for patients with high risk of metastatic RCC, with most studies showing good specificity but suboptimal sensitivity. At present its best role is for patients with equivocal findings on conventional imaging. In this setting an abnormal PET scan may indicate metastatic disease and could strongly influence further evaluation and management (Hoh et al, 1998; Ramdave et al, 2001; Nimeh et al, 2002; Jadvar et al, 2003; Kang et al, 2004; Lawrentschuk et al, 2006; Griffin et al, 2007; Powles et al, 2007; Bouchelouche et al, 2008; Nieh, 2009). PET/CT combined with radiolabeled monoclonal antibody to CA-9 is also being explored in this population for molecular imaging of clear cell RCC (Brouwers et al, 2002; Divgi et al, 2007; Zhang, 2008). Biopsy of the primary tumor and/or potential metastatic sites is also selectively required as part of the staging process.

Prognosis

Important prognostic factors for cancer-specific survival in patients with localized RCC include specific clinical signs or symptoms, tumor-related factors, and various laboratory findings (Table 49–17) (Lane and Kattan, 2008). Overall, tumor-related factors such as pathologic stage, tumor size, nuclear grade, and histologic subtype have the greatest utility on an independent basis. However, an integrative approach, combining a variety of factors that have proved to have independent value on multivariate analysis, appears to be most powerful (Kattan et al, 2001; Zisman et al, 2001b; Frank et al, 2002; Sorbellini et al, 2005; Lane and Kattan, 2008; Parker et al, 2009; Isbarn and Karakiewicz, 2009). Patient-related factors, such as age, CKD, and co-morbidity, have a significant impact on overall survival in some patients and should be a primary consideration during treatment planning for patients with localized RCC (Hollingsworth et al, 2006; Berger et al, 2009; Pettus et al, 2008).

Clinical findings suggestive of a compromised prognosis in patients with presumed localized RCC include symptomatic presentation, weight loss of more than 10% of body weight, and poor performance status (Gelb, 1997; Srigley et al, 1997; Kattan et al, 2001; Zisman et al, 2001b; Kim HL et al, 2003; Kontak and Campbell, 2003; Schips et al, 2003; Patard et al, 2004a). Anemia, thrombocytosis, hypercalcemia, albuminuria, elevated serum alkaline phosphatase, C-reactive protein, lactate dehydrogenase, or erythrocyte sedimentation rate, and other paraneoplastic signs or symptoms have also correlated with poor outcomes for patients with RCC (Gelb, 1997; Srigley et al, 1997; Symbas et al, 2000; Jacobsen et al, 2000; O’Keefe et al, 2002; Kim HL et al, 2003; Kontak and Campbell, 2003; Vaglio et al, 2003; Patard et al, 2004a; Karakiewicz et al, 2007; Magera et al, 2008). Although abnormal values are more common in patients with advanced RCC, some of these abnormalities, including hypercalcemia, anemia, and elevated erythrocyte sedimentation rate, were independent predictors of cancer-specific mortality in patients with localized clear cell RCC after accounting for other major prognostic factors (Magera et al, 2008).

Pathologic stage has proved to be the single most important prognostic factor for RCC (Thrasher and Paulson, 1993; Delahunt, 1998; Kontak and Campbell, 2003; Leibovich et al, 2005a; Lane and Kattan, 2008; Kanao et al, 2009). The RCC TNM staging system clearly distinguishes between patient groups with different predicted cancer-specific outcomes (Table 49–18), confirming that the extent of locoregional or systemic disease at diagnosis is the primary determinant of outcome for this disease (Bassil et al, 1985; Hermanek and Schrott, 1990; Kontak and Campbell, 2003; Lane and Kattan, 2008). Previous studies confirmed that the 2002 modification of the TNM system provided better prognostic ability than its predecessors (Ficarra et al, 2005; Frank et al, 2005a), and it is anticipated that the 2010 TNM staging system will impart further improvements (Frank et al, 2005a; Leibovich et al, 2005a; Thompson et al, 2005a).

Several studies demonstrate 5-year survival rates of 70% to 90% for organ-confined disease and document a 15% to 20% reduction in survival associated with invasion of the perinephric fat (Kontak and Campbell, 2003; Leibovich et al, 2005a; Lane and Kattan, 2008). Renal sinus involvement is classified along with perinephric fat invasion as T3a, and several studies suggest that these patients may be at even higher risk for metastasis related to increased access to the venous system (Bonsib et al, 2000; Uzzo et al, 2002; Thompson et al, 2005a; Margulis et al, 2007b; Bedke et al, 2009; Bertini et al, 2009). Collecting system invasion has also been shown to confer poorer prognosis in otherwise organ-confined RCC (Uzzo et al, 2002; Klatte et al, 2007a). Several reports have shown that most patients with direct or metastatic ipsilateral adrenal involvement, which is found in 1% to 2% of cases, eventually succumb to systemic disease progression, suggesting a hematogenous route of dissemination or a highly invasive phenotype (Sagalowsky et al, 1994; Sandock et al, 1997; Paul et al, 2001b; Von Knobloch et al, 2004; Siemer et al, 2005; Thompson et al, 2005a). The most recent staging system now reclassifies tumor as T4 if there is direct invasion of the adrenal gland or otherwise as M1, to reflect this poor prognosis (Guinan et al, 1997; Thompson et al, 2005a).

Venous involvement was once thought to be a very poor prognostic finding for RCC, but several reports demonstrate that many patients with tumor thrombi can be salvaged with an aggressive surgical approach. These studies document 45% to 69% 5-year survival rates for patients with venous tumor thrombi as long as the tumor is otherwise confined to the kidney (Novick et al, 1990; Thrasher and Paulson, 1993; Staehler and Brkovic, 2000; Quek et al, 2001; Zisman et al, 2003; Leibovich et al, 2005b; Blute et al, 2007; Haferkamp et al, 2007; Klatte et al, 2007b; Wotkowicz et al, 2008; Subramanian et al, 2009). At one extreme, Golimbu and associates (1986) reported 84% 5-year survival in the best of circumstances—tumor thrombus limited to the main renal vein and tumor otherwise confined to the kidney. Patients with venous tumor thrombi and concomitant lymph node or systemic metastases have markedly decreased survival, and those with tumor extending into the perinephric fat have intermediate survival (Montie et al, 1991; Glazer and Novick, 1996; Gettman et al, 1999; Naitoh et al, 1999; Sweeney et al, 2002b; Bissada et al, 2003; Kim HL et al, 2004c; Moinzadeh and Libertino, 2004; Leibovich et al, 2005b; Klatte et al, 2007b; Wotkowicz et al, 2008). The most recent version of the TNM system advocates capturing all such adverse features during the staging process.

The prognostic significance of the cephalad extent of tumor thrombus has been controversial, and it is difficult to compare various series because of selection biases and related covariables (Leibovich et al, 2005b; Wotkowicz et al, 2008). In several series the incidence of advanced locoregional or systemic disease increased with the cephalad extent of the tumor thrombus, accounting for the reduced survival associated with tumor thrombus extending into or above the level of the hepatic veins (Sosa et al, 1984; Thrasher and Paulson, 1993; Kim HL et al, 2004a; Wotkowicz et al, 2008). However, other data suggest that the cephalad extent of tumor thrombus is not of prognostic significance as long as the tumor is otherwise confined (Libertino et al, 1987; Glazer and Novick, 1996; Staehler and Brkovic, 2000; Blute et al, 2007). Direct invasion of the wall of the vein appears to be a more important prognostic factor than level of tumor thrombus and is now classified as pT3c independent of the level of tumor thrombus (Hatcher et al, 1991; Zini et al, 2008).

The major drop in prognosis comes in patients whose tumor extends beyond the Gerota fascia to involve contiguous organs (stage T4), which is rarely associated with 5-year survival, and in patients with lymph node or systemic metastases (Thrasher and Paulson, 1993; Thompson et al, 2005a). Lymph node involvement has long been recognized as a dire prognostic sign because it is associated with 5- and 10-year survival rates of 5% to 30% and 0% to 5%, respectively (Bassil et al, 1985; Phillips and Taneja, 2004). Systemic metastases also portend a particularly poor prognosis for RCC, traditionally with 1-year survival of less than 50%, 5-year survival of 5% to 30%, and 10-year survival of 0% to 5%, although these numbers have improved modestly in the era of targeted treatments (Motzer et al, 1999; Motzer and Russo, 2000; Négrier et al, 2002; Sella et al, 2003; Rini et al, 2009). Patients presenting with synchronous metastases fare worse, with many patients dying of disease progression within a year (Motzer et al, 1999, 2004; Mekhail et al, 2005; Leibovich et al, 2005b; Rini et al, 2009). For patients with asynchronous metastases the metastasis-free interval has proved to be a useful prognosticator because it reflects the tempo of disease progression (Maldazys and deKernion, 1986; Négrier et al, 2002; Motzer et al, 2004; Mekhail et al, 2005; Rini et al, 2009). Other important prognostic factors for patients with systemic metastases include performance status, number and sites of metastases, anemia, hypercalcemia, elevated alkaline phosphatase or lactate dehydrogenase levels, thrombocytosis, and sarcomatoid histology (Maldazys and deKernion, 1986; Motzer et al, 1999; Motzer and Russo, 2000; Négrier et al, 2002; Leibovich et al, 2003b; Motzer et al, 2004; Mekhail et al, 2005; Escudier et al, 2007; Choueiri et al, 2007; Lane and Kattan, 2008). The presence of bone, brain, and/or liver metastases, and multiple metastatic sites have been associated with further compromise in prognosis (Négrier et al, 2002; Leibovich et al, 2005b; Mekhail et al, 2005; Escudier et al, 2007). These factors have been used to effectively categorize patients with metastatic RCC as low, intermediate, and poor risk, with corresponding differences in median survival (Table 49–19) (Motzer et al, 1999; Boumerhi et al, 2003; Motzer et al, 2004; Mekhail et al, 2005; Choueiri et al, 2007; Escudier et al, 2007). These risk groups provide important information for determining the likelihood of benefit a patient may expect to receive after cytoreductive nephrectomy and/or resection of other metastatic disease.

Another significant prognostic factor for RCC is tumor size, which has proved to be an independent prognostic factor for both organ-confined and invasive RCC (Kattan et al, 2001; Kontak and Campbell, 2003; Lane and Kattan, 2008). Giuliani and colleagues (1990) reported 5-year survival rates of 84% for patients with tumor diameter less than 5 cm, 50% for tumors between 5 and 10 cm, and 0% for tumors more than 10 cm in diameter. To a large extent, this is due to a strong correlation between tumor size and pathologic tumor stage, but several studies have subsequently demonstrated that tumor size can function as an independent prognostic factor (Guinan et al, 1995b; Kattan et al, 2001; Sorbellini et al, 2005; Crispen et al, 2008b; Nguyen and Gill, 2009). Larger tumors are more likely to exhibit clear cell histology and high nuclear grade, and both of these factors correlate with a compromised prognosis (Frank et al, 2003b; Lane et al, 2007a; Thompson et al, 2009). A review of 1771 patients with organ-confined RCC showed 10-year cancer-specific survival rates of 90% to 95%, 80% to 85%, and 75% for patients with pT1a, pT1b, and pT2 tumor, respectively (Patard et al, 2004a). Many other studies have also shown a particularly favorable prognosis for the unilateral pT1a tumors that are now being discovered with increased frequency. In series from the Cleveland Clinic and the Mayo Clinic, such tumors were associated with greater than 95% 5-year cancer-specific survival rates, whether they were managed with nephron-sparing surgery or radical nephrectomy (Butler et al, 1994; Lerner et al, 1996; Cheville et al, 2001; Gill et al, 2007; Lane and Gill, 2007; Crispen et al, 2008b).

Other important prognostic factors for RCC include nuclear grade, histologic subtype and symptomatic presentation. Several grading systems for RCC have been proposed on the basis of nuclear size and morphology and presence or absence of nucleoli. Unfortunately, interobserver variability is common in the assignment of nuclear grade; there is no ideal classification system that can overcome the subjectivity of this exercise. Nevertheless, almost all the proposed grading systems have provided prognostic information for RCC, and nuclear grade has proved in most cases to be an independent prognostic factor when subjected to multivariate analysis (Goldstein, 1997; Ficarra et al, 2001; Kattan et al, 2001; Zisman et al, 2001b; Lohse et al, 2002, 2005; True, 2002; Kontak and Campbell, 2003; Patard et al, 2004c; Lang et al, 2005; Lane and Kattan, 2008; Ficarra et al, 2009).

Fuhrman’s classification system has been the most generally adopted grading system for RCC. In Fuhrman and colleagues’ original report (1982) the 5-year survival rates for grades 1 to 4 were 64%, 34%, 31%, and 10%, respectively, and nuclear grade proved to be the most significant prognostic factor for organ-confined tumors in this series. Subsequent reports have demonstrated correlations between Fuhrman’s nuclear grade and tumor stage, tumor size, venous tumor thrombi, and lymph node and systemic metastases (Bretheau et al, 1995; Lang et al, 2005; Lohse et al, 2005; Ficarra et al, 2009). Significant differences have been consistently observed between low-grade (grades 1/2) and high-grade (grade 3/4) tumors with some difficulty distinguishing the intermediate grades. Based on these studies, some have recommended changing to a three-tiered system, but this remains a matter of investigation (Medeiros et al, 1997; Lang et al, 2005; Lane and Kattan, 2008; Zhou 2009). In addition, although significant differences according to nuclear grade have been reported in series that have included patients with all types of RCC or clear cell RCC alone, the relevance of the Fuhrman classification system to evaluation of other subtypes of RCC is not entirely clear. Papillary RCC may be better subgrouped into type 1 and type 2, and oncocytic neoplasms may be better classified as chromophobe RCC, hybrid oncolytic tumors, and oncocytomas, although whether these provide better prognostic information about cancer recurrence has not been determined (see Pathology).

Histologic subtype also carries prognostic significance, although, again, primarily at the ends of the spectrum. The presence of sarcomatoid differentiation or collecting duct, renal medullary, or unclassified histologic subtype denotes a poor prognosis (Carter et al, 1992; Davis et al, 1995b; Chao et al, 2002b; Escudier et al, 2002a; Mian et al, 2002; Polascik et al, 2002; Mejean et al, 2003; Cheville et al, 2004; Nanus et al, 2004; Tokuda et al, 2006; Kobayashi et al, 2008; Zhou 2009). Several studies now suggest that clear cell RCC may have a worse prognosis on average compared with papillary or chromophobe RCC, although there are clearly poorly differentiated tumors in each of these subcategories that can be lethal (Moch et al, 2000; Amin et al, 2002; Lau et al, 2002; Cheville et al, 2003; Krejci et al, 2003; Beck et al, 2004; Patard et al, 2005; Lane and Kattan, 2008; Crispen et al, 2008b; Dall’Oglio et al, 2008; Lam et al, 2008; Margulis et al, 2008; Klatte et al, 2008b; Rothman et al, 2009). Finally, several subtypes of RCC are predictably indolent, including multiloculated cystic clear cell RCC and mucinous tubular and spindle cell carcinoma.

For patients with clinically localized disease, mode of presentation (incidental vs. symptomatic) can be combined with other predictive elements to better stratify patients after primary surgical management (Kattan et al, 2001; Patard et al, 2004a; Sorbellini et al, 2005; Karakiewicz et al, 2007; Lane and Kattan, 2008). In addition, patients with systemic symptoms suggestive of metastatic spread have significantly poorer outcomes than those with only local symptoms, such as hematuria or flank pain (Kattan et al, 2001; Sorbellini et al, 2005).

Dozens of genes that may have prognostic or therapeutic significance for patients with RCC have been identified using high-throughput technologies (Takahashi et al, 2003; Zhao et al, 2006; Lane et al, 2008b). Gene expression profiling (cDNA microarrays) can quantify the levels of thousands of individual messenger RNA transcripts within an individual tumor sample. Alterations in gene expression can then be correlated with the amount and location of specific gene products (proteins) using immunohistochemical staining of cancer specimens (Kim HL et al, 2004a; Liu et al, 2004; Parker et al, 2009). Construction of tissue microarrays can facilitate the screening of hundreds of tumors, but interpretation of results can be challenging due to tumor heterogeneity and the selection of only a small amount of tissue for analysis (Kim HL et al, 2004a; Liu et al, 2004; Leibovich et al, 2007). Furthermore, when evaluating the potential value of a new marker, it is important to consider its contribution after accounting for other known prognostic factors (George and Bukowski, 2007; Tunuguntia and Jorda, 2008).

Several molecular markers appear to serve as independent prognostic factors for RCC and have provided important insights into tumor biology (see Tumor Biology) (Bui et al, 2001; Han et al, 2003; Crispen et al, 2008b; Nogueira and Kim, 2008; Parker et al, 2009). One such factor is CA-IX, which is regulated by the VHL gene and overexpressed in most clear cell RCCs (Bui et al, 2003, 2004; Leibovich et al, 2007). Although initial studies indicated that decreased expression of CA-IX is independently associated with poorer survival in patients with metastatic RCC (Bui et al, 2003; Kim et al, 2005), this association does not appear to apply for patients with localized disease (Kim et al, 2005; Leibovich et al, 2007). CA-IX also may serve as a marker for response to systemic therapy, making CA-IX immunostaining of particular value for patients with advanced disease (Bui et al, 2004; Atkins et al, 2005; Cho et al, 2007). B7-H1 is a T-cell coregulatory molecule that is a strong independent predictor of disease progression for RCC (Thompson et al, 2006; Parker et al, 2009). This association holds even after accounting for other molecular factors and the major clinical and pathologic predictors (Krambeck et al, 2007; Parker et al, 2009). Increased proliferative index as assessed by Ki-67 has also been correlated with reduced survival in clear cell RCC (Bui et al, 2004; Klatte et al, 2009; Parker et al, 2009). Although initial data indicated that Ki-67 expression was a surrogate for histologic necrosis, more recent studies have found Ki-67 to be an independent predictor and have incorporated it into predictive algorithms (Tollefson et al, 2007; Klatte et al, 2009; Parker et al, 2009). Other factors that appear to be useful include cell cycle regulators, such as the tumor suppressor TP53 (Kim HL et al, 2004a; Shvarts et al, 2005b; Klatte et al, 2009); various growth factors and their receptors, including members of the VEGF family (Jacobsen et al, 2000; Rivet et al, 2008; Phyoc et al, 2008; Klatte et al, 2009); adhesion molecules; and other factors, such as survivin (Parker et al, 2006, 2009; Byun et al, 2007; Krambeck et al, 2007).

Several investigators have now developed tools that combine various prognostic factors, and this has greatly improved our predictive capacity for patients with RCC (see Table 49–19) (Kattan et al, 2001; Zisman et al, 2001b; Frank et al, 2002; Sorbellini et al, 2005; Lane and Kattan, 2008). For instance, Kattan and colleagues (2001) have combined manner of presentation (incidental vs. local or systemic symptoms), tumor histology, tumor size, and pathologic stage to develop a nomogram that predicts cancer-free survival after nephrectomy. Tumor grade was not included in this analysis because its role for non–clear cell RCC has not been clearly defined. A subsequent analysis from this same group focused only on patients with clear cell RCC and incorporated tumor grade, assessment of tumor necrosis, and vascular invasion to further improve prognostication (Fig. 49–17) (Sorbellini et al, 2005). Such nomograms provide an individual assessment of risk that clinicians can use during patient counseling (www.nomograms.org). Although several predictive algorithms incorporate histologic necrosis (Frank et al, 2002; Sorbellini et al, 2005), the utility of this predictor has been called into question in some recent studies because its assessment has not been standardized and it is not routinely reported at many centers (Isbarn and Karakiewicz, 2009).

Pantuck and colleagues (2001a) have also performed an integrated analysis of prognostic factors for patients with all stages of RCC with encouraging results. A sophisticated multivariate analysis revealed three independent prognostic factors that were most robust for predicting outcomes, namely, TNM stage, performance status, and tumor grade (Zisman et al, 2001b). The UCLA integrated staging system (UISS) was subsequently modified to identify patients with localized or metastatic disease at low, intermediate, and high risk of disease progression and has been validated internally and externally (Zisman et al, 2002b; Patard et al, 2004c; Cindolo et al, 2006, 2008; Parker et al, 2009). Molecular factors such as TP53, Ki-67, VEGF family members, and CA-IX have also been incorporated into UISS-based algorithms to predict outcomes for patients with localized or metastatic RCC (Kim et al, 2005; Klatte et al, 2009).

Another predominant model that provides individualized information for patients with clear cell RCC is the SSIGN score, which incorporates 1997 TNM stage, tumor size, nuclear grade, and presence of tumor necrosis to predict recurrence and survival after radical nephrectomy (Frank et al, 2002). Estimated cancer-specific survival according to the SSIGN score was based on data from more than 1800 patients and has subsequently been validated in several other datasets (Ficarra et al, 2006, 2009; Fujii et al, 2008; Zigeuner et al, 2010). Thompson and colleagues (2007d) have also developed a dynamic outcome prediction model that provides patients with cancer-specific survival rates that improve as the disease-free interval following surgery increases. Most recently, the group at Mayo Clinic has developed a sequential approach in which the predicted outcomes for patients at various risk of recurrence according to clinical and pathologic factors can be further stratified based on molecular data incorporated into a BioScore (Parker et al, 2009). The expression of B7-H1, survivin and Ki-67 each added independent predictive ability after accounting for either the UISS or SSIGN score, especially for patients at intermediate or high risk of recurrence (Parker et al, 2009).

TNM staging systems and prognostic algorithms have different purposes. The TNM staging system is used to provide a universal language for communication between clinicians and patients and is based solely on the anatomic extent of cancer dissemination. A wealth of literature now supports the notion that algorithms that incorporate multiple predictive elements, such as nomograms and artificial neural networks, outperform risk assessment based on expert opinion or simpler models, such as classic staging systems (Dawes et al, 1989; Ross et al, 2002; Isbarn and Karakiewicz, 2009; Shariat et al, 2009). The development and use of these predictive tools can help guide counseling and follow-up of patients with RCC and identify patients more likely to benefit from specific interventions.