Abstract 1575 Background:We have previously reported that serum concentration of serum-soluble tumor necrosis factor receptor 2 (sTNF-R2) predicts the clinical outcome of patients with aggressive non-Hodgkin’s lymphoma (NHL) treated with CHOP without rituximab (Goto et al. 2005). We extracted the diffuse large B-cell lymphoma (DLBCL) patients from aggressive NHL, used in our previous study, and extended our observation period. We found that serum sTNF-R2 level was a strong independent prognostic factor even for long-term observation in the CHOP era (8-year overall survival (OS); high serum sTNF-R2 group: 17.6%, low sTNF-R2 group: 69.8%, P<0.0001). Based on the result, we analyzed serum sTNF-R2 for patients with DLBCL who had received the CHOP plus rituximab (R-CHOP) regimen. Purpose:We aim to re-assess the prognostic significance of serum sTNF-R2 for DLBCL patients treated with R-CHOP regimen, and to assess sTNF-R2 with subtype of DLBCL, such as germinal center B cell type (GCB type) and non-GCB type which were classified by Hans et al. method (Hans et al. 2004). Methods:Consecutive 154 untreated patients with DLBCL prospectively participated in this study between 2002 and 2008. The patients were treated with R-CHOP. Results:In all patients with DLBCL, the median of serum sTNF-R2 level was 13.72 ng/ml (range 2.66–112.5 ng/ml). Various poor prognostic indicators, such as elderly age, poor performance status, increased LDH, multiple extranodal involvement sites, advanced disease, and existence of B-symptoms, were strongly associated with high serum sTNF-R2 levels. Serum sTNF-R2 levels significantly correlated with the increasing grade of international prognostic index (IPI), revised-IPI (R-IPI) (Sehn et al. 2007), and subtype, respectively (P<0.0001, P<0.0001, P=0.0058). We established the cut-off value for sTNFR2 at 20 ng/ml which was determined by ROC analysis. Patients with high sTNF-R2 (20 ng/ml and over) at onset had significantly lower OS rate (5-year: 29%), than that with low sTNF-R2 (5-year: 83%), respectively (P <0.0001). Multivariate analysis involving R-IPI, subtypes of DLBCL and serum sTNF-R2 revealed that sTNFR-2 significantly correlated with OS and progression free survival (PFS) (OS: P = 0.0001, PFS: P= 0.0085). Additionally, as the OS rate was about 50% even in the poor prognosis group of R-IPI, we tried to divide risk categories of R-IPI into low and high sTNF-R2 groups, resulting actually in five groups: very good risk group, good risk group with high sTNF-R2, good risk group with low sTNF-R2, poor risk group with high TNF-R2, and poor risk group with low sTNF-R2. A strong significant difference was observed. (5-year OS; very good risk group: 100%, good risk group with high sTNF-R2: 20%, good risk group with low sTNF-R2: 85%, poor risk group with high TNF-R2: 28.6%, and poor risk group with low sTNF-R2: 77.8%.) Conclusion:These results suggest that a high serum sTNF-R2 level predicts a poor prognosis in DLBCL and may be a useful biomarker for selecting appropriate treatment. Upfront high dose chemotherapy or alternative therapy of first line chemotherapy instead of R-CHOP might be well indicated for DLBCL patients with high serum sTNF-R2. Disclosures:No relevant conflicts of interest to declare.