Prognosis and treatment of plasmablastic lymphoma in the United States: a multicenter retrospective study.

Introduction

Plasmablastic lymphoma (PBL) is a rare subtype of large B-cell lymphoma first described in 1997 in a small cohort of people living with HIV (PLWH) [ 1 ]. In this initial cohort, it presented as a tumor in the oral cavity and was almost uniformly fatal. Since then, this entity has been identified post-transplant, in the setting of other immunosuppressed states, and even in otherwise immunocompetent patients [ 2 ].

PBL arises from the post-germinal center plasmablast, an activated B cell that ranges from immunoblastic to plasmacytic in morphology [ 3 ]. Immunophenotypically, PBL is characterized by expression of plasma cell markers (CD38, CD138) and most commonly a lack of B-cell markers (CD19, CD20, PAX5) [ 4 ]. MYC rearrangement and EBV infection are believed to contribute to the pathogenesis of PBL, but their roles remain incompletely elucidated [ 3 ].

While first described as a lesion in the oral cavity/jaw [ 1 ], PBL also presents in nodal and other extranodal sites. Many patients are diagnosed with advanced-stage disease that follows an aggressive clinical course [ 5 ]. While the prognosis of PBL was initially reported to be dismal, with a median overall survival of 8–15 months, recent studies have suggested improved outcomes [ 6 , 7 ]. The mainstay of initial management has been multiagent cytotoxic chemotherapy. Using CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) and CHOP-like regimens, 1- and 2-year overall survival of 50–60% has been reported in small case series ( n = 35) [ 8 ]. The addition of biological agents commonly used for the management of plasma cell disorders, such as the proteasome inhibitor bortezomib or the CD38-directed monoclonal antibody daratumumab, may confer some benefit, but has not yet been tested in a randomized fashion [ 9 , 10 ]. Moreover, no well-defined standard of care exists. While NCCN guidelines consider CHOP to be inadequate and favor higher intensity regimens [ 11 ], recent studies have called this into question [ 6 , 12 , 13 ].

Given its rarity, the literature on PBL is primarily limited to case reports and small case series, with larger multicenter and database studies appearing in recent years [ 5 , 12 ]. Still, significant unknowns remain surrounding the underlying biology and treatment of PBL. We aimed to perform a large, multicenter, retrospective study to understand disease characteristics, prognostic factors, and treatment-related outcomes in a contemporary cohort of PBL patients treated in the United States.

Methods

Study design

In this multicenter, retrospective cohort study, we identified 344 patients with PBL from 21 academic centers in the United States. Each center identified patients retrospectively using electronic medical records and submitted their anonymized data to the study center. Patients were included if they were ≥ 18 years old and diagnosed with PBL between 1/2005 and 12/2022. Patients were excluded if their diagnosis failed to meet the World Health Organization (WHO), revised 4 th edition diagnostic criteria [ 14 ] for PBL or their data were deemed incomplete. Patients included in the final analysis were grouped into one of four cohorts according to immune status (Fig. 1 ): patients with a previous history of HIV or HIV diagnosed concurrently with PBL were classified as HIV-PBL; patients with prior organ transplantation were classified as PBL arising as a post-transplant lymphoproliferative disorder (PTLD-PBL); those without HIV or prior transplant, but with known immunosuppression (e.g., underlying lymphoproliferative disorder, previous chemotherapy for malignancy, primary immunodeficiency, iatrogenic immunodeficiency, or autoimmune disease on current or prior immunosuppressive therapy with prednisone +/- biologic agents) were classified as PBL arising in the setting of other immunosuppressed states (OIS-PBL). All remaining patients who did not meet the criteria mentioned above were classified as immunocompetent PBL (IC-PBL). For a more detailed description of OIS-PBL, see Supplementary Table 1 . Fig. 1 Consort diagram. HIV-PBL: PBL in a person living with HIV; PTLD-PBL: PBL arising as a post-transplant lymphoproliferative disorder; OIS-PBL: PBL arising in the setting of some other immunosuppressed state; IC-PBL: PBL in an otherwise immunocompetent patient.

Objectives and definitions

The primary objective of this study was to determine the overall survival (OS) of the entire cohort. OS was defined as the time from PBL diagnosis to death from any cause or censoring at the time of last follow-up (FU). Secondary objectives included determining progression-free survival (PFS), non-relapse mortality (NRM), and treatment-related mortality (TRM), as well as assessing the prognostic impact of patient-, disease-, and treatment-related factors on survival outcomes. PFS was defined as the time from initiation of treatment to relapse/progression, death from any cause, or censoring at time of last FU. TRM was defined as death not due to PBL within 30 days of the most recent treatment.

Patients were staged with CT or combined PET/CT. Bone marrow involvement was assessed either via bone marrow biopsy or PET/CT as per institutional practice. Involvement of the oral cavity/jaw, nasopharynx, oropharynx, paranasal sinuses, or orbit was defined as disease occurring in the head and neck. Immunophenotype was characterized by immunohistochemistry (IHC). EBV status was assessed either via IHC for latency membrane protein-1 (LMP1) or in situ hybridization (ISH) to EBV-encoded RNA (EBER). MYC rearrangement was assessed via fluorescence in situ hybridization (FISH), with rearrangement at any locus considered a positive result.

As detailed in Supplementary Table 2 , we categorized chemotherapeutic backbone regimens received in the frontline (1 L) setting into four groups: (1) standard-intensity CHOP/CHOP-like regimens, (2) dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (EPOCH), (3) high intensity regimen with hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone, methotrexate, and cytarabine (Hyper-CVAD) or cyclophosphamide, vincristine, doxorubicin, high-dose methotrexate/ifosfamide, etoposide, and high-dose cytarabine (CODOX-M/IVAC), and everything else as (4) “other”. Second-line (2 L) chemotherapeutic regimens are also detailed in Supplementary Table 2 .

Statistical analysis

Continuous variables were summarized using medians with ranges and compared using the Wilcoxon rank-sum test. Categorical variables were summarized using frequencies and compared using the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\chi }}$$\end{document} χ 2 test.

Unadjusted OS and PFS were calculated via the Kaplan-Meier method. Cox regressions were used to investigate variables associated with survival. Disease classification-specific survival figures were adjusted by inverse generalized propensity score-based weighting [ 15 , 16 ]. The PTLD-PBL group was excluded from the propensity score models due to the small sample size. Factors included in multivariable Cox regression and the multinomial logistic propensity score model included age, sex, year of diagnosis, race, ethnicity, immune status, ECOG performance status, Ann Arbor stage, extranodal disease, LDH elevation, IPI score, MYC rearrangement, EBV positivity by either LMP1 or EBER, 1 L chemotherapeutic regimen, use of a proteasome inhibitor (PI) in the 1 L, and use of CNS prophylaxis (PPx). The impact of consolidative autologous stem cell transplant (ASCT) was assessed in a separate sensitivity analysis. We used multiple imputation to account for missing data in the multivariable models [ 17 ].

Approval for this study was granted by the Institutional Review Board (IRB) of The University of Pennsylvania (Philadelphia, USA) and by the IRBs of all participating institutions. The study protocol was approved with exemption by the IRB, which granted waiver of informed consent due to the retrospective nature of the study. All methods were performed in accordance with relevant guidelines and regulations.

Results

Patient characteristics

Three hundred seventy-five patients were identified from 21 institutions. After excluding 31 patients who did not meet eligibility criteria, 344 patients were included in the final analysis: 48% had HIV-PBL ( n = 164), 6% had PTLD-PBL ( n = 19), 10% had OIS-PBL ( n = 36), and 36% had IC-PBL ( n = 125) (Fig. 1 ). The median age at diagnosis was 53 years (range 19–91) (Table 1 ). Most patients were male (78%, n = 270); 67% of patients were White ( n = 230), 17% were Black/African American ( n = 59), and 33% were Hispanic/Latino ( n = 114). The median age at diagnosis was significantly younger in the HIV-PBL cohort (46 years) compared with the PTLD-, OIS-, and IC-PBL cohorts (55, 67, and 68 years, respectively; p < 0.001). Table 1 Patient characteristics, clinical presentation, and treatment patterns by immune status cohort. Total n = 344 (%) HIV-PBL n = 164 (%) PTLD-PBL n = 19 (%) OIS-PBL n = 36 (%) IC-PBL n = 125 (%) p -value* Age at diagnosis (y) Median (Range) < 60 ≥ 60 53 (19-91) 212 (62) 132 (38) 46 (19-76) 150 (91) 14 (9) 55 (23-75) 11 (58) 8 (42) 67 (21-88) 10 (28) 26 (72) 68 (20-91) 41 (33) 84 (67) <0.001 Sex Male Female 270 (78) 74 (22) 141 (86) 23 (14) 13 (68) 6 (32) 22 (61) 14 (39) 94 (75) 31 (25) 0.003 Year of diagnosis 2005–2009 2010–2014 2015–2019 2020–2022 29 (8) 75 (22) 158 (46) 82 (24) 23 (14) 40 (24) 75 (46) 26 (16) 0 (0) 6 (32) 7 (36) 6 (32) 2 (6) 5 (14) 15 (42) 14 (38) 4 (3) 24 (19) 61 (49) 36 (29) 0.003 Race White Black/African American Other Unknown 230 (67) 59 (17) 34 (10) 21 (6) 90 (55) 44 (27) 23 (14) 7 (4) 16 (85) 1 (5) 1 (5) 1 (5) 28 (77) 5 (14) 1 (3) 2 (6) 96 (77) 9 (7) 9 (7) 11 (9) <0.001 Ethnicity Hispanic/Latino Non-Hispanic/Latino Other/Unknown 114 (33) 206 (60) 24 (7) 62 (38) 94 (57) 8 (5) 2 (11) 16 (84) 1 (5) 13 (36) 19 (53) 4 (11) 37 (30) 77 (62) 11 (8) 0.143 ECOG Performance Status ≤ 1 2+ Unknown 227 (66) 74 (22) 43 (12) 104 (64) 35 (21) 25 (15) 13 (68) 2 (11) 4 (21) 25 (69) 9 (25) 2 (6) 85 (68) 28 (22) 12 (10) 0.773 Ann Arbor Stage I II III IV Unknown 52 (15) 33 (10) 22 (6) 227 (66) 10 (3) 15 (9) 18 (11) 13 (8) 114 (70) 4 (2) 4 (21) 0 (0) 0 (0) 13 (68) 2 (11) 6 (16) 2 (6) 4 (11) 24 (67) 0 (0) 27 (22) 13 (10) 5 (4) 76 (61) 4 (3) 0.071 Nodal Disease Yes No Unknown 215 (63) 112 (33) 17 (4) 112 (68) 44 (27) 8 (5) 8 (42) 11 (58) 0 (0) 23 (64) 13 (36) 0 (0) 72 (58) 44 (35) 9 (7) 0.046 Extranodal Disease Yes No Unknown 316 (92) 24 (7) 4 (1) 149 (91) 12 (7) 3 (2) 17 (89) 2 (11) 0 (0) 33 (92) 3 (8) 0 (0) 117 (94) 7 (5) 1 (1) 0.837 LDH Elevation >1x ULN Not Elevated Unknown 177 (51) 124 (36) 43 (13) 100 (61) 48 (29) 16 (10) 7 (37) 7 (37) 5 (26) 21 (58) 13 (36) 2 (6) 49 (39) 56 (45) 20 (16) 0.009 IPI at diagnosis, mean (SD)* 2.51 (1.32) 2.50 (1.14) 2.38 (1.39) 2.88 (1.34) 2.42 (1.52) 0.423 MYC Rearrangement + − Unknown 45 (13) 33 (10) 266 (77) 26 (16) 9 (5) 129 (79) 2 (11) 2 (11) 15 (78) 7 (19) 6 (17) 23 (64) 10 (8) 16 (13) 99 (79) 0.045 CNS Involvement Yes No Unknown 11 (3) 329 (96) 4 (1) 6 (4) 155 (95) 3 (1) 0 (0) 19 (100) 0 (0) 0 (0) 36 (100) 0 (0) 5 (4) 119 (95) 1 (1) 0.529 Head/Neck Involvement Yes No Unknown 94 (27) 246 (72) 4 (1) 53 (32) 108 (66) 3 (2) 2 (11) 17 (89) 0 (0) 3 (8) 33 (92) 0 (0) 36 (29) 88 (70) 1 (1) 0.008 EBV+ by LMP or EBER + − Unknown 194 (56) 112 (33) 38 (11) 25 (15) 118 (72) 21 (13) 5 (26) 13 (68) 1 (5) 23 (64) 13 (36) 0 (0) 59 (47) 50 (40) 16 (13) <0.001 Chemo Regimen in 1 L** CHOP/CHOP-like EPOCH Hyper-CVAD or CODOX-M/IVAC Other 44 (14) 220 (64) 26 (8) 19 (6) 17 (10) 110 (67) 17 (10) 9 (5) 1 (5) 15 (79) 1 (5) 1 (5) 10 (28) 19 (53) 0 (0) 2 (6) 16 (13) 76 (61) 8 (6) 7 (6) 0.073 Additional Agents Used in 1 L Proteasome Inhibitor Rituximab Daratumumab 105 (31) 60 (17) 12 (4) 47 (29) 31 (19) 7 (4) 5 (26) 4 (21) 2 (11) 8 (22) 11 (31) 2 (6) 45 (36) 14 (11) 1 (1) 0.028 CNS PPx in 1 L IT MTX and/or AraC HD MTX None Unknown 90 (26) 23 (7) 179 (52) 52 (15) 50 (30) 16 (10) 74 (45) 24 (15) 2 (11) 0 (0) 11 (58) 6 (31) 5 (14) 0 (0) 25 (69) 6 (17) 33 (26) 7 (6) 69 (55) 16 (13) 0.024 Consolidative RT in 1 L Yes No Unknown 63 (18) 270 (78) 11 (3) 21 (13) 142 (86) 1 (1) 3 (16) 14 (74) 2 (10) 6 (17) 30 (83) 0 (0) 33 (26) 84 (68) 8 (6) 0.014 Consolidative ASCT in 1 L Yes No Unknown 27 (8) 307 (89) 10 (3) 13 (8) 150 (91) 1 (1) 0 (0) 17 (89) 2 (11) 4 (11) 32 (89) 0 (0) 10 (8) 108 (86) 7 (6) 0.580 SD Standard deviation. *Bolded p -values are statistically significant. P -values do not account for missing data rows. SD Standard deviation. **More detailed description in Supplementary Table 2 .

In the HIV-PBL cohort, the median CD4 count at time of PBL diagnosis was 147 cells/μL (range 1–986); 30% ( n = 49) had a CD4 count <100 cells/μL. Fifty-eight percent ( n = 95) of patients were previously on antiretroviral therapy (ART) with a median CD4 count of 177 cells/μL. Thirty-six percent ( n = 59) were ART-naïve and commenced ART at the time of PBL diagnosis, with a median CD4 count of 102 cells/μL.

In the PTLD-PBL cohort, the median time from transplant to PBL diagnosis was 7.9 years (range 1–25). Of the 36 patients classified as having OIS-PBL, roughly half ( n = 17) had an underlying lymphoproliferative disorder (LPD) (CLL, n = 6; FL, n = 3; DLBCL, n = 3; MZL, n = 2; WM, n = 1; HL, n = 1; MALT lymphoma, n = 1), with a median time from diagnosis of LPD to PBL of 12 years (range 6–26).

Most patients were diagnosed with advanced-stage PBL [stage III ( n = 22, 6%) or IV ( n = 227, 66%)]. The majority had nodal disease ( n = 215, 63%), and most had extranodal involvement ( n = 316, 92%), with the most common extranodal site being the GI tract ( n = 107, 31%). Bone marrow involvement was identified in 30% of patients who underwent a bone marrow biopsy ( n = 63/212). Of the eighty-eight patients who had BM involvement assessed via PET/CT alone (i.e., without a BM biopsy), only 9% were found to have disease involving the BM ( n = 8). Only 3% of patients presented with CNS involvement ( n = 11). Twenty-seven percent presented with disease located in the head and neck ( n = 94), with significantly more in the HIV-PBL cohort compared with the PTLD-, OIS-, and IC-PBL cohorts (32%, vs 11%, 8%, and 29%, respectively; p = 0.008).

The most common immunophenotype was CD38+ (84%), CD138+ (87%), CD19- (83%), and CD20- (90%). Median Ki-67 was 90%. Fifty-eight percent of tested samples had a detectable MYC rearrangement ( n = 45/78), and this genomic abnormality was significantly more common in the HIV-PBL cohort at 74% ( n = 26/35; p = 0.045). Likewise, 64% of tested samples were EBV+ by EBER ISH, with the OIS- and IC-PBL cohorts significantly less likely to be EBER+ compared with the HIV- and PTLD-PBL cohorts ( p < 0.001).

Survival outcomes

Median time until death or censoring for the entire cohort was 3.4 y (range 0–17). Nineteen patients were lost to FU. Median OS was 5.0 y (95% CI: 3.2–8.6), with 1y- and 2y- OS rates of 70% and 59%, respectively. Median PFS was 1.4 y (95% CI: 1.0–3.0), with 1y- and 2y- PFS rates of 54% and 47%, respectively. The median OS after 1 st relapse/progression was 0.6 y (95% CI: 0.42–0.85). There were 170 deaths. The most common cause of death was PBL ( n = 104), followed by infection ( n = 25, with 12 being classified as TRM), and second primary malignancies (NSCLC, n = 2; HNSCC, n = 1; esophageal carcinoma, n = 1; mucinous adenocarcinoma, n = 1). The cause of death was unknown in 29 patients.

When comparing survival outcomes by immune status (Fig. 2a–b ), the PTLD-PBL cohort had the poorest outcomes, with a median OS of 1.1 y (95% CI: 0.43-not reached) and a median PFS of 1.0 y (95% CI: 0.3–3.9). In contrast, the HIV-PBL cohort had the best outcomes, with a median OS of 7.2 y (95% CI: 4.4–14.4) and a median PFS of 1.8 y (95% CI: 0.8–4.5). Median OS was 2.3 y (95% CI: 1.1–5.0) with a median PFS of 1.0 y (95% CI: 0.5–4.0) in the OIS-PBL cohort. For the IC-PBL cohort, median OS was 4.1 y (95% CI: 2.4-not reached) with a median PFS of 1.4 y (95% CI: 0.8–4.1). Fig. 2 Survival outcomes by immune status cohort. Kaplan–Meier curves comparing: unadjusted overall survival by immune status ( a ); unadjusted progression-free survival by immune status ( b ); propensity score-adjusted overall survival by immune status ( c ); propensity score-adjusted progression-free survival by immune status ( d ). Median survival (years) with 95% CI is denoted next to each curve, with p-values representing significant difference in median survival between cohorts.

Propensity score adjusted survival

We present unadjusted and propensity score-adjusted characteristics of the HIV-PBL, IC-PBL, and OIS-PBL cohorts in Supplementary Table 3 .

Propensity score-adjusted OS medians for the non-PTLD-PBL cohorts were: 6.1 years (Interquartile Range [IQR]: 0.6-not reached) for HIV-PBL, 4.6 years (IQR: 1.0-13.8) for IC-PBL, and 4.4 years (IQR: 1.2-not reached) for OIS-PBL (Fig. 2c ). Propensity score-adjusted Cox regressions did not find significant OS differences among the three groups (HR = 1.0, 95% CI: 0.6–1.6 for IC-PBL; HR = 1.1, 95% CI: 0.6–2.0 for OIS-PBL, relative to the HIV-PBL cohort). Adjusted PFS medians were 1.4 years (IQR: 0.4–14.4) for HIV-PBL, 1.3 years (IQR: 0.5-not reached) for IC-PBL, and 3.1 years (IQR: 0.7-not reached) for OIS-PBL (Fig. 2d ). Propensity score-adjusted PFS models did not find significant differences among groups (HR = 0.9, 95% CI: 0.6–1.4 for IC-PBL; HR = 0.8, 95% CI: 0.4–1.5 for OIS-PBL, relative to HIV-PBL).

Adjusted median NRM was not reached for any of the three cohorts (Fig. 3 ). Twenty-fifth percentile mortality was 14.4 years for HIV-PBL, and not reached for the IC-PBL or OIS-PBL cohorts. Subdistribution hazard ratios were 1.0 (95% CI: 0.6–1.5) for IC-PBL and 0.9 (95% CI: 0.5–1.6) for OIS-PBL, relative to HIV-PBL. Fig. 3 Propensity score adjusted non-relapse mortality (NRM) by immune status cohort. Cumulative incidence of death due to causes other than PBL, by immune status cohort.

Sensitivity analysis

Substantial differences between the cohorts impeded our ability to achieve an optimal balance of demographic and clinical factors among the three adjusted arms (Supplementary Table 3 ). For example, the oldest patient was 72 in the HIV-PBL cohort versus 91 in the IC-PBL cohort. To ensure that residual confounding was not driving our findings, we estimated a logistic propensity score model using data only from patients in the HIV-PBL and IC-PBL cohorts who were 73 or younger. The OIS-PBL cohort was excluded from this analysis due to its small sample size. In this sensitivity analysis, we achieved better adjusted demographic balance across groups, and the results (Supplementary Table 4 ) were similar to those of the three-group propensity score-adjusted model.

Prognostic factors

The results of the survival analysis are detailed in Table 2 . In the multivariable Cox regression model utilizing multiple imputation methods, only advanced age, ECOG ≥2, advanced stage, and LDH elevation were associated with worse OS. BM involvement was associated with worse OS (HR = 1.96, 95% CI: 1.17–3.28, p = 0.011) and PFS (HR = 1.42, 95% CI: 1.02–2.26, p = 0.038) on a separate sensitivity analysis. Ki-67 was associated with neither OS (HR = 1.01, 95% CI: 0.99–1.02, p = 0.064) nor PFS (HR = 1.01, 95% CI: 0.99–1.01, p = 0.126) on separate sensitivity analyses. We sought to more thoroughly assess the impact of several factors on survival outcomes, including CD4 count and MYC /EBV status. In the HIV-PBL cohort, a CD4 count ≥ 100 cells/μL at the time of PBL diagnosis was not associated with improved OS on multivariable analysis. Likewise, initiation of ART at the time of PBL diagnosis impacted neither OS nor PFS within the HIV-PBL cohort. Further, we assessed the impact of MYC translocation and EBV status on survival outcomes in the entire cohort: neither was an independent predictor of OS; however, EBV positivity was associated with improved PFS (HR = 0.57, 95% CI: 0.39–0.84, p = 0.004). The data on MYC rearrangement should be precautionary, as only 23% of patients were assessed for this abnormality. Table 2 Survival analysis using multiply imputed data. Risk Factor Univariable Analysis, OS Cox Regression (Multiple Imputation), OS HR 95% CI p -value* HR 95% CI p -value* Cohort (vs HIV-PBL) PTLD-PBL IC-PBL OIS-PBL 1.51 1.17 1.59 0.82–2.78 0.83–1.64 0.99–2.54 0.187 0.370 0.055 1.91 0.80 0.93 0.92–3.99 0.47–1.37 0.49–1.79 0.083 0.418 0.832 Age (years) 1.02 1.01–1.03 <0.001 1.02 1.00–1.03 0.018 Female sex 1.52 1.08–2.15 0.018 1.26 0.84–1.88 0.263 Year of diagnosis (vs 2005–2009) 2010–2014 2015–2019 2020–2022 1.49 1.88 1.61 0.79–2.81 1.02–3.46 0.83–3.12 0.220 0.042 0.162 1.10 1.65 1.51 0.50–2.4 0.79–3.41 0.67–3.38 0.817 0.180 0.322 Race (vs White) Black/African American Other 1.58 0.49 1.09–2.30 0.24–1.00 0.016 0.049 1.43 0.63 0.91–2.23 0.28–1.40 0.118 0.253 Ethnicity (vs Hispanic/Latino) Non-Hispanic/Latino 1.35 0.96–1.89 0.087 1.08 0.72–1.62 0.717 ECOG 2+ (vs <2) 2.59 1.82–3.68 <0.001 1.91 1.14–3.2 0.014 Stage III/IV (vs I/II) 2.52 1.64–3.86 <0.001 1.75 1.02–3.02 0.043 Extranodal disease 1.57 1.16–2.13 0.004 1.27 0.84–1.93 0.258 LDH elevation 3.04 2.12–4.36 <0.001 2.63 1.65–4.20 <0.001 IPI (per one point increase) 1.45 1.25–1.67 <0.001 1.05 0.84–1.32 0.643 MYC rearrangement 1.34 0.27–6.68 0.712 1.26 0.26–6.06 0.761 EBV+ by either LMP1 or EBER 0.59 0.42–0.84 0.004 0.71 0.46–1.10 0.126 1 L Regimen (vs CHOP) EPOCH High-Intensity Other 0.77 1.01 1.33 0.49–1.21 0.52–1.97 0.64–2.76 0.254 0.983 0.448 0.65 1.36 0.89 0.39–1.07 0.62–3.01 0.43–1.83 0.091 0.444 0.742 PI in 1 L 1.05 0.76–1.46 0.766 0.99 0.68–1.43 0.950 CNS PPx (vs none) IT MTX/AraC HD IV MTX 0.76 0.87 0.52–1.10 0.50–1.52 0.149 0.621 0.81 0.69 0.51–1.28 0.34–1.39 0.363 0.295 Consolidative RT in 1 L 0.60 0.38–0.93 0.024 0.93 0.57–1.52 0.768 *Bolded p -values are statistically significant.

Frontline treatment

Of the 313 patients who received chemotherapy in the frontline setting, most received EPOCH-based regimens (70%; n = 220), 14% received CHOP/CHOP-like regimens ( n = 44), and only 8% received Hyper-CVAD or CODOX-M/IVAC ( n = 13, each). Many patients received agents in addition to chemotherapy in the 1 L setting, with 34% receiving bortezomib ( n = 105), 19% receiving rituximab ( n = 60), and 4% receiving daratumumab ( n = 12). Three patients received single-agent biologic treatment (bortezomib, n = 2; rituximab, n = 1). There was no apparent benefit to either OS or PFS with the use of EPOCH or other high-intensity chemotherapy regimens over standard intensity CHOP/CHOP-like regimens in both the univariable and multivariable models (Table 2 ). This lack of OS and PFS benefit with use of EPOCH (compared with CHOP) was observed in a separate subgroup analysis of the HIV-PBL cohort as well. The overall and complete response rates (ORR, CRR) associated with EPOCH were 64% and 52%, respectively, versus 66% and 52% for CHOP. Of note, EPOCH was associated with higher rates of TRM than CHOP: 4% of EPOCH-treated patients died of TRM ( n = 8) vs 2% of CHOP-treated patients ( n = 1), p < 0.001. Furthermore, the use of a proteasome inhibitor failed to show either OS or PFS benefit on MVA (OS: HR = 0.99, 95% CI: 0.68–1.43, p = 0.950; PFS: HR = 1.10, 95% CI: 0.79–1.53, p = 0.591).

Radiation, autologous stem cell transplant, and CNS prophylaxis

Eighteen percent of patients ( n = 63) received RT as consolidation in the 1 L setting, while 8% were consolidated with a hematopoietic ASCT at first remission ( n = 27). Both RT and ASCT consolidation were independently associated with improved OS on UVA (Table 2 ); however, only ASCT remained associated with improved OS (HR = 0.43, 95% CI: 0.19–0.97, p = 0.041) and PFS (HR = 0.49, 95% CI: 0.26–0.94, p = 0.032) when added to the multivariable Cox model in a separate sensitivity analysis. However, to account for immortal time bias, we conducted a sensitivity analysis in which we modeled time from relapse or progression to death or censoring among those who relapsed or progressed. In this analysis, ASCT (which was included as a time varying covariate) was not associated with OS (UVA: HR = 0.82, 95% CI: 0.38–1.79, p = 0.629, MVA: HR = 0.92, 95% CI: 0.34–2.43, p = 0.859).

Thirty-three percent of patients ( n = 113) received CNS prophylaxis in the form of intrathecal methotrexate/cytarabine (MTX/AraC; n = 83), high-dose MTX ( n = 23), or both ( n = 7). None of these treatment modalities were associated with improvements in OS or PFS in the multivariable Cox regression models. Only one of the patients who received CNS PPx ( n = 113) experienced relapse involving the CNS, versus three of the patients who did not receive CNS PPx ( n = 231). Of the eleven patients who presented with CNS involvement, three were alive at time of last follow-up ( n = 3), and eight were dead ( n = 8).

Second-line treatment

Of the 138 patients who relapsed or progressed after receiving frontline treatment, 105 were treated with 2 L therapies. The most commonly used 2 L regimens included ICE (ifosfamide, carboplatin, and etoposide, n = 30), DHAP/similar (dexamethasone, high-dose cytarabine, and platinol, n = 23), low-intensity myeloma-like regimens ( n = 19), and GemOx/similar (gemcitabine and oxaliplatin, n = 8) (Supplementary Table 2 ). The ORRs for each of these regimens were: 50%, 39%, 26%, and 38%, respectively.

Furthermore, 51% of these 105 patients were treated with biological agents alone or in addition to chemotherapy, such as daratumumab ( n = 24), bortezomib ( n = 27), rituximab ( n = 9), or the immunomodulatory drug lenalidomide ( n = 13). Several patients were consolidated in the second-line setting with either an autologous ( n = 3) or allogeneic ( n = 3) hematopoietic cell transplant (HCT). Of note, four patients received chimeric antigen receptor (CAR) T-cell therapy in the 2 L setting (CD19-directed, n = 3; unknown target, n = 1), with only one alive at last FU (treated with lisocabtagene maraleucel).

Discussion

We performed a multicenter, retrospective study to better understand the underlying biology, survival outcomes, and optimal treatment of PBL in a contemporary cohort of patients in the US. To our knowledge, this is the largest real-world cohort of PBL patients to date.

The main finding of this study was the markedly improved median OS in our cohort (5.0 y) compared with a recently published international cohort (1.7 y), albeit with similar median PFS (1.4 y vs 8.4 mo) [ 12 ]. This discrepancy may be explained by both improvements in supportive care and increased availability of novel biological agents used in the relapsed/refractory setting in our modern cohort. Di Ciaccio et al. enrolled from 1999 to 2020, with only 35% of patients diagnosed after 2015, whereas we enrolled from 2005 to 2022, with 70% diagnosed after 2014. It seems plausible that improvements in supportive care (e.g., infection prophylaxis, growth factor support) as well as enhanced efficacy and tolerability of ART regimens in our more contemporary cohort treated in a different geographic region explain the improved OS. Furthermore, while our study (as well as that of Di Ciaccio et al.) demonstrates a lack of benefit of novel biological agents when used in the frontline setting, it is possible that their increased use in the second-line setting in our more contemporary cohort is what caused increased OS without impacting PFS. While we did not formally assess this hypothesis, it represents an area of future investigation.

The results of our study are broadly consistent with the current understanding of PBL. Known lymphoma risk factors that are usually captured in the IPI, such as advanced age, stage III/IV, ECOG ≥ 2, and LDH elevation, were associated with poor OS in our cohort [ 18 ]. Consistent with Di Ciaccio et al., we also found EBV negativity to be associated with poor PFS [ 12 ]. While we did not collect data on EBV PCR status, this represents an area of future investigation.

While there is mixed evidence regarding the prognostic impact of MYC rearrangement in PBL [ 12 , 13 , 19 , 20 ], our results suggest that MYC + /EBV- PBL may represent an aggressive disease subtype; however, this requires further exploration, especially as data on MYC rearrangement was missing in many patients. EBV-positivity has been associated with different genetic features in Burkitt lymphoma (BL) when compared to EBV-negative cases. Recently, investigators identified that EBV + BL is characterized by fewer driver mutations, higher aberrant somatic hypermutation rates, and significantly more breakpoints upstream of MYC [ 21 ]. Therefore, interaction of EBV and MYC in PBL may result in different pathobiology underlying the observed differential outcomes. Furthermore, the association between female sex and poor OS on UVA (albeit with no association on MVA) is intriguing and warrants further investigation, particularly since males tend to have worse survival outcomes in most other lymphoma subtypes [ 22 , 23 ]. However, this observation may be related to the fact that male sex was predominant in the HIV-PBL cohort, which, in general, had better outcomes.

Despite improvement in median OS, PBL remains a difficult-to-treat disease, as demonstrated by a 5-year OS rate of 50% in our cohort. Survival outcomes differed by immune status: HIV-PBL had the best outcomes (median OS 7.2 y, median PFS 1.8 y), while PTLD-PBL had the worst outcomes (median OS 1.1 y, median PFS 1.0 y). This may be due to a partially reversible immunodeficiency in PLWH, especially as about a third of the HIV-PBL cohort was previously ART-naïve. The poor survival outcomes of PTLD-PBL suggest this may represent an aggressive subtype, possibly related to the mostly irreversible, underlying, severe iatrogenic immunodeficiency in post-transplant patients. The underlying histopathologic features of this entity represent an area of future investigation. While there were various underlying immunosuppressed states in the OIS-PBL cohort, many were LPDs, raising the possibility that the PBL diagnosis represented a transformation event. Further investigation into the biology and clinical outcomes of this subgroup is warranted.

The optimal treatment of PBL remains poorly defined due to the rarity of the disease and paucity of prospective clinical trials. While early studies of HIV-associated non-Hodgkin lymphomas suggested superior outcomes with EPOCH compared with CHOP and prompted NCCN guidelines to favor higher-intensity regimens, more recent retrospective studies have increasingly suggested no benefit to such an approach in PBL [ 6 , 11 , 12 , 24 ]. Our results similarly found no benefit with the use of EPOCH or other high-intensity regimens in the 1 L setting over CHOP and therefore do not support the use of higher-intensity regimens, which are subject to higher rates of treatment-related morbidity and mortality. However, none of these are randomized prospective studies.

Similarly, despite promising results in small, retrospective case series [ 7 ], proteasome inhibitors failed to show frontline benefit in a recent multicenter study [ 12 ], consistent with the findings in our cohort. Nevertheless, our study was not powered to definitively address the merit of biological agents classically used for plasma cell malignancies, such as bortezomib, daratumumab, and lenalidomide, in the management of PBL. However, the improved OS despite a similar PFS after 1 L therapy in our US cohort could indicate a role for these agents in PBL. Prospective and ideally randomized clinical trials will be necessary to address this question. At present, a prospective randomized clinical trial in the US explores the combination of daratumumab with EPOCH [ 25 ], while another study sponsored by the Fondazione Italiana Linfomi studies the combination of daratumumab, bortezomib, and dexamethasone for relapsed/refractory (r/r) PBL [ 26 ].

RT and ASCT consolidation were associated with improvements in OS in our univariable model. However, RT failed to show OS benefit in our multivariable model, and ASCT did not show OS benefit when included as a time varying covariate in a sensitivity analysis. Thus, RT and ASCT do not appear to benefit OS, and their use should be determined on a case-by-case basis.

Likewise, the use of CNS prophylaxis was not associated with OS and had no apparent effect on the risk of CNS recurrence, although the numbers were too small for firm conclusions. This is in line with recent observations in DLBCL that indicate a lack of benefit of CNS prophylaxis irrespective of modality [ 27 , 28 ]. Thus, the decision to use CNS prophylaxis in PBL should be highly individualized.

Several small studies have investigated the management of r/r PBL with chemotherapeutic regimens such as ICE, with or without agents like daratumumab and lenalidomide [ 29 – 32 ]. We gathered data on 105 patients who underwent treatment in the r/r setting, and a comparison of survival outcomes using various 2 L regimens, as well as the impact of use of biologic agents in the 2 L setting will be addressed in future analyses. Further, while CAR T-cell therapy has revolutionized the treatment landscape of both DLBCL and multiple myeloma, its use in PBL has been limited. To our knowledge, there are only three published cases using CAR T-cell therapy in r/r PBL. One achieved a complete metabolic response, followed by disease progression at 5 months post-CD19-directed CAR T-cell infusion [ 33 ]. Two additional patients achieved complete remissions in response to BCMA-directed CAR T-cell products, with CR documented at 6 and 14 months post-infusion, respectively [ 34 , 35 ]. Herein, we report four patients treated with CAR T-cell therapy (CD19-directed, n = 3; unknown target, n = 1) with disappointing outcomes. It is unclear if the expression of CD19 and/or BCMA played a role in treatment choice. More data is needed regarding the safety, efficacy, and optimal target antigen of CAR T-cell therapies in the management of r/r PBL.

There are several limitations to this study. Namely, given the retrospective nature of data collection, there was considerable variation in data availability between participating institutions. While some clinicopathologic variables were missing, we addressed missing data bias by using multiple imputation for multivariable analyses.

Further, no centralized pathologic review was done to confirm a diagnosis of PBL in each patient; however, all participating institutions were major academic centers with experienced hematopathology departments. Nonetheless, the fact that 10% of our cohort was CD20+ raises the possibility that our cohort included cases that would now be classified as DLBCL, NOS according to the WHO 5 th edition, thus contributing to erroneously improved outcomes. We investigated whether exclusion of the 31 patients with CD20 positivity resulted in a significant change in outcomes via a sensitivity analysis using multiply imputed data. After excluding CD20+ cases, we did not observe a substantial change in median OS (5.1 y) and median PFS (1.4 y). Furthermore, we did not see survival benefit with the use of EPOCH (versus CHOP) or with the use of proteasome inhibitors in the 1 L setting.

Despite the limitations of our study, it has several strengths, including our large sample size and granular, patient-level data. This allowed us to control for possible confounders by using propensity score adjustment and, therefore, significantly reduce bias. Further, this study expands the existing literature on this rare entity, augments findings of a recently published cohort from Australia, the UK, Canada, and Singapore, and at the same time contrasts them with our contemporary US cohort [ 12 ].

The prognosis of PBL has significantly improved over recent decades, particularly for HIV-PBL. While no optimal treatment has been established, more intense cytotoxic regimens may not confer a significant survival benefit. Although biological agents commonly used for plasma cell dyscrasias have been suggested to be of benefit based on small case series, their frontline use was not associated with improved outcomes in our study, and their role in PBL remains to be defined. This can only be done definitively in well-designed, prospective, and randomized clinical trials. Better understanding of the pathogenesis that underlies PBL based on oncogenic events and interactions with the host immune environment may also lead to more rationally targeted therapies and further improved outcomes.

Supplementary information

Supplemental Tables and Figures

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at 10.1038/s41408-026-01457-3.

Acknowledgements

We acknowledge the contributions of all individuals at participating centers involved in study design and completion. Further, we acknowledge the patients from whom data was abstracted. This study was supported by AIDS Malignancy Consortium (AMC) Grant UM1CA121947 and Fox Chase Cancer Center NCI Grant P30CA006927. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.

Author contributions

MH and SKB conceived the study. MH collected, analyzed, and interpreted the data, and drafted, reviewed, and edited the manuscript. SKB coordinated enrollment from participating institutions, analyzed and interpreted the data, and reviewed and edited the manuscript. BLE led statistical design of the study, analyzed and interpreted the data, and reviewed and edited the manuscript. ZF and AN assisted in study design and reviewed and edited the manuscript. VC and EC assisted with IRB approval and coordination of data use agreements between the host site and participating institutions. RES, SST, GJL, JEA, RP, SA, JS, TV, RB, ARG, JJC, EH, CD, IAN, NMS, NNTD, JSS, AER, NK, AD, VS, PGR, AA, GS, AV, CM, AJO, GS, EA, SD, JMR, GM, TF, MP, DMA, WB, CG, VB, and MH collected data, reviewed, and edited the manuscript.

Data availability

De-identified patient data is available upon request from the corresponding author and/or senior author, via email.

Competing interests

The authors declare no competing interests.

References

1. Delecluse HJ Anagnostopoulos I Dallenbach F Hummel M Marafioti T Schneider U Plasmablastic lymphomas of the oral cavity: a new entity associated with the human immunodeficiency virus infection Blood 1997 89 1413 20 10.1182/blood.V89.4.1413 9028965 Delecluse HJ, Anagnostopoulos I, Dallenbach F, Hummel M, Marafioti T, Schneider U, et al. Plasmablastic lymphomas of the oral cavity: a new entity associated with the human immunodeficiency virus infection. Blood. 1997;89:1413–20. 9028965

2. Morscio J Dierickx D Nijs J Verhoef G Bittoun E Vanoeteren X Clinicopathologic comparison of plasmablastic lymphoma in HIV-positive, immunocompetent, and posttransplant patients: single-center series of 25 cases and meta-analysis of 277 reported cases Am J Surg Pathol 2014 38 875 86 10.1097/PAS.0000000000000234 24832164 Morscio J, Dierickx D, Nijs J, Verhoef G, Bittoun E, Vanoeteren X, et al. Clinicopathologic comparison of plasmablastic lymphoma in HIV-positive, immunocompetent, and posttransplant patients: single-center series of 25 cases and meta-analysis of 277 reported cases. Am J Surg Pathol. 2014;38:875–86. 24832164 10.1097/PAS.0000000000000234

3. Castillo JJ Bibas M Miranda RN The biology and treatment of plasmablastic lymphoma Blood 2015 125 2323 30 10.1182/blood-2014-10-567479 25636338 Castillo JJ, Bibas M, Miranda RN. The biology and treatment of plasmablastic lymphoma. Blood. 2015;125:2323–30. 25636338 10.1182/blood-2014-10-567479

4. Vega F Chang CC Medeiros LJ Udden MM Cho-Vega JH Lau CC Plasmablastic lymphomas and plasmablastic plasma cell myelomas have nearly identical immunophenotypic profiles Mod Pathol 2005 18 806 15 10.1038/modpathol.3800355 15578069 Vega F, Chang CC, Medeiros LJ, Udden MM, Cho-Vega JH, Lau CC, et al. Plasmablastic lymphomas and plasmablastic plasma cell myelomas have nearly identical immunophenotypic profiles. Mod Pathol. 2005;18:806–15. 15578069 10.1038/modpathol.3800355

5. Hansen AR Vardell VA Fitzgerald LA Epidemiologic characteristics, treatment patterns, and survival analysis of plasmablastic lymphoma in the United States: a SEER and NCDB analysis Clin Lymphoma Myeloma Leuk 2024 24 e152 160 10.1016/j.clml.2023.12.014 38262787 Hansen AR, Vardell VA, Fitzgerald LA. Epidemiologic characteristics, treatment patterns, and survival analysis of plasmablastic lymphoma in the United States: a SEER and NCDB analysis. Clin Lymphoma Myeloma Leuk. 2024;24:e152–160. 38262787 10.1016/j.clml.2023.12.014

6. Tchernonog E Faurie P Coppo P Monjanel H Bonnet A Algarte Genin M Clinical characteristics and prognostic factors of plasmablastic lymphoma patients: analysis of 135 patients from the LYSA group Ann Oncol 2017 28 843 8 10.1093/annonc/mdw684 28031174 Tchernonog E, Faurie P, Coppo P, Monjanel H, Bonnet A, Algarte Genin M, et al. Clinical characteristics and prognostic factors of plasmablastic lymphoma patients: analysis of 135 patients from the LYSA group. Ann Oncol. 2017;28:843–8. 28031174 10.1093/annonc/mdw684

7. Castillo JJ Guerrero-Garcia T Baldini F Tchernonog E Cartron G Ninkovic S Bortezomib plus EPOCH is effective as frontline treatment in patients with plasmablastic lymphoma Br J Haematol 2019 184 679 82 10.1111/bjh.15156 29527667 Castillo JJ, Guerrero-Garcia T, Baldini F, Tchernonog E, Cartron G, Ninkovic S, et al. Bortezomib plus EPOCH is effective as frontline treatment in patients with plasmablastic lymphoma. Br J Haematol. 2019;184:679–82. 29527667 10.1111/bjh.15156

8. Castillo JJ Winer ES Stachurski D Perez K Jabbour M Milani C Prognostic factors in chemotherapy-treated patients with HIV-associated plasmablastic lymphoma Oncologist 2010 15 293 9 10.1634/theoncologist.2009-0304 20167839 PMC3227958 Castillo JJ, Winer ES, Stachurski D, Perez K, Jabbour M, Milani C, et al. Prognostic factors in chemotherapy-treated patients with HIV-associated plasmablastic lymphoma. Oncologist. 2010;15:293–9. 20167839 10.1634/theoncologist.2009-0304 PMC3227958

9. Noy A Barta SK Kwon D Baiocchi R Ramos JC Rubinstein P Daratumumab with dose-adjusted EPOCH is feasible in newly diagnosed plasmablastic lymphoma: AIDS malignancy consortium 105 Blood 2024 144 870 10.1182/blood-2024-211388 Noy A, Barta SK, Kwon D, Baiocchi R, Ramos JC, Rubinstein P, et al. Daratumumab with dose-adjusted EPOCH is feasible in newly diagnosed plasmablastic lymphoma: AIDS malignancy consortium 105. Blood. 2024;144:870.

10. Ryu YK Ricker EC Soderquist CR Francescone MA Lipsky AH Amengual JE Targeting CD38 with daratumumab plus chemotherapy for patients with advanced-stage plasmablastoid large B-cell lymphoma J Clin Med 2022 11 4928 10.3390/jcm11164928 36013165 PMC9409851 Ryu YK, Ricker EC, Soderquist CR, Francescone MA, Lipsky AH, Amengual JE. Targeting CD38 with daratumumab plus chemotherapy for patients with advanced-stage plasmablastoid large B-cell lymphoma. J Clin Med. 2022;11:4928. 36013165 10.3390/jcm11164928 PMC9409851

11. NCCN guidelines version 3.2024. HIV-related B-cell lymphomas. Accessed October 17, 2024. https://www.nccn.org/professionals/physician_gls/pdf/b-cell.pdf .

12. Di Ciaccio PR Polizzotto MN Cwynarski K Gerrie AS Burton C Bower M The influence of immunodeficiency, disease features, and patient characteristics on survival in plasmablastic lymphoma Blood 2024 143 152 65 10.1182/blood.2023021348 37832030 Di Ciaccio PR, Polizzotto MN, Cwynarski K, Gerrie AS, Burton C, Bower M, et al. The influence of immunodeficiency, disease features, and patient characteristics on survival in plasmablastic lymphoma. Blood. 2024;143:152–65. 37832030 10.1182/blood.2023021348

13. Castillo JJ Furman M Beltrán BE Bibas M Bower M Chen W Human immunodeficiency virus-associated plasmablastic lymphoma: poor prognosis in the era of highly active antiretroviral therapy Cancer 2012 118 5270 7 10.1002/cncr.27551 22510767 Castillo JJ, Furman M, Beltrán BE, Bibas M, Bower M, Chen W, et al. Human immunodeficiency virus-associated plasmablastic lymphoma: poor prognosis in the era of highly active antiretroviral therapy. Cancer. 2012;118:5270–7. 22510767 10.1002/cncr.27551

14. Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues, Revised 4th Edition. IARC; 2017.

15. McCaffrey DF Ridgeway G Morral AR Propensity score estimation with boosted regression for evaluating causal effects in observational studies Psychol Methods 2004 9 403 25 10.1037/1082-989X.9.4.403 15598095 McCaffrey DF, Ridgeway G, Morral AR. Propensity score estimation with boosted regression for evaluating causal effects in observational studies. Psychol Methods. 2004;9:403–25. 15598095 10.1037/1082-989X.9.4.403

16. Ridgeway G, McCaffrey D, Morral A, Cefalu M, Burgette L, Pane J, et al. Toolkit for Weighting and Analysis of Nonequivalent Groups: A Tutorial for the Twang Package.

17. Raghunathan TE Lepkowski JM Van Hoewyk J Solenberger P A multivariate technique for multiply imputing missing values using a sequence of regression models Surv Methodol 2001 27 85 95 Raghunathan TE, Lepkowski JM, Van Hoewyk J, Solenberger P. A multivariate technique for multiply imputing missing values using a sequence of regression models. Surv Methodol. 2001;27:85–95.

18. International Non-Hodgkin’s Lymphoma Prognostic Factors Project. A predictive model for aggressive non-Hodgkin’s lymphoma N Engl J Med 1993 329 987 94 10.1056/NEJM199309303291402 8141877 International Non-Hodgkin’s Lymphoma Prognostic Factors Project. A predictive model for aggressive non-Hodgkin’s lymphoma. N Engl J Med. 1993;329:987–94. 8141877 10.1056/NEJM199309303291402

19. Jessa R Chien N Villa D Freeman CL Slack GW Savage KJ Clinicopathological characteristics and long-term outcomes of plasmablastic lymphoma in British Columbia Br J Haematol 2022 199 230 8 10.1111/bjh.18399 35961783 Jessa R, Chien N, Villa D, Freeman CL, Slack GW, Savage KJ, et al. Clinicopathological characteristics and long-term outcomes of plasmablastic lymphoma in British Columbia. Br J Haematol. 2022;199:230–8. 35961783 10.1111/bjh.18399

20. Witte HM Hertel N Merz H Bernd HW Bernard V Stolting S Clinicopathological characteristics and MYC status determine treatment outcome in plasmablastic lymphoma: a multicenter study of 76 consecutive patients Blood Cancer J 2020 10 63 10.1038/s41408-020-0327-0 32471974 PMC7260224 Witte HM, Hertel N, Merz H, Bernd HW, Bernard V, Stolting S, et al. Clinicopathological characteristics and MYC status determine treatment outcome in plasmablastic lymphoma: a multicenter study of 76 consecutive patients. Blood Cancer J. 2020;10:63. 32471974 10.1038/s41408-020-0327-0 PMC7260224

21. Thomas N Dreval K Gerhard DS Hilton LK Abramson JS Ambinder RF Genetic subgroups inform on pathobiology in adult and pediatric Burkitt lymphoma Blood 2023 141 904 16 10.1182/blood.2022016534 36201743 PMC10023728 Thomas N, Dreval K, Gerhard DS, Hilton LK, Abramson JS, Ambinder RF, et al. Genetic subgroups inform on pathobiology in adult and pediatric Burkitt lymphoma. Blood. 2023;141:904–16. 36201743 10.1182/blood.2022016534 PMC10023728

22. Radkiewicz C Bruchfeld JB Weibull CE Jeppesen ML Frederiksen H Lambe M Sex differences in lymphoma incidence and mortality by subtype: a population-based study Am J Hematol 2023 98 23 30 10.1002/ajh.26744 36178436 PMC10092431 Radkiewicz C, Bruchfeld JB, Weibull CE, Jeppesen ML, Frederiksen H, Lambe M, et al. Sex differences in lymphoma incidence and mortality by subtype: a population-based study. Am J Hematol. 2023;98:23–30. 36178436 10.1002/ajh.26744 PMC10092431

23. Pfreundschuh M Age and sex in non-Hodgkin lymphoma therapy: it’s not all created equal, or is it? Am Soc Clin Oncol Educ Book 2017 37 505 11 10.1200/EDBK_175447 28561693 Pfreundschuh M. Age and sex in non-Hodgkin lymphoma therapy: it’s not all created equal, or is it? Am Soc Clin Oncol Educ Book. 2017;37:505–11. 28561693 10.1200/EDBK_175447

24. Barta SK Xue X Wang D Tamari R Lee JY Mounier N Treatment factors affecting outcomes in HIV-associated non-Hodgkin lymphomas: a pooled analysis of 1546 patients Blood 2013 122 3251 62 10.1182/blood-2013-04-498964 24014242 PMC3821722 Barta SK, Xue X, Wang D, Tamari R, Lee JY, Mounier N, et al. Treatment factors affecting outcomes in HIV-associated non-Hodgkin lymphomas: a pooled analysis of 1546 patients. Blood. 2013;122:3251–62. 24014242 10.1182/blood-2013-04-498964 PMC3821722

25. Tan CR Barta SK Lensing SY Noy A A multicenter, open-label feasibility study of daratumumab with dose-adjusted EPOCH in newly diagnosed plasmablastic lymphoma with or without HIV Blood 2019 134 1595 10.1182/blood-2019-129702 Tan CR, Barta SK, Lensing SY, Noy A. A multicenter, open-label feasibility study of daratumumab with dose-adjusted EPOCH in newly diagnosed plasmablastic lymphoma with or without HIV. Blood. 2019;134:1595.

26. Study to Evaluate Combined Treatment of Daratumumab, Bortezomib and Dexamethasone in PBL Patients, NCT04915248 .

27. Klanova M Sehn LH Bence-Bruckler I Cavallo F Jin J Martelli M Integration of cell of origin into the clinical CNS International Prognostic Index improves CNS relapse prediction in DLBCL Blood 2019 133 919 26 10.1182/blood-2018-07-862862 30617197 PMC6396175 Klanova M, Sehn LH, Bence-Bruckler I, Cavallo F, Jin J, Martelli M, et al. Integration of cell of origin into the clinical CNS International Prognostic Index improves CNS relapse prediction in DLBCL. Blood. 2019;133:919–26. 30617197 10.1182/blood-2018-07-862862 PMC6396175

28. Orellana-Noia VM Reed DR McCook AA Sen JM Barlow CM Malecek MK Single-route CNS prophylaxis for aggressive non-Hodgkin lymphomas: real-world outcomes from 21 US academic institutions Blood 2022 139 413 23 10.1182/blood.2021012888 34570876 PMC8777199 Orellana-Noia VM, Reed DR, McCook AA, Sen JM, Barlow CM, Malecek MK, et al. Single-route CNS prophylaxis for aggressive non-Hodgkin lymphomas: real-world outcomes from 21 US academic institutions. Blood. 2022;139:413–23. 34570876 10.1182/blood.2021012888 PMC8777199

29. Dittus C Miller JA Wehbie R Castillo JJ Daratumumab with ifosfamide, carboplatin and etoposide for the treatment of relapsed plasmablastic lymphoma Br J Haematol 2022 198 e32 e34 10.1111/bjh.18228 35506301 Dittus C, Miller JA, Wehbie R, Castillo JJ. Daratumumab with ifosfamide, carboplatin and etoposide for the treatment of relapsed plasmablastic lymphoma. Br J Haematol. 2022;198:e32–e34. 35506301 10.1111/bjh.18228

30. Cheng L Song Q Liu M Wang Y Yi H Qian Y Case report: successful management of a refractory plasmablastic lymphoma patient with tislelizumab and lenalidomide Front Immunol 2021 12 702593 10.3389/fimmu.2021.702593 34322131 PMC8312258 Cheng L, Song Q, Liu M, Wang Y, Yi H, Qian Y, et al. Case report: successful management of a refractory plasmablastic lymphoma patient with tislelizumab and lenalidomide. Front Immunol. 2021;12:702593. 34322131 10.3389/fimmu.2021.702593 PMC8312258

31. Marrero WD Cruz-Chacon A Castillo C Cabanillas F Successful use of bortezomib-lenalidomide combination as treatment for a patient with plasmablastic lymphoma Clin Lymphoma Myeloma Leuk 2018 18 e275 e277 10.1016/j.clml.2018.04.011 29753690 Marrero WD, Cruz-Chacon A, Castillo C, Cabanillas F. Successful use of bortezomib-lenalidomide combination as treatment for a patient with plasmablastic lymphoma. Clin Lymphoma Myeloma Leuk. 2018;18:e275–e277. 29753690 10.1016/j.clml.2018.04.011

32. Ando K Imaizumi Y Kobayashi Y Niino D Hourai M Sato S Bortezomib- and lenalidomide-based treatment of refractory plasmablastic lymphoma Oncol Res Treat 2020 43 112 6 10.1159/000504608 31842017 Ando K, Imaizumi Y, Kobayashi Y, Niino D, Hourai M, Sato S, et al. Bortezomib- and lenalidomide-based treatment of refractory plasmablastic lymphoma. Oncol Res Treat. 2020;43:112–6. 31842017 10.1159/000504608

33. Raychaudhuri R Qualtieri J Garfall AL Axicabtagene ciloleucel for CD19+ plasmablastic lymphoma Am J Hematol 2020 95 E28 E30 10.1002/ajh.25682 31725923 Raychaudhuri R, Qualtieri J, Garfall AL. Axicabtagene ciloleucel for CD19+ plasmablastic lymphoma. Am J Hematol. 2020;95:E28–E30. 31725923 10.1002/ajh.25682

34. Raghunandan S Pauly M Blum WG Qayed M Dhodapkar MV Elkhalifa M BCMA CAR-T induces complete and durable remission in refractory plasmablastic lymphoma J Immunother Cancer 2023 11 e006684 10.1136/jitc-2023-006684 37137553 PMC10163502 Raghunandan S, Pauly M, Blum WG, Qayed M, Dhodapkar MV, Elkhalifa M, et al. BCMA CAR-T induces complete and durable remission in refractory plasmablastic lymphoma. J Immunother Cancer. 2023;11:e006684. 37137553 10.1136/jitc-2023-006684 PMC10163502

35. Feng S Xiong Y Liu W Liu H Sui W Zheng P BCMA CAR-T induces complete and durable remission in plasmablastic lymphoma synchronous transformation of chronic lymphocytic leukemia: case report and literature review Crit Rev Oncol Hematol 2025 205 104551 10.1016/j.critrevonc.2024.104551 39510436 Feng S, Xiong Y, Liu W, Liu H, Sui W, Zheng P, et al. BCMA CAR-T induces complete and durable remission in plasmablastic lymphoma synchronous transformation of chronic lymphocytic leukemia: case report and literature review. Crit Rev Oncol Hematol. 2025;205:104551. 39510436 10.1016/j.critrevonc.2024.104551