Background: Up to 30% of Hodgkin lymphoma (HL) patients are refractory or relapse (R/R) after first treatment and their prognosis is poor. We have developed a refined CD30-CAR-T (HSP-CAR30) targeting a proximal epitope within the CD30 molecule to overcome soluble CD30 and generated products enriched in memory T-cells to ensure efficient engraftment, persistence and enhancement of antitumor efficacy (Alvarez-Fernandez et al, 2021). Here, we report the results of our Phase 1 study evaluating HSP-CAR30 for the treatment of R/R HL and CD30+ T-cell non-Hodgkin lymphoma (T-NHL) (NCT04653649). Aims: Primary endpoints were to assess safety of HSP-CAR30 and to establish maximum tolerated dose (MTD) recommended for the following Phase 2. Secondary objectives include best response rates after infusion. Methods: We conducted a phase 1 dose-escalation study in 11 patients with R/R HL or CD30+ T-NHL. HL patients were R/R to treatments including chemotherapy, brentuximab and anti-PD-1 antibodies, while T-NHL patients were R/R to at least 2 chemotherapy treatments. T-cells were transduced with a lentivirus encoding a second-generation 4-1BB costimulated CAR, containing a scFv directed against an epitope from the proximal non-cleavable part of CD30 protein. Three cell-dose levels were evaluated: DL1 (3x106/kg), DL2 (5x106/kg) and DL3 (10x106/kg) CAR30+ T-cells. Results: From February 2021 to December 2021, 11 patients (9 HL and 2 CD30+ T-NHL) were enrolled and underwent leukapheresis. Of these, 10 patients received HSP-CAR30: 3 patients at DL1, 3 at DL2 and 4 at DL3. Demographic characteristics and baseline disease features are summarized in the Table. Median age was 49.9 years (range 21–65). Median number of prior lines of treatment was 4.6 (range 3–7). One patient did not received treatment due to lack of T-cell expansion. All patients received LD before infusion, fludarabine/bendamustine in HL patients (n=8) and fludarabine/cyclophosphamide in T-NHL (n=2). Mean HSP-CAR30 expression was 94.79±3,38% (±SD). Memory T-cell subset comprised 93.07±4,8% (±SD) in CD4+ and 91.64±4,9% (±SD) in CD8+. Mean time to HSP-CAR30 cell peak level across all doses was 29 days (range 6–98). CAR+ T-cells were detectable by flow cytometry up to 11 months after infusion. HSP-CAR30 infusion was well tolerated; there were no dose limiting toxicities (DLTs). Relevant adverse events are shown in the Table. Grade 1 cytokine release syndrome (CRS) was observed in 6 (60%) patients. No patient developed neurotoxicity. Self-limited skin rash was seen in 4 (40%) patients. One patient with history of cytomegalovirus (CMV) infections had CMV pneumonia. Another patient developed pulmonary tuberculosis. At data cutoff (February 21st, 2021), the median follow-up was 204 days (60–351). Best objective response was 100%, including 5 (50%) patients with complete response (CR), all with HL (DL1=1; DL2=3; DL3= 1). Three patients have died of progressive disease (2 T-NHL and 1 HL). There were no non-relapse mortality events. Median PFS and median overall survival (OS) was not reached. Six-month PFS for HL patients was 75%. Image:Summary/Conclusion: This is the first European academic CART clinical trial evaluating a T-cell memory-enriched CART 30. Our Phase 1 study provides evidence for feasibility and safety of HSP-CAR30. Additionally, HSP-CAR30 has shown promising efficacy in heavily treated HL patients that is being explored in a phase 2 trial already started.

Rituximab hypersensitivity reactions are rare but are one of the main causes of rituximab elimination from antilymphoma immunochemotherapy treatments. While the clinical picture may be indistinguishable from other infusion-related reactions, hypersensitivity reactions (HSR) do not disappear and instead become more intense with subsequent administrations. Objective. To describe the use of the 12-step protocol for desensitization to intravenous rituximab in clinical practice and the complementary study of a possible IgE-mediated HSR in the context of B-cell lymphoma treatment. Methods. A 12-step rituximab desensitization protocol was performed prospectively within clinical practice in 10 patients with a history of severe infusion reactions or in patients who had a repeated reaction at subsequent doses despite taking more intense preventive measures. Skin prick tests were performed at the time of reaction and at a later time to eliminate false negatives due to possible drug interference. Results. Overall, with the desensitization protocol, 70% of patients were able to complete the scheduled immunochemotherapy. Two patients had to discontinue the therapy due to clinical persistence and the third due to lymphoma progression. Intradermal tests with 0.1% rituximab were positive in only 20% of cases, demonstrating a mechanism of hypersensitivity. Conclusions. The 12-step desensitization protocol is very effective and assumable within healthcare practice. There is a need to determine the mechanism underlying the infusion reaction in a large proportion of cases due to the risk of future drug exposure.

Background:Pola‐G‐Len has the potential to enhance anti‐tumor immune response in R/R FL. Here, we report a pre‐planned interim analysis of a phase Ib/II study (NCT02600897) of Pola‐G‐Len in pts with R/R FL.Aims:To assess the safety and efficacy of induction and maintenance with Pola‐G‐Len in pts with R/R FL.Methods:This is an ongoing, open‐label, multicenter study of pts with R/R FL (excluding grade [Gr] 3b) who have received ≥1 prior anti‐CD20 antibody‐containing chemo‐immunotherapy regimen. The study comprises an initial 3+3 dose‐escalation (DE) phase (to determine the recommended phase II dose [RP2D] of both Pola and Len for the Pola‐G‐Len regimen) followed by an expansion phase to assess the RP2D of Pola and Len. Pts received induction treatment with 6x 28‐day (D) cycles (C) of: G 1000 mg IV (C1: D1, D8, D15; C2‐6: D1); Pola 1.4 mg/kg or 1.8 mg/kg (DE) or RP2D (expansion) IV (D1); and Len 10–20 mg (DE) or RP2D (expansion) PO (D1‐21). Pts with complete response (CR)/partial response (PR)/stable disease (SD) at the end of induction (EOI) received G 1000 mg (D1 every 2mo, for 24mo), and Len (10 mg, D1‐21 monthly, 12mo). Primary endpoints were C1 dose‐limiting toxicities (DLTs), safety/tolerability, CR rate at EOI (modified Lugano criteria).Results:At the interim data cut‐off (6 July 2018), 52 pts were enrolled: 9 discontinued the study (adverse events [AE], n = 3; death due to PD, n = 4; pt withdrawal, n = 1; other, n = 1). At baseline, the median pt age was 62 (range 32–87) years; 60% were male; 58% had FLIPI Gr 3–5; 79% had received ≥2 prior therapy lines; 50% were refractory to their last treatment; 17% had bulky disease (≥7 cm). Two DLTs were reported in the cohort receiving Pola 1.8 mg/kg + Len 10 mg during the DE period (Gr 4 lipase/amylase elevation; asymptomatic, resolved with supportive care; Gr 3 thrombocytopenia leading to a delay in the initiation of cycle 2). Therefore, Pola 1.4 mg/kg + Len 20 mg was selected as the RP2D for expansion. Gr ≥3 AEs were experienced by 75% of pts: neutropenia (46%), thrombocytopenia (17%), anemia (12%) and infections (12%) were the most common AEs. AEs leading to Len dose reduction or interruption occurred in 31% and 52% of pts, respectively. One Gr 5 AE was reported (septic shock after PD in pt receiving subsequent therapy). The RP2D was determined as Pola 1.4 mg/kg + Len 20 mg. Preliminary efficacy data suggest high activity, with an independent review committee‐assessed Modified Lugano response rate of 89% and a CR rate of 67% (Table). Median progression‐free survival was not reached (median follow‐up duration 8.95 mo in the efficacy‐evaluable population).imageSummary/Conclusion:The safety profile of Pola‐G‐Len is consistent with known profiles of the individual drugs. Response rates at EOI with Pola‐G‐Len are promising, with high CR compared with available R/R FL treatments.

Background: Recently, a new prognostic model (T cell score) has been developed by the International T cell Project Network that identifies a group of peripheral T cell lymphoma otherwise specified (PTCL-NOS) with very unfavorable outcomes. However, validation in independent cohorts remains as an essential step for models to gain widespread use. Aims: To validate T cell score, based on series of three hospitals from Spain. Furthermore, we explore the usefulness of T score in other subtypes of peripheral T cell lymphomas (PTCLs). Methods: T cell score is based on four covariates: serum albumin, ECOG-performance status, stage and absolute neutrophil count (ANC). The T cell score is categorized in three risk groups as describe in the original paper. T cell score was applied retrospectively to series of all PTCLs in three Institutions. Validation was performed following the guidelines by Royston et al. Overall survival (OS) and progression-free survival (PFS) were calculated using Kaplan-Meier estimators, comparison between categories performed by log-rank test and Cox PH regression, and the effect of covariates reported with 95% confidence interval (95 CI). Results: Between 1993 and 2018, 129 cases of PTCLs were identified for developing the prognostic model. The median age was 56 years (range 15-91), 62 % of patients (pts) were male, and advanced stage disease was found in 77%. PTCLs subtypes: PTCL-NOS 52 pts (40%), AITL 24 pts (26%), ALCL (ALK+)10 pts (8%), ALCL (ALK-)17 pts (13%) and others 16 pts (12%). 77% of PTCL had advanced disease and 35% had ECOG ≥ 2. 71% of PTLC had albumin <35 g/l, 68% >6.5x109 cells/l. 84% of PTCLs were treated with anthracyclines. At a median follow-up of 102 months, the median OS and PFS was 30 and 9.5 months, respectively. Three groups of PTCL-NOS were separated for OS at low risk (LR, 6 patients, 11.5%, score 0), intermediate risk (IR, 29 patients, 55.8%, score 1–2), and high risk (HiR, 17 patients, 32.7%, score 3–4) (figure 1). The three risk groups had a 5-year OS of 100%, 42 % [95 CI 24–61], and 6% [95 CI 0–17], respectively for patients at LR, IR and HiR respectively (p < 0.001) (Fig 1). The model also proved to be a potential tool for PFS; the 3-year PFS was 33% [95 CI 0-70], 38 % [95 CI 19-56] and 0, respectively for patients at LR, IR and HiR respectively (p < 0.001; data not shown).Summary/Conclusion: Validation of the T cell score was performed in our series of PTCL-NOS from three hospitals. T cell score clearly identified three risk groups for OS. Further validation of the T cell score in other independent series and analyze its potential application in other subtypes of PTCL, need to be performed.

Posttransplant high-dose cyclophosphamide (PTCy) effectively prevents GvHD after haploidentical SCT. However, its use in HLA-matched SCT has been less explored. Fifty-six consecutive patients who underwent allo-SCT for hematological malignancies have been included in this prospective single-center protocol. Donors have been HLA-identical siblings, fully-matched unrelated or 1-allele-mismatched unrelated donors in 30%, 32%, and 37% of cases, respectively. Nine patients have received a TBI-containing MAC regimen, while the remaining (84%) received RIC platforms based on Fludarabine plus Busulfan/Melphalan. Due to the high graft failure (GF) rate (21%) in a preliminary analysis in the allo-RIC cohort ( n = 29), protocol amendments have been implemented, with no further cases of GF after the introduction of mini-thiotepa (0/18). The overall incidence of grade II–IV acute GvHD is 24% (95% CI: 17–31%) with four steroid-refractory cases. Severe chronic GvHD has occurred in only 1 of 43 evaluable cases. The 1-year NRM and relapse are 18% (95% CI: 12–26%) and 30% (18–42%) and the OS and DFS are 78% and 64%, respectively. These outcomes support the feasibility of using PTCy as a SOC outside the haplo-setting, albeit mini-thiotepa (3 mg/kg) was incorporated in the standard allo-RIC platforms to prevent GF. Despite the limitations of a single-center experience and the short follow-up, these protocols show promising results with particular benefit in reducing the occurrence of moderate-to-severe GvHD.

Background:Post‐transplantation cyclophosphamide (PTCy) is an effective graft vs host disease (GvHD) prophylaxis strategy in haploidentical stem cell transplant (HaploSCT), with comparable survival outcomes to HLA‐matched SCT. Due to its promising results, some centers are including this strategy in 9/10 or 10/10 HLA‐matched related or unrelated donor SCT (MatchSCT) settings.Aims:PTCy has shown to reduce NRM by decreasing GvHD compared to classical immunosuppressive agents, although there are a few studies focusing on the etiology and outcomes of infections after PTCy‐based HaploSCT and MatchSCT.Methods:We selected 75 consecutive alloSCT adult patients who received PTCy‐based GvHD prophylaxis in our center from December 2013 to August 2018. Patients were divided in 2 groups: (1)HaploSCT, n = 32; (2)MatchSCT, n = 43. Baseline characteristics are shown in Table1. For each patient, we reported every episode of severe infection, including gram‐positive and gram‐negative bacteremias, pneumonia, viral hemorrhagic cystitis, viral LRTI, CMV and VEB reactivations and/or disease and proven/probable invasive fungal infections (IFI) at different time points after SCT: early (≤30 days post‐SCT) and intermediate/late infections (≥31 days post‐SCT).Results:Patients developed a total of 127 severe infectious episodes; 57 of them (45%) were early infections and 70 (55%) were intermediate/late cases, which distributed evenly between the 2 groups. All these infections are shown in Table 2. Bacteremia episodes were the most common cause of severe infection, accounting for 35% of documented infections (51% in group 1 and 49% in group 2). Gram‐positive bacteria CVC infections were the most frequent cause of bacteremia (56%). 72% of all bacterial infections were diagnosed ≤30d post‐SCT. Four episodes of proven/probable IFI were diagnosed ≤100 days after SCT, 3 of them during the early phase post‐SCT. Unsurprisingly, 79% of all viral infections were diagnosed after day +30 (49% HaploSCT and 51% MatchSCT). Most of them were CMV reactivations, with only one case of CMV disease. Among the 50 patients sequentially monitored for EBV viral load (32 in group 1 and 18 in group 2), 3 cases of EBV reactivation were diagnosed in group 1 and received rituximab as preemptive treatment. A total of 16 viral hemorrhagic cystitis episodes were identified (12 in group 1 and 4 in group 2): BK virus‐related (n = 13) and adenovirus (n = 3). Infection‐related mortality (IRM) was the cause of death in 9% of the patients in both groups, with a median follow‐up for survivors of 821 days for group 1 (range 278–1940) and 339 days for the group 2 (range, 64–959). The 1‐year cumulative incidence of IRM was 8.9% (95% CI, 2.7–15) vs 9% (95% CI, 3–14.8) in group 1 and 2, respectively. NRM occurred in 4 (12.5%) and 6 (14%) patients in each group, with a 1‐year cumulative incidence of 12.5% (95% CI, 7–18%) vs 17% (95% CI, 7–27), in group 1 and 2, respectively (p = 0.68). Fifty‐five patients (73.3%) were alive at last follow‐up, with a 1‐year overall survival of 72% (95% CI, 64–80) in group 1 and 80% (95% CI, 72–87) in group 2 (p = 0.5).Summary/Conclusion:In our study, IRM was the main cause of 1‐year NRM, although their incidence was low in both groups. With PTCy the 1‐year IRM seems much inferior than with other approaches. In a real life setting this strategy reduces IRM and NRM not only by reducing GvHD, but also by allowing early protective immune reconstitution. The small sample sizes did not allow to reliably compare the types of infections in different engraftment time periods, which will be the objective of future studies.image

GCB N=35 ABC N=11 UNC N=9 P-value (GCB vs. ABC) Introduction: Diffuse large B-cell lymphoma (DLBCL) can be divided according to cell-of-origin (COO) in germinal center B-cell–like (GCB) and activated B-cell–like (ABC). Although the importance of COO determined with the Lymph2Cx assay is well established in DLBCL arising in immunocompetent individuals, there are no reports on its use in HIV-infected patients. We aimed to study the characteristics and prognostic impact of COO subtypes in a series of HIV-related DLBCL using the Lymph2Cx assay and to compare the results with those obtained with Hans algorithm. Methods: A series of 55 patients with the diagnosis of HIV-related DLBCL (N = 48), high-grade B-cell lymphoma (HGBL) with MYC and BCL2 and/or BCL6 rearrangements (N = 3), or HGBL NOS (N = 4) was studied. The following clinical parameters were collected: age, gender, ECOG, extranodal and bulky disease, B-symptoms, Ann-Arbor stage, LDH and beta2-microglobulin, HCV and HBV serology, history of opportunistic infection and of AIDS-defining illness, onset of combination antiretroviral therapy, CD4-counts, and HIV-loads. IHC and FISH studies were performed on tissue microarrays. RNA was extracted from FFPE samples with RecoverAll kit (Ambion, Carlsbad, CA), and digital GEP was determined with the Lymph2Cx assay (NanoString Technologies, Seattle, WA). Cohen's kappa was calculated to measure the agreement between COO given by Hans algorithm and Lymph2Cx assay. Results: Allocation of COO subtypes with the Lymph2Cx assay was 63.6% GCB, 20% ABC, and 16.4% unclassified. The only clinical feature significantly associated with a defined COO subtype was B-symptoms (ABC = 81.8% vs GCB = 28.6%, P = .003) and detectable HIV-loads tended to be more frequently observed in ABC (90%) than in GCB (58.1%, P = .066). Regarding IHC and FISH characteristics (Table 1), MYC rearrangements were only detected in GCB cases and expression of CD10 and BCL6 tended to be associated with GCB. Hans algorithm and Lymph2Cx assay differently assigned COO subtypes (κ = −0.288, P = .029) showing that 44.1% of the GCB cases had a non-GC phenotype according to Hans. The median follow-up of living patients was 8.5 years. Only patients treated with RCHOP were considered in survival analyses (N = 47). COO subtypes had neither impact on OS nor PFS, independently of being determined with Lymph2Cx assay or Hans. Conclusions: In HIV-related lymphomas, COO subtypes were discordantly assigned with Hans and Lymph2Cx assay and COO subtypes showed no impact on outcomes, independently of the method applied. Keywords: GCB lymphoma subtype; human immunodeficiency virus (HIV); immunodeficiency-associated lymphomas

Introduction: About 60-70% of DLBCL patients are currently cured with immunochemotherapy. However, patients with early failure (EF), either primary refractory or early progressions, show a dismal outcome independently of standard prognostic factors, including the cell of origin (COO). The aim of the present study was to analyze the genetic profile of EF patients, in order to identify predictive factors of response to R-CHOP and understand mechanisms of resistance. Methods: We selected 121 patients (65 M/56F; median age, 63 years; median overall survival 9.2 years) diagnosed with de novoDLBCL not otherwise specified at GELCAB centers, treated with R-CHOP. Thirty four cases (28%) either primary refractory or relapsed within 12 months were considered EF. Median follow-up for surviving patients was 4.7 yrs. Genomic profiling included targeted next generation sequencing of 106 representative genes, Cytoscan HD arrays and gene expression profiling (GEP). COO was established by gene expression based assays. Results: Main clinical features predicting EF are detailed in the table. EF cases had more frequently mutations in KLHL6 and gains in 12q (CDK4), 11q (ETS1) and 5p (TERT). Regarding signaling pathways, alterations in the immune surveillance (CD58, B2M, CIITA), TP53/CDKN2A and NOTCH (NOTCH1&2, SGK1 and FBXW7) pathways also predicted EF (table). Of note, among GCB cases, those with EF showed more frequently mutations of KLHL6, MYC, HIST1H1E and SGK1, and gains in Chr 12. In multivariate analysis only TP53/CDKN2A (HR 2.2; P = .03) and NOTCH (HR 2.5; P = .01) pathways were independently associated with EF. Finally, GEP performed in 41 tumors showed that EF cases had over-representation of pathways related to inflammatory responses mediated by IFNα, IFNγ, IFN-response factors and IL-1, IL6-, IL-8, together with signatures related to leukocyte infiltration. Conclusions: Different genetic profile aberrations, including alterations in TP53/CDKN2A and NOTCH pathways, can predict EF after R-CHOP better than standard prognostic factors, including COO. Further studies are necessary to validate these findings in prospective series. Keywords: diffuse large B-cell lymphoma (DLBCL); gene expression profile (GEP); R-CHOP.