Prognostic features of pathologic stage T1 renal cell carcinoma after radical nephrectomy

ADULT UROLOGY

PROGNOSTIC FEATURES OF PATHOLOGIC STAGE T1 RENAL CELL CARCINOMA AFTER RADICAL NEPHRECTOMY WEBER K. LAU, JOHN C. CHEVILLE, MICHAEL L. BLUTE, AMY L. WEAVER,

AND

HORST ZINCKE

ABSTRACT Objectives. To assess the effect of renal cell carcinoma (RCC) subtype, tumor size, and Fuhrman grade on clinical outcome in patients with pathologic T1 (pT1) RCC treated with radical nephrectomy. Methods. Between 1970 and 1998, 840 patients underwent radical nephrectomy for pT1 RCC. Tumors were subtyped and graded. Univariate and multivariate Cox proportional hazards models were fitted to assess the features associated with metastasis-free survival (MFS) and cancer-specific survival (CSS). We identified a range of tumor sizes of clear cell RCC in which a transition occurred from low to high risk. Cox proportional hazards models were then fitted by using size cutoffs. Results. The mean follow-up (⫾ SD) was 9.4 ⫾ 6.6 years among the patients alive at latest follow-up. At 10 years, the CSS and MFS for clear cell RCC (n ⫽ 682) were 89.1% and 88.6%, respectively; for papillary RCC (n ⫽ 122), they were 95.5% and 93.8%; and for chromophobe RCC (n ⫽ 33), they were both 100%. The differences in CSS (P ⫽ 0.013) and MFS (P ⫽ 0.023) between clear cell RCC and the other subtypes were statistically significant. For clear cell RCC, tumor size and Fuhrman grade were independently associated with CSS and MFS (P ⬍0.001). A transition in risk occurred for tumor sizes between 4.5 and 5.0 cm, and the tumor size cutoff of 5.0 cm had the highest concordance index for predicting CSS and MFS. Conclusions. RCC subtype is a strong independent prognostic variable for patients with pT1 RCC treated with radical nephrectomy. For clear cell RCC, Fuhrman grade and tumor size are independently associated with CSS and MFS. UROLOGY 59: 532–537, 2002. © 2002, Elsevier Science Inc.

A

ccording to the 1997 TNM classification of renal cell carcinoma (RCC), pathologic T1 (pT1) tumors are organ-confined tumors up to 7 cm in diameter.1 With the advent of improved radiographic imaging, more cases of renal cancer are diagnosed at this earlier stage. Most of these incidentally detected RCCs are small, but some have spread to regional lymph nodes and distant organs.2 The identification of patients who are at risk of RCC recurrence and progression is of utmost importance to allow close surveillance and early treatment. The prognostic value of tumor size has not been endorsed by all investigators.3–5 However, a survival advantage has been reported among patients who have small RCCs after radical nephrectomy.2,6 –11 Other features of RCC that are important for identifying patients at risk of metastasis From the Departments of Urology, Laboratory Medicine, and Pathology and Section of Biostatistics, Mayo Clinic, Rochester, Minnesota Reprint requests: Horst Zincke, M.D., Department of Urology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905 Submitted: November 10, 2000, accepted (with revisions): November 5, 2001

532

© 2002, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED

and cancer death are RCC subtype and nuclear grade.12–19 We report our outcome with radical nephrectomy for TNM Stage T1 RCC in 840 patients. We also examined the effect of RCC subtype, tumor size, and Fuhrman grade on the clinical outcome of these patients and determined a size cutoff that allows subclassification of pT1 tumors that identifies patients at increased risk of cancer death. MATERIAL AND METHODS Between 1970 and 1998, 840 consecutive patients (532 men, 308 women; mean age ⫾ SD, 63.3 ⫾ 11.3 years) underwent radical nephrectomy for sporadic pT1 RCC at our institution. Patients were excluded if they had lymph node and distant metastasis at presentation, had undergone previous surgical procedures, or had bilateral synchronous tumors, multiple RCC tumors, hereditary RCC, such as von HippelLindau syndrome, familial papillary RCC, or RCC related to tuberous sclerosis. Among the 840 patients with Stage pT1 RCC, 94 (11.2%) had negative results from lymphadenectomy and were considered pT1N0, and the remaining 746 patients were pT1Nx. Patients were considered symptomatic at presentation if they had a palpable mass in the flank or abdomen; discomfort on the ipsilateral or contralateral side; constitutional symptoms, including rash, weight loss, and fatigue; 0090-4295/02/$22.00 PII S0090-4295(01)01589-8

gross hematuria (urine that was brown or red with clots or renal colic); gastrointestinal symptoms, such as early satiety or decreased appetite; or acute onset of left varicocele. A total of 480 patients (57.1%) were symptomatic. All slides were reviewed by one pathologist (J.C.C.) for histologic subtype and grade without knowledge of patient outcomes. Tumor grading was done according to the classification described by Fuhrman et al.19 Fuhrman grade was determined as the worst grade within a tumor; the worst grade was the highest grade occupying at least 1 high power field (400⫻ Olympus BX40 microscope, Olympus Optical, Tokyo, Japan). The 1997 American Joint Committee on Cancer/Union Internationale Contre le Cancer (International Union Against Cancer) (AJCC/IUCC) TNM classification defined pathologic Stage T1.1 Tumor size was determined as the maximal diameter of the surgical specimen of a primary lesion and was available for all surgical specimens. All patients were evaluated postoperatively every 3 to 6 months for the first 2 years and every 6 months thereafter. Follow-up consisted of physical examination, chest radiography, abdominal computed tomography or ultrasonography, blood chemistry panel, and, if indicated, radionuclide bone scanning. The duration of follow-up was calculated from the date of nephrectomy to death or to the latest follow-up. Cancer-specific survival (CSS) and metastasis-free survival (MFS) curves were estimated by the Kaplan-Meier method. Associations of prognostic factors (age, sex, RCC subtype, maximal tumor size, pathologic grade, and preoperative concentrations of serum creatinine) with survival were assessed using Cox proportional hazards models and summarized with risk ratios and 95% confidence intervals (CIs). Multivariate models were fitted using a stepwise variable selection method with a P value of 0.05 for entry into the model. A patient’s risk of death from RCC adjusted for the duration of follow-up was estimated by fitting a Cox model to calculate the martingale residual for each subject. The relationship between a patient’s adjusted risk of death and maximal tumor size was evaluated visually from a scatterplot. A smooth curve was fitted to the data to summarize the pattern of the plot. From this plot, a range of tumor sizes in which a transition occurred from low to high risk was identified. Cox models were then fitted on the basis of dichotomizing tumor size into two groups using cutoffs for tumor size identified from this graphic assessment. The predictive ability of each size cutoff was expressed by a concordance index as described by Harrell et al.20 for time-to-event models with censoring. The concordance index ranges from 0 to 1; a value of 0.5 indicates the same predictive ability as that due to chance alone, and a value of 1 indicates perfect predictive ability. All calculated P values were two-sided, and P ⬍0.05 was considered statistically significant.

RESULTS Among the 840 patients with pT1 RCC, 682 (81.2%) had clear cell (conventional) RCC, 122 (14.5%) had papillary RCC, 33 (3.9%) had chromophobe RCC, and 3 (0.4%) had RCC not otherwise specified. The results are reported separately for clear cell, papillary, and chromophobe RCC. Patient and tumor characteristics are summarized in Table I by histologic subtype. The median follow-up among the patients alive at latest follow-up with clear cell, papillary, and chromophobe RCC was 8.2, 7.8, and 7.1 years, respectively. UROLOGY 59 (4), 2002

Among the 682 patients with clear cell RCC, 62 of the 319 deaths were due to disease (mean time to death 5.8 ⫾ 5.0 years; range 0.5 to 26). Local recurrence was observed in 7 patients, and 69 patients had distant metastasis. At 10 and 15 years, the CSS rate was 89.1% and 87.3%, respectively, and the MFS rate was 88.6% and 84%, respectively (Fig. 1). The associations of the clinical and pathologic features studied with CSS and MFS are summarized in Table II for the patients with clear cell RCC. Tumor size and Fuhrman grade were univariately associated with CSS and MFS (both P ⬍0.001). On the basis of a stepwise multivariate model, both tumor size and Fuhrman grade were independently associated with CSS and MFS (both P ⬍0.001). The relationship between a patient’s adjusted risk of death from RCC (adjusted for duration of follow-up) and the maximal size of tumor is depicted in Figure 2 for patients with clear cell RCC. A transition from low (less than expected) to high (greater than expected) risk was observed for tumor sizes between 4.5 and 5.0 cm. The results of evaluating cutoffs for tumor size by dichotomizing size into two groups are summarized in Table III. The tumor size cutoff of 5.0 cm had the highest concordance index and risk ratio for predicting CSS in patients with clear cell RCC for all patients combined, as well as within the subset with lowgrade tumors (Fuhrman grades 1 and 2) and highgrade tumors (Fuhrman grades 3 and 4). After adjusting for grade (grades 1 and 2 versus grades 3 and 4), the risk ratio for CSS with a cutoff of 5.0 cm was 5.0 (95% CI 2.7% to 9.3%). For patients with tumor grades 1 and 2, the CSS at 10 years was 97.1% for tumors smaller than 5 cm and 85.3% for tumors from 5 to 7 cm. For patients with tumors with Fuhrman grades 3 and 4, the CSS at 10 years was 92.2% for tumors smaller than 5 cm and 58.9% for tumors 5 to 7 cm (Fig. 3). Among the 122 patients with papillary RCC, 54 patients died during follow-up. Local recurrence was observed in 1 patient, and 4 of the 6 patients who developed distant metastases died of tumor. At 10 years, the CSS rate was 95.5% and the MFS rate was 93.8%. The tumor size and Fuhrman grade of each of the 4 patients who died of papillary RCC were as follows: 5 cm (grade 2), 5.5 cm (grade 2), 5.5 cm (grade 3), and 7 cm (grade 3). Of the 33 patients with chromophobe RCC, 14 were dead at latest follow-up. None of the patients with chromophobe RCC had local recurrence or distant metastases. When all patients with pT1 papillary and chromophobe RCC were compared with all patients with pT1 clear cell RCC, patients with clear cell RCC had a significantly worse CSS (P ⫽ 0.013) and MFS (P ⫽ 0.023). 533

TABLE I. Clinical and pathologic characteristics of study patients Characteristic Age (yr) Mean ⫾ SD Range Sex (%) Male Female Kidney (%) Left Right Tumor size (cm) Mean ⫾ SD Median Preoperative serum creatinine (mg/dL) Mean ⫾ SD Median Range Fuhrman grade (%) G1 G2 G3 G4 Follow-up among patients alive at latest follow-up Patients (n) Mean duration ⫾ SD (yr) Median duration (yr) Range of duration (yr)

Clear Cell RCC (n ⴝ 682)

Papillary RCC (n ⴝ 122)

Chromophobe RCC (n ⴝ 33)

63.0 ⫾ 11.3 26–89

64.6 ⫾ 10.4 36–84

64.4 ⫾ 14.0 21–85

410 (60.1) 272 (39.9)

100 (82.0) 22 (18.0)

21 (63.6) 12 (36.4)

314 (46.0) 368 (54.0)

62 (50.8) 60 (49.2)

14 (42.4) 19 (57.6)

4.4 ⫾ 1.5 4.5

4.1 ⫾ 1.6 4.0

4.6 ⫾ 1.5 4.5

1.2 ⫾ 0.8 1.1 0.5–11.2

1.3 ⫾ 1.1 1.1 0.7–12.7

1.3 ⫾ 1.0 1.1 0.7–6.3

122 438 111 11

(17.9) (64.2) (16.3) (1.6)

363 9.7 ⫾ 6.8 8.2 0.01–30.3

5 84 32 1

(4.1) (68.9) (26.2) (0.8)

68 8.5 ⫾ 5.8 7.8 0.5–27.1

2 26 4 1

(6.1) (78.8) (12.1) (3.0)

19 8.0 ⫾ 6.5 7.1 0.01–23.1

KEY: RCC ⫽ renal cell carcinoma.

FIGURE 1. Overall CSS and MFS of patients with Stage pT1 clear cell RCC. Number of patients still at risk at selected points is indicated in parentheses.

COMMENT In our study, RCC subtype was significantly associated with CSS and MFS in patients with pT1 RCC. In addition, tumor size and Fuhrman grade were significantly associated with CSS and MFS in patients with clear cell RCC, and the tumor cutoff of 5 cm identified patients at low and high risk of 534

tumor progression and death from RCC. Pathologic stage, nuclear grade, and histologic RCC subtype are considered important factors associated with prognosis for RCC.6 –27 In 1997, the AJCC/ IUCC increased the tumor size cutoff for pT1 tumors from 2.5 to 7 cm. This change, as well as the significance of tumor size in predicting the behavior of RCC, is controversial. In a recent study of 209 patients with clear cell RCC by Moch et al.,5 no prognostically significant tumor size cutoff was identified in tumors confined to the kidney. Similarly, Targonski et al.7 found no difference in survival for patients with 1992 T1N0M0 and T2N0M0 RCC dichotomized by tumor size cutoffs at 2.5, 5, 7.5, and 10 cm. When patients with lymph node or distant metastases were included in their analysis, a 5-cm cutoff maximized the value of tumor size in predicting survival. Other studies have also shown no significant correlation between tumor size and prognosis.3,4 What is common among all these studies is the small size of the cohorts and the small number of patients who died of RCC, limiting the power to detect differences in outcome among patients with different tumor sizes. Two recent, larger studies that combined all subtypes of RCC in their analyses identified that a tumor size cutoff of UROLOGY 59 (4), 2002

TABLE II. Associations with cancer-specific survival and metastasis-free survival for patients with clear cell renal cell carcinoma Associations with CSS Feature Age at surgery* Male Preoperative serum creatinine† Fuhrman grade 3 or 4‡ Tumor size§

Univariate 1.19 1.15 1.43 5.72 1.70

(0.95–1.51) (0.69–1.94) (0.83–2.45) (3.46–9.47) (1.42–2.05)

Multivariate NS NS NS 4.38 (2.62–7.32) 1.59 (1.31–1.92)

Associations with MFS Univariate 1.02 1.38 1.16 5.37 1.61

(0.83–1.27) (0.83–2.29) (0.66–2.03) (3.32–8.67) (1.36–1.92)

Multivariate NS NS NS 4.18 (2.56–6.81) 1.50 (1.26–1.79)

KEY: CSS ⫽ cancer-specific survival; MFS ⫽ metastasis-free survival; NS ⫽ not statistically significant (P ⬎ 0.05). Data presented as the risk ratio, with the 95% confidence interval in parentheses. * Risk ratio per 10-year increase in age. † Risk ratio per a doubling in the preoperative concentration of serum creatinine; since the distribution of creatinine concentrations was positively skewed, the log2 transformation was applied to the data. ‡ The reference group for the risk ratio is Fuhrman grade 1 or 2. § Risk ratio per 1-cm increase in tumor size.

FIGURE 2. Relationship between a patient’s risk of death from RCC (adjusted for duration of follow-up) and maximal size of tumor in patients with Stage pT1 clear cell RCC. Transition from low (less than expected) to high (greater than expected) risk is shown for tumor sizes between 4.5 and 5.0 cm.

4.0 or 4.5 cm was a better predictor of patient outcome than a cutoff of 7 cm.9,11 Our study supports the finding that a 7-cm tumor size cutoff for pT1 RCC is too large for patients with clear cell RCC, and a tumor size cutoff of 5.0 cm is most predictive of patient outcome. The tumor size cutoff of 5 cm stratifies patients with clear cell RCC into those with low and high risks of tumor progression and death from RCC. The results of our study confirm that the histologic subtype of RCC is a significant predictor of patient outcome. At 10 years, the CSS rate for clear cell RCC was 89.1% and the MFS rate was 88.6% compared with a CSS rate of 95.5% and an MFS rate of 93.8% for papillary RCC and 100% for chromophobe RCC. Of the 122 patients with papillary RCC, 6 developed metastases (4 of these patients died). None of the 33 patients with chromophobe RCC developed metastases. Therefore, because of the low number of patient deaths, no tumor size UROLOGY 59 (4), 2002

cutoff was predictive of outcome in papillary and chromophobe RCC. Also, tumor size appears less important in papillary and chromophobe RCC than in clear cell RCC in predicting tumor behavior. Our findings are similar to those of previous studies that showed that papillary and chromophobe RCCs are less aggressive than clear cell RCC.12,28 Therefore, it is critically important that RCC is subtyped appropriately, and histologic subtype must be included in nomograms that are developed as prognostic tools for patients with RCC. Grading of RCC is another well-established prognostic marker.5,13,15,17–19,22,27 Studies of Fuhrman grade, as well as other nuclear grading systems, have identified tumor grade as an independent variable for determining patient outcome. However, like tumor size, the prognostic significance of grade depends on the histologic subtype. For clear cell RCC, our study showed that patients with high-grade tumors (grades 3 and 4) have a poorer CSS and MFS than patients with low-grade tumors (grades 1 and 2). However, because of the small number of deaths from papillary and chromophobe RCC, we could not identify an association of grade and outcome. Others have suggested that the nuclear grade is of uncertain significance in non-clear cell RCC. We are reviewing our database of more than 3000 RCCs to identify factors of prognostic significance in papillary and chromophobe RCC. CONCLUSIONS RCC histologic subtype is an important prognostic feature. In clear cell RCC, tumor size and Fuhrman grade are strong independent prognostic variables for patients who have pT1 RCC treated with radical nephrectomy. The tumor size cutoff of 5.0 cm is useful for subclassification of T1 tumors into T1a (less than 5 cm) and T1b (5 to 7 cm). On the 535

TABLE III. Tumor size cutoffs overall and by Fuhrman grade for predicting cancer-specific survival for patients with clear cell renal cell carcinoma Overall Size Cutoff (cm) ⬍3.5 ⬍4.0 ⬍4.5 ⬍5.0 ⬍5.5 ⬍6.0

vs. vs. vs. vs. vs. vs.

ⱖ3.5 ⱖ4.0 ⱖ4.5 ⱖ5.0 ⱖ5.5 ⱖ6.0

c Index* 0.62 0.63 0.66 0.72 0.67 0.65

RR (95% CI) 7.78 4.99 3.89 6.11 3.45 3.50

(2.44–24.83) (2.15–11.59) (2.11–7.18) (3.31–11.27) (2.08–5.71) (2.13–5.76)

Fuhrman Grades 1 and 2 c Index* 0.64 0.62 0.65 0.71 0.66 0.66

RR (95% CI) 7.20 3.42 3.12 4.95 3.05 3.32

(1.72–30.13) (1.32–8.85) (1.45–6.72) (2.30–10.66) (1.54–6.04) (1.67–6.59)

Fuhrman Grades 3 and 4 c Index* 0.52 0.55 0.60 0.64 0.58 0.57

RR (95% CI) 2.85 5.31 3.21 4.87 2.28 2.04

(0.39–21.05) (0.72–39.14) (1.11–9.33) (1.67–14.20) (1.05–4.95) (0.98–4.25)

KEY: c ⫽ concordance; CI ⫽ confidence interval; RR ⫽ risk ratio. * The c index summarizes the predictive ability of each cutoff and ranges from 0 (no predictive ability) to 1 (perfect predictive ability); a value of 0.5 indicates predictive ability no different from that due to chance alone.

FIGURE 3. CSS in patients with Stage pT1 clear cell RCC, stratified by tumor size (less than 5 cm versus 5 cm or greater) and Fuhrman grade (low, grades 1 and 2; high, grades 3 and 4).

basis of this study, stratification of patients with T1 clear cell RCC into prognostic subgroups would be useful. It would facilitate better patient counseling, design of appropriate protocol for follow-up, and selection for suitable clinical trials with patients at risk of disease recurrence. REFERENCES 1. Sobin LH, and Fleming ID, for the Union Internationale Contre le Cancer and the American Joint Committee on Cancer: TNM Classification of Malignant Tumors, 5th ed (1997). Cancer 80: 1803–1804, 1997. 2. Tsukamoto T, Kumamoto Y, Takahashi A, et al: Tumor size of renal cell carcinoma: its clinical implication. Urol Int 48: 378 –383, 1992. 3. Nativ O, Sabo E, Raviv G, et al: The impact of tumor size on clinical outcome in patients with localized renal cell carcinoma treated by radical nephrectomy. J Urol 158: 729 –732, 1997. 4. Yoshino S, Kato M, and Okada K: Evaluation of the prognostic significance of microvessel count and tumor size in renal cell carcinoma. Int J Urol 5: 119 –123, 1998. 5. Moch H, Gasser T, Amin MB, et al: Prognostic utility of the recently recommended histologic classification and revised TNM staging system of renal cell carcinoma: a Swiss experience with 588 tumors. Cancer 89: 604 – 614, 2000. 536

6. Guinan PD, Vogelzang NJ, Fremgen AM, et al, for the Members of the Cancer Incidence and End Results Committee: Renal cell carcinoma: tumor size, stage and survival. J Urol 153: 901–903, 1995. 7. Targonski PV, Frank W, Stuhldreher D, et al: Value of tumor size in predicting survival from renal cell carcinoma among tumors, nodes and metastases stage 1 and stage 2 patients. J Urol 152: 1389 –1392, 1994. 8. Lerner SE, Hawkins CA, Blute ML, et al: Disease outcome in patients with low stage renal cell carcinoma treated with nephron sparing or radical surgery. J Urol 155: 1868 – 1873, 1996. 9. Zisman A, Pantuck AJ, Chao D, et al: Reevaluation of the 1997 TNM classification for renal cell carcinoma: T1 and T2 cutoff point at 4.5 cm rather than 7 cm better correlates with clinical outcome. J Urol 166: 54 –58, 2001. 10. Kattan MW, Reuter V, Motzer RJ, et al: A postoperative prognostic nomogram for renal cell carcinoma. J Urol 166: 63– 67, 2001. 11. Igarashi T, Tobe T, Nakatsu HO, et al: The impact of a 4 cm cutoff point for stratification of T1N0M0 renal cell carcinoma after radical nephrectomy. J Urol 165: 1103–1106, 2001. 12. Amin MB, Corless CL, Renshaw AA, et al: Papillary (chromophil) renal cell carcinoma: histomorphologic characteristics and evaluation of conventional pathologic prognostic parameters in 62 cases. Am J Surg Pathol 21: 621– 635, 1997. 13. Medeiros LJ, Gelb AB, and Weiss LM: Renal cell carcinoma: prognostic significance of morphologic parameters in 121 cases. Cancer 61: 1639 –1651, 1988. 14. Mancilla-Jimenez R, Stanley RJ, and Blath RA: Papillary renal cell carcinoma: a clinical, radiologic, and pathologic study of 34 cases. Cancer 38: 2469 –2480, 1976. 15. Gelb AB, Shibuya RB, Weiss LM, et al: Stage I renal cell carcinoma: a clinicopathologic study of 82 cases. Am J Surg Pathol 17: 275–286, 1993. 16. Storkel S, Thoenes W, Jacobi GH, et al: Prognostic parameters in renal cell carcinoma—a new approach. Eur Urol 16: 416 – 422, 1989. 17. Grignon DJ, Ayala AG, el-Naggar A, et al: Renal cell carcinoma: a clinicopathologic and DNA flow cytometric analysis of 103 cases. Cancer 64: 2133–2140, 1989. 18. Green LK, Ayala AG, Ro JY, et al: Role of nuclear grading in stage I renal cell carcinoma. Urology 34: 310 –315, 1989. 19. Fuhrman SA, Lasky LC, and Limas C: Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol 6: 655– 663, 1982. UROLOGY 59 (4), 2002

20. Harrell FE Jr, Lee KL, and Mark DB: Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 15: 361–387, 1996. 21. Strohmeyer T, and Ackermann R: Classic and modern prognostic indicators in renal cell carcinoma: review of the literature. Urol Int 47: 203–212, 1991. 22. Thrasher JB, and Paulson DF: Prognostic factors in renal cancer. Urol Clin North Am 20: 247–262, 1993. 23. Hafez KS, Fergany AF, and Novick AC: Nephron sparing surgery for localized renal cell carcinoma: impact of tumor size on patient survival, tumor recurrence and TNM staging. J Urol 162: 1930 –1933, 1999. 24. Frank I, Blute ML, Weaver A, et al: Stage- and gradespecific long-term outcome prediction after radical nephrec-

UROLOGY 59 (4), 2002

tomy for unilateral renal cell carcinoma. J Urol 163(suppl): 175–176, 2000. 25. Kinouchi T, Saiki S, Meguro N, et al: Impact of tumor size on the clinical outcomes of patients with Robson Stage I renal cell carcinoma. Cancer 85: 689 – 695, 1999. 26. Eschwege P, Saussine C, Steichen G, et al: Radical nephrectomy for renal cell carcinoma 30 mm or less: long-term follow-up results. J Urol 155: 1196 –1199, 1996. 27. Frank W, Guinan P, Stuhldreher D, et al: Renal cell carcinoma: the size variable. J Surg Oncol 54: 163–166, 1993. 28. Crotty TB, Farrow GM, and Lieber MM: Chromophobe cell renal carcinoma: clinicopathological features of 50 cases. J Urol 154: 964 –967, 1995.

537