The 2002 American Joint Committee on Cancer (AJCC) tumor classification for renal cell carcinoma (RCC) does not take into account concomitant venous invasion and extrarenal tumor extension (ERE). In the current study, the authors evaluated the prognostic significance of venous tumor thrombus (VTT) and its extent, the presence and location of ERE, and a combination of both features on survival after the surgical management of patients with pathologic T3 (pT3) RCC.
With Institutional Review Board approval, the institutional nephrectomy database of 3470 patients treated at the University of Texas M. D. Anderson Cancer Center from 1990 to 2006 was searched for pT3 RCC patients who were treated with partial or radical nephrectomy. Patients with nonmetastatic, lymph node-negative RCC and a minimum follow-up of 6 months were included in the analysis.
A total of 419 patients with pT3N0/NxM0 RCC and a mean follow-up of 40.8 months met the study inclusion criteria. In multivariate Cox regression analyses, the 2002 AJCC primary tumor classification was not found to be an independent predictor of cancer-specific mortality. A total of 211 patients with ERE only (50.4%) and 72 patients with VTT only (17.2%) were found to have a similar risk of death from RCC (hazards ratio [HR] of 1.018; P = .957), whereas 136 patients harboring both features (32.5%) were found to be significantly more likely to die from RCC (HR of 2.660; P < .001). The authors proposed a new primary tumor classification in which they grouped patients with both ERE and VTT (which was found to be an independent predictor of cancer-specific survival) into a separate staging category, and demonstrated improved prognostic ability when compared with the 2002 AJCC classification (c indexes of 0.625 vs 0.580, respectively).
Over the past 20 years, the incidence of renal cell carcinoma (RCC), the third most common malignancy of the genitourinary system, has been steadily increasing.1 In addition to the increased radiologic detection of “incidental” renal tumors, the diagnostic rate of advanced tumors and death from disease in patients with RCC has also increased.2 In parallel, treatment paradigms for advanced RCC have also changed. A growing understanding of the biology of RCC, the acceptance of aggressive surgical approaches in the setting of advanced disease, and the development of more effective systemic therapies has altered survival rates.3
Based on peer reviewed observations in the literature, the American Joint Committee on Cancer (AJCC) TNM staging system for classifying RCC has been modified as needed to reflect changes in clinical presentation, diagnosis, and treatment outcomes. The current primary tumor classification categorizes tumors with an invasion of perinephric fat, renal sinus fat, or the ipsilateral adrenal gland as pT3a; tumors with renal vein thrombi or vena cava thrombi below the level of the diaphragm as pT3b; and tumors with thrombi extending above the level of the diaphragm as pT3c.4 Recently, several centers have clearly demonstrated that tumors with direct ipsilateral invasion of the adrenal gland should be staged as pT4 because they behave more aggressively than tumors involving perinephric or renal sinus fat.5–7 Moreover, controversy exists regarding the prognostic significance of the cranial extension of venous tumor thrombus (VTT), the location of extrarenal extension (ERE), and the concomitant presence of both pathologic features.8–13
By examining the prognostic features and cancer-specific outcomes of a large cohort of patients treated within the past 15 years for RCC at a tertiary referral center, this study seeks to reassess and suggest possible improvements to the 2002 AJCC pT3 classification of RCC.
MATERIALS AND METHODS
All studies were undertaken with the approval and institutional oversight of the Institutional Review Board for the Protection of Human Subjects at the University of Texas M. D. Anderson Cancer Center. Covering the years 1990 to the present, the institutional nephrectomy database of 3470 entries was searched for patients with pT3 RCC who were treated with partial or radical nephrectomy. Patients with fewer than 6 months of follow-up, those who underwent incomplete resection, those with non-RCC pathology, or those with metastatic disease at the time of nephrectomy were excluded from the analysis. The presence or absence of synchronous metastases was evaluated routinely with computed tomography, magnetic resonance imaging, and chest X-rays. Additional studies, such as nuclear bone scans and brain imaging, were obtained when indicated.
The clinical and pathologic variables assessed for each patient are shown in Table 1. RCC staging was assigned according to the AJCC 2002 TNM classification.4 Tumor grade was determined using the Fuhrman grading system, whereas tumor histology was classified according to 2004 World Health Organization criteria.14 ERE was defined as tumor infiltrating the perinephric fat directly adjacent to the renal capsule and/or tumor involvement of renal sinus fat. Lymphovascular invasion was defined as the presence of tumor cells within an endothelium-lined space. All pathologic findings were reviewed by experienced genitourinary pathologists specializing in RCC.
Table 1. Clinical and Pathologic Characteristics of 419 Patients with pT3N0/NxM0 RCC
SD indicates standard deviation; LN, lymph node, ERE, extrarenal extension; VTT, venous tumor thrombus.
Mean age ± SD, y
60.2 ± 10.4
Tumor classfication (%)
Treatment modality (%)
LN status (%)
Tumor thrombus level (%)
Location of ERE (%)
Renal sinus fat
Involvement of pelvicalyceal system (%)
ERE only (%)
VTT only (%)
VTT and ERE (%)
Mean tumor size ± SD, cm
8.8 ± 4.2
Fuhrman grade (%)
Sarcomatoid differentiation (%)
Lymphovascular invasion (%)
Follow-up was comprised of a physical examination, serum chemistry evaluation, liver function tests, chest radiography, and abdominal computed tomography, which were performed semiannually for the first 2 years and annually thereafter.
The Fisher exact test and the chi-square test were used to evaluate the association between clinical and pathologic variables. Differences in variables with a continuous distribution across dichotomous or ranked categories were assessed using the Mann-Whitney U test or the Kruskal-Wallis nonparametric analysis of variance, respectively. The Kaplan-Meier method was used to calculate survival functions, and differences were assessed with the log-rank statistic. Univariate and multivariate survival analyses were performed using the Cox proportional hazards regression model. Predictive accuracy was quantified using the Harrell concordance index. Statistical significance in this study was set as P ≤ .050. All reported P values are 2-sided. All analyses were performed with SPSS software (version 13.0; SPSS Inc., Chicago, Ill).
A total of 419 patients who were treated with radical or partial nephrectomy for pT3N0/NxM0 RCC met protocol inclusion criteria. Of these 419 patients, 90 (21.5%) died of RCC, 50 (11.9%) died of other causes, 45 (10.7%) were alive with disease at the time of last follow-up, and 234 (55.8%) were alive with no evidence of disease recurrence. The mean follow-up for the entire cohort was 40.8 months (median, 27.4 months; range, 6.1–178.2 months) and was 41.5 months (median, 28.4 months; range, 6.1–178.2 months) for surviving patients. Clinical and pathologic patient characteristics used in outcome analyses are listed in Table 1. ERE was identified in 347 of the 419 study patients (82.8%) and in 136 of 208 pT3b/c RCC patients (65.4%).
The mean disease-specific survival in 2 patients with pT3N0/NxM0 RCC directly invading the adrenal gland was 24.1 months, and both patients had died of RCC 3 years after undergoing nephrectomy. Because of a small number of patients and based on several reports demonstrating that patients with direct invasion of the adrenal gland have survival comparable to those with tumors involving adjacent organs (pT4 disease), patients with direct adrenal gland invasion were excluded from analyses.5, 7, 15
Kaplan-Meier analyses of the risk groupings in the pT3 2002 AJCC primary renal tumor classification revealed decreased disease-specific survival in pT3b patients compared with pT3a RCC patients. Trend analysis revealed that, taken as a whole, the pT3 2002 AJCC RCC classification did not effectively stratify patients according to their risk of death from RCC (Fig. 1).
In univariate Cox proportional hazards regression analyses of all pT3 patients (Table 2), tumor size, tumor grade, the presence of VTT, and 2002 AJCC primary tumor classification were found to be predictive of death from RCC. It is interesting to note that tumor histology, the presence of sarcomatoid differentiation, lymphovascular invasion, invasion of the pelvicalyceal system, level of VTT (renal vein vs subdiaphragmatic vs supradiaphragmatic), and location of ERE (perinephric fat vs renal sinus fat vs both) were not found to be significant predictors of disease-specific outcome. When adjusted for tumor size, tumor grade, and the presence of VTT, the 2002 AJCC primary tumor classification was no longer correlated with cancer-specific mortality (P = .147).
Table 2. Univariate and Multivariate Cox Regression Analysis of Clinical and Pathologic Features For The Prediction of Disease-Specific Mortality in 419 Patients with pT3N0/NxM0 RCC Treated with Partial or Radical Nephrectomy
Patients with ERE without concomitant VTT (Fig. 2) were found to have similar cancer-specific survival when compared with patients with VTT and no ERE (hazards ratio [HR] of 1.018; P = .957), whereas patients harboring both pathologic features had statistically inferior cancer-specific outcomes (HR of 2.660; P < .001) when compared with patients with VTT or ERE alone. A new primary tumor classification that combines patients with VTT only and patients with ERE only into 1 prognostic grouping and patients with both VTT and ERE into a separate category was found to effectively stratify patients into statistically significantly different risk groups for the prediction of death from RCC (Fig. 3). Application of the Harrell concordance index to the proposed pT3 RCC classification demonstrated improved predictive ability when compared with tumor classification within the 2002 TNM system (c indexes of 0.625 vs 0.580, respectively). In a multivariate Cox proportional hazards regression analysis, which adjusted for the effects of tumor size, tumor grade, and the presence of VTT, the proposed primary tumor classification was found to be the only independent predictor of cancer-specific survival (Table 3).
Table 3. Proposed Primary Tumor Classification in Multivariate Cox Regression Analysis of Clinical and Pathologic Features for the Prediction of Disease-Specific Mortality in 419 Patients With pT3N0/NxM0 RCC Treated with Partial or Radical Nephrectomy
Within the emerging framework of nomograms and algorithms for the prediction of overall survival and cancer-specific survival, the pathologic stage of RCC remains the most important determinant of disease-specific outcome.16 Reflecting changing trends in RCC presentation, treatment, and outcomes, the AJCC TNM tumor classification has been continuously modified to improve its accuracy. Recently, it has been suggested that the current stratification of patients within the pT3 RCC classification does not stratify patients according to disease-specific survival.15, 17, 18 Several large, single-center experiences have clearly established that cancer-specific survival is significantly lower for patients with direct invasion of the adrenal gland compared with patients with pT3a and pT3b RCC. Moreover, there was no significant difference in survival noted between patients with direct adrenal invasion and T4 disease.5–7 Although the small number of patients in the current study precludes meaningful statistical analysis, we confirmed these observations; 2 patients with pT3N0/NxM0 RCC and direct invasion of the adrenal gland who were identified in our cohort had a mean survival of 24.1 months and none were still alive at the time of last follow-up, 3 years after nephrectomy.
We have found that within the pT3N0/NxM0 cohort of patients, the presence of VTT but not its extent (renal vein vs subdiaphragmatic vs supradiaphragmatic) was independently correlated with cancer-specific survival. These findings are supported by single institution reports by Blute et al. and Kim et al., who demonstrated that the risk of cancer recurrence and cancer-specific death in patients with localized RCC and venous involvement did not differ when stratified by the level of VTT.9, 10
With regard to the prognostic significance of ERE location, the data from the current study do not parallel the findings of Thompson et al., who reported significantly decreased cancer-specific survival among 43 patients with invasion of the renal sinus fat (21%) compared with 162 patients in whom only the perinephric fat was invaded (79%).13 These differences may be explained by a larger cohort of patients and a higher frequency of renal sinus fat invasion in the current series. In agreement with the findings of Bonsib et al.,12 we documented the presence of renal sinus fat invasion in 55% of patients with ERE, whereas Thompson et al. identified renal sinus fat invasion in only 20% of their pT3a cohort.13
Although it has been demonstrated previously that ERE portends a worse prognosis among patients with VTT, the current AJCC primary tumor classification for RCC does not differentiate between patients with VTT only and patients with both VTT and ERE. We found that patients with ERE only and patients with VTT without concomitant ERE were at a similar risk of death from RCC and had overlapping 5-year and 10-year disease-specific survival rates. Conversely, 136 patients with both VTT and ERE (33%) were found to be at a significantly increased risk of death from RCC.
Insight into the significance of concomitant ERE and VTT served as a basis for a staging system proposed by Thompson et al. in which patients with locally advanced RCC were reclassified into 5 different subgroups: pT3aN (VTT level 0 without ERE), pT3bN (ERE only), pT3cN (VTT level 0 with ERE or VTT levels I–III without ERE), and pT3dN (VTT levels I–III with ERE or VTT level IV).17 Likewise, based on a multicenter European experience, Ficarra et al. proposed a 3-tier classification system for locally advanced RCC: pT3aN (ERE or VTT below the diaphragm), pT3bN (VTT below the diaphragm associated with ERE), and pT4N (adrenal gland invasion or invasion of Gerota fascia or VTT above the diaphragm).15 Derived from observations in our patient cohort, the pT3 RCC classification proposed in the current study is much simpler in its clinical application and demonstrates significantly improved prognostic accuracy when compared with the 2002 AJCC primary tumor classification. Moreover, the concordance index of our proposed pT3 classification is comparable to a more complex system proposed by Thompson et al.17 Unlike the current AJCC RCC staging system, our proposed classification system retains its independent predictive value in multivariate analysis of prognostic factors shown to be significant in patients with nonmetastatic, lymph node-negative pT3 RCC.
One limitation of the current study merits discussion. Pathology specimens obtained from patients in the current series were not submitted for reevaluation; consequently, important prognostic factors such as tumor necrosis and lymphovascular invasion could not be assessed in all patients. Nonetheless, variables such as sarcomatoid differentiation and pelvicalyceal and renal sinus invasion, which to our knowledge are not available in other series, were recorded prospectively by genitourinary pathologists at our center.
A multicenter collaborative effort incorporating multiple institutions with significant experience in the treatment of RCC patients is needed to validate the prognostic significance, reliability, and reproducibility of the new pT3 classification system proposed in the current study.
A survival analysis of a large group of patients who underwent radical nephrectomy for RCC within the modern era suggests that the current stratification within the pT3 staging category does not correlate with disease-specific outcome. The results of the current study demonstrate that the concomitant presence of ERE and VTT identifies a subgroup of patients with inferior outcomes compared with patients harboring only 1 of these pathologic features. A clinically simple reclassification within the pT3 RCC staging category, proposed herein, demonstrates significantly improved prognostic accuracy and retains its independent predictive value among prognostic factors shown to impact disease-specific outcome in RCC patients.