MTV‐guided radiotherapy bridging in large B‐cell lymphoma patients receiving CD19 CAR T‐cell therapy

To the Editor, The majority of relapsed or refractory (r/r) large B-cell lymphoma (LBCL) patients intended to be treated with CD19 CAR T receive bridging therapy (BT) between leukapheresis and infusion to control the disease during the weeks of cell manufacturing. The optimal BT approach remains unclear and significantly varies between centres. There is strong rationale for using radiotherapy (RT) as preferred bridging: (A) high efficacy in a largely RT-naïve and chemo-refractory patient population, (B) avoiding overlapping toxicity seen with chemotherapy bridging, particularly haematotoxicity which might potentiate post-CAR T aplasia and infectious complications,1 (C) potential immunomodulatory synergistic and priming effects of RT,2 (D) targeted eradication of sanctuary sites of disease associated with lesion-specific CAR T resistance.3 Despite these potential advantages, only 7%–18% of patients receive RT as BT,4-7 due to challenges in delivering RT with rapid turnaround in advanced stage, refractory patients and the lack of expertise and standardised protocols. The RadiothErapy priMIng for CAR-T (REMIT) trial (NCT04726787) was designed as a single-arm phase II study across three UK CAR T centres, assessing feasibility of metabolic tumour volume (MTV)-guided RT as preferred bridging modality in patients with r/r LBCL approved to receive tisagenlecleucel (tisa-cel) as per licenced indication. Baseline MTV was centrally assessed and used to guide peer-reviewed RT planning, defining full-dose (FD) and low-dose (LD) clinical target volumes (CTV). FD-CTV covered sites where durable disease control is important, that is, symptomatic or bulky masses, primary refractory disease or sites at high risk of lesion-specific CAR T resistance.3 For lesions with MTV ≤400 cc, 20 Gy in 5 or 10 fractions (#) was planned, 30 Gy/10–15# for lesions with MTV >400 cc. The cut-off of 400 cc was chosen based on previous experience in first-line LBCL,8 and similar cut-offs have been used in other datasets.9, 10 The rationale for LD-CTV (4 Gy/2#) was to cover other disease sites for maximum debulking and potential priming effects, while minimising the dose to organs at risk. The primary endpoint was the percentage of patients starting lymphodepletion (LD) without RT-related delays of >1 day from the planned start date (details of the statistical analysis are provided in Supporting Information S1). Six patients were recruited between September 2022 and March 2023 (20 planned), but the trial was terminated early due to a significant decline in tisa-cel use and eventual withdrawal of the product from the UK market. An additional cohort of seven patients was included in the analysis who were approved for axicabtagene ciloleucel (axi-cel) and planned for comparable RT bridging at the lead centre during the same time period, fulfilling the same eligibility criteria (as part of a Service Evaluation not requiring separate consent). Baseline characteristics of the 13 patients are given in Table 1. 77% had advanced stage disease, 46% elevated Lactate Dehydrogenase (LDH) and 31% had a total MTV ≥100 cc at the time of CAR T approval/trial enrolment. Of the 13 patients who started MTV-guided RT bridging, one REMIT patient did not proceed to CAR T-cell infusion due to cardiac comorbidities (unrelated to RT). Median time from leukapheresis to infusion was 36 days (Interquartile range (IQR): 35–44), 37 days for the tisa-cel (REMIT trial) and 35 days for the axi-cel (non-trial) cohort. The median time from leukapheresis to the start of RT was 6.5 days (IQR: 5–19). REMIT (Tisa-cel) N = 6 Non-trial (Axi-cel) N = 7 All N = 13 Details on RT techniques and anatomical locations are provided in Table 1. Seven patients received RT to >1 site (median 2, IQR: 1–2, range: 1–4), eight patients received both full- and low-dose RT. RT toxicities were observed in only 3/12 patients, and all were grade 1 (two mucositis, one nausea). Adverse events recorded on the REMIT trial are provided in Table S1. No RT-related delays of starting LD were reported (primary end-point). Median time from last dose of RT to start of LD was 12 days (IQR: 5.5–18). Accordingly, response to RT was assessed at a very early time point (pre-LD positron emission tomography (PET) scan). In-field response was seen in nine of 11 evaluable patients (81.8%); four with complete metabolic response (CMR) and five partial metabolic response. Five patients with in-field response progressed out of field (defined as relapse fully located outside the planning target volume). During the BT phase, 10/12 patients maintained an Eastern Cooperative Oncology Group (ECOG) performance status of 0/1, and median LDH levels did not increase, indicating effective disease stabilisation through RT. Patients' absolute lymphocyte count decreased from median 0.63 (IQR: 0.52–0.8) at baseline to 0.40 (IQR: 0.25–0.59) at the time of LD (p = 0.003), suggesting an additive lymphodepleting effect of RT to fludarabine/cyclophosphamide conditioning. The best overall response rate after CAR T-cell infusion was 83.3% (all CMR), which was maintained at 6 months (80.0% for the REMIT cohort; 85.7% for the non-trial cohort). With a median follow-up of 13.3 months, none of the patients who achieved a response to CAR T progressed; there were no deaths reported in any of the cohorts, and no patient received further treatment before progression. The 1-year progression-free survival (PFS) in the total cohort of infused patients was 83.3% (95% confidence interval (CI): 48.2%–95.6%; Figure 1; PFS for the two cohorts separately is provided in Figure S1). The incidence of cytokine release syndrome was 91.7% with 8.3% grade ≥3 events. Immune effector cell-associated neurotoxicity syndrome (ICANS) occurred in 25% (8.3% grade ≥3 ICANS), in line with previous toxicity outcomes.6, 12, 13 At 1 month post-infusion, grade ≥3 neutropenia and thrombocytopenia were seen in 8% and 33% of patients, which persisted at 6 months for 1/11 patients. Our preliminary results suggest feasibility of using a rapid turnaround, individualised, dose- and field-adapted bridging RT for r/r LBCL planned for CD19 CAR T. The minimal RT-related toxicity seen in our cohort indicates that it is safe to deliver RT to most anatomical sites and wider fields, without delaying CAR T-cell infusion. However, it is difficult to draw firm conclusions due to the limited sample size of our study, and larger datasets are needed to confirm these findings. Since implementing a tumour burden-adapted bridging RT protocol at our centre, the use of single-modality RT has increased to around 30% of all CAR T patients, the majority having advanced stage disease, elevated LDH or other risk features. We have recently shown that RT bridging is associated with favourable outcomes both in limited and advanced stage patients, whether RT is applied as comprehensive or focal treatment.13 MTV assessment and high-precision RT techniques are currently restricted to specialised centres; however, the concept of a combined full- and low-dose RT protocol for CAR T bridging could be more widely adopted across referring centres. Full integration of radiation oncologists into the CAR T multidisciplinary team and close collaboration with local centres are key to achieve this. The local control rate in our cohort was excellent, including sites irradiated with low dose. CAR T outcomes in this small cohort were favourable, in line with previously published results of RT-bridged LBCL patients. The question of whether the choice of BT has an impact on CAR T-cell toxicity and efficacy outcomes is still not clear. Favourable outcomes seen in patients receiving RT as a single modality could be largely driven by selection bias, with high-risk, fast-progressing patients deemed to require systemic BT.7 However, synergistic and priming effects could play an additional role, as supported by pre-clinical data.14 We believe that RT should be strongly considered as the preferred BT over systemic treatment in patients where this modality is deemed feasible, regardless of stage and number of disease sites, to mitigate the risk of cumulative haematotoxicity and long-term immunosuppression leading to significant non-relapse mortality after CD19 CAR T.15 In this regard, we advocate for the development of a consensus protocol of individualised dose-adapted RT bridging with a view to allow more patients access to this effective BT and potentially improve long-term outcomes after CAR T-cell therapy. AK, NGM, SFB, SG, TM, JLB, MN, WO, PEMP, RB, JF and RP were involved in the study concept and design. AK, NGM, SB and SG analysed the data and wrote the manuscript. SFB, TM, LS, SK, WO, RJ, EK, RS, PEMP, RB, JF, RP and JLB contributed to collecting and analysing the data. All authors were involved in revising the manuscript and have read and approved the final version of the manuscript. The authors would like to thank the patients, their relatives and caregivers and investigators and staff involved in this analysis. This work was supported by Novartis (CCTL019CGB02T). A.K. has served on advisory boards and received honoraria from Kite/Gilead, Novartis, Abbvie, Roche and BMS. M.N. received honoraria from Kite/Gilead. R.S. has served on advisory boards and received honoraria from Kite/Gilead and Novartis. R.J. has served on advisory boards and received honoraria from Kite/Gilead. W.O. has received honoraria from Roche, Takeda, Pfizer, Servier, Kite/Gilead, MSD, Novartis, Beigene, Astra Zeneca, Syneos, Autolus, Kyowa Kirin, Abbvie, Incyte, BMS, Janssen. The remaining authors declare no conflict of interest. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.