SEARCH

SEARCH BY CITATION

Keywords:

  • kidney neoplasms;
  • renal cell carcinoma;
  • cancer staging;
  • survival analysis

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Authors from Iowa City report on the incidence of RCC; they compared the rate of these tumours at autopsy and felt that the decrease found was a result of better antemortem detection, and an increase with time in the frequency of clinically detected renal cancer.

A study from New York attempted to determine whether size, or transcapsular extension irrespective of size, was more likely to produce an adverse outcome. They analysed their database of 717 such tumours between 1988 and 2002, and found that absolute tumour size was the more significant of the two findings.

OBJECTIVE

To determine which factor was more predictive of adverse outcome in our institutional experience with T2N0M0 and T3N0M0 renal cortical tumours (RCTs) treated surgically, as the current Tumour-Node-Metastasis (TNM) staging system for RCTs differentiates between tumours of >7.0 cm but confined to the renal capsule (T2) and tumours that extend through the renal capsule regardless of size (T3a).

MATERIALS AND METHODS

We analysed our institutional database of surgical urological oncology for all patients with T2N0M0 and T3aN0M0 RCT treated with partial or radical nephrectomy from 1988 to 2002. All patients with preoperative metastasis, bilateral or multifocal tumours, nonsporadic disease or benign histology were excluded from analysis. A follow-up of ≥ 6 months from the time of surgery was required for inclusion. Primary outcomes included local and distant recurrence, and death.

RESULTS

In all, 717 patients had a partial or radical nephrectomy for RCT during the study period. After exclusion criteria were applied, 21 patients with T2N0M0 and 97 with T3aN0M0 tumours were eligible; the median (mean, range) age was 63 (16.6–88.3) years and follow-up 30.5 (40.8, 6–162) months. The estimated 5-year disease-free survival was 68% and 85% for T2N0M0 and T3aN0M0 RCT, respectively (P = 0.002). The 5-year disease-specific survival was 81% and 94% for the T2N0M0 and T3aN0M0 groups, respectively (P = 0.085).

CONCLUSION

Patients with T3aN0M0 tumours appear to have better disease-free and disease-specific survival than those with T2N0M0 disease, which suggests that tumour invasion through the renal capsule is not as significant as the absolute tumour size.


Abbreviations
AJCC

American Joint Commission on Cancer

RCT

renal cortical tumour.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The purpose of cancer staging systems is to predict a patient's prognosis based on clinical and pathological findings. An accurate cancer staging system is extremely valuable for many reasons. With accurate prognostic information, a physician is able to determine suitable treatment options, effectively counsel patients about various treatments and develop an appropriate surveillance scheme after treatment. Physicians rely on cancer staging systems as integral tools in patient care; the ability of a staging system to accurately reflect clinical outcomes is paramount.

The American Joint Commission on Cancer (AJCC) TNM staging system for renal cortical tumours (RCTs) was last modified in 2002, and is based on gross tumour characteristics such as size, local extent, vascular invasion, nodal involvement and presence of metastatic disease [1]. In 1997, the TNM staging system was revised, resulting in a change in the T2N0M0 classification from tumours confined to the kidney and >2.5 cm, to those confined to the kidney and >7.0 cm [2]. The classification of T3aN0M0 RCT has been in place since 1978, when the AJCC TNM staging system was first created. The T3aN0M0 staging classification is defined as a tumour that extends through the renal capsule and into perinephric tissue, but not beyond Gerota's fascia. Currently, the TNM classification scheme implies that invasion through the renal capsule predicts a worse clinical outcome than a tumour of >7.0 cm with the tumour confined within the renal capsule. Robson et al.[3] first proposed that tumour invasion into the perirenal fat was associated with lower survival rates than for tumours confined to the kidney. Using the 2002 TNM staging classification, a small tumour extending through the renal capsule is assigned a higher stage and worse prognosis than a large tumour confined to the kidney. Although recent studies have proposed modifying the staging system by integrating additional pathological and clinical factors, no published study has examined the prognostic validity of the T2N0M0 and T3aN0M0 classifications [4,5]. In an effort to determine whether absolute size or renal capsular invasion is more closely associated with adverse outcome, we reviewed our institution's experience with patients treated surgically for T2N0M0 and T3aN0M0 RCT.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

We reviewed the Institutional Review Board-approved Database of Surgical Urological Oncology for patients with RCTs treated surgically at Columbia University between 1988 and 2002. Renal cortical tumours were classified according to the 2002 TNM staging scheme; T2N0M0 tumours were defined as >7.0 cm in diameter but confined within the renal capsule, and T3aN0M0 tumours as extending through the renal capsule but not beyond Gerota's fascia, independent of size. Tumours were also classified based on histological subtype and Fuhrman nuclear grade.

Exclusion criteria included bilateral tumours, multifocal disease, benign histology, nonsporadic RCC, positive surgical margin on final pathology, and presence of nodal involvement or distant metastatic disease. Patients were followed after surgery with physical examinations, chest radiographs and abdominal and pelvic CT. A follow-up of ≥ 6 months was required for inclusion in the analysis, with follow-up defined as the time from nephrectomy to the date of death or the date of last contact. Survival was analysed using the Kaplan-Meier method, with the log-rank test used with α = 0.05 to determine the statistical significance between T2N0M0 and T3aN0M0 stages for disease-free or disease-specific survival. Survival was also analysed on the basis of Fuhrman grade and tumour size.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

In all, 717 patients had a radical or partial nephrectomy for renal masses at our institution between 1988 and 2002. Patients with benign histology (75), no available tumour stage (20), T1 (298), T3b (87), T3c (10) and T4 tumours (11) were excluded from the analysis. In all, 43 patients had a T2 and 173 a T3a tumour. Patients with preoperative metastasis (22), bilateral tumours (16) or multifocal disease (20) were also excluded from the analysis. An additional 40 patients were excluded as they had insufficient follow-up. After the exclusion criteria were applied, 21 patients with T2N0M0 and 97 with T3aN0M0 RCT were included in the dataset for analysis; the clinical and pathological characteristics are shown in Table 1. There was no statistically significant difference (P > 0.05) between the T2N0M0 and T3aN0M0 groups on the basis of age, sex, surgical technique, histology and Fuhrman grade. Because the T2N0M0 and T3aN0M0 groups were well-matched, observed differences in recurrence and survival rates are more probably explained by tumour stage.

Table 1.  Characteristics of patients with T2N0M0 and T3aN0M0 RCT
CharacteristicT2N0M0T3aN0M0TotalP
  • *

    two-tailed t-test;

  • †chi-square test.

No. of patients2197118 
Median (range) age, years61.0 (36.0–79.1)63.9 (16.6–88.3) 63.8 (16.6–88.3)0.729*
Sex, n (%) 0.635
Male16 (76)66 (68) 82 (69)
Female 5 (24)31 (32) 36 (31)
Surgical technique, n (%) 0.186
Radical20 (95)78 (80) 98 (83)
Partial 1 (5)19 (20) 20 (17)
Histology, n (%) 0.848
Conventional17 (81)73 (75) 90 (76)
Papillary 2 (10)13 (13) 15 (13)
Chromophobe 2 (10)10 (10) 12 (10)
Mixed  1 (1)  1 (1)
Fuhrman grade 0.467
1 3 (14)13 (13) 16 (14)
2 8 (38)49 (51) 57 (48)
3 6 (29)27 (28) 33 (28)
4 2 (10) 1 (1)  3 (3)
Unclassified 2 (10) 7 (7)  9 (8) 
Median (range): size, cm 8.5 (7.4–18.0) 5.3 (1.1–17.0)  
follow-up, months27 (6–113)31 (6–162) 30.5 (6–162) 

The median follow-up was similar in the two groups (Table 1). At the time of analysis, seven T2N0M0 patients (33%) had had a recurrence; one developed a local recurrence, five metastatic disease and one had both. In the T3aN0M0 group, six patients (6%) had disease recurrence, all with distant metastases. In the T2N0M0 group, 13 (62%) had no evidence of disease, four (19%) were alive with disease, three (14%) were dead from RCC and one (5%) from an unknown cause; in the T3aN0M0 group, the respective values were 84 (87%), two (2%) and four (4%), with three (3%) dead from an unrelated cause and four (4%) from an unknown cause.

The estimated 5-year disease-free and disease-specific survival for the two groups are shown in Fig. 1a,b; survival analysis using Fuhrman grade showed that it was a significant predictor of disease-free survival (P = 0.028; Fig. 1c). When tumour size was used as a categorical variable for survival analysis, a threshold of 8.0 cm was a statistically significant predictor of disease-free survival (P = 0.015). When analysed as a continuous variable with Cox regression analysis, tumour size was almost significant as a predictor of disease-free survival (P = 0.076).

image

Figure 1. Kaplan-Meier survival curves: a, the disease-free curves for patients with T2N0M0 (green) and T3aN0M0 (red) RCT; the 5-year disease-free survival was 68% (mean 66 months, 95% CI 39–93) in the T2N0M0 and 85% (mean 139 months, 95% CI 123–154) in the T3aN0M0 group (P = 0.002); b, the disease-specific curves for patients with T2N0M0 (green) and T3aN0M0 (red) RCT. The 5-year survival was 81% (mean 96 months, 95% CI 78–113) in the T2N0M0 and 94% (mean 149 months, 95% CI 136–162) in the T3aN0M0 group (P = 0.085); c, disease-free curves stratified by Fuhrman grade (1, red; 2, green; 3, light red; 4, light green) for 109 patients with T2N0M0 or T3aN0M0 RCC (P = 0.028).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Since its inception, the AJCC TNM staging system has been the most widely used predictor of kidney cancer recurrence and mortality. Given its pivotal role in guiding treatment and surveillance after surgery, the accuracy of the staging system should be periodically reviewed. Although an ideal prognostic model would incorporate all factors known to be predictive of clinical outcome, a comprehensive model is difficult to achieve. Furthermore, as prognostic models become more inclusive and complex, they may become more difficult to use by clinicians. In an effort to improve a practical tool, we chose to work within the framework of the TNM classification system and examine the ability of the T2N0M0 and T3aN0M0 classifications to accurately predict the clinical course of RCC in a modern patient cohort.

There has been debate about the current T2N0M0 and T3aN0M0 classifications for RCC. Levy et al.[6] examined the clinical course of 286 patients with RCTs who had a nephrectomy. Patients with T2N0M0 and T3aN0M0 disease had an equivalent rate of postoperative metastasis and time to development of metastasis. About 27% of patients with T2N0M0 cancer and 30% with T3aN0M0 disease developed metastasis, both with a median time to diagnosis of 32 months. The distinction between lower and higher stages was more pronounced, with 7% of T1 patients and 49% of T3b developing metastasis. More recently, the Memorial Sloan Kettering Cancer Center group reviewed their experience with RCC and developed a postoperative prognostic nomogram [7]. Interestingly, stage T2N0M0 tumours were assigned a greater point value than T3aN0M0 tumours in this nomogram, corresponding to a worse predicted clinical outcome for the T2N0M0 group.

Based on our institution's experience with surgically managed clinically localized RCTs, the current T2N0M0 and T3aN0M0 classification does not accurately predict disease recurrence and death. Indeed, the current classification inversely correlates with risk of recurrence and death. In the 2002 TNM staging system, it is implied that on average, patients with T2N0M0 disease should have a more favourable prognosis and a better chance of cure than those with T3aN0M0 disease. In contrast to this model, the present T2N0M0 patients had a worse clinical course than the T3aN0M0 patients. The 5-year disease-free survival was 68% in the T2N0M0 and 85% in the T3aN0M0 group, and the T2N0M0 patients failed sooner after surgery than the T3aN0M0 patients (P = 0.002). The 5-year disease-specific survival was 81% and 94% in the T2N0M0 and T3aN0M0 groups, respectively (P = 0.085). Although the difference in disease-specific survival was not quite significant, the discrepancy between disease-free and disease-specific survival suggests that currently, the T2N0M0 and T3aN0M0 classifications do not accurately reflect prognosis.

Similar to previously published studies, our analysis showed the prognostic value of Fuhrman grade in predicting clinical outcome [8,9]. When analysing survival based on tumour size alone, we determined that comparing tumours of ≤ 8.0 cm in greatest dimension to those ≥ 8.0 cm gave a significant difference in disease-free survival (P = 0.015). When tumour size was analysed as a continuous variable using Cox regression analysis, tumour size was almost significant as a predictor of disease-free survival.

Although renal capsular invasion is routinely assessed during pathological examination of a nephrectomy specimen, the classification of renal capsular invasion has been questioned [10]. The presence of renal capsular invasion is often difficult to determine accurately and no definitive pathological criteria have been established to aid in this decision process. In this respect, the classification of renal capsular invasion may not be a reliable prognostic variable. The questionable nature of renal capsular invasion status combined with the prognostic value of tumour size suggests that a reclassification of the TNM system relying more on absolute tumour size may be useful.

Our direct comparison of T2N0M0 and T3aN0M0 survival suggests that revision of the current RCT staging system is warranted. As shown by the present analysis, renal capsular invasion did not correlate with a worse prognosis than a larger tumour confined to the kidney. The study has several limitations. The analysis was based on one institution's experience with patients treated surgically for T2N0M0 and T3aN0M0 RCT and subsequently, the final dataset may be limited by its small size and potential lack of power. However, these findings might be confirmed with further analysis at other institutions with a considerable experience in the surgical management of RCC. Despite the inherent limitations of the study, we think that the findings warrant further consideration and reassessment of the accuracy of the T2N0M0 and T3aN0M0 classification in a larger study.

Our findings indicate that further examination of the T2N0M0 and T3aN0M0 classification, as defined by the 2002 AJCC TNM staging system for RCT, is warranted. Collaboration with other institutions would increase the power of the analysis and a more definitive trend may be detected. We conclude that formulating a definitive modification to the RCT staging system accounting for the superior predictive value of absolute size over renal capsular invasion would merit consideration during the next AJCC revision of the TNM classifications.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

A.M.M. and S.M.G. contributed equally in the production of this manuscript. A.M.M. was supported by a Clinical Research Fellowship award from the Doris Duke Charitable Foundation.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES