Is There a Need to Further Subclassify pt2 Renal Cell Cancers as Implemented by the Revised 7th TNM Version?

EUROPEAN UROLOGY 59 (2011) 258 263 available at www.sciencedirect.com journal homepage: www.europeanurology.com Kidney Cancer Is There a Need to Further Subclassify pt2 Renal Cell Cancers as Implemented by the Revised 7th TNM Version? Sandra Waalkes a,1, Frank Becker b,1, Andres J. Schrader c,d,1, Martin Janssen b, Gerd Wegener e, Axel S. Merseburger a, Mark Schrader c, Rainer Hofmann d, Michael Stöckle b, Markus A. Kuczyk a, * a Department of Urology and Urologic Oncology, Hannover University Medical School, Hannover, Germany b Department of Urology, University Clinics of Saarland, Homburg/Saar, Germany c Department of Urology, Ulm University Medical School, Ulm, Germany d Department of Urology, Philipps University, Marburg, Germany e Clinical Cancer Registry, Hannover University Medical School, Hannover, Germany Article info Article history: Accepted October 5, 2010 Published online ahead of print on October 15, 2010 Keywords: Renal cell cancer Prognosis Survival TNM 2010 Abstract Background: The recently modified TNM classification of renal cell carcinoma (RCC) (7th edition) has implemented a subdivision of pt2 tumours into stage pt2a (>7 or 10 cm) versus pt2b disease (>10 cm). Objective: Our aim was to evaluate whether this subdivision of pt2 RCC is justified due to a clinical prognosis divergence between the two groups (pt2a vs pt2b) Design, setting, and participants: In total, 5122 patients were subjected to either radical nephrectomy or nephron-sparing surgery at three centres in Germany (University Hospitals of Hannover, Homburg/Saar, and Marburg). Patients were reclassified into stage pt2a and pt2b according to the maximum tumour diameter as suggested by the 7th revised version of the TNM classification system. Measurements: The t test and Fisher exact test were applied to evaluate the comparability of the two groups (pt2a vs pt2b) regarding several additional patients and tumourspecific characteristics of known prognostic relevance for RCC. Univariable (Kaplan-Meier analysis) and multivariable statistical analyses (Cox proportional hazards regression model) were applied to identify a possible difference between the two groups (pt2a vs pt2b) regarding cancer-specific survival (CSS). Results and limitations: Applying the new TNM classification, 579 previously pt2-staged patients were divided into 445 (76.9%) with pt2a and 134 (23.1%) with pt2b tumours. Kaplan-Meier curves revealed no significant difference in CSS between pt2a and pt2b patients; 5-yr CSS was 79.0% and 74.1%, respectively (p = 0.38). When applying multivariable analysis, unlike tumour grade and N/M status, pt2 subclassification failed to independently predict survival in RCC patients. Conclusions: The new subclassification of pt2 RCC into two different subgroups as suggested by the latest modification of the TNM system does not yield additional/ prognostic information. # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. 1 These authors contributed equally to this publication. * Corresponding author. Department of Urology, Hannover Medical School, Carl-Neuberg-Str. 1, D-30625 Hannover, Germany. Tel. +49 511 532 3650; Fax: +49 511 532 5634. E-mail address: kuczyk.markus@mh-hannover.de (M.A. Kuczyk). 0302-2838/$ see back matter # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2010.10.005

EUROPEAN UROLOGY 59 (2011) 258 263 259 1. Introduction The Union Internationale Contre le Cancer and the American Joint Committee on Cancer (AJCC) proposed the first TNM staging system for renal cell carcinoma (RCC) in 1978 [1]. The TNM criteria define the anatomic extent of disease and stage. They have been shown to correlate with prognosis and to provide important information for optimal patient management. On the basis of new evidence, the TNM system was subsequently modified to improve its prognostic accuracy in 1987, 1997, 2002, and, most recently, in 2010 [1 4]. Because primary tumour size is a key component of the TNM staging system, it remains one of the most important prognostic factors for RCC [5]. The prognostic significance of tumour diameter for localised disease has therefore been investigated in a number of previous studies [6 10]. As a consequence, many modifications of the TNM staging system were primarily made to take different tumour size cut-off points into account. The current 7th TNM edition was in part established on data from a study by Frank et al, who evaluated the optimal cut-off point to stratify localised RCC [6]. This led to an additional substratification of T2 tumours into T2a (>7 but 10 cm) and T2b (>10 cm). The aim of this large retrospective multicentre study was to evaluate the new TNM staging system with respect to delineation of stage T2a versus T2b tumours on patients clinical outcome following surgery. 2. Patients and methods 2.1. Patient selection and pathologic features A total of 5122 RCC patients treated from 1975 to 2009 by radical nephrectomy or nephron-sparing surgery at three university hospitals in Germany (Departments of Urology in Hannover, Homburg/Saar, and Marburg) were included. Information on patients and tumour characteristics, such as age, sex, tumour diameter, stage, presence of regional lymph node or distant metastases, histologic subtype, and Fuhrman grade, was obtained from institutional databases. Tumours were reclassified according to the current TNM version by recognising their largest diameter (in centimetres) as indicated in the pathologic reports (pt2a: tumour diameter >7 cm but 10 cm; pt2b: >10 cm). 2.2. Follow-up The duration of the follow-up was calculated from date of surgery to the date of death or last follow-up. Death was assessed as either cancer related or unrelated. The primary end point of this study was cancerspecific survival (CSS). Information about the exact date as well as cause of death for each patient was received from the patient s general practitioner, a close family member, or the patient s hospital records if she or he had been followed up or died in one of our institutions. Followup assessment ended in November 2009. Until then, all patients data were updated regularly at least every 6 mo. 2.3. Statistical methods Continuous variables were reported as mean value and standard deviation (SD) or median value and interquartile ranges (IQRs) in the case of parametric or nonparametric distribution, respectively. Normal distribution of continuous parameters such as age, tumour size, and follow-up period were tested using the Kolmogorov-Smirnov test. Fisher exact tests were conducted to assess the differences in covariate distributions between pt2 categories. Age and tumour size were correlated applying the Spearman test. Kaplan-Meier estimates of survival time were calculated, and subgroups were compared with the log-rank test statistic. Multivariable Cox regression models were used to assess the association between survival and pt2 subgroups adjusted for different clinical and patient covariates (i.e., age, sex, tumour grade, histologic subtype, metastatic status). SPSS 17.0 (SPSS Inc, Chicago, IL, USA) was used for statistical assessment. In all tests, a two-sided p value <0.05 was considered to indicate significance. 3. Results 3.1. Patient and tumour characteristics In total, 579 of 5122 patients (11.3%) were classified as harbouring RCCs of stage pt2 (76.3% clear cell, 9.7% papillary, and 3.5% chromophobe subtypes; 10.5% unclassified or other RCC subtype). Following reclassification according to the 7th TNM edition, 445 patients (76.9%) and 134 patients (23.1%) were classified as pt2a or pt2b, respectively (Table 1). The mean age for the total cohort was 59.3 yr (12.7). In total, 353 (61%) were male and 226 (39%) were female. The mean tumour size of all patients was 9.5 cm (2.9 cm): 8.2 cm and 13.6 cm in the subgroups of patients with pt2a and pt2b tumours, respectively. Patients revealing tumours classified as pt2b were significantly younger than those with pt2a disease (mean: 57 vs 60 yr; p = 0.015; t test); age correlated negatively with tumour size (r = 0.09; p = 0.025; Spearman). Using univariable analyses, the two groups were comparable in terms of the remaining patients and tumour characteristics of potential prognostic relevance: Neither sex nor histologic subtype, nodal status/distant metastasis, or tumour grade correlated with the newly defined pt2 subgroups (Table 1). 3.2. Survival analysis The median follow-up for the entire cohort was 49 mo (IQR:18 102mo):56mo(IQR:21 109mo)and40mo (IQR: 11 87 mo) for pt2a and pt2b tumours, respectively (p = 0.003; Mann-Whitney test). The median follow-up for all patients alive and free of recurrence (n =302)was64 mo (IQR: 27 119 mo): 73 mo (IQR: 29 128 mo) and 43 mo (IQR: 19 87 mo) for pt2a and pt2b tumours, respectively (p =0.001). At the time of the last follow-up evaluation, 124 patients (23.4%) had died due to tumour progression: 95 patients (21.3%) and 29 patients (21.6%) with pt2a or pt2b disease, respectively. In addition, 2.9% and 5.2%, respectively, of all patients experienced recurrence but were still alive (p = 0.75; Fisher exact test). Applying Kaplan-Meier analysis to compare the two pt2 subgroups, no significant difference in CSS could be demonstrated (p = 0.38; log-rank test). The calculated 5-yr CSS for patients with pt2a and pt2b RCC was 79.0% and 74.1%, respectively (Fig. 1a).

260 EUROPEAN UROLOGY 59 (2011) 258 263 Table 1 Patient and tumour characteristics * Variable pt2a pt2b RR 95% CI p value Test Total (%) 445 (76.9) 134 (23.1) Age, yr, mean ( SD) 60.0 (12.1) 56.7 (14.5) 0.7 6.1 0.015 t test Sex, No. (%) 0.76 Fisher exact Male 273 (61.3) 80 (59.7) 1 Reference Female 172 (38.7) 54 (40.3) 1.05 0.78 1.43 Histology, No. (%) 0.064 Fisher exact Clear cell 347 (78.2) 94 (70.1) 1 Reference Non clear cell 97 (21.8) 40 (29.9) 1.37 0.99 1.88 LN metastases, No. (%) y 0.83 Fisher exact N 421 (94.6) 126 (94.0) 1 Reference N+ 24 (5.4) 8 (6.0) 1.09 0.58 2.0 Pulmonary/visceral metastasis, No. (%) y 0.75 Fisher exact M 396 (89.0) 121 (90.3) 1 Reference M+ 49 (11.0) 13 (9.7) 0.9 0.54 1.49 Metastasis (N and/or M+), No. (%) y 0.78 Fisher exact N, M 379 (85.2) 116 (86.6) 1 Reference N+ and/or M+ 66 (14.8) 18 (13.4) 0.91 0.59 1.42 Grade, No. (%) 0.59 Fisher exact G1 61 (13.7) 23 (17.2) 1 Reference G2 270 (60.7) 74 (55.2) 0.79 0.53 1.18 0.25 G3 92 (20.7) 29 (21.6) 0.88 0.55 1.4 0.63 G4 21 (4.7) 8 (6.0) 1.0 0.51 2.0 1.0 CI = confidence interval; M = distant visceral/pulmonal disease; N = nodal disease; RR = relative risk; SD = standard deviation. * Comparability of both groups (pt2a vs pt2b) according to all variables evaluated except for patients age (patients with tumours classified as pt2b were significantly younger compared with those diagnosed with pt2a disease). y At time of renal surgery. Because the presence of metastasis at diagnosis and/or at the time of surgery is surely the most important confounder concerning overall survival (Table 2), we subsequently compared the CSS of the two pt2 subcategories excluding those patients with nodal and/or distant metastasis. Interestingly, adjusting for metastasis at the time of surgery, we still could not show significant differences in CSS Table 2 Multivariable analysis of patients and tumour characteristics with regard to their prognostic impact on cancerspecific long-term survival (Cox regression analysis) Variable HR 95% CI p value Age, yr * 1.02 1.0 1.04 0.02 Sex 0.06 Female 1 Reference Male 1.46 0.99 2.17 Grade 0.06 G1 1 Reference G2 1.41 0.76 2.62 0.28 G3 2.20 1.13 4.29 0.02 G4 2.10 0.76 5.64 0.14 Metastasis y <0.001 N0/M0 1 Reference N+ and/or M+ 5.49 3.70 8.15 Histology z 0.2 n-ccrcc 1 Reference ccrcc 0.75 0.49 1.16 pt2 subdivision 0.49 pt2a 1 Reference pt2b 1.16 0.76 1.79 CI = confidence interval; HR = hazard ratio. * Continuous variable. y At time of renal surgery: N1M0, N1M1, N0M1. z Clear cell renal cell carcinoma (ccrcc) versus other histologic classifications (n-ccrcc). between pt2 subcategories (5-yr CSS: 85.2% and 80.1%, respectively; p = 0.1, log-rank test; Fig. 1b). Applying univariable analysis, in contrast to sex (p = 0.06), T2 subcategories (p = 0.38), and histologic subtype (p = 0.46), age (p = 0.02, favouring younger patients), nodal (p < 0.001) or visceral involvement (p < 0.001), and tumour grade (p = 0.001; Cox regression) correlated with CSS. In contrast, employing multivariable analysis including these potential prognostic factors for patients with kidney cancer, only age (favouring younger patients; p = 0.02, Cox regression analysis) and metastatic disease (N+ and/or M+; p < 0.001) could be confirmed as significant independent prognostic markers for pt2 RCC patients (Table 2). The classification of patients into pt2 subgroups (T2a vs T2b) as recommended by the latest TNM edition again did not qualify as a variable of independent prognostic importance (p = 0.49; Cox regression). Using multivariable analysis and focussing on the subgroup of patients with localised disease (pt2, N0, M0), unlike age (p = 0.001), pt2 subclassification (p = 0.10) again could not be identified as an independent prognostic marker for tumour-specific survival. However, applying multivariable analysis including all patients and exchanging pt2 subclassification with tumour size as a continuous variable, in addition to age (p = 0.01), the tumour diameter was identified as an independent prognostic parameter (p = 0.01). 4. Discussion Tumour size is a key component of the TNM staging system and remains one of the most important prognostic factors for RCC. Through the years the TNM system has been

[()TD$FIG] EUROPEAN UROLOGY 59 (2011) 258 263 261 Fig. 1 Clinical outcome did not correlate with pt2 subcategories (Kaplan-Meier method). (a) The 5-yr cancer-specific survival (CSS) rate for the whole cohort of pt2 patients was 78.0% and did not differ significantly between pt2a and pt2b tumours (79.0% vs 74.1%; p = 0.38; log-rank test). (b) For the subgroup of patients without regional lymph node or distant metastases (N0/M0), the 5-yr CSS rates were 85.2% and 80.1% for stage pt2a and pt2b, respectively (p = 0.10; log-rank test). revised repeatedly. In the 1987 edition of the TNM classification [3], the cut-off point separating T1 from T2 was set at 2.5 cm. This change was based on the assumption that smaller tumours have a very low risk of metastases. However, this categorisation was found to be limited because only a few tumours were <2.5 cm [7]. Several additional studies showed prognostic significance for various other cut-off points [8 11], irrespective of the TNM classification. Ficarra et al proposed a tumour size break point of 5.5 cm for patients with organ-confined disease [12]. In 2007, Klatte et al [13] evaluated a group of 706 patients with pt2 RCC and identified 11 cm as the ideal tumour size cut-off. Patients with a tumour size of 11 cm had a significantly longer CSS compared with patients with tumours >11 cm. However, according to Karakiewicz et al and Moch et al, tumour size could be shown to be

262 EUROPEAN UROLOGY 59 (2011) 258 263 significantly associated with outcome when modelled continuously, irrespective of specific cut-off points [14,15]. The probability of death increased 3.5 times for each doubling of tumour size [16]. In the 7th edition of the AJCC Cancer Staging Manual, the results of a total of 37 166 patients (for 2001 2002) from the National Cancer Data Base (NCDB; http://www.facts. org/cancer/ncdb/index.html) regarding the impact of size on CSS of T2 patients are presented. The 5-yr CSS for organconfined tumours >7 cm but 10 cm was 57% compared with 47.5% for tumours with a diameter >10 cm. With a median 5-yr-survival rate of 79.0% for T2a patients versus 74.1% for T2b, we were not able to confirm a significant difference in survival for patients with <10 cm or >10 cm T2 RCC. Furthermore, patients with pt2 RCC in our study population seemed to live significantly longer than the patients included in the NCDB database. The rationale for changes in the TNM system is usually based on new evidence from large multicentre clinical studies. In contrast, the current T2 subclassification was based on a study published by Frank et al [6] in 2005 analysing 544 patients from the Mayo Clinic. Their data indicated that patients with tumours >7 cm and <10 cm had a worse outcome than those with pt1b RCC. Furthermore, patients with tumours 10 cm had a worse prognosis than those with 7- to 10-cm tumours but a better outcome than those with pt3a RCC. Only recently, including patients from 16 different centres, Novara et al [5] were able to validate the 7th edition of the TNM staging system for RCC as published in this journal. They concluded that the latest changes of the primary tumour staging system for kidney cancer are a powerful predictor of CSS. Regarding the subclassification of pt2 tumours, they observed a significant difference in 5-yr CSS of 85.4% and 70% for patients with pt2a and pt2b RCC, respectively. In contrast to these two studies with a similar size and a slightly shorter follow-up time, we were not able to confirm a significant difference between pt2a and pt2b tumours. We cannot explain why our results differ from the previously published studies by Novara et al and Frank et al [5,6]. However, there are several differences in patients and tumour characteristics that might account for the different outcome. In our trial the mean tumour size for all pt2 tumours was 9.4 cm compared with 10.4 cm in the study published by Frank et al [6]. Mean tumour sizes of the pt2 subgroups investigated by Frank et al and Novara et al have not been specified in their publications. It could be conceivable that the meantumourdiameterof pt2a tumours might have been bigger and that of pt2b tumours smaller in our trial. Comparing the number of patients in each pt2 subgroup in the study by Frank et al [6], the pt2 subgroups, surprisingly, were almost equal in size (49.4% pt2a and 50.6% pt2b). In contrast, in the study by Novara et al [5] (74.1% pt2a and 24.9% pt2b) and in our analysis (76.9% pt2a and 23.1% pt2b), the distribution between the pt2 subgroups was very similar, with almost three times more patients in thept2agroup.afurther point is that the follow-up periods differed between the two pt2 subgroups. We cannot completely rule out that this fact might have influenced our results. However, neither Novara et al nor Frank et al [5,6] indicated the follow-up period for their pt2 subgroups. Hence a potential influence of the length of follow-up periods in Frank s and Novara s studies as an explanation for different outcomes of those trials and our own cannot be ruled out either. Our study had several important limitations. First and foremost are the limitations inherent in retrospective analysis, the lack of central pathologic review, and a relatively small number of pt2 patients. Thus we cannot completely rule out that our study was underpowered to observe different outcomes between patients with pt2a and pt2b disease. This is supported by the fact that tumour diameter as a continuous variable is significantly associated with CSS in multivariable analysis, in contrast to the novel T2 subcategories. A further limitation was that patients revealing tumours classified as pt2b were significantly younger than those with pt2a disease. Because age was included in the multivariable analysis, it is unlikely that this fact might have influenced our results concerning the significance of T2 subdivision. 5. Conclusions The current study evaluates the prognostic importance of the new subdivision of pt2 RCC. It was not possible to confirm a substantially different outcome for the substratification of T2 into T2a and T2b in this large patient collective. Although our data were generated by the evaluation of a database that should be representative because of its size, additional investigations should be awaited before the new subdivision of pt2 RCC is abandoned. Author contributions: Markus A. Kuczyk 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: Kuczyk, Waalkes, Schrader. Acquisition of data: Waalkes, Schrader, Becker, Janssen, Wegener, Stöckle, Hofmann. Analysis and interpretation of data: Schrader, Waalkes. Drafting of the manuscript: Waalkes, Schrader, Becker, Kuczyk. Critical revision of the manuscript for important intellectual content: Kuczyk, Schrader, Merseburger. Statistical analysis: Schrader. Obtaining funding: None. Administrative, technical, or material support: Wegener. Supervision: Kuczyk. 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. Acknowledgement statement: We thank Dr. Jens Dreyhaupt, Institute for Medical Biometrie, Ulm University, for supervising our statistical computations.

EUROPEAN UROLOGY 59 (2011) 258 263 263 References [1] Hermanek P, Henson DE, Hutter RVP, Sobin LH, editors. TNM supplement. A commentary on uniform use. Berlin, Germany: Springer; 1993. [2] Greene FL, Page DL, Fleming ID, Fritz A, Balch CM, editors. AJCC cancer staging manual. ed 6. New York, NY: Springer; 2002. [3] Sobin LH, Hermanek P, editors. TNM classification of malignant tumours. ed 4. Berlin, Germany: Springer; 1987. [4] Sobin LH, Wittekind C, editors. TNM classification of malignant tumours. ed 5. New York, NY: Wiley-Liss; 1997. [5] Novara G, Ficarra V, Antonelli A, et al. Validation of the 2009 TNM version in a large multi-institutional cohort of patients treated for renal cell carcinoma: are further improvements needed? Eur Urol 2010;58:588 95. [6] Frank I, Blute ML, Leibovich BC, et al. pt2 classification for renal cell carcinoma. Can its accuracy be improved? J Urol 2005;173:380 4. [7] Moch H, Gasser T, Amin MB, Torhorst J, Sauter G, Mihatsch MJ. Prognostic utility of the recently recommended histologic classification and revised TNM staging system of renal cell carcinoma: a Swiss experience with 588 tumors. Cancer 2000;89:604 14. [8] Gettman MT, Blute ML, Spotts B, Bryant SC, Zincke H. Pathologic staging of renal cell carcinoma: significance of tumor classification with the 1997 TNM staging system. Cancer 2001;91:354 61. [9] Hafez KS, Fergany AF, Novick AC. Nephron sparing surgery for localized renal cell carcinoma: impact of tumor size on patient survival, tumor recurrence and TNM staging. J Urol 1999;162: 1930 3. [10] Kinouchi T, Saiki S, Meguro N, et al. Impact of tumor size on the clinical outcomes of patients with Robson State I renal cell carcinoma. Cancer 1999;85:689 95. [11] 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 rather than 7 cm. better correlates with clinical outcome. J Urol 2001;166:54 8. [12] Ficarra V, Guille F, Schips L, et al. Proposal for revision of the TNM classification system for renal cell carcinoma. Cancer 2005;104: 2116 23. [13] Klatte T, Patard JJ, Goel RH, et al. Prognostic impact of tumor size on pt2 renal cell carcinoma: an international multicenter experience. J Urol 2007;178:35 40, discussion 40. [14] Karakiewicz PI, Lewinshtein DJ, Chun FK-H, et al. Tumor size improves the accuracy of TNM predictions in patients with renal cancer. Eur Urol 2006;50:521 9. [15] Moch H, Artibani W, Delahunt B, et al. Reassessing the current UICC/AJCC TNM staging for renal cell carcinoma. Eur Urol 2009;56:636 43. [16] Delahunt B, Kittelson JM, McCredie MR, Reeve AE, Stewart JH, Bilous AM. Prognostic importance of tumor size for localized conventional (clear cell) renal cell carcinoma: assessment of TNM T1 and T2 tumor categories and comparison with other prognostic parameters. Cancer 2001;94:658 64.