10599 Background: The International Prognostic Index (IPI), the standard tumor risk model in patients (pts) with diffuse large cell lymphoma (DLCL), has 5 factors (age, LDH, performance status, extra nodal sites, and stage). We previously proposed a simpler two-factor model, based on LDH and B2M (JCO10;1989). B2M is a component of HLA class I antigens, expressed in lymphocytes, and a known prognostic indicator in some lymphoid malignancies. The benefits of this model are: objective measures; lab method widely available; simplicity; biologic marker. Methods: We applied the model to a large cohort of DLCL patients with prospective baseline B2M, and treated with doxorubicin-based (chemo) regimens, with and without rituximab. 718 pts with DLCL were sequentially treated at MDACC by IRB approved chemo protocols from 1988–2000. In 2001, rituximab plus chemo (RCHOP) became standard. 311 DLCL pts were sequentially treated with RCHOP from 2001–2005. Cox regression analyses for univariate and multivariate models of IPI factors and B2M were done. Kaplan-Meier survival projections were in three risk categories: low (normal [nl] LDH and B2M); intermediate (either LDH or B2M > nl); or high (LDH and B2M > nl). Results: In both treatment groups, IPI factors and B2M were significant as univariate factors. In the RCHOP group, however, the IPI multivariate model showed age, stage, and extra-nodal sites were not significant risk factors, while B2M and LDH remained highly significant (p<0.01). The 5 year survival projections by risk category were: (*) combines intermediate low and intermediate-high categories Conclusions: This simple two-factor model predicts risk for patients with DLCL, treated with or without rituximab, comparably to the IPI. B2M should be considered an important prognostic indicator in DLCL, particularly in rituximab treated patients. Exploratory analyses to revise the IPI model are indicated. [Table: see text] No significant financial relationships to disclose.