Author reply.

Abstract

With interest we read the comments of de Manzoni et al. We wish to thank them for their response, and by this means adding to our list of references. Indeed we have quoted two studies from the same group, but we believe that when a novel classification system is applied for the first time, it is not only important to determine whether this system has prognostic value, but also whether it is reproducible (interobserver or intraobserver agreement). Therefore, we felt the need to refer to both of these articles.1, 2 With regard to their second point, the data published in Chirurgia Italiana are not readily available to us. It remains interesting to learn that de Manzoni et al. found an independent prognostic value for the Lauren classification, although the original publication did not and neither have we. Concerning the Goseki classification, it is not surprising to see that it is associated with the Lauren classification; the Goseki classification is based on two aspects of gastric carcinoma, namely tubular differentiation and the amount of intracytoplasmic mucus. The Lauren classification differentiates between intestinal and diffuse type, macroscopically meaning well and poor tubular differentiation. Roy et al.3 found no prognostic value for the Goseki classification, in agreement with the authors4 of the comments. On this point we should like to point out that our data were collected prospectively (especially regarding survival), based on a trial with a clear protocol definition and strict quality control,5 whereas their data were not.6 Moreover, our findings are in agreement with the original reports of Lauren7 and Goseki et al.8 We also have demonstrated our results (scoring according Goseki classification) to be reproducible.9 With regard to the third point: we also performed a univariate analysis in which the Goseki classification had prognostic value, whereas the Lauren classification did not. Even though confounding factors may play a role in the multivariate analysis, we believe because the Lauren and Goseki systems were colinear, our multivariate analysis is valid. There is still no consensus regarding which histologic classification is the most important prognosticator for gastric carcinoma. The fact that so many systems are in use is a clear sign that none is perfect or at least good enough for clinical use. Although the Goseki classification system is new and promising, its place remains to be determined. Ilfet Songun M.D., Ph.D.*, Johannes H. J. M. van Krieken M.D., Ph.D. , * Department of Surgery, Leiden University, Medical Center, Department of Pathology, Leiden University, Medical Center, Leiden, The Netherlands