Are epidemiological data on lymphoma incidence comparable? Results from an application of the coding recommendations of WHO, InterLymph, ENCR and SEER to a cancer registry dataset

Purpose: The REAL classification of 1994 and the subsequent WHO classification of 2001 can be considered a breakthrough of international harmonization of lymphoma characterization, terminology and codification. These efforts promised to produce internationally comparable cancer registry data in the future. However, in practice discrepancies of usage of these classifications occurred which hamper comparability of registration outcome and must be taken into account by epidemiologic research. Methods: In order to analyze such discrepancies, we used the assignment recommendations of the World Health Organisation 2008, InterLymph 2010, European Network of Cancer Registry 2009 and Surveillance, Epidemiology, and End Results Program 2010 for lymphoid neoplasms in groups and major NHL groups. We used data of the Federal State Cancer Registry of Baden-Wuerttemberg 2010–2011 to test differences in incidence outcome when evaluated according to the different recommendations of these institutions. Results: Depending on the recommendations of the above institutions, extraction of lymphoid neoplasms provided 4021, 4295, 3873 and 3848 incident cases, respectively. Case numbers for some major NHL groups diverge substantially by recommendation. Conclusions: Epidemiologists must be aware of potential discrepancies in coding conventions of cancer registries and have to consider them in comparative data analyses. Cancer registries should make transparent which recommendations were applied for lymphoma codification, currently and in the past. Conversion rules should be offered to ascertain proper mapping of lymphoma entities which were coded under varying coding practices over time.