Fractionated initial infusion and booster dose of ARI0002h, a humanised, BCMA-directed CART-cell therapy, for patients with relapsed or refractory multiple myeloma (CARTBCMA- HCB-01): a single-arm, multicentre, academic pilot study

www.thelancet.com/oncology Vol 24 August 2023 913 Articles Lancet Oncol 2023; 24: 913–24 Published Online July 3, 2023 https://doi.org/10.1016/ S1470-2045(23)00222-X See Comment page 828 Hospital Clínic de Barcelona. IDIBAPS. University of Barcelona, Barcelona, Spain (A Oliver-Caldés MD, M Español-Rego MS, S Varea MSc, L Rosiñol MD, N Tovar MD, R Jiménez MD, L G Rodríguez-Lobato MD, E Olesti PhD, M Calvo-Orteu BS, J Delgado MD, J Sáez-Peñataro MD, M Juan MD, M Pascal PhD, A Urbano-Ispizua MD, C Fernández de Larrea MD); Hospital Universitario de Salamanca, Instituto de Investigación Biomédica de Salamanca (IBSAL), Centro de Investigación del Cancer (IBMCC-USAL, CSIC), Salamanca, Spain (V González-Calle PhD, L López-Corral PhD, Miriam López-Parra MD, M V Mateos PhD); Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria Pascual Parrilla, University of Murcia, Murcia, Spain (V Cabañas PhD, A Sánchez-Salinas MD, J M Moraleda MD); Clínica Universidad de Navarra, Centro de Investigación Médica Aplicada (CIMA), Instituto de Investigación Sanitaria de Navarra (IDISNA), CIBERONC, Pamplona, Pamplona, Spain (P Rodríguez-Otero MD, A Zabaleta PhD, S Inogés MD, A López-Díaz de Cerio PhD, B Paiva PhD, F Prósper MD); Hospital Universitario Virgen del Rocío, Instituto de Fractionated initial infusion and booster dose of ARI0002h, a humanised, BCMA-directed CAR T-cell therapy, for patients with relapsed or refractory multiple myeloma (CARTBCMAHCB-01): a single-arm, multicentre, academic pilot study Aina Oliver-Caldés, Verónica González-Calle, Valentín Cabañas, Marta Español-Rego, Paula Rodríguez-Otero, Juan Luis Reguera, Lucía López-Corral, Beatriz Martin-Antonio, Aintzane Zabaleta, Susana Inogés, Sara Varea, Laura Rosiñol, Ascensión López-Díaz de Cerio, Natalia Tovar, Raquel Jiménez, Miriam López-Parra, Luis Gerardo Rodríguez-Lobato, Andrés Sánchez-Salinas, Eulàlia Olesti, Maria Calvo-Orteu, Julio Delgado, José Antonio Pérez-Simón, Bruno Paiva, Felipe Prósper, Joaquín Sáez-Peñataro, Manel Juan, José M Moraleda, María-Victoria Mateos, Mariona Pascal, Alvaro Urbano-Ispizua, Carlos Fernández de Larrea Summary Background Chimeric antigen receptor (CAR) T-cell therapy is a promising option for patients with heavily treated multiple myeloma. Point-of-care manufacturing can increase the availability of these treatments worldwide. We aimed to assess the safety and activity of ARI0002h, a BCMA-targeted CAR T-cell therapy developed by academia, in patients with relapsed or refractory multiple myeloma. Methods CARTBCMA-HCB-01 is a single-arm, multicentre study done in five academic centres in Spain. Eligible patients had relapsed or refractory multiple myeloma and were aged 18–75 years; with an Eastern Cooperative Oncology Group performance status of 0–2; two or more previous lines of therapy including a proteasome inhibitor, an immunomodulatory agent, and an anti-CD38 antibody; refractoriness to the last line of therapy; and measurable disease according to the International Myeloma Working Group criteria. Patients received an initial fractionated infusion of 3 × 10⁶ CAR T cells per kg bodyweight in three aliquots (0·3, 0·9, and 1·8 × 10⁶ CAR-positive cells per kg intravenously on days 0, 3, and 7) and a non-fractionated booster dose of up to 3 × 10⁶ CAR T cells per kg bodyweight, at least 100 days after the first infusion. The primary endpoints were overall response rate 100 days after first infusion and the proportion of patients developing cytokine-release syndrome or neurotoxic events in the first 30 days after receiving treatment. Here, we present an interim analysis of the ongoing trial; enrolment has ended. This study is registered with ClinicalTrials.gov, NCT04309981, and EudraCT, 2019-001472-11. Findings Between June 2, 2020, and Feb 24, 2021, 44 patients were assessed for eligibility, of whom 35 (80%) were enrolled. 30 (86%) of 35 patients received ARI0002h (median age 61 years [IQR 53–65], 12 [40%] were female, and 18 [60%] were male). At the planned interim analysis (cutoff date Oct 20, 2021), with a median follow-up of 12·1 months (IQR 9·1–13·5), overall response during the first 100 days from infusion was 100%, including 24 (80%) of 30 patients with a very good partial response or better (15 [50%] with complete response, nine [30%] with very good partial response, and six [20%] with partial response). Cytokine-release syndrome was observed in 24 (80%) of 30 patients (all grade 1–2). No cases of neurotoxic events were observed. Persistent grade 3–4 cytopenias were observed in 20 (67%) patients. Infections were reported in 20 (67%) patients. Three patients died: one because of progression, one because of a head injury, and one due to COVID-19. Interpretation ARI0002h administered in a fractioned manner with a booster dose after 3 months can provide deep and sustained responses in patients with relapsed or refractory multiple myeloma, with a low toxicity, especially in terms of neurological events, and with the possibility of a point-of-care approach. Funding Instituto de Salud Carlos III (co-funded by the EU), Fundación La Caixa, and Fundació Bosch i Aymerich. Copyright © 2023 Elsevier Ltd. All rights reserved. Introduction The survival of patients with multiple myeloma has improved as a result of the incorporation of the combination of proteasome inhibitors, immunomodulating agents, and anti-CD38 monoclonal antibodies into the standard of care since 2008. Still, a large proportion of patients continue to relapse and multidrug resistance remains an important challenge leading to poor outcomes in patients with relapsed or refractory multiple myeloma.1,2 Chimeric antigen receptor (CAR) T-cell therapy has emerged as a promising option for relapsed or refractory multiple myeloma. Several CAR T-cell products targeting BCMA using different approaches in terms of origin of Articles 914 www.thelancet.com/oncology Vol 24 August 2023 Biomedicina de Sevilla (IBIS/ CSIC), University of Seville, Seville, Spain (J L Reguera MD, J A Pérez-Simón PhD); Department of Experimental Hematology, Instituto de Investigación SanitariaFundación Jiménez Díaz, University Autonomous of Madrid, Madrid, Spain (B Martin-Antonio PhD) Correspondence to: Dr Carlos Fernández de Larrea, Hospital Clínic de Barcelona, 08036 Barcelona, Spain [email protected] the antigen-recognition domain (murine, humanised, human, or llama), co-stimulatory domain (4-1BB and CD28), and transduction method (lentiviral, retroviral, or transposon)3 are under clinical investigation. Nevertheless, only two BCMA CAR T-cell therapies, idecabtagene vicleucel and ciltacabtagene autoleucel, have been approved by the US Food and Drug Administration after at least four previous lines of therapy and the European Medicines Agency after at least three for the treatment of multiple myeloma, including proteasome inhibitors, immuno modulating agents, and anti-CD38 monoclonal antibodies.4,5 The idecabtagene vicleucel study reported an overall response rate of 73%, with 33% of patients having complete responses (26% stringent complete responses), with a median progression-free survival of 8·8 months (95% CI 5·6–11·6) and some expected CAR T cell-related adverse events: cytokine-release syndrome in 107 (84%) of 128 patients (5% grade ≥3), immune effector cellassociated neurotoxicity syndrome (ICANS) in 23 (18%; 3% grade ≥3), and cytopenias (neutropenia 117 [91%], anaemia 89 [70%], and thrombocytopenia 81 [63%]).6 In a phase 3 clinical trial, idecabtagene vicleucel showed improved responses and progression-free survival compared with standard care in patients exposed to three families of drugs: proteasome inhibitors, immunomodulating agents, and anti-CD38 antibodies.7 The 2-year update8 of the ciltacabtagene autoleucel phase 3 trial showed an overall response rate of 97·9%, with 82·5% being stringent complete responses in 97 patients with multiple myeloma treated with 0·75 × 10⁶ CAR T cells per kg bodyweight. Median progression-free survival and overall survival were not reached after a median follow-up of 27·7 months. Cytokine-release syndrome was reported in 92 (95%) of 97 patients (4% grade ≥3) and neurotoxicity in 20 (21%; nine [9%] grade ≥3), including both ICANS and other neurotoxicities. Parkinsonism symptoms occurred in six (6%) patients, with one related death and two deaths due to other causes. Parkinsonism-like neurotoxicity decreased to less than 1% after patient management strategies.9 A study10 of the first allogeneic BCMA-CAR T-cell therapy (ALLO-715) reported an overall response rate of 71%, including 25% or more with complete response, with a very short median time (5 days [range 0–20]) from enrolment to lymphodepletion. Our academic institution developed two CAR T-cell constructs; one directed against CD19 (ARI-0001; varnimcabtagene autoleucel) and another against BCMA (ARI0002h). The CART19-BE-01 multicentre clinical trial with ARI-0001 for adult and paediatric CD19-positive malignancies led to its approval as hospital exemption in Spain in February, 2021, making it the first academic CAR T-cell therapy in clinical use in Europe to our knowledge.11–13 From our previous experience treating CD19 malignancies with ARI-0001, we observed how administering the initial dose in a fractionated manner might diminish severity of Research in context Evidence before this study Effective therapies with acceptable safety profiles are needed to improve outcomes of patients with relapsed or refractory multiple myeloma, particularly for patients with disease that is refractory to immunomodulatory drugs, proteasome inhibitors, and anti-CD38 monoclonal antibodies, who have few treatment options available. We searched PubMed on Feb 15, 2023, for publications in English, with no date restriction, using the keywords “myeloma”, “B-cell maturation antigen”, “chimeric antigen receptor T (CART)” AND “fractionated dose” OR “booster dose”. Although multiple trials assessing BCMA-targeted chimeric antigen receptor (CAR) T-cell therapies and related adoptive cellular approaches are available, we found no published studies that used a fractionated dose or a booster dose. Added value of this study In this single-arm, multicentre study using ARI0002h (an autologous CAR T-cell product targeting BCMA that is under development at Hospital Clinic de Barcelona, an academic institution in Spain), a fractionated dose of ARI0002h followed by a booster dose 100 days later led to early, deep, and durable responses in adult patients with relapsed or refractory multiple myeloma who had received at least two previous treatment regimens, including a proteasome inhibitor, an immunomodulatory drug, and an anti-CD38 monoclonal antibody. This trial reports novel findings in the field of immunotherapy for multiple myeloma. First, we reduced the potential immunogenicity of the CAR by using a humanised single-chain variable fragment for the recognition of BCMA, instead of one of full animal origin (mouse or llama), as is the case for the already approved commercial CAR T cells. Second, we lowered toxicity by using a fractionated administration scheme— no cases of neurotoxicity were observed. Third, a second dose was planned after 100 days in patients with response and no relevant toxicity, providing fresh and active CAR T cells to consolidate or deepen the response. Finally, this trial incorporated the concept of point-of-care treatment in cellular therapy, showing that a strategy based on two production sites and five treating hospitals is possible in a public health system. Implications of all the available evidence ARI0002h is a viable treatment option for patients with relapsed or refractory multiple myeloma, with an activity and safety profile comparable to approved CAR T-cell products. A point-of-care strategy to facilitate real access and administration of the product is crucial, given limited patient access to these treatments in the USA and Europe because of the lack of manufacturing capacity and reimbursement. Articles www.thelancet.com/oncology Vol 24 August 2023 915 immune-related side-effects without reducing efficacy, although that study was not specifically designed to compare outcomes of a fractionated versus nonfractionated infusion.12,14 ARI0002h is a humanised 4-1BB-based BCMA-CAR T-cell therapy, lentivirally transduced on autologous T cells obtained by peripheral blood leukapheresis, that has proven antitumour activity in preclinical in vitro and in vivo approaches.15 We aimed to investigate the safety and activity of ARI0002h in patients with relapsed or refractory multiple myeloma after an initial fractionated infusion and a non-fractionated booster dose, administered at least 3 months after the first infusion in patients with some degree of response and no limiting side-effects. Methods Study design and participants CARTBCMA-HCB-01 is a single-arm, open-label study done in five academic centres in Spain. Eligible patients had relapsed or refractory multiple myeloma; were aged 18–75 years; had an Eastern Cooperative Oncology Group performance status of 0–2; two or more previous lines of therapy including a proteasome inhibitor, an immuno modulating agent, and an anti-CD38 antibody; refractoriness to the last line of therapy; and measurable disease (serum and urine monoclonal protein >10 g/L or 200 mg/24 h and involved free light chain >100 mg/L) according to the International Myeloma Working Group (IMWG) criteria;16 and a life expectancy of more than 3 months. Exclusion criteria included previous BCMAdirected therapy and a non-adequate organ system function including an estimated glomerular filtration rate below 50 mL/min. The full list of exclusion criteria, study protocol, and participating sites are summarised in the appendix (pp 2–5). Patients provided written informed consent. The study protocol was approved by the Ethics Committee of Hospital Clínic de Barcelona and was done in accordance with the Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects. Procedures Clinical coordination and vector viral production were done in Hospital Clínic de Barcelona (Barcelona, Spain). Two centres were responsible for CAR T-cell production: Hospital Clínic de Barcelona and Clínica Universidad de Navarra (Pamplona, Spain; appendix p 6). Three CAR T development strategies were adopted to improve outcomes. Firstly, the murine single-chain variable fragment (scFv), obtained from the J22.9 antibody, was humanised to reduce immunogenicity. We did preclinical experiments to show non-inferiority between constructs containing the humanised scFv versus murine scFv.15 Second, the first dose was fractioned into 3 aliquots to reduce toxicity. Finally, we planned a second infusion (booster dose) of ARI0002h at least 3 months after the first infusion, as an experimental attempt to improve CAR T-cell persistence and response. Patient characteristics, including sex, were defined by electronic medical records. Race and ethnicity data were not collected. We assessed patients for eligibility and proceeded to leukapheresis in their respective centres. Fresh apheresis products were sent to one of the two production centres. The target dose was 3 × 10⁶ CARpositive cells per kg bodyweight for the first infusion and a second dose of up to 3 × 10⁶ CAR-positive cells per kg was also obtained, when possible. The final product was cryopreserved. Bridging therapy was allowed in the period between apheresis and lymphodepletion according to investigator’s choice. Lymphodepletion was administered intravenously on days –6 to –4 (before infusion) and included fludarabine (30 mg/m²/day; total dose 90 mg/m²) and cyclophosphamide (300 mg/m²/day; total dose 900 mg/m²). The first CAR T infusion was split into three administrations of 0·3 (10%), 0·9 (30%), and 1·8 × 10⁶ (60%) CAR-positive cells per kg intravenously on days 0, 3, and 7, with at least 24 h between doses in all cases. If adverse events occurred between administrations, remaining doses were adjusted until resolution. The booster dose of up to 3 × 10⁶ CAR-positive cells per kg was administered in a single intravenous infusion after day 100 in patients with some degree of response and no limiting side-effects after the first dose, including grade 3–4 cytokine-release syndrome or ICANS and other adverse events of interest (persistent cytopenias or macrophage activation syndrome). Lymphodepletion was readministered with the same scheme only in patients without CAR T-cell persistence in peripheral blood. We followed up patients for 36 months, or until progression or death. Patients could be removed from the study on the basis of patient decision. Laboratory monitoring was planned in the following days and months from infusion: days 1, 10, 14, 21, 28, 35, 42, 56, 70, 84, and 100; and months 4, 6, 7, 8, 9, 10, 11, 12, 15, 18, 21, 24, 27, 30, 33, and 36. Bone marrow aspirates assessing measurable residual disease by next generation flow cytometry at a sensitivity of 1 × 10–⁶ were planned on days 28 and 100 and months 6, 12, 18, and 24. Assessment was done using two-tube eight-colour flow cytometry according to the EuroFlow platform, using a FACSCanto (BD Biosciences, [San Jose, CA, USA]) flow cytometer and Infinicyt 2.0 software (Cytognos SL, Salamanca, Spain; appendix p 7). We considered any detectable level of measurable residual disease greater than 1 × 10–⁶ to be positive. PET-CT scans were planned at screening and on day 100 and month 12 to assess plasmacytomas. Multiple myeloma disease evaluation was planned on days 28, 56, and 100; and months 4–12, 15, 18, 21, 24, 27, 30, 33, and 36. Responses were assessed according to IMWG criteria (appendix pp 6–7).17 Adverse events were monitored in all follow-up visits, including a daily monitoring from day of infusion until hospital discharge. Cytokine-release See Online for appendix Articles 916 www.thelancet.com/oncology Vol 24 August 2023 syndrome and ICANS were assessed according to the American Society for Transplantation and Cellular Therapy consensus.18 Intravenous immunoglobulins could be administered according to local guidelines when IgG concentrations were below 400 mg/dL. Samples were obtained from peripheral blood and bone marrow for correlative studies (appendix p 10). BCMA expression on bone marrow plasma cells was measured by flow cytometry (PE anti-human CD269 [BCMA] antibody; Biolegend [San Diego, CA, USA]) at baseline and in measurable residual disease-positive disease. Molecules of BCMA were quantified using the BD QuantiBRITE Beads (BD Biosciences; molecules/ cell). The kinetics of ARI0002h in the peripheral blood were measured by quantitative PCR (qPCR) assessing the time course of vector transgene Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element to elucidate the number of ARI0002h copies per cell. Inmunogenicity against ARI0002h was assessed by flow cytometry on Attune Next (Invitrogen, Thermofisher Scientific [Waltham, MA, USA]) flow cytometer to assess the presence of human anti-human antibodies in the peripheral blood (appendix pp 8–9). Outcomes The primary safety endpoint was the proportion of patients who developed cytokine-release syndrome, ICANS, or both in the first 30 days after ARI0002h administration. The primary activity endpoint was overall response rate in the first 100 days of infusion, defined as achievement of at least partial response.17 Secondary endpoints were complete response rates at 100 days and 6 months from infusion, overall response at 6 and 12 months, time to best response and time to complete response, measurable residual disease negativity rates at day 100 and month 6, plasmacytoma assessment by PET-CT at day 100, duration of response, progression-free survival, progression-free survival at 12 months, overall survival, presence of infusion reactions, tumour lysis syndrome, neurotoxicity besides ICANS, prolonged cytopenias defined as the reduction of neutrophil or platelet peripheral blood counts, grade 3 or 4 after 4 weeks from infusion, ARI0002h persistence, BCMA expression, and soluble BCMA. Quality-of-life assessment was also prespecified in the protocol but results are not provided since data are not available. Duration of response was defined as the time between first response and disease progression. Progression-free survival was defined as the time between infusion of ARI0002h and disease progression or death. Overall survival was defined as the time between infusion of ARI0002h and death due to any cause. Time to best response was defined as time between infusion and deepest response achieved according to IMWG response criteria. Time to complete response was defined as time between infusion and achievement of complete response in applicable patients. Statistical analysis We initially proposed to recruit patients with the objective to treat 30 patients. We assumed that some patients would not be treated with ARI0002h because of early progression before or after apheresis. We estimated a pre-treatment patient loss of 20%; therefore, we reasoned that 36 patients would be necessary to achieve the 30 administrations expected. Sample size calculation was not done. Response analyses were done in all patients who received at least one fraction of ARI0002h. We had two hypotheses: (1) overall response rate in the first 100 days would be at least 80%, which would allow us to rule out with 95% confidence that the real overall response rate is below 65% in a population similar to the one studied; and (2) overall response rate in the first 100 days would be at least 70%, which would allow us to rule out with 95% confidence that the real overall response rate is below 54% in a population similar to the one studied. The analysis population for safety was patients who received at least one cycle of lymphodepletion. Here, we present results of the planned interim analysis with a cut-off date of Oct 20, 2021 to initiate regulatory review by the Spanish National Competent Authority (SNCA) and thus receive authorisation for hospital exemption. Interim data is provided to the Spanish National Competent Authority for the rolling review on a periodic basis, including information regarding safety and activity after the administration of the booster dose. At the cut-off date of the planned Figure 1: Trial profile 44 patients assessed for eligibility 9 excluded 4 non-compliance of selection criteria 3 logistical issues 2 other treatment options were considered 35 enrolled 5 did not receive intervention 2 progression or death before apheresis 3 progression or death before final product release 30 received ARI0002h 14 discontinued 12 progression 2 death without progression 16 continued 30 included in analyses Articles www.thelancet.com/oncology Vol 24 August 2023 917 interim analysis, general follow-up and follow-up after the booster dose was short. To provide scientifically relevant data after a longer follow-up of patients, we did also a post-hoc analysis with a cut-off date of May 15, 2022. Duration of response, progression-free survival, overall survival, time to best response, and time to complete response were analysed using the Kaplan-Meier method. Survival calculations between independent groups were tested with the log-rank test. When appropriate (eg, for response), 95% CIs were calculated using the Wilson score method. For survival times, 95% CIs were calculated with the log method. Reasons for censoring were last follow-up without progression or death without progression for duration of response, last follow-up without progression or death for progression-free survival, and last follow-up without death for overall survival. Post-hoc analyses of vein-to-vein time according to receipt of bridging therapy (yes or no), response and progression-free survival according to receipt of the full first dose of ARI0002h (yes or no), expansion of CAR T-cells in peripheral blood after booster dose according to receipt of lymphodepletion (yes or no), duration of response and progression-free survival according to the achievement of complete response, progression-free survival and overall survival according to the presence of plasmacytomas were done. p values from post hoc comparisons are provided, which should be considered informative and descriptive, but not conclusive. We calculated p values of categorical variables with χ² or Fisher’s exact test according to sample size. We calculated p values of continuous variables, including paired sample comparisons, with the Student’s t test, Wilcoxon–Mann–Whitney, or Wilcoxon signed-rank test according to their adherence to the Gaussian distribution tested with the Saphiro– Wilk test. If appropriate, p values were adjusted for multiple comparisons with the Benjamini–Hochberg method. A p value of less than 0·05 was considered statistically significant. Statistical analyses were done with SAS System (version 9.4) and R (version 4.3.0). This study is registered with ClinicalTrials.gov, NCT04309981, and EudraCT, 2019-001472-11. Patients treated with ARI0002h (n=30) Age, years 61 (53–65) Sex Female 12 (40%) Male 18 (60%) Median time since diagnosis, years 4·7 (3·7–9·1) Heavy chain isotype IgG 14 (47%) IgA 8 (27%) Bence Jones 7 (23%) IgD 1 (3%) Light chain isotype κ 15 (50%) λ 15 (50%) ISS stage I 5/25 (20%) II 8/25 (32%) III 12/25 (48%) ECOG performance status 0 18/29 (62%) 1 9/29 (31%) 2 2/29 (7%) Plasmacytomas 14 (47%) Extramedullary plasmacytomas 6 (20%) High-risk cytogenetics* 10 (33%) TP53 alterations 7 (23%) t(4;14) 4 (13%) t(14;16) 1 (3%) Number of previous lines 3·5 (2·8–5·0) Triple exposed† 30 (100%) Triple refractory† 20 (67%) Penta exposed‡ 11 (37%) Penta refractory‡ 7 (23%) (Table 1 continues in next column) Patients treated with ARI0002h (n=30) (Continued from previous column) Refractory to the last line 30 (100%) Previous drug exposure Bortezomib 30 (100%) Carfilzomib 15 (50%) Ixazomib 2 (7%) Lenalidomide 30 (100%) Lenalidomide refractory 22 (73%) Thalidomide 17 (57%) Pomalidomide 17 (57%) Daratumumab 30 (100%) Daratumumab refractory 27 (90%) Previous autologous stem-cell transplantation 28 (93%) Previous allogeneic stem-cell transplantation 4 (13%) Previous autologous and allogeneic stem-cell transplantation 4 (13%) Bone marrow plasma cells 11·0% (1·0–32·5) Serum monoclonal protein, g/L 12·5 (0–30·4) Urine monoclonal protein, g/24 h 0·08 (0–1·08) Differential sFLC, mg/L 443·2 (154·9–1144·7) Data are n (%), n/N (%), or median (IQR). ECOG=Eastern Cooperative Oncology Group. ISS=international staging system. sFLC=serum free light chain. *Some patients had more than one high-risk cytogenetic abnormality: one patient had del(17p) plus t(4;14) and another patient had del(17p) plus t(14;16). †To a proteasome inhibitor (bortezomib or carfilzomib), an immunomodulatory drug (lenalidomide or pomalidomide), and an anti-CD38 monoclonal antibody (daratumumab). ‡To bortezomib, carfilzomib, lenalidomide, pomalidomide, and an anti-CD38 monoclonal antibody. Table 1: Baseline characteristics of patients treated with ARI0002h Articles 918 www.thelancet.com/oncology Vol 24 August 2023 Role of the funding source The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. Results Between June 2, 2020, and Feb 24, 2021, we assessed 44 patients for eligibility (figure 1). We enrolled 35 patients, of whom 33 (94%) underwent leukapheresis and 30 (86%) received ARI0002h (figure 1). The median manufacturing time of ARI0002h was 10 days (IQR 9–10), with a mean transduction rate of 56% (SD 25). All final products for the first infusion were successfully obtained at the first attempt except one, which was produced with a second apheresis. Median turnaround time, defined as days from apheresis reception to product liberation, was 30 days (IQR 26–36; range 19–45). In a patient requiring urgent treatment, the product was released in as fast as 19 days. The median age of patients who received ARI0002h was 61 years (IQR 53–65), 12 (40%) were female, and 18 (60%) were male (table 1). 14 (47%) of 30 patients presented with plasmacytomas at inclusion, with six (20%) presenting with a true extramedullary, non-paraskeletal location (table 1). Assessment of disease before and after bridging therapy showed either no change or progression of the serum M-protein or free light chain in six patients (43%); in eight (57%) a decrease was observed, with no patients having a complete response (appendix p 11). All 30 patients received fractions 1 and 2. Five (17%) patients did not receive the third fraction of the first dose of ARI0002h because of cytokine-release syndrome. 24 (86%) of 28 eligible patients received the second administration of ARI0002h (booster dose) in a single infusion of 100% of the dose. Two patients were not eligible for the booster dose because of extramedullary progression on day 100 (n=1) and death (cranial injury) in month 4 (n=1). Four (14%) of the 28 eligible patients did not receive the booster dose because of prolonged cytopenias, diagnosis of a secondary neoplasm, macrophage activation syndrome, and lymphocytosis due to CAR T-cell expansion (n=1 each; appendix p 12). In 19 (79%) of the 24 patients who were reinfused, 3 × 10⁶ CAR-positive cells per kg were available for the booster dose; 1·8 × 10⁶ CAR-positive cells per kg were available for three patients and 1·2 × 10⁶ CAR-positive cells per kg were available for two patients (appendix p 12). Median time to booster dose was 4 months (IQR 3–5; range 3–7). Eight (33%) of 24 patients received a second lymphodepletion regimen before the booster dose, based on CAR T-cell detection in peripheral blood. At the interim analysis (data cutoff Oct 20, 2021; median follow-up 12·1 months [IQR 9·1–13·5]), 10 (33%) of 30 (95% CI 19–51) patients had discontinued: eight patients developed disease progression and two died without progression (one after a cranial injury in month 4 and one after severe SARS-CoV-2 pneumonia in month 9). No dis continuations occurred due to manufacturing failures. An additional death occurred in a patient with disease progression. No deaths occurred within the first month of treatment. Cytokine-release syndrome was observed in 24 (80%) of 30 patients, in each case after receiving at least the second fraction of 30% (total dose 40%), with no grade 3 or higher events (15 [63%] grade 1 and nine [38%] grade 2; table 2). Median time to onset of cytokine release syndrome was 7 days (IQR 5–8) from the first fraction (10%), with a median duration of symptoms of 2 days (0–14). Tocilizumab was administered in 19 (63%) of 30 patients, mainly for persistent grade 1 cytokinerelease syndrome, and three (10%) patients received steroids. Because of cytokine-release syndrome, a clinical decision was made that five patients would not receive the third fraction of 1·8 × 10⁶ CAR-positive cells per kg. No cases of ICANS or late neurologic events were observed. None of the patients who received the booster dose had cytokine-release syndrome, ICANS, or any adverse events of special interest after the booster dose. One mild infusion reaction and one moderate tumour lysis syndrome were reported (table 2). Prolonged cytopenias were reported in 20 (67%) of 30 patients (table 2). The median duration of grade 4 neutropenia was 35 days (95% CI 26–44) and time to complete resolution of cytopenias was 4 months (95% CI 3–5) for neutropenia, 12 months (6–18) for thrombocytopenia, and 3 months (1–13) for anaemia. All patients recovered without requiring a stem-cell boost. The overall safety profile is shown in table 3. Infections were reported in 20 (67%) of 30 patients. 45 infectious episodes (seven [16%] grade ≥3) were reported, with most patients experiencing an average of 1–2 infection episodes (appendix p 13). Respiratory tract infections were the most common, with 24 (53%) episodes, including three (10%) of 30 patients having SARS-CoV-2 infections. Other relevant adverse events were a new diagnosis of colon adenocarcinoma, considered non-related to ARI0002h, and one reactivation of hepatitis B virus, considered related to ARI0002h. Three cases of macrophage activation syndrome occurred, two of them in combination with a grade 1 cytokine-release syndrome; all cases resolved completely. Grade 1 Grade 2 Grade 3–4 Cytokine release syndrome 15/24 (63%) 9/24 (38%) 0 Immune effector cell-associated neurotoxicity syndrome 0 0 0 Infusion reaction 1/30 (3%) 0 0 Tumour lysis syndrome 0 1/30 (3%) 0 Persistent cytopenias 0 0 20/30 (67%) Data are n (%). Adverse events of special interest are depicted per MedDRA preferred term. Table 2: Adverse events of special interest Articles www.thelancet.com/oncology Vol 24 August 2023 919 Overall response rate during the first 100 days from infusion was 100%, including 24 (80%) of 30 patients with a very good partial response or better (15 [50%] with complete response, nine [30%] with very good partial response, and six [20%] with partial response). Median time to complete response was 3·8 months (IQR 1·0–11·6). On day 28, measurable residual disease by next generation flow cytometry was evaluable in 22 (73%) of 30 samples, with 21 (95% [95% CI 78–99]) patients presenting a negative result (appendix p 14). On day 100, 24 (92% [95% CI 76–98]) of 26 evaluable samples were negative (appendix p 14). In a post-hoc analysis using a data cutoff date of May 15, 2022 (median follow-up 18 months [IQR 15–20]), the overall response rate was 100% (95% CI 89–100) with 20 patients with complete response (67% [49–81]), eight with very good partial response (27% [14–44]), and two with partial response (7% [2–21]; figure 2A). 18 (60%) of 20 patients with complete response had a stringent complete response. Responses at month 6 were 14 (47% [95% CI 30–64]) of 30 patients with complete response, 11 (37% [22–55]) with very good partial response, two (7% [2–21]) with partial response, two (7% [2–21]) with progressive disease, and one (3% [1–17]) death without relapse. Responses at different timepoints are shown in the appendix (pp 15–16). Median time to best response was 3·3 months (95% CI 1·0–3·4), with some patients improving responses after 100 days (figure 2B). The differences in overall response rate at 100 days and at this follow-up (median 18 months) show that responses deepen over time, with five (25%) of 20 patients having a complete response after month 3 (100 days), and five (17%) of 30 patients having their best response after 6 months (one converted to very good partial response and four converted to complete response; figure 2B). 20 (80%) of 25 evaluable patients were negative for measurable residual disease at 6 months, and 16 (80%) evaluable patients were negative for measurable residual disease at 12 months (appendix p 14). 14 (58% [95% CI 30–76]) of 24 patients who received the booster dose had a stringent complete response before reinfusion, six (25% [12–45]) patients maintained the response after reinfusion (all very good partial response) and four (17% [7–36]) patients had an improved response after the booster dose, two patients with very good partial response and two with partial response converted to complete response (figure 2C). At the May 15, 2022, data cutoff date, median duration of response was not reached (95% CI 12·9–not reached; figure 3A) and median overall survival was not reached (8·0–not reached; figure 3B). Two patients died in response in months 4 and 9. 12-month overall survival was 86·5% (95% CI 75·1–99·7). Median progression-free survival was 14·5 months (95% CI 12·8–not reached; figure 3B). 14 (47%) patients had a progression-free survival event: 12 developed disease progression and two died without progression. 12-month progression-free survival was 70% (95% CI 55–89). Grade 1–2 Grade 3 Grade 4 Grade 5 Blood and lymphatic system disorders Anaemia 16 (53%) 2 (7%) 1 (3%) 0 Neutropenia 9 (30%) 8 (27%) 13 (43%) 0 Thrombocytopenia 13 (43%) 8 (27%) 6 (20%) 0 Asthenia 9 (30%) 0 0 0 Pyrexia 7 (23%) 2 (7%) 0 0 Oedema peripheral 3 (10%) 0 0 0 Febrile neutropenia ·· 1 (3%) 0 0 Lymphocytosis ·· 0 1 (3%) 0 Lymphopenia ·· 0 1 (3%) 0 Gastrointestinal and hepatobiliary disorders Diarrhoea 5 (17%) 1 (3%) 0 0 Hepatobiliary disorders Hepatotoxicity 3 (10%) 0 0 0 Immune system disorders Cytokine release syndrome 24 (80%) 0 0 0 Haemophagocytic lymphohistiocytosis 3 (10%) 0 0 0 Respiratory tract infection 3 (10%) 0 0 0 Upper respiratory tract infection 6 (20%) 0 0 0 Investigations Hypocalcaemia 7 (23%) 0 0 0 Hypomagnesaemia 6 (20%) 0 0 0 Headache 5 (17%) 0 0 0 Back pain 4 (13%) 0 0 0 Alanine aminotransferase increased 4 (13%) 1 (3%) 0 0 Aspartate aminotransferase increased 3 (10%) 1 (3%) 0 0 Blood lactate dehydrogenase increased 3 (10%) 0 0 0 Gamma-glutamyltransferase increased 3 (10%) 0 0 0 Hypokalaemia 3 (10%) 0 0 0 Psychiatric disorders Anxiety 3 (10%) 0 0 0 Infections and infestations COVID-19 ·· 0 0 1 (3%) Leishmaniasis ·· 1 (3%) 0 0 Rhinovirus infection ·· 1 (3%) 0 0 Septic shock ·· 1 (3%) 0 0 Severe acute respiratory syndrome ·· 1 (3%) 0 0 Staphylococcal bacteraemia ·· 1 (3%) 0 0 Injury; poisoning and procedural complications Head injury ·· 0 0 1 (3%) Nervous system disorders Seizure ·· 0 1 (3%) 0 Renal and urinary disorders Acute kidney injury ·· 1 (3%) 0 0 Vascular disorders Hypertension ·· 1 (3%) 0 0 For grades 1–2, only adverse events with an occurrence of 10% or more are shown. For grades 3–5, all adverse events are shown. Adverse events are reported according to Common Terminology Criteria of Adverse Events (version 5.0). Table 3: Adverse events (n=30) Articles 920 www.thelancet.com/oncology Vol 24 August 2023 Figure 2: Activity of ARI0002h (A) Overall response rate and response evaluation at consecutive timepoints. (B) Swimmer plot with the response of each individual patient after first infusion (n=30). Data cutoff was May 15, 2022. (C) Swimmer plot with the response of each patient after booster dose (n=24). Day 28 Day 100 Month 6Month 12 Overall response 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Booster dose 10 61 612 14 18842 Time since booster dose (months) 1 3 2 5 8 6 10 9 12 11 13 14 16 7 17 15 20 18 19 22 24 25 27 26 A Patients (n) 25 15 5 30 20 10 0 Unknown Death Progressive disease Partial response Very good partial response Complete response Patients Patients Month 22 Month 20 Month 18 Month 16 Month 14 Month 12 Month 10 Month 8 Month 6 Month 4 Day 100 Inclusion Day 28 Day 56 B C Time to first evaluation Death Progressive disease Partial response Very good partial response Complete response Time since inclusion Articles www.thelancet.com/oncology Vol 24 August 2023 921 A PET-CT scan at day 100 showed a metabolic response in 13 (93% [95% CI 69–98) of 14 patients with a plasmacytoma at baseline. The six patients with plasmacytomas who had disease progression died, and all patients without plasmacytomas who had disease progression were alive at data cutoff. ARI0002h detection in peripheral blood by PCR showed a median persistence of 5·0 months (95% CI 3·8–6·2). 15 (52%) of 29 patients with available samples on day 100, seven (28%) of 25 with available samples at month 6, and four (20%) of 20 with available samples at month 12 had detectable CAR T cells in peripheral blood (appendix p 19). The peak of expansion was observed on day 14 for most patients (range 7 days to 6 months; appendix p 20). Mean copies per genome at the peak of expansion were 11·1 (SD 14·2). Nine (75%) of 12 patients with an available sample at relapse, three (33%) still had detectable CAR T cells in peripheral blood. After the booster dose, 12 (50%) of 24 patients presented a low-grade expansion of CAR T cells immediately after administration, with a mean peak of 4·0 copies per genome (SD 9·2; appendix p 20). In a post-hoc analysis, we found no correlation between previous lymphodepletion and expansion of CAR T cells (three [38%] of eight patients receiving lymphodepletion expanded vs nine [56%] of 16 without lymphodepletion; p=0·67). The mean number of BCMA molecules per cell on malignant bone marrow plasma cells by flow cytometry at inclusion was 1306·5 (SD 889·1). The change in the number of BCMA molecules per cell in six paired samples available at relapse showed a decrease from mean 1784 (1229·8) molecules per cell to 1001·5 (916·1) molecules per cell (p=0·054). None of the six samples had a complete loss of BCMA expression. Soluble BCMA was detectable in peripheral blood of all patients at inclusion with a mean of 89·3 ng/mL (SD 124·6). A significant decrease was observed in all patients on days 28 and 100 (p<0·0025 at both timepoints), and the 12 patients who relapsed had detectable soluble BCMA at the end-of-treatment sample (appendix p 21). Positivity of human anti-human antibodies was not sustained for each individual patient (appendix p 22). We detected human anti-human antibodies in 21 (70%) of 30 patients (appendix p 22). Eight (25%) of 12 patients who relapsed had available human anti-human antibody measure ments, of whom only two (25%) were positive. In a post-hoc analysis, median vein-to-vein time was 43 days (IQR 35–54), with differences among patients who did or did not receive bridging therapy (54 days [IQR 44–58] vs 36 days [IQR 30–43]; p=0·0006). We found no differences in response rates or progression-free survival between patients who received only the first two fractions of the first dose versus those who received the full first dose (complete response: three [60%] of five vs 17 [68%] of 25; p=0·73; median progression-free survival: 14·5 months [95% CI 12·8–not reached] vs not reached [12·1–not reached]; p=0·83; post-hoc analysis). In a post-hoc Figure 3: Duration of response and survival of patients treated with ARI0002h (A) Duration of response (n=30). (B) Progression-free survival. (C) Overall survival. Median follow-up for survival was 18 months (IQR 15–20). Data cutoff was May 15, 2022. A Censored 30 (0) Progression-free survival (%) 100 80 60 40 20 0 B Overall survival (%) 100 80 60 40 20 0 C Number at risk (number censored) Patients remaining in response (%) 100 80 60 40 20 0 Median duration of response not reached (95% CI 12·9–not reached) Median progression-free survival 14·5 months (12·8–not reached) Median overall survival not reached (95% CI 8·0–not reached) Time since first infusion (months) 0 (18)4 (14)9 (9)17 (4)24 (2)26 (1)29 (0) 0 21181512963 30 (0)Number at risk (number censored) 0 (16)0 (16)9 (7)19 (2)25 (0)28 (0)30 (0) 0 21181512963 24 4 (12) 30 (0)Number at risk (number censored) 0 (22)16 (9)23 (3)26 (1)28 (0)30 (0) 01 8151296 32 1 6 (16) Time since first response (months)