Background: Plasmablastic lymphoma (PBL) is a rare and aggressive disease, and data on outcomes from large series are limited. A prognostic score combining age, ECOG, bone marrow involvement (BMi) and CD138 expression was presented at the 66th ASH Annual Meeting (Martin-Moro F, Blood 2024; 144 supp. 1: 4478). Our aims were to compare PBL cohorts from Europe (EU) and Latin America (LATAM), and to validate the prognostic score in an independent international cohort.

Methods: The design of the PBL International Prognostic Index (PBL-IPI) was performed within the Spanish Lymphoma Group (GELTAMO); the training cohort was amplified to 106 patients. Baseline variables with independent impact on event-free survival (EFS) multivariate (MV) analysis were included in the PBL-IPI. Scores were assigned by B coefficients (B) and cases were clustered based on EFS. The index was applied for both progression-free survival (PFS) and overall survival (OS). The prognostic impact of the variables included in the score and the PBL-IPI were analyzed in an independent LATAM cohort of 114 patients from the Argentine Group for Treatment of Malignant Hemopathies (GATLA) and the Latin American Lymphoma Study Group (GELL).

PBL characteristics, therapeutic approaches and prognosis were compared between the training and validation cohorts.

Results: Characteristics at PBL onset were different between the training and the validation cohorts. Patients in the EU cohort were older (median age 57 vs 43 and IQR 43-73 vs 35-54; P <0.001. Age >60 45% vs 16%; P <0.001) and presented with worse performance status (ECOG >1 45% vs 33%; P=0.07) at PBL diagnosis than in the LATAM cohort. HIV positivity was less common in EU than in LATAM (47% vs 71%; P <0.001). No differences were observed among sex (males 85% vs 84%; P=0.89), extranodal involvement (85% vs 84%; P=0.8) and BMi (30% vs 24%; P=0.34). CD20 (13% vs 12%; P=0.9), CD138 (86% vs 85%; P=0.9) and EBER (58% vs 70%; P=0.13) positivity were also equivalent among cohorts. The first-line approach differed among EU/LATAM cohorts: DA-EPOCH 33%/61%, Hyper-CVAD or ESHAP or CHOEP 5%/4%, CHOP 37%/27.5%, and non-curative 25%/7.5%. Adding targeted therapies to frontline was more common in EU than in LATAM: rituximab 10% vs 0% (P=0.001), bortezomib 35% vs 8% (P <0.001) and daratumumab/brentuximab-vedotin 6% vs 2% (P=0.1). Autologous stem-cell transplant as consolidation was also more common in EU than in LATAM (10% vs 2%; P=0.02). Two-year PFS/OS was 46%/41% in the training cohort and 49%/61% in the validation cohort. During follow-up, 43% patients progressed and 63% died in EU, against 51% and 47% in LATAM. Median follow-up in alive patients was 66.3 and 33.4 months in each cohort, respectively.

According to EFS MV analysis in the 106 Spanish patients the variables included in the PBL-IPI were ECOG >1 (HR 3.7, B 1.31, P<0.001; 2 points), CD138 expression (HR 2.3, B 0.82, P=0.04; 1 point), BMi (HR 2.1, B 0.75, P=0.01; 1 point), and age >60 (HR 1.9, B 0.63, P=0.03; 1 point). Seventy-five patients with information about all PBL-IPI variables were divided into four groups according to their risk: 7% low (0 points: 2-year PFS/OS 100%/100%), 25% low-intermediate (1 point: 2-year PFS/OS 58%/72%), 21% intermediate-high (2 points: 2-year PFS/OS 40%/42%), and 47% high (3-5 points: 2-year PFS/OS 18%/9%).

In the validation LATAM cohort (N=114) univariate analyses for PFS/OS among PBL-IPI variables were: ECOG >1 HR 2.4/3.2 (P=0.001/<0.001), CD138 expression HR 4.3/1.6 (P=0.1/0.4), BMi HR 1.8/2.3 (P=0.04/0.004), and age >60 HR 1/1.9 (P=1/0.07). The PBL-IPI was applied to 57 patients with available data in the validation cohort: low risk 3.5% (2-year PFS/OS 100%/100%), low-intermediate risk 35% (2-year PFS/OS 75%/89%), intermediate-high 16% (2-year PFS/OS 57%/62%), and 45.5% high (2-year PFS/OS 26%/26%).

Conclusions: This study represents the largest international series of PBL to date, comparing European and Latin American cohorts. Relevant differences were observed in baseline characteristics, therapeutic strategies, and outcomes. The PBL-IPI demonstrated prognostic utility for predicting PFS and OS; however, its performance varied across cohorts, suggesting that regional differences—such as age distribution, HIV prevalence, and treatment availability—may influence the weight of individual prognostic factors. Further refinement or regional calibration of the score may enhance its applicability in diverse populations.

This content is only available as a PDF.
2025
Sign in via your Institution