It is debated whether non-Hodgkin's lymphomas originating in Waldeyer's ring (WR NHL) behave as NHL originating in lymph nodes or share common features with extranodal lymphomas originating in mucosa associated lymphatic tissue (MALT). We analyzed data from a population based NHL registry on patterns of dissemination at diagnosis, response to treatment, patterns of failure and survival of 77 primary Waldeyer's ring Non-Hodgkin's lymphomas (WR NHL) patents. Data of completely staged patients with diffuse large cell lymphomas (DLCL) originating in WR (n=44) were compared with those of patients retrieved from the same registry with DLCL originating in lymph nodes or stomach (the latter as prototype of a lymphoma originating in MALT). Primary WR NHL had favorable risk scores according to the International Prognostic Index (IPI), and responded well to therapy: a complete response (CR) rate of 74% was observed. Disease free survival (DFS) and overall survival (OS) were poor, however (47% and 31% at 10 years, respectively). The comparison of DLCL originating in WR, lymph nodes and stomach revealed that WR and gastric NHL patients shared a restricted pattern of dissemination at diagnosis, in contrast to patients with DLCL originating in lymph nodes. Although not all patients were completely restaged at relapse, analysis of patterns of failure suggested that the gastro-intestinal tract is a preferential site for recurrences, both for WR and gastric DLCL patients. CR rates of WR, nodal and gastric DLCL patients were 77%, 55% and 55% respectively (P=0.03), OS of the three patient subgroups did not differ (33%, 27% and 37% at 10 years). DFS of WR DLCL patients was similar to nodal DLCL patients but inferior to gastric DLCL patients (47%, 48% and 73% at 10 years respectively, P=0.006). After Cox regression analysis the relative relapse risk for patients with WR DLCL when compared to patients with DLCL originating in lymph nodes was 2.01 (C.I. 0.99-4.01, P=0.05), and 3.46 (C.I. 1.32-9.00, P=0.01) when compared to patients with gastric DLCL. The clinical picture of primary WR NHL emerging from this population based study is in agreement with data form hospital based studies. In the comparison of WR DLCL, nodal DLCL and gastric DLCL, the observed patterns of dissemination suggest similarities between WR DLCL and gastric DLCL. The frequent relapses after CR observed for WR DLCL patients, however, indicate that these lymphomas clinically behave as nodal DLCL, and should be treated accordingly.