A New Prognostic Index Including Baseline Clinico-Biological Variables Predicts Outcome in Plasmablastic Neoplasms

Background: Plasmablastic neoplasms are rare and aggressive tumors that exhibit biological complexity and poor prognosis despite intensive approaches. Our aims were to evaluate the prognostic impact of clinical and biological variables in a series of plasmablastic neoplasms and to design a prognostic score. Besides, we analyzed the first-line therapeutic approach.

Methods: Multicentric retrospective study within the Spanish Lymphoma Group (GELTAMO) including de novo plasmablastic neoplasms. Clinical, biological and therapeutic variables were analyzed, highlighting the evaluation of 17 markers by immunochemistry or flow cytometry. The baseline parameters which demonstrated independent impact (P<0.05) on event-free survival (EFS) according to univariate (UV) Cox model were included in multivariate (MV) analysis. Variables that were independently related to EFS after successive MV assessments were included in the final index, with scores assigned to each parameter based on their B coefficient. Patients were clustered in prognostic groups by Kaplan Meier curves. The prognostic index was also applied to overall survival (OS).

Results: Ninety patients were included (72 plasmablastic lymphomas and 18 primary effusion lymphomas) with a median age of 58 (IQR 45-73), 82% males and 78% with advanced stage. Forty percent were HIV positive at lymphoma onset (in 37% the diagnosis of both entities was concurrent). The median follow-up was 15 months (IQR 5-57). Baseline parameters independently related to EFS in UV analysis were age >60 (HR 1.9 CI95% 1.1-3.1; P=0.01), ECOG >1 (HR 3 CI95% 1.8-5.2; P<0.001), advanced stage (HR 2.7 CI95% 1.3-5.8; P=0.009), elevated lactate dehydrogenase (HR 2.4 CI95% 1.3-4.2; P=0.004), elevated beta-2 microglobulin (HR 2.7 CI95% 1.1-6.8; P=0.04), extranodal sites >1 (HR 2.1 CI95% 1.2-3.5; P=0.007), bone marrow involvement (BMi) (HR 2.8 CI95% 1.6-5.1; P<0.001) and CD138 expression (HR 2.1 CI95% 1.4-6; P=0.04). Neither the Bcl2+/Myc+ phenotype (HR 0.9 CI95% 0.4-2; P=0.8), nor the p53 overexpression (HR 1.1 CI95% 0.4-3; P=0.8), nor the MYC rearrangement (HR 0.7 CI95% 0.3-1.6; P=0.4) were related to worse EFS. After EFS MV analysis the variables included in the prognostic index were age >60 (HR 2.2 CI95% 1.2-4.3; P=0.02), ECOG >1 (HR 3.4 CI95% 1.7-6.9; P=0.001), BMi (HR 2.5 CI95% 1.2-4.9; P=0.01) and CD138 expression (HR 2.7 CI95% 1.1-7; P=0.01); each variable received a 1 point score by B coefficient (0.8, 1.2, 0.9 and 1, respectively). Cases with 3 and 4 points were clustered together due to equivalent EFS (P=0.9), thus, patients were divided in four risk groups with the following distribution: low 7% (0 points; 2-year EFS 100% and OS 100%), low-intermediate 32% (1 point; 2-year EFS 58% and OS 77%), intermediate-high 36% (2 points; 2-year EFS 19% and OS 32%) and high 25% (3-4 points; 2-year EFS 0% and OS 0%). The IPI was less useful in predicting prognosis, with no difference between low-intermediate and intermediate-high risk groups (UV HR 0.8 CI95% 0.3-1.9; P=0.8). According to frontline approach 68/90 were treated with curative intention. Among them 2/68 received daratumumab-chemotherapy (both relapsed) and 2/68 brentuximab vedotin-chemotherapy (no relapses); all 4 cases remained alive. Patients who received V-DA-EPOCH (11/68) presented a slightly better prognosis than cases treated with DA-EPOCH without bortezomib (18/68) (2-year EFS 55% vs 38%; P=0.4), and greater than those treated with CHOP (18/68) and V-CHOP/V-CAP (14/68) in which 2-year EFS was 27% (P=0.1). The remaining 3/68 received other chemotherapy protocols. After first line, 32/68 achieved complete remission (CR), among them 28% (9/32) received an autologous stem-cell transplant (auto-SCT) as consolidation, without differences in prognosis against the 23/32 who did not proceed to auto-SCT (UV EFS HR 0.8 CI95% 0.1-4; P=0.8. UV OS HR 0.4 CI95% 0.1-3.6; P=0.4); cases who received auto-SCT were younger (median age 41 vs 51; P=0.6) and presented higher rates of high-risk IPI (14% vs 7%; P=0.5).

Conclusions: A new prognostic score combining age, ECOG, BMi and CD138 expression predicts outcome in plasmablastic neoplasms dividing patients into four prognostic groups. Our next step is to validate this index in an independent cohort. In this series, the frontline approach including targeted therapies such as brentuximab vedotin seemed to be promising, while auto-SCT after first CR did not improve outcomes.