Prognostic and Therapeutic Impact of the Histopathologic Definition of Parenchymal Epithelial Renal Tumors

Transcription

1 EUROPEAN UROLOGY 58 (2010) available at journal homepage: Platinum Priority Collaborative Review Kidney Cancer Editorial by Matthew K. Tollefson and R. Houston Thompson on pp of this issue Prognostic and Therapeutic Impact of the Histopathologic Definition of Parenchymal Epithelial Renal Tumors Vincenzo Ficarra a, *, Matteo Brunelli b, Liang Cheng c, Ziya Kirkali d, Antonio Lopez-Beltran e, Guido Martignoni b, Rodolfo Montironi f, Giacomo Novara a, Hein Van Poppel g a Department of Surgical and Oncological Sciences, Urologic Unit, University of Padua, Italy b Department of Pathology, University of Verona, Italy c Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, IN, USA d Dokuz Eylul University School of Medicine, Department of Urology, Izmir, Turkey e Department of Pathology, Reina Sofia University Hospital and Faculty of Medicine, Cordoba, Spain f Section of Pathological Anatomy, Polytechnic University of the Marche Region, School of Medicine, United Hospitals, Ancona, Italy g Department of Urology, University Hospitals Leuven, Leuven, Belgium Article info Abstract Article history: Accepted August 2, 2010 Published online ahead of print on August 10, 2010 Keywords: Renal cell carcinoma Renal core biopsy Partial nephrectomy Surgical margins Pathologic stage Tumor histotypes Fuhrman nuclear grading Necrosis Microvascular invasion Sarcomatoid dedifferentiation Integrated staging systems Molecular markers Cytogenetic markers Please visit europeanurology to read and answer questions on-line. The EU-ACME credits will then be attributed automatically. Context: In the last few years, the treatment of renal cell carcinoma (RCC) has progressed significantly, and some histopathologic issues have become important for selection and follow-up after medical and surgical therapies. Objective: The aim of this collaborative article is to review the most recent literature on the role of traditional histopathologic features obtained from renal core biopsy or nephrectomy specimens in the management of confined, locally advanced, and metastatic RCC. Evidence acquisition: A nonsystematic review of the literature was performed in April 2010 using the Medline database. Multiple free-text searches were performed for the following items: renal cell carcinoma, clear cell, papillary, chromophobe, histologic* subtype*, histotype*, nuclear grade*, necrosis, sarcomatoid differentiation, biopsy, molecular marker*, and cytogenetic marker*. A total of 2369 records were retrieved from Medline, and 263 full-text studies were considered and partially included in the present review. A panel of experts reached consensus on the main subheadings of this paper. Evidence synthesis: Core needle biopsies can provide important information that is useful to avoid the overtreatment of benign tumors and to help plan watchful waiting or minimally invasive treatments in selected patients. Tumor histotype is fundamental in the pathologic report. In the context of integrated prognostic systems, the combination of the most important clinical and pathologic factors (TNM stage, Fuhrman nuclear grade, presence of necrosis, microvascular invasion, and sarcomatoid dedifferentiation) allows us to reach a high prognostic accuracy. These models can be used to select patients suitable for adjuvant protocols, to design an appropriate follow-up schedule, and to provide careful patient counseling. Molecular and cytogenetic markers should be further evaluated. Conclusions: The histopathologic definition of parenchymal epithelial renal tumors is fundamental to plan the management and follow-up of patients with locally confined, locally advanced, and metastatic RCC. # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Department of Surgical and Oncological Sciences, Urologic Unit, University of Padua, via Giustiniani, 2, Padova, Italy. address: vincenzo.ficarra@unipd.it (V. Ficarra) /$ see back matter # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 656 EUROPEAN UROLOGY 58 (2010) Introduction Over the past several years, the treatment of renal cell carcinoma (RCC) has significantly improved, and some histopathologic features have become important for selection and follow-up after both medical and surgical therapies. Current guidelines consider active surveillance, radiofrequency ablation (RFA), and cryotherapy as management options for small renal masses (SRM) in elderly patients with significant comorbidities who are unfit for surgical therapy [1,2]. However, a recent meta-analysis showed that only a few patients underwent renal tumor biopsies (RTBs) before or after minimally invasive treatments or before or during active surveillance [3]. Therefore, the role of RTB in selecting candidates for nonsurgical treatment of SRMs and in defining success or failure after minimally invasive ablative techniques remains undefined. Parallel to the significant increase of incidentally detected small renal tumors, a trend in favor of nephron-sparing surgery (NSS) has been observed. The guidelines recommend NSS for renal tumors 4 cm, although the role of elective NSS in treating T1b tumors is more controversial [1,2]. Open or laparoscopic radical nephrectomy (RN) remains the gold standard of treatment for patients with tumors and suitable for NSS or patients with nonmetastatic T2 3 RCC. The most critical issue is to identify those patients with a high risk of progression who might benefit from multimodal therapies. Finally, targeted therapies have improved the prospects of patients with metastatic disease. The selection among the first-line available targeted therapies is mainly based on the risk-group stratification proposed by Motzer et al, which was based only on clinical variables [4]. Considering that >90% of metastatic cases treated with targeted therapies received upfront cytoreductive RN, the potential role of other integrated prognostic systems including pathologic data should be evaluated. The purpose of this collaborative literature review is to define the role of traditional histopathologic features obtained from RTB or nephrectomy specimens in the management of locally confined, locally advanced, and metastatic RCC. 2. Evidence acquisition A nonsystematic review of the literature was performed in April 2010 using the Medline database. Multiple free-text searches were performed, in the fields title and abstract for the following items dating from January 1, 1997: renal cell carcinoma, clear cell, papillary, chromophobe, histologic* subtype*, histotype*, nuclear grade*, necrosis, sarcomatoid differentiation, biopsy, molecular marker*, and cytogenetic marker*. The search used these parameters following the last consensus consultation held in Rochester, Minnesota, organized by the Union Internationale Contre le Cancer and the American Joint Committee on Cancer (AJCC) and published in 1997 [5]. No limitations regarding the language of publication or the type of publication were imposed. A total of 2369 records were retrieved from Medline. All records were evaluated by two of the authors to identify the relevant papers. Finally, 263 full-text studies were considered and partially included. Some relevant recent papers published after April 2010 were added according to the discretion of the authors. All the authors shared the responsibility for evidence synthesis, and a final expert consensus was reached regarding the subheadings of this review. 3. Evidence synthesis 3.1. Histologic information from renal tumor biopsies Percutaneous RTB still has a limited role in the clinical workup of patients with renal parenchymal tumors [1]. Improvements in the technique used to perform core biopsies under computed tomography (CT) or ultrasound guidance and the use of a coaxial needle has reduced the risk of tumor spread and complications significantly. In a 2007 systematic review, Volpe et al reported an overall risk of tumor seeding in <0.01% of cases, and bleeding and other complications were uncommon [6]. In recent years, an increasing number of clinicians have used RTBs preoperatively to identify benign tumors and thus avoid surgical overtreatment and to select patients suitable for active surveillance or minimally invasive treatment [2,7]. As with prostate biopsy, cores that are torn or <10 mm are considered unsatisfactory, and further cores must be taken until at least two good-quality cores have been obtained [8]. These cores should be obtained from different locations, one from the central and one from the peripheral area in tumors <4 cm, and two more from peripheral areas in larger tumors [9]. This protocol can reduce the finding of necrosis in cores taken from the central part of tumors >4 cm. Needle cores should be placed in tissue cassettes between two nylon meshes according to the pre-embedding methods described by Rogatsch et al [10]. Each core should be numbered and appropriately identified according to the site of the biopsy. Most of the data available in the literature are obtained after hematoxylin and eosin staining [6,7,11]. Only a few authors used immunohistochemical (IHC) stains on the core specimens in the preclinical study [12]. The final histopathologic report generally includes histotypes for benign and malignant tumors and grades for RCCs. The most critical concerns of RTB are the percentages of unsatisfactory biopsies and the relative diagnostic accuracy. Recent reports show that 0 21% of core biopsies were insufficient for diagnosis (Table 1). Suboptimal amounts of tissue, sampling of necrotic areas, and the presence of blood or normal kidney parenchyma can produce core biopsies that are unsatisfactory for pathologic evaluation. There was an overall diagnostic accuracy (defined as the percentage of biopsies with sufficient material for proper distinction between benign and malignant tumors) of >90%. Interesting results have also been reported in the identification of tumor histotype on biopsy cores, with an accuracy ranging from 86% to 100%. Specifically, oncocytomas continue to be

3 EUROPEAN UROLOGY 58 (2010) Table 1 Results reported in the most recent clinical studies evaluating the renal core biopsies in patients with renal tumors Reference Cases, n Pathology confirmed, % Failure, % Nonconclusive, % Accuracy in cancer diagnosis, % Accuracy in histologic subtype diagnosis, % Accuracy in grade diagnosis, % Hara et al [13] NA NA Johnson et al [14] NA NA Caoili et al [15] NA NA Rybicki et al [16] NA NA Neuzillet et al [8] Eshed et al [17] NA NA Lebret et al [18] NA Volpe et al [19] NA 84 NA NA Schmidbauer et al [20] NA Wang et al [21] NA 91 NA NA Overall (11) NA = not available. a relevant diagnostic challenge because cells with deeply eosinophilic cytoplasms are also found in numerous RCCs, including clear cell carcinoma with granular cytoplasm, oncocytic papillary RCC, and eosinophilic variants of chromophobe RCC. Although an adequate number of biopsies and a good interpretation of the morphologic features usually minimize diagnostic errors, the use of IHC and chromosomal analyses on formalin-fixed, paraffin-embedded tissues can be necessary [22]. An IHC panel including CD10, parvalbumin, AMACR, CK7, and S100A1 seems the most promising. Fluorescence in situ hybridization analysis using centromeric probes to evaluate the gains of chromosomes 7 and 17 and the losses of chromosomes 1, 2, 3p, 6, 10, 17, and Y can be helpful in selected cases (Fig. 1 and Table 2) [23 29]. Patients with low-grade tumors (grades 1 2) who are elderly or unfit for surgery can be considered candidates for cryotherapy, RFA, or surveillance programs. In contrast, Table 2 Immunohistochemical and cytogenetical panel for the differential diagnosis of renal cell tumors * FISH = fluorescent in situ hybridization; IHC = immunohistochemical; RCC = renal cell carcinoma. * Orange: positive (IHC) or abnormal (FISH) in >80% of the tumors; yellow: positive (IHC) or abnormal (FISH) in between 25% and 75% of the tumors; green: positive (ICH) or abnormal (FISH) in <25% of the tumors. patients with high-grade (grades 3 4) tumors should be candidates for a surgical approach [30]. Therefore, accuracy in the evaluation of tumor Fuhrman grades in core biopsies is of paramount relevance. Unfortunately, recent clinical series showed accuracy rates in the evaluation of tumor grade between 69% and 76% [8,20,31]. This percentage increases to 85% in preclinical series using in-bench core biopsies of nephrectomy specimens [9]. The low accuracy can be explained by grade heterogeneity within the same tumor that was observed in about 25% of cases [32] and by the interobserver and intraobserver variability in assigning tumor grade. The application of nuclear grading in patients with non clear cell RCC should be reconsidered [33,34]. As a whole, the main limitation of many clinical studies is the lack of definitive histology of the surgical specimen as a reference standard. Considering studies published after 2001, there was pathologic confirmation in 20 70% of cases, with the exception of Schmidbauer s study, where all patients who had a biopsy subsequently had surgery [20] (Table 1). Unfortunately, very limited data exist on RTB after minimally invasive treatments of SRMs such as cryoablation and RFA. Although most studies used biopsy to identify the histologic subtype of the tumor before treatment, most follow-up protocols were based on radiologic imaging with enhanced CT or magnetic resonance imaging. Specifically, the presence of contrast-enhancing lesions after ablation is regarded as a potential indicator of disease persistence or recurrence, based on the belief that ablated areas do not enhance because the tissues no longer have a blood supply to deliver the contrast material. The accuracy of RTBs in this setting is unknown because patients without suspicion of recurrence were not biopsied; nonetheless, biopsy seems to have a role following RFA and also in the case of negative follow-up imaging [35]. To date, the improvement in the technique of RTBs has significantly reduced the risks of tumor seeding and complications. RTBs are useful to characterize renal masses that are indeterminate at imaging to avoid surgical overtreatment in benign cases. Unfortunately, the most problematic distinction is between oncocytomas and clear cell RCC with granular cytoplasm, oncocytic papillary RCC,

4 658 EUROPEAN UROLOGY 58 (2010) Fig. 1 Renal cell neoplasms (major histotypes): immunohistochemical and fluorescent in situ hybridization analysis. (A) Clear cell renal cell carcinoma (hematoxylin and eosin [H&E]); (B) immunoreactivity for CD10 and (C) loss of 3p (orange: CEP3; green: 3p); (D) papillary renal cell carcinoma (H&E); (E) immunoreactivity for CK7 and (F) gains of chromosomes 7 and 17 (orange CEP17; green CEP7); (G) chromophobe renal cell carcinoma (H&E), (H) immunoreactivity for parvalbumin and (I) loss of chromosome 6 (green: CEP 6); (J) renal oncocytoma (H&E), (K) immunoreactivity for S100A1, and (L) normal numerical complement of both chromosomes 2 and 6 (orange: CEP 2; green: CEP 6). and eosinophilic variants of chromophobe RCC. RTBs should be used more to define the success or failure of minimally invasive treatments (ie, cryotherapy and RFA) and sometimes in the follow-up of patients who are in active surveillance protocols. However, further studies are required to evaluate the diagnostic accuracy of RTBs after ablative therapies. The limited accuracy in tumor grading is an important limitation for the use of RTB in selecting patients suitable for active surveillance or minimally invasive treatment Histologic information from partial nephrectomy specimen The correct handling of the partial nephrectomy (PN) specimen by urologists and pathologists is essential to retrieve all useful information for diagnosis and prognosis. The specimen should be immediately brought fresh and without any cut from the operating room to the pathology laboratory to avoid any artifact that could cause inadequate sampling, in particular regarding the evaluation of surgical margins and neoplastic infiltration of the perinephric fat. If, for surgical reasons, it is necessary to separate the perinephric adipose tissue, it must be sent for pathologic evaluation indicating the zone of tumor contact. The pathologist should ink the surgical margin on the NSS specimen and the surface of the whole specimen, particularly in areas where the radicality is dubious. To reduce the risk of local recurrence due to incomplete resection, the traditional standard surgical technique of NSS involved the excision of an additional 1-cm margin of

5 EUROPEAN UROLOGY 58 (2010) healthy renal parenchyma [36,37]. Some authors have demonstrated that negative surgical margins can be achieved by reducing the safety margin to 5mm [38 40]. The European Association of Urology (EAU) guidelines recommend the presence of a minimal tumor-free surgical margin of healthy renal parenchyma surrounding the resected tumor, without specifying the exact minimum thickness of the healthy parenchyma to be taken [1]. However, the prognostic significance of positive surgical margins (PSMs) remains controversial. Recently, Bensalah et al showed that only tumor location (central vs peripheral) and indication (imperative vs elective) turned out to be independent predictors of local recurrence [41]. In contrast, Bernhard et al showed that a PSM was an independent predictor of ipsilateral recurrence after PN together with tumor size >4 cm and tumor bilaterality [42]. In this scenario, some authors have supported simple enucleation instead of traditional enucleoresection of the renal tumors [43]. A recent histopathologic study performed on simple enucleation specimens confirmed that a neoplastic pseudocapsule infiltration was detected in 33% of cases and, similar to most PN series, 14% of patients showed a neoplastic infiltration beyond the pseudocapsule into the renal parenchyma. However, the authors described the presence of a precious layer of 1-mm tissue (range: mm) of renal parenchyma beyond the tumor pseudocapsule. This layer could explain the presence of negative surgical margins in patients with tumors extending beyond the pseudocapsule [44]. Taking into account the minimum thickness of the safety margin in the most recent NSS series, it is likely that in a significant percentage of cases the margin that resulted was <1 mm, substantially similar to the one reported by Minervini et al after simple enucleation. Intraoperative evaluation by frozen sections and/or gross analysis should be limited to those cases where it would modify the therapeutic approach. The microscopic analysis on frozen sections is usually done on small specimens obtained from the surgical bed in NSS when there is some suspicion of cancer. Many potential pitfalls exist in the interpretation of cytologic and architectural features of neoplastic or crushed benign renal tubules [45]. It was shown that definitive histologic examination detected higher positive margin rates both in laparoscopic and open PN series [46,47]. To date, the presence of a minimal tumor-free surgical margin of healthy renal parenchyma surrounding the resected tumor has been considered sufficient to minimize the risk of local recurrence and/or disease progression. Therefore, information regarding the tumor pseudocapsule status and the thickness of the safety area around the tumor has a place only in histopathologic and/or clinical studies but is not relevant in clinical practice. Although the presence of PSMs has a controversial role as a predictor of local recurrence, it must be reported in the final pathologic report to evaluate the surgical radicality during PN. Patients with PSMs are candidates for a careful watchful waiting program, thus avoiding RN Histopathologic factors stratifying the outcome of patients with renal cell carcinoma Pathologic stage, tumor histologic subtype, nuclear grading, presence of coagulative necrosis, microvascular invasion, and sarcomatoid differentiation are the most important traditional histopathologic prognostic factors included in the final pathology report of the PN and RN specimen Pathologic stage The TNM classification is the most common staging system in patients with RCC. Renal tumors were first categorized by the TNM staging system in 1974, which was subsequently revised several times in the following years to improve its prognostic accuracy on the basis of new evidence [48]. The last modified update was published in 2009 (Table 3) [49]. In comparison with the 6th edition of TNM, T2 cancers are subclassified into two subgroups: T2a 10 cm and T2b >10 cm. Tumors with invasion in the renal vein or in the perinephric fat were classified as T3a, whereas those with adrenal involvement were identified as T4 cancers. All the other categories remained unchanged [49]. To date, no validation studies are yet available. Although in the last years several reports have proposed changes to the TNM classification both for localized and locally advanced stages, only a few modifications were included in the most recent version of the TNM. Therefore, the optimal stratification of patients with RCC using the TNM staging system remains controversial, and further revisions should be considered Tumor histologic subtypes Until 1997, only a few studies had evaluated the prognostic impact of histologic subtypes on treatment strategy and patient survival. The Heidelberg [50] and Rochester [5] classifications identified a group of RCCs that lacked Table 3 Changes in classification of renal cell carcinoma using the TNM staging system Tumor 2002 version 2009 version T1a 4 cm, limited to kidney 4 cm, limited to kidney T1b >4 cm and 7 cm, limited to kidney >4 cm and 7 cm, limited to kidney T2 >7 cm, limited to kidney T2a >7 cm and 10 cm limited to kidney T2b >10 cm, limited to kidney T3a Perinephric or sinus fat or adrenal extension Perinephric or sinus fat invasion or renal vein involvement T3b Renal vein or vena cava involvement below diaphragm Vena cava involvement below diaphragm T3c Vena cava involvement above diaphragm Vena cava involvement above diaphragm or infiltration of the wall of the vena cava T4 Outside Gerota s fascia Outside Gerota s fascia and adrenal extension

6 660 EUROPEAN UROLOGY 58 (2010) morphologic or genetic markers that would allow them to be defined as clear cell, papillary, chromophobe, or collecting duct carcinomas. Subsequently, cytogenetic and molecular research permitted the definition of new pathologic entities, such as renal medullary carcinoma; Xp11 translocation carcinoma; carcinoma associated with neuroblastoma; and mucinous, tubular, and spindle cell carcinomas (Fig. 2), which were identified as histologic subtypes in the classification proposed by the World Health Organization (WHO) in 2004 [51]. In the past decades, the three most common histologic subtypes of RCC have become familiar: clear cell (80%), papillary (10%), and chromophobe (<5%). Specifically, data in the literature show that papillary and chromophobe RCC (ie, non clear cell RCC) often have favorable pathologic stages and better nuclear grades, as well as a lower risk of metastasizing, compared with clear cell RCC. Patients with clear cell RCC had significantly lower cancer-specific survival probabilities compared with those with either papillary or chromophobe cancers, whereas the outcomes of papillary or chromophobe cancers were similar. Five-yr cancer-specific survival probabilities ranged from 43% to 83% in clear cell RCC, from 61% to 90% in papillary RCC, and from 80% to 100% in chromophobe RCC [52]. In 1997, Delahunt and Eble divided papillary RCC into two different subtypes (types 1 and 2).Type 1 consists of papillae and tubular structures covered by a single layer of small cells with pale or basophilic cytoplasm and small oval nuclei; in these tumors foamy macrophages and psammoma bodies are easily found. Type 2 consists of papillae covered by pseudostratified or multilayered large cells with eosinophilic cytoplasm and large irregular nuclei [53]. Multiple studies evaluated such subtyping, demonstrating that type 2 papillary RCC were usually of higher stage and grade compared with type 1 and displayed different cytogenetic features [54 56]. Specifically, greater expression of vascular endothelial growth factor (VEGF)-R2 in the tumor epithelium, VEGF-R3 in both tumor epithelium and endothelium, loss of chromosome 1p, loss of 3p, and gain of 5q were observed in type 2 [56]. This subtyping remains difficult, and different percentages of type 1 and type 2 are reported in the large series. Indeed the correct allocation of Fig. 2 Recently described new entities: (A) clear cell papillary renal cell carcinoma; (B) renal cell carcinoma with prominent leiomyomatous proliferation; (C) tubulocystic renal cell carcinoma; (D) oncocytic papillary renal cell carcinoma.

7 EUROPEAN UROLOGY 58 (2010) type 2 papillary tumors requires an experienced pathologist. In the meantime there is no consensus yet about the prognostic relevance of this subclassification. Most of the available studies could not demonstrate an independent predictive role of papillary subtype for disease recurrence and cancer-specific survival [57]. Other studies have reconfirmed that type 2 papillary RCC had a worse outcome in multivariable analysis [55]. In 2003, after pathologic revision of the slides of 2528 patients who had undergone PN or RN for RCC at the Mayo Clinic, it was suggested that histologic subtype had an independent predictive value, regardless of both pathologic stage and nuclear grade [58]. However, this was not reconfirmed in a multicenter international study evaluating 4063 patients, but in this study there was no pathologic slide review performed [59]. More recently, a NCI Surveillance Epidemiology and End Results Program (SEER) study confirmed that the histotype was an independent predictor of cancer-specific survival, although the variable added little to the overall accuracy of the prognostic model [60]. The role of pathologic slide review in recognizing the histologic subtypes is illustrated by the finding that about 20% of the cancers originally identified as papillary RCC were reclassified as clear cell RCC [52]. Literature data highlight that 5-yr cancer-specific survival probabilities of patients with papillary RCC ranged from 80% to 90% when a pathologic slide revision was performed [52,58,61], whereas these probabilities were as low as 70% in the study from Patard et al [59]. It is noteworthy that a better concordance was observed among original and reviewed pathologic diagnoses of the histologic subtypes for patients treated after the publication of the Heidelberg and Rochester classifications. Histologic subtypes are also very useful in the treatment planning for patients with metastatic RCC. Histologic subtypes other than clear cell RCC are shown to be poorly responsive to immunotherapy and chemoimmunotherapy [1]. In the era of targeted therapies, the different histologic subtypes have major relevance in selecting patients for enrollment in randomized controlled trials (RCTs) and in the evaluation of their efficacy in first- and second-line treatment. For this reason, the pathologist must not only be asked to diagnose the most common types but also to recognize the rare types of renal carcinoma (many of which have been identified in recent years). Among these newly recognized tumors, mucinous tubular, spindle cell, and tubulocystic carcinoma were previously considered patterns of low-grade collecting duct carcinoma, but they are now recognized as separate tumor entities. Carcinomas associated with translocations of TFE3 and TFEB compose a significant percentage not only of pediatric carcinomas but also of adult RCCs. Renal carcinoma associated with end-stage renal disease is now recognized as having distinct morphologic patterns and behavior. Finally, there is a group of rare recently described carcinomas, such as clear cell papillary carcinoma, oncocytic papillary renal cell carcinoma, follicular renal carcinoma, and leiomyomatous renal cell carcinoma [62] Grading system Currently, the four-tiered Fuhrman grade classification, described for the first time by Susan Fuhrman in 1982, represents the most frequently used system in Europe and North America [63] (Table 4). Although several studies have demonstrated an independent predictive role of the Fuhrman nuclear grading system in patients with clear cell RCC [64], other authors have suggested a lower prognostic relevance of nuclear grades in papillary RCC [33] and chromophobe RCC [34]. Specifically, among the parameters of the Fuhrman system, neither the morphometric measures of nuclear size nor pleomorphism is correlated with outcome in papillary RCC. Nucleolar pleomorphism was significantly associated with outcome only in univariable analysis [33]. However, Klatte et al, in a series of 158 cases of papillary RCC, showed recently that Fuhrman nuclear grading provides higher prognostic accuracy (C index: 74.7%) in comparison with nucleolar grade (C index: 67.8%) [65]. With regard to chromophobe RCC, none of the three grading components was correlated with outcome when tested separately, which raises questions about the validity of Fuhrman grading for this histologic subtype [34]. Notably, when tumors in both the Sika-Paotonu et al and Delahunt et al series were classified on the basis of nuclear size alone according to the break points proposed by Fuhrman, all the tumors were classified as grade 1. In contrast, assessment of focal nucleolar prominence on its own showed 58% of papillary RCCs and 24% of chromophobe RCCs to exhibit grade 3 characteristics [33,34]. The widespread diffusion and popularity of the Fuhrman system can probably be attributed to the system s simplicity, proven correlation with other pathologic variables, and wellestablished prognostic role. Data coming from large multiinstitutional studies demonstrated quite distinct risks of mortality for a single category of patients identified by Fuhrman nuclear classification [59,66,67]. In these series, the 5-yr survival probabilities were 86 89% for grade 1 tumors, Table 4 Morphologic characteristics of Fuhrman nuclear grading Grade Morphologic features 1 Tumors consist of cells with small (approximately 10 mm), round, uniform nuclei with inconspicuous or absent nucleoli 2 Tumors consist of cells with larger nuclei (approximately 15 mm), with irregular morphology and small nucleoli when examined under high power (400 magnification) 3 Tumors consist of cells with even larger nuclei (approximately 20 mm) with irregular outlines and large prominent nucleoli that are evident even at low power (100 magnification) 4 Tumors differ from grade 3 lesions in that they contain bizarre multilobed nuclei and heavy chromatin clumps

8 662 EUROPEAN UROLOGY 58 (2010) % for grade 2 tumors, 50 60% for grade 3 tumors, and 28 30% for grade 4 tumors [59,66]. Despite the excellent discriminant proprieties of four-tiered nuclear grading, simplified two-tiered (G1 2 vs G3 4) or three-tiered (G1 2 vs G3 vs G4) Fuhrman systems have been proposed [66,68]. In fact, several investigators indicated that a four-tiered system for Fuhrman grade assignment could result in a worse interobserver agreement. Simplified schemes consisting of two or three nuclear grade strata could reduce the complexity of the four-tiered system, resulting in higher agreement among different pathologists, fewer misclassifications, and, consequently, homogeneous distribution of the cases in the different series [67]. The use of a simplified nuclear grading system in clinical practice must indeed be supported by the evidence of an overlapping predictive accuracy in comparison with the conventional four-tiered system. Such a relevant statistical point has been evaluated for the first time in a large multiinstitutional cohort of European patients treated in 14 different institutions. In this study, the authors showed that Fuhrman nuclear grade was an independent predictor of cancer-specific survival in patients with RCC, regardless of the coding system. The addition of nuclear grade to the multivariate model resulted in similar increases in predictive accuracy (0.8% for four-, three-, or two-tiered systems) [67]. From a statistical point of view, this study shows that the simplified Fuhrman grading system is as informative and accurate as the classical four-tiered system. More recently, Sun et al validated the earlier multiinstitutional study using a population-based cohort of patients with clear cell RCC who were treated surgically between 1988 and 2004 and included in nine SEER cancer registries in the United States. They confirmed that the simplified two-tiered or three-tiered nuclear systems performed similarly to the conventional four-tiered Fuhrman system in a large North American population and that the inclusion of the different four-, three-, and two-tier grading systems resulted in similar increases in prognostic accuracy (+1.2%, +1.2%, +1% for the four-, three-, and two-tier, respectively), compared with a base model including the other clinicopathologic features (patients age and gender, year and type of surgery, tumor stage and size) [69]. The main difference between this last American study and the previous European retrospective studies concerns the application of the Fuhrman nuclear grading system regardless of tumor histologic subtypes. Rioux-Leclercq et al analyzed a large population of European patients with different tumor histotypes [67]. In contrast, Sun et al restricted their analysis to patients with clear cell RCC [69]. However, the identification of the most appropriate system to be used universally remains an open question. Both simplified systems were based on the concept that grade 1 and 2 tumors could be clustered. This is now well accepted by most pathologists, considering the fact that the morphologic differences in nuclear dimensions between the two categories and the visibility of nucleoli under high power can be considered venial. Few studies including very large numbers of patients were able to demonstrate statistically significant survival differences between grade 1 and 2 RCC. Indeed most studies reported overlapping or slightly different cancer-specific survival probabilities for grade 1 and grade 2 tumors [64]. The most critical point in the decision process for the choice of a two- or a three-tiered system concerns grade 3 and grade 4 tumors. Looking at the 5-yr cancer-specific survival probabilities, the differences between grade 3 and grade 4 tumors are consistently significant and clinically more relevant than those observed between grade 1 and grade 2 cancers. Specifically, the reported 5-yr cancerspecific survival probabilities ranged from 45% to 65% in grade 3 cancers and from 25% to 40% in grade 4 cancers, respectively [64]. Finally, no well-conducted study demonstrated that the interobserver agreement was higher using a two-tiered rather than a three-tiered system. The use of a simplified version of the Fuhrman nuclear grading in clinical practice requires further clarification and preferably a consensus between pathologists and urologists Coagulative necrosis The most common form of necrosis observed in RCC tumors is coagulative necrosis. Reports from the Mayo Clinic clearly demonstrated the prognostic role of microscopic tumor necrosis with regard to both recurrence-free survival [70] and cancer-specific survival [71] in clear cell RCC. These data have been subsequently validated in several external series, applying data obtained after both pathologic slide review [72] and routine pathology [73]. However, other reports from single institution [74,75] or multicenter collaborations [76] failed to demonstrate an independent predictive role for microscopic tumor necrosis evaluated on a presence or absence basis. Notably, Klatte et al evaluated the prognostic accuracy of microscopic tumor necrosis in clear cell RCC evaluating its extent in a continuous fashion. The authors evaluated a small prospective series of 227 patients with clear cell RCC, applying a specific protocol to measure the extent of microscopic tumor necrosis. They found that tumor necrosis expressed continuously was an independent predictor of cancer-specific survival and that recursive partitioning-based survival tree analysis identified 20% necrosis extent as the ideal cut point. Compared with tumors without necrosis, those with 20% necrosis had a 6.4-fold higher risk of RCC-specific mortality. Those patients with necrosis in >20% of the tumor had a 2.2-fold and 14.1% increased risk of cancer-related death compared with those with necrosis in 20% and those without necrosis, respectively [75]. The data on the prognostic role of tumor necrosis in other histologic subtypes of RCC are more limited. With regard to papillary RCC, Klatte et al evaluated a series of 158 papillary RCCs treated at the University of California, Los Angeles (UCLA), demonstrating that tumor necrosis was an independent predictor of cancer-specific survival once adjusted for the effect of Eastern Cooperative Oncology Group (ECOG) performance status, symptoms at presentation, T stage, and presence of metastases [56]. However, similar findings were not reconfirmed by Pignot et al, who found a significant association of tumor necrosis with disease-free survival only in univariate analysis [55], or by Sengupta

9 EUROPEAN UROLOGY 58 (2010) et al, who also failed to identify any survival difference in univariate analysis [77]. Regarding chromophobe RCC, only Amin et al demonstrated the independent predictive role of tumor necrosis, evaluating a multicenter series of 145 patients. They found that tumor necrosis was an independent predictor of a composite end point including local recurrence, metastasis, or death, once adjusted for the effect of T stage, tumor size, and sarcomatoid differentiation [78]. Further data from the Mayo Clinic suggest an association between tumor necrosis and cancer-specific survival, with 10-yr cancer-specific survival probabilities of 68% in patients harboring tumor necrosis, compared with probabilities of 90% in patients without this pathologic feature ( p = 0.003) [77]. However, the authors were unable to reconfirm the data in multivariable analysis, due to the small number of cancer-related deaths observed in their cohort of chromophobe RCC Microvascular invasion Microvascular invasion (MVI) has been defined as the presence of neoplastic cells invading the vessel wall or neoplastic emboli in the intratumoral vessel lumen [79].MVI is an important prognostic factor in various malignancies such as liver, testis, penile cancer, bladder, and upper tract urothelial carcinoma, and in some of these, it has been included in the AJCC TNM staging criteria where its presence upstages the patient. However, the prognostic role in RCC is controversial. Several single-center studies evaluated its prognostic roles, demonstrating the association of MVI with higher stage, higher grade, and higher risk of lymph node and distant metastases [80 83]. However, only a few studies demonstrated an independent predictive role for either disease recurrence or cancer-related death once adjusted for the effect of the other clinical and pathologic covariates [84 86]. On the whole, further validation studies are needed Sarcomatoid dedifferentiation Sarcomatoid dedifferentiation represents transformation to a higher grade malignancy characterized by a spindle cell histologic appearance with ultrastructural or immunohistochemical evidence of epithelial and mesenchymal differentiation. Sarcomatoid dedifferentiation may represent the final common pathway of cancer progression for all subtypes of RCC [87]. Several synonyms are used to refer to the tumors with this histologic feature, including carcinosarcoma, metaplastic carcinoma, and spindle cell carcinoma. Defined as a diagnostic category in the 1981 and 1998 WHO classifications, sarcomatoid carcinoma is now considered an extreme form of dedifferentiation of RCC [88]. Data from the literature indicate that the presence of sarcomatoid differentiation in RCC is associated with a median survival <1 yr, with reported 5- and 10-yr survival rates as low as 22% and 13%, respectively [89]. The Mayo Clinic team reported that cancer-specific survival rates at 2 yr following surgery for patients who had clear cell, papillary, and chromophobe RCC with sarcomatoid differentiation were 30%, 40%, and 25%, respectively, compared with 84%, 96%, and 96%, respectively, for those patients who had RCC of the same histologic subtype without sarcomatoid differentiation [90]. Currently, histologic subtype according to the WHO classification should be included in the final pathologic report. Adequate sampling and a good understanding of the histopathologic characters minimize diagnostic errors, but the use of immunohistochemical and chromosomal analysis on formalin-fixed paraffin-embedded tissues is useful. In patients with clear cell RCC, nuclear grading, the presence or absence of coagulative necrosis, and sarcomatoid dedifferentiation should be reported. Fuhrman classification is the most used and recommended system to assign nuclear grade in patients with RCC. Little histologic and prognostic differences are present between Fuhrman grades 1 and 2. It could be argued that these categories should be combined to simplify classification. Combining grades 3 and 4 is more controversial, considering the more relevant histologic and prognostic differences between these other categories. However, the modality for applying the Fuhrman nuclear grading system in clinical practice as well as the choice between a two- or a three-tiered system should be based on a consensus of experts in the field of RCC, including urologists, pathologists, and statisticians. Recent studies have questioned the application of nuclear grading in papillary and chromophobe RCC. In patients with papillary RCC, the nucleolar grading seems to be better correlated with prognosis, whereas in patients with chromophobe RCC, Fuhrman grading is inappropriate. For newly recognized tumors, the Fuhrman grading system has to be applied, although larger series are needed to assess its usefulness as a prognostic variable. Still controversial is the role of the coagulative necrosis in non clear cell RCC. In contrast, the prognostic role of sarcomatoid differentiation has also been confirmed in non clear cell RCC Integrated prognostic systems In the past decade, various prognostic systems integrating the most important clinical and pathologic prognostic factors have been proposed to improve the prognostic accuracy in patients with both confined and advanced RCC [91] (Table 5). The most important integrated staging systems are the UCLA Integrated Staging System (UISS) [68] and the stage, size, grade, and necrosis (SSIGN) score [71]. These systems have been externally validated in the context of multicenter series [92] and European series [72,73]. More recently, a multi-institutional panel proposed and validated another significant nomogram able to predict cancer-specific survival in patients surgically treated for RCC [93]. This last nomogram seems to be significantly more accurate in comparison with previous ones proposed by Kattan et al [94] and Sorbellini et al [84]. The UISS is able to stratify patients with RCC according to pathologic stage (TNM, 1997), Fuhrman nuclear grade, and ECOG performance status. The system distinguishes patients with either localized or metastatic RCC into subgroups with a low, intermediate, or high risk of progression and mortality

10 664 EUROPEAN UROLOGY 58 (2010) Table 5 Integrated staging systems available to predict prognosis of patients with renal cell carcinoma System Stage Histologic type Variables End points External validation UCLA Integrated Staging System [68] All All TNM OS Yes Grading CSS Performance status ECOG Stage, size, grade, and necrosis (SSIGN) score) [71,92] N0M0 [84] Clear cell RCC TN(M) DFS Yes All [64] Size CSS Grading Necrosis Karakiewicz nomogram [93] All All TNM CSS Yes Size Grading Symptoms CSS = cancer-specific survival; DFS = disease-free survival; ECOG = Eastern Cooperative Oncology Group; OS = overall survival; RCC = renal cell carcinoma. [68]. Patard et al performed an external validation in a multicenter study including 4202 patients treated in eight different American and European academic centers. In 3119 patients with nonmetastatic RCC, the values of the C index ranged from to 0.863, whereas in the 1083 patients with metastatic RCC, the values of the C index ranged from to [92]. The UISS was used to select high-risk patients who were to be included in two RCTs (Adjuvant Sorafenib or Sunitinib for Unfavorable Renal Carcinoma [ASSURE] and Sunitinib Trial in Adjuvant Renal Carcinoma [STAR]) that are evaluating the efficacy of sunitinib and sorafenib as adjuvant therapies. In 2007, Karakiewicz et al published a new nomogram that was able to estimate the 1-, 2-, 5-, and 10-yr cancerspecific survival probabilities of patients undergoing PN or RN for RCC. The nomogram was generated analyzing a series of 2530 patients and validated on an external cohort of 1422 patients with RCC. The prognostic accuracy of the nomogram in the prediction of 1-, 2-, 5-, and 10-yr cancer-specific survival was as high as 87.8%, 89.2%, 86.7%, and 88.8%, respectively [93]. In 2002, Frank et al proposed a different algorithm to predict the cancer-specific survival of patients with clear cell RCC undergoing RN. Differing from all the other models, this algorithm included only pathologic variables. According to the different scores, the patients were stratified into 10 different subgroups with different cancer-related prognoses. The value of the C index was [71]. In the first external validation, performed in a small series after slides revision, the C index was as high as 0.88 [72]. More recently, Zigeuner et al reported a second external validation of the SSIGN score performed in a single-center data set of 1862 cases based on routine pathology reports. The C index in this last external validation was [73]. The practical implication is that this algorithm is no longer destined to remain confined to centers of excellence but may become readily available to clinicians practicing in institutions where dedicated uropathologists are not available [95]. A modified version of the SSIGN score was proposed by Leibovich et al with the aim of predicting the disease-free survival of patients undergoing RN for clear cell RCC. The following variables were included in the score: pathologic stage of the primary tumor (pt, 2002), locoregional lymph node involvement (N), pathologic size of the primary tumor (<10 cm vs 10 cm), Fuhrman nuclear grade, and presence of microscopic tumor necrosis. The value of the C index of the score was as high as [96]. Recently, the Mayo Clinic group presented a dynamic version of the SSIGN score that was able to predict cancer-specific survival, taking into account the disease-free interval from surgery to follow-up. In this study, 1-, 5-, and 10-yr cancer-specific survival was predicted in patients 6, 12, 24, 36, and 60 mo after RN, demonstrating a decrease in the risk of cancer death during follow-up [97]. Finally, the SSIGN score has been used to select patients who are to be recruited into the SORCE trial that evaluates the efficacy of sorafenib as an adjuvant treatment in intermediate- and high-risk localized RCC. To date, no study has compared these three prognostic systems in the context of clear and non clear cell RCC. According to the EAU guidelines, the use of the prognostic integrated systems is not recommended in clinical practice [1]. However, these tools significantly improve the prognostic accuracy of the single parameters and should be taken into consideration in counseling preoperative and postoperative patients and in planning the follow-up schedule. The same prognostic models should be used for the definition of selection criteria for and interpretation of the results of ongoing trials evaluating the effect of new targeted therapies in the adjuvant setting and in patients with metastasis receiving cytoreductive nephrectomy. Concerning patients with metastasis undergoing cytoreductive nephrectomy, the UISS, SSIGN score, and Karakiewicz nomogram should be considered for patient stratification and the interpretation of results of targeted therapies. It is possible that these integrated prognostic tools can identify good and poor responders better than the Motzer criteria, which are based only on laboratory parameters Molecular and genetic prognostic factors The introduction of targeted therapies has enlarged our knowledge of the main molecular pathways involved in the mechanisms of growth and metastases of kidney cancer. Several studies have evaluated the diagnostic and prognostic role of cytogenetic and molecular markers in RCC [98]. With regard to clear cell RCC, the largest study on the

11 EUROPEAN UROLOGY 58 (2010) prognostic role of molecular markers was recently reported by the UCLA group, who tested the expression of 29 different markers mainly related to the hypoxia-inducible and mammalian target of rapamycin (mtor) pathways in the surgical specimens of 170 patients with clinically localized RCC undergoing nephrectomy [99]. The authors showed that expression of Ki-67, p53, endothelial VEGFR-1, epithelial VEGFR-1, and epithelial VEGFR-D were independent predictors of disease-free survival once adjusted for the effects of standard clinical and pathologic variables, including ECOG performance status, pathologic stage of the primary tumor, and Fuhrman grade. Interestingly, a nomogram was generated to predict disease-free survival, including molecular, clinical, and pathologic factors, which yielded a prognostic accuracy as high as 90% [99]. Other studies evaluating the prognostic role of the cytogenetic profile of clear cell RCC have demonstrated that loss of 9p was an independent predictor of survival once adjusted for the standard clinical and pathologic variables [100,101]. In 2009, Klatte et al proposed a nomogram to predict the 3-yr disease-specific survival including the TNM staging system, the Fuhrman nuclear grading, and the loss of 9p. The prognostic accuracy was 0.89 [101]. To date, the literature on serum and urine markers as well as on molecular markers for biopsy specimens is limited [98]. With regard to the other histologic subtypes of RCC, the studies on cytogenetic and molecular markers are sparse. Specifically, Klatte et al evaluated a series of 158 patients with papillary RCC treated at UCLA, suggesting that the loss of 1p, 3p, and 9p was associated with reduced cancerspecific survival in univariable analysis [56]. The value of novel markers is required within the framework of existing prognostic factors and integrating models. The highly accurate nature of the available models including clinical and pathologic variables raises the bar for novel biomarkers because it is relatively difficult to improve accuracy beyond 90%. To date, the potential role of cytogenetic and molecular markers suggested by the available research is encouraging, especially for clear cell RCC. Further clinical research aiming at validating the available findings and exploring new markers is urgently needed to achieve wider knowledge on outcomes and the probabilities of responding to different therapies, as well as to investigate the possible role of new therapeutic targets. The lack of models capable of accurately predicting the probability of response to targeted therapies represents an important unmet need in the field of RCC prognostic factors. Specifically, the VHL gene, HIF1-alfa, CAIX, downstream molecules of VEGF, and mtor pathways are the more promising biomarkers evaluating the therapy response. However, despite their promising characteristics, all these markers await further confirmation of their added value in comparison with traditional predictors. 4. Conclusions Histopathologic definition of parenchymal epithelial renal tumors is a fundamental step in planning the management and follow-up of patients with locally confined, locally advanced, and metastatic RCC. Needle core renal biopsies can provide new information to avoid the overtreatment of benign tumors and plan watchful waiting or minimally invasive treatments in elderly patients with significant comorbidity. The combination of the most important clinical and/or pathologic factors in the context of the integrated prognostic systems can reach a high prognostic accuracy. These models can be used to select patients suitable for adjuvant protocols, to plan the more appropriate follow-up, and to perform careful patient counseling. Molecular and cytogenetic markers should be further evaluated in the context of clinical research. Various molecular and cytogenetic prognostic markers have been evaluated. Almost all are ongoing to demonstrate their potential new information beyond the traditional integrating systems. Probably the main interest for these markers lies in their ability to predict the response to the targeted therapies. Author contributions: Vincenzo Ficarra had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ficarra, Novara. Acquisition of data: Ficarra, Novara. Analysis and interpretation of data: Ficarra, Novara. Drafting of the manuscript: Ficarra, Novara. Critical revision of the manuscript for important intellectual content: Ficarra, Brunelli, Cheng, Kirkali, Lopez-Beltran, Martignoni, Montironi, Novara, Van Poppel. Statistical analysis: None. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: None. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/ affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None. References [1] Ljungberg B, Hanbury DC, Kuczyk MA, et al. Guidelines on renal cell carcinoma. European Association of Urology Web site. [2] Campbell SC, Novick AC, Belldegrun A, et al. Guideline for management of the clinical T1 renal m ass. J Urol 2009;182: [3] Kunkle DA, Egleston BL, Uzzo RG. Excise, ablate or observe: the small renal mass dilemma a meta-analysis and review. J Urol 2008;179: [4] Motzer RJ, Bacik J, Murphy BA, et al. Interferon-alfa as a comparative treatment for clinical trials of new therapies against advanced renal cell carcinoma. J Clin Oncol 2002;20: [5] Union Internationale Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC). Workshop on Diagnosis and Prognosis of Renal Cell Carcinoma. Rochester, Minnesota, March 21 22, Cancer 1997; 80: