Comparison Between aaIPI, LAB‐PI, NCCN‐IPI and GELTAMO‐IPI to Predict Prognosis in Elderly Patients with Diffuse Large‐B Cell Lymphoma who Undergo R‐CHOP/R‐miniCHOP

Background: Elderly patients with diffuse large B-cell lymphoma (DLBCL) need thorough evaluation at diagnosis to select the most suitable approach. Different scores are used to stratify DLBCL cases according to their prognosis, but data are scarce in old patients. The Laboratory Prognostic Index (LAB-PI) [Martin-Moro. ASH 2022; abstract 320] has emerged as an interesting tool for this group of patients, due to its simplicity and accessibility, as it only includes three variables routinely tested in peripheral blood at DLBCL diagnosis: lactate dehydrogenase, hemoglobin, and beta-2 microglobulin. The aim was to compare the prognostic usefulness of validated prognostic indexes in old DLBCL patients, with a special focus on the LAB-PI score. Methods: Retrospective multicenter study (on behalf of the Spanish Lymphoma Group GELTAMO) including de novo DLBCL patients with ≥70 years old at diagnosis who received first-line therapy (N = 386). Four prognostic scores were calculated prior to treatment initiation: aaIPI, LAB-PI, NCCN-IPI, and GELTAMO-IPI. Descriptive statistics were applied to compare the different prognostic scores, which were analyzed for both event-free survival (EFS) and overall survival (OS) by Kaplan Meier curves and by the concordance C-index. In each score, risk clusters which showed no EFS difference, calculated by univariate hazard ratio (UV HR) Cox regression analysis, were grouped. Results: The series was composed of 386 DLBCL patients with a median age at diagnosis of 76 years (IQR 73–80) and a male:female ratio of 0.8:1. Patients presented with advanced stage were 254/386 (66%). The distribution of the prognostic groups (low, low-intermediate, high-intermediate, and high) among the four indexes was heterogeneous, as shown in Figure 1A when comparing LAB-PI with aaIPI and NCCN-IPI. No statistical EFS difference was seen between low and low-intermediate risk groups for aaIPI (UV HR 1.4, 95% CI: 0.8–2.6) and LAB-PI (UV HR 1.2, 95% CI: 0.8–1.9), so these clusters were grouped in each index. Three hundred and seventy-four patients (97%) were treated with anthracycline-based regimens (R-CHOP or R-miniCHOP). The median follow-up of the cohort was 34.2 months (IQR 15–61). EFS and OS curves according to each prognostic score are presented in Figure 1B. According to C-index, the most useful score to predict EFS was the LAB-PI (0.66). Keywords: Aggressive B-cell non-Hodgkin lymphoma, Diagnostic and Prognostic Biomarkers, Risk Models No conflicts of interests pertinent to the abstract.