Clinical risk scores do not accurately identify a very high risk population with diffuse large B cell lymphoma-an analysis of 386 Portuguese patients.

The identification of high-risk patients deserving alternative first-line treatments to R-CHOP is a research priority in diffuse large B cell lymphoma (DLBCL). Despite the increasing recognition of biological features underlying aggressive behavior, clinical scores remain the basis for prognostic evaluation and treatment stratification in DLBCL. We performed a retrospective analysis of patients with DLBCL uniformly treated with immunochemotherapy with the aim of assessing the discriminative power of the NCCN international prognostic index (IPI) and the GELTAMO-IPI scores in risk group stratification and compared them with the IPI. Additionally, we investigated if bulky disease, gender, beta-2 microglobulin (β2m), body mass index, and B-symptoms have independent prognostic impact. We confirmed the discriminative ability of the three prognostic scores in terms of progression-free survival and overall survival and found that the NCCN-IPI performs better in the identification of a high-risk population compared to the IPI and the GELTAMO scores. In an attempt to improve the prognostic power of the NCCN-IPI we analyzed additional clinical variables. Bulky disease and elevated β2m were found to be independent predictors of prognosis when controlling for the NCCN-IPI risk groups. However, they seem to bring no incremental power to the latter in the identification of poor outcome patients. We support the use of the NCCN-IPI for the clinical identification of high-risk patients in DLBCL. Future studies to unravel the biological heterogeneity within NCCN-IPI groups are needed to improve risk prediction and design targeted therapies for poor prognosis patients.