Victor R. Reuter

This WHO/ISUP system is an attempt to develop as broad a consensus as possible in the classification of urothelial neoplasms, building upon earlier works and classification systems. It is meant to serve as a springboard for future studies that will help refine this classification, thus enabling us to provide better correlation of these lesions with their biologic behavior using uniform terminology.

PURPOSE: We determined the frequency of tumor multifocality in patients with renal cortical tumors, characterized clinical and pathological features associated with multifocality and evaluated its effect on patient survival. MATERIALS AND METHODS: Between July 1989 and July 2002, 1,071 radical nephrectomies were performed at our institution. Specimens were examined grossly and microscopically for multifocal tumors. Preoperative imaging was reviewed to determine whether multifocality was suspected prior to operation. Multivariate analysis was performed to identify clinical and pathological factors associated with multifocality. RESULTS: Of 1,071 radical nephrectomy specimens 57 (5.3%) had pathological evidence of tumor multifocality. Bilateral synchronous renal cortical tumors were present in 6 of the 57 multifocal cases (11%). A total of 19 cases (33%) had evidence of multifocality on preoperative imaging and, therefore, occult multifocality undetected on preoperative imaging was present in 3.5% of radical nephrectomies (38 of 1,071). Primary tumors in the multifocal group were most commonly conventional clear cell carcinoma, followed by papillary carcinoma. Of multifocal cases 74% had the same histological subtype in all tumors. Multivariate analysis demonstrated that bilaterality, papillary subtype, advanced tumor stage and lymph node metastasis were associated with multifocality. At a median follow up of 40.5 months overall survival, disease-free survival, and disease-free probability were not significantly different between the multifocal and unifocal groups. CONCLUSIONS: We report a 5.3% frequency of multifocal renal cortical tumors and a 3.5% frequency of clinically unsuspected multifocal tumors. Multifocality had no apparent effect on recurrence or survival in patients who underwent radical nephrectomy.

We appreciate the comments raised by Drs. Harnden and Southgate. We agree with their statement that “we need to improve our ability to correctly predict the biological behavior of individual tumours.” The authors describe, for example, their use of cytokeratin 20 to predict recurrence and nonprogression of papillary urothelial tumors. While their initial results may be promising, we are all too well aware of the studies that fill up our journals each month retrospectively correlating various markers with tumor behavior. However, it is only a rare biomarker that maintains its powerful prognostication across multiple studies from different institutions that it becomes incorporated in routine patient care. For this reason, we disagree with the comments that “until these (markers) become accepted as intrinsic to international classification systems, their full potential will not be realized.” Incorporating every new promising, but unproven, marker within our classification systems would result in anarchy with each proponent of a specific classification system favoring their own “pet” marker. These classification systems would need to be constantly revised as new markers arise and older ones prove to be unreliable. The authors state that “the need for a consensus classification was paramount decades ago.” We disagree that the need for consensus classifications of tumor is an antiquated concept. In order to determine whether various markers provide independently prognostic information beyond that provided by routine histologic classification, it is necessary for pathologists to use the same terminology when describing various lesions. The purpose of our classification system was not, as the authors suggested, “to replace one morphological classification by another” since a unified classification system for urothelial neoplasms did not exist. While we need to continue to investigate various biomarkers to enhance our prognostication of urothelial neoplasms, until the “magic markers” are discovered that would supplant our hematoxylin and eosin classification, we still need to have a unified approach to the classification of these lesions based on the hematoxylin and eosin appearance. A common language to classify urothelial neoplasms is needed not only to increase the clarity of our research endeavors, but also to improve patient care. Jonathan I. Epstein M.D. Mahul B. Amin M.D. Victor R. Reuter M.D. Fathollah K. Mostofi M.D.