Investigators are zeroing in on the relationship between histologic features and post-op survival
By Bradley C. Leibovich, MD
Despite current therapies, renal cell carcinoma (RCC) is often lethal. Many factors influence outcomes, including stage, grade, subtype, necrosis, and molecular attributes. Prognosis can be difficult to determine. Comprehensive staging strategies that incorporate multiple patients and tumor factors are necessary to predict survival accurately and optimize research strategies. Recently designed algorithms allow for convenient and up-to-date estimates of outcome in this rapidly evolving field.
Staging
Researchers at Mayo Clinic have been able to estimate postoperative survival on a large cohort of RCC patients using the primary tumor/regional nodes/metastases (TNM) staging protocol. The dedicated team of the Mayo Nephrectomy Registry has gathered outcome data from 5,400 patients from 1970-2004, with only 3% lost to follow-up. A single pathologist had read more than 14,000 slides.
The most extensively used classification scheme, the TNM algorithm categorizes RCC patients into stages I through IV. Those free of nodal involvement and metastases are either stage I or II, depending upon tumor grade (T1, stage I and T2, stage II). All patients with T3 or N1 classification are considered stage III. Patients with T4 N1 malignancies are stage IV, as are those with T4 tumor extension, N2 nodal involvement, and/or metastases.
Within the Mayo cohort, there is a clear 10-year survival advantage conferred by lower TNM stage. Fewer than 20% of stage IV patients were alive at 10 years compared with more than 50% of stage II patients and 60% of stage I.
Grade
There are several grading schemes for RCC, each significantly related to stage and prognosis. Unfortunately, there is no agreed-upon grading standard. Mayo Clinic uses a nuclear grading system: Grade 1 tumors have small, round nuclei with inconspicuous nucleoli. Grade 2 tumors contain round to slightly irregular nuclei with mildly enlarged nucleoli. Grade 3 tumors have round to irregular nuclei with prominent nucleoli, whereas grade 4 tumors contain enlarged pleomorphic or giant cells (Clin Lab Med. 2005;25:433-464).
As with RCC stage, tumor grade directly correlated with outcome in the cohort at the Mayo Clinic. Patients with tumor grades 1 or 2 had significantly higher overall 10-year post-surgical survival rates (80% or greater) compared with grade 3 (40%) or grade 4 (less than 20%).
Other histologic features
RCC constitutes a group of parenchymal tumors that are highly heterogeneous with respect to morphology and associated natural history. By the International Union Against Cancer (UICC)/American Joint Committee on Cancer consensus (AJCC) histologic classification, eight RCC subtypes can be distinguished according to morphologic, histochemical, and molecular characteristics: Five malignant (RCC) forms—clear cell (or conventional), papillary, chromophobe, collecting duct, and RCC not otherwise specified; and three benign forms—metanephric adenoma (adenofibroma), papillary renal cell adenoma, and oncocytoma. Clear cell and papillary RCC subtypes are both derived from the proximal tubule cells whereas chromophobe RCC is derived from collecting duct intercalated cells in the renal cortex.
Whether histologic type is an independent prognostic variable is controversial. A retrospective study from eight international centers examined outcomes data for 4,163 patients with the three most common forms of RCC: clear cell, papillary, and chromophobe, although these subtypes were not verified by central pathologic review. Histologic subtype was not shown to independently predict survival. The authors concluded that histologic subtype (as defined by the 1997 UICC/AJCC) did not add to prognostic information provided by other variables such as TNM stage, Fuhrman grade, and Eastern Cooperative Oncology Group (ECOG) performance status (J Clin Oncol. 2005;23:2763-2771).
Data from our center contrast with these findings. A retrospective review in 2003 by John Cheville, MD, and his colleagues, which looked at outcomes among 2,385 surgically treated sporadic unilateral RCC patients, showed that cancer-specific survival at five years was significantly diminished in patients with clear cell (69%) versus papillary or chromophobe subtypes (87% each). The difference remained significant when stratified by stage and nuclear grade (Am J Surg Pathol. 2003;27:612-624). Our current collection of data includes 3,103 RCC patients who have been followed for 15 years. Univariate analysis revealed a three-fold increase in mortality among clear cell compared with chromophobe and papillary subtypes. After adjusting for TNM stage, grade, and histologic necrosis, clear cell was associated with a nearly two-fold higher mortality than chromophobe and papillary subtypes. No significant difference in survival between papillary and chromophobe subtypes was observed.
Tumor necrosis
Tumor necrosis is a negative prognostic sign among patients with clear cell or chromophobe subtypes. In a review of 3,009 surgically treated RCC cases at Mayo, 28% of clear cell RCC patients had histologic evidence of tumor necrosis, and these patients had a 5.7-fold increased risk of death relative to those with the same subtype but no necrosis. On multivariate analysis adjusting for stage, grade and tumor size, necrosis was still associated with a two-fold increased risk of death compared with non-necrotic clear cell tumors. Necrosis was most prevalent in papillary tumors (47%), but no significantly increased risk was associated with necrosis within this subtype (Cancer. 2005;104;511-520).
Sarcomatoid differentiation
Sarcomatoid differentiation is one of the most reliable indications of an aggressive tumor. In a review of 2,381 RCC cases at Mayo, sarcomatoid differentiation was present in 5%. Mean duration of survival for patients with this adverse pathologic feature was 1.4 years. Sarcomatoid features were an independent predictor of RCC death regardless of histologic subtype, stage, size, and necrosis (Am J Surg Pathol. 2004;28:435-441).
Prognostic algorithms
Outcome prediction models and algorithms integrate many of these findings and assist the clinician in important ways. Algorithms provide a foundation for risk stratification, patient counseling, assessment of treatment efficacy, optimal trial design, and development of surveillance programs. Mayo researchers have developed algorithms to predict outcomes in patients with clear cell RCC. One of these algorithms is SSIGN (Stage, SIze, Grade, and Necrosis). SSIGN scoring can predict survival among patients with clear cell RCC treated with radical nephrectomy. In a multivariate analysis of 1,801 surgically treated clear cell cases, features significantly associated with death were determined and regression coefficients were used to generate a scoring algorithm. Low SSIGN scores correlated with improved survival.
A similar algorithm called the progression score is being used to identify clear cell RCC patients at high risk for progression who make good candidates for clinical trials of adjuvant therapy. Based on retrospective data of 1,671 clear cell patients who underwent surgery at Mayo Clinic between 1970 and 2000 for localized disease, metastasis-free survival has been shown to correlate with a low progression score. A phase III trial of sorafenib, an oral multitargeted kinase inhibitor, is enrolling patients with high or intermediate metastasis risk based on this algorithm.
Molecular markers
Molecular prognostic markers have the potential to greatly refine our ability to predict prognosis for RCC patients and may help identify therapeutic targets. As our identification strategies mature, molecules thought to play roles in adhesion, angiogenesis, and tumor growth represent the next wave of prognostic information. For example, B7-H1 is a cell surface glycoprotein that involved in T-cell co-stimulation. Aberrant expression of B7-H1 by RCC tumors may allow tumors to proliferate. Recently, Mayo Clinic investigators demonstrated that RCC marker B7-H1 was associated with higher mortality compared with B7-H1-negative tumors and currently are investigating the potential of targeting B7-H1 therapeutically.
Bradley C. Leibovich, MD, is associate professor of urology and director of urologic oncology at the Mayo Clinic in Rochester, Minn.