Prognosis of DLCL patients is variable and associated with well-defined risk factors. In the past decade several pretreatment variables have been incorporated into prognostic models to predict the death risk of individual patients. The International Prognostic Index (IPI), developed in an international consensus study, has been one of the most widely accepted of these models. In our study we applied some of the major prognostic models proposed for DLCLs in a cohort of 111 patients uniformly treated with a CHOP-like regimen in order to compare their sensitivity and specificity. We also evaluated the possibility of improving the IPI with the inclusion, from among the variables analysed, of serum beta-2 microglobulin level (beta-2M). The sensitivity, reflecting the ability to predict all failures in the cohort of patients as a whole, has improved from 45 to 73 per cent when the beta-2M-IPI model is compared with IPI, without a significant loss of specificity. Based on these results, the beta-2M-IPI may be useful for identifying the subset of patients with very poor prognoses. Therefore, the use of the serum beta-2M value in addition to the IPI may help in selection of the patients with DLCL at higher risk for treatment failure, and identification of those who may require specifically tailored therapeutic approaches.
Copyright 2000 John Wiley & Sons, Ltd.