CAR T-cell therapy has significantly improved survival for relapsed/refractory large B cell lymphoma (LBCL), but bridging therapy may be required as patients await CAR T infusion.¹ Bridging radiation therapy (RT) can control locally progressive disease, alleviate symptomatic disease, and debulk tumors with high metabolic volume without increasing the toxicities of subsequent CAR T-infusion.¹

Our group has previously presented the outcomes of bridging therapy in LBCL patients undergoing CD19-targeted CAR T-cell therapy.² As a continuation, we now present our multi-institutional experience in assessing the impact of bridging RT in comparison to other systemic regimens.

Following Institutional Review Board approval, a retrospective study was conducted for consecutive LBCL patients who received either tisagenlecleucel (tisa-cel) or axicabtagene ciloleucel (axi-cel) CAR T-cell therapy between 2017 and 2021. Methodology details are provided in the Supplementary Data.

Patient characteristics are summarized in Table 1S. One hundred twenty five patients received bridging therapy: 28 patients received RT alone, 11 patients received combined modality therapy (CMT), and 86 patients received systemic therapy (ST) alone—Table 2S provides a summary of the systemic regimens. There was a greater proportion of patients with advanced age (≥60 years) in the RT group (p = .02).

Following CAR T-cell therapy infusion, the best overall response rate (ORR) was 77% (n = 96), with a complete response (CR) in 51% (n = 64) and a partial response (PR) in 26% (n = 32). No significant difference was detected in ORR between axi-cel versus tisa-cel patients, but CR was better in patients who received axi-cel (OR = 2.3, p = .02).

Cytokine release syndrome (CRS) and immune effector cell neurotoxicity syndrome (ICANS) occurred in 80% (n = 100) and 51% (n = 64), respectively, with grade 3 or higher CRS and ICANS reported in 5% (n = 6) and 24% (n = 30), respectively. No statistically significant differences were observed based on type of bridging therapy.

The median follow-up after CAR T-cell therapy infusion was 9.2 months (IQR: 2.7–26.8 months). The median event-free survival (EFS) was 5.4 months for the axi-cel group, compared to 2.9 months for those who received tisa-cel (p = .03). The median overall survival (OS) was not significantly different between axi-cel and tiso-cel groups (20.6 months vs. 42.2 months; p = .9) (Figure 1SA,B).

Based on the type of bridging regimen, the median OS was 14 months for the RT group, not reached for the CMT group, and not reached for the ST group (p = .01), while the median EFS was 3.5 months for the RT group, 3.0 months for the CMT group, and 4.0 months for the ST group (p = .9) (Figure 2SA,B). No statistically significant difference was detected in the duration of response (DOR) based on the type of bridging therapy (p = .6).

Furthermore, for patients with stage I disease, no significant difference was observed in OS or EFS based on the type of bridging therapy. We then identified 37 patients (30%) who had response data based on the PET scan conducted after bridging therapy before CAR T-cell infusion. The median OS among responders (n = 15) was not reached, compared to 18.8 months for non-responders (n = 22) (p = .3), while the median EFS among responders was not reached, compared to 4.0 months for non-responders (p = .03) (Figure 3SA,B). There was no difference in response to therapy based on the class of bridging therapy (p = 1.0).

Among the 39 patients who underwent bridging RT (including RT and CMT cohorts), a total of 45 sites were irradiated. The median dose/fractionation was 24 Gy (range, 10–37.5 Gy) and 8 fractions (range, 4–15 fractions). Thirty-seven sites (82%) were irradiated for symptom palliation, and 8 sites (18%) were irradiated to control asymptomatic disease. Sites of RT include head and neck (n = 13, 29%), CNS (n = 10, 20%), chest (n = 7, 16%), pelvis (n = 6, 13%), abdomen (n = 3, 6%), extremities (n = 3, 6%), and paraspinal area (n = 3, 6%) (Table 3S).

Out of the 35 sites evaluated, 21 sites (60%) were bulky (≥5 cm) at the time of RT. The 1-year in-field PFS rate was 82% (Figure 1A). Seven sites experienced local recurrence, with median time to in-field progression of 5.3 months (range, 1.8–12.1 months), and 4 of these sites were bulky at time of RT. No significant difference in in-field response was detected between axi-cel and tisa-cel recipients. The 1-year out-of-field PFS rate was 35% (Figure 1B). Twenty-three patients experienced out-of-field progression, with a median time to progression of 3.5 months (range, 1.4–14.7 months); 9 of these sites were bulky at time of RT. There was no significant difference in in-field or out-of-field recurrence between bulky sites and those that were not. No significant difference was detected in ORR, CR, or DOR post-CAR T infusion in patients with bulky lesions between those who received only ST regimens (n = 40) and those who received RT alone or CMT (n = 21).

Sixteen patients underwent bridging RT prior to apheresis due to clinical urgency. On univariate analysis, patients receiving RT due to clinical urgency prior to apheresis had a lower likelihood of achieving overall response post-CAR T compared to those receiving bridging RT post-apheresis (OR = 0.2, p = .04). This finding is most likely confounded by the high tumor burden and unfavorable baseline characteristics pre-CAR T infusion leading to rapidly progressive disease necessitating an urgent bridging RT before apheresis. There was no statistically significant difference in in-field response between patients who received bridging RT before apheresis and those who received it after apheresis. The mean ± SD of pre-RT ALC was 0.46 ± 0.51/μL, and for post-RT/apheresis, it was 0.39 ± 0.41 K/μL. The absolute ALC Δ RT/apheresis was 0.44 ± 0.51 K/μL. We found no statistically significant difference in ALC count pre-RT and post-RT/apheresis (p = .8). Post-RT/apheresis ALC showed no significant association with post-CAR T ORR or in-field response post-RT. No significant association was also found between the drop in ALC Δ RT/apheresis and ORR post-CAR T-cell therapy.

Among the 39 patients who were treated with bridging RT, 21 patients were treated with comprehensive RT field to 25 sites with a median dose of 24.5 Gy (range, 18–36 Gy), while 18 patients received focal RT to 20 sites with a median dose of 22 Gy (range, 10–37.5 Gy) (p = .5). No significant difference in OS was found between high-dose RT (BED₁₀ >30 Gy) and low-dose RT (BED₁₀ ≤30 Gy) groups. There was also no statistically significant difference in BED₁₀ in sites that experienced local failure and sites that remained locally controlled.

Patients who received focal RT were more likely to have an IPI of ≥3 (p = .006), and advanced-stage disease (p < .001). No significant difference was detected among patients with elevated LDH, poor ECOG PS, CNS disease, age (≥60 years), double/triple hit or expressor status, type of CAR T product, and bulky disease. The median survival among patients who received comprehensive RT was 21.5 months, and for focal RT was 13.1 months (p = .1) (Figure 4S).

We found that the patients who receiving bridging RT had lower median OS compared to the CMT and ST groups (p = .01), despite having comparable baseline characteristics at the time of apheresis, including International Prognostic Index (IPI) risk factors and bulky disease, except for advanced age (≥60 years) (p = .02). Previously reported series of patients who received bridging RT had similarly unfavorable baseline characteristics—for example, our study had high IPI in 18/28 compared to Pinnix et al.³ who reported 6/11 with high IPI. The majority of the patients who received bridging RT in our cohort were referred for RT primarily for palliative purposes to alleviate symptomatic disease, including 36% of patients in the CMT group who received RT for progression of systemic therapy. On the other hand, Roddie et al.⁴ have reported that 44% (25/62) of patients with high IPI in the RT-only group, translating to superior outcomes in overall response rates and survival rates post-CAR T compared to the CMT and ST groups. Since the initial reports on bridging RT, there has been significant improvement in systemic therapy bridging options, which may help explain the results as RT cannot be expected to improve OS in patients with advanced stage disease as its role in these patients is restricted to local control. Indeed, no significant difference was detected in OS or EFS among patients with stage I disease based on the type of bridging therapy, which argues against RT being an inferior bridging modality for patients with limited disease burden. Within the RT group, comprehensive bridging RT patients trended towards improved long-term outcomes compared to focal bridging RT patients, particularly for limited-stage disease. We show that RT continues to offer high rates of in-field response in this highly refractory population. However, patients with more than one site of disease may be better served with CMT or ST bridging rather than RT bridging alone in terms of OS outcomes.

Currently, no established consensus exists for dose/fractionation for the purposes of bridging RT. Therefore, we are conducting a pilot trial, the first of its kind, to investigate once-weekly RT using artificial intelligence-driven adaptive technology to optimize bridging RT dose and field in terms of logistics, time, cost, and toxicities (NCT06004167). Furthermore, our results from the sensitivity analysis showed that RT is a reasonable bridging strategy in both CNS and non-CNS sites (See Supplement).

Due to the theoretical concern that RT might impact adequate T-cell collection and CAR T-cells' efficacy, it has been recommended that bridging RT be administered after apheresis. However, in the present study, we report 16 patients who received bridging RT prior to apheresis due to clinical urgency. To the best of our knowledge, this study is the first to investigate the impact of bridging RT prior to apheresis, which underscores the successful T-cell collection in those patients; however, conclusive statements regarding whether bridging RT affects CAR T levels and T-cell fitness cannot be drawn.

We acknowledge the potential limitations in this study, including the retrospective aspect; variability in RT dose/fractionation; lack of PET quantitative parameters; and selection of bridging therapy, which was determined at the discretion of the physician; and the heterogeneity of the patient population, which encompassed patients with CNS lymphoma and those who received two distinct CAR T-cell products. Furthermore, in contrast to our findings in the salvage setting post-CAR T failure,⁵ we didn't observe a dose–response relationship favoring higher doses for improved OS, similar to a previously reported series,⁶ possibly due to the small sample size and the small number of local failures.

In conclusion, our findings support that ST and CMT bridging may be better than RT for the vast majority of patients given their advanced stage disease. While RT demonstrates high in-field response rates in this highly refractory population, patients with multiple disease sites may benefit more from CMT or ST bridging strategies than from RT alone for improved OS outcomes. Nonetheless, bridging RT, whether alone or in combination with ST, is safe and effective for patients with locally progressive LBCL prior to CAR T-cell therapy. Patients who are suitable for comprehensive RT trend towards superior OS compared to focal RT, and those patients who are not would likely better be served by adding systemic therapy. Exploring the use of bridging RT prior to apheresis for clinical urgency and its impact on CAR T-cells' efficacy requires further research considering factors such as the volume of irradiated bone marrow and the irradiated tumor size. Despite improved outcomes with systemic bridging options in our retrospective study, further investigation is warranted to optimize RT dose, fractionation, field size, and concurrent therapies to enhance its effectiveness for a broader patient population.

CAR T 细胞疗法显着提高了复发/难治性大 B 细胞淋巴瘤 (LBCL) 的生存率，但由于患者等待 CAR T 输注，可能需要桥接疗法。 ¹桥接放射治疗 (RT) 可以控制局部进展性疾病、缓解症状性疾病并消除高代谢量肿瘤，且不会增加后续 CAR T 输注的毒性。 ¹

我们的小组之前曾介绍过接受 CD19 靶向 CAR T 细胞治疗的 LBCL 患者的桥接治疗结果。 ²作为延续，我们现在介绍我们在评估桥接放疗与其他全身治疗方案的影响方面的多机构经验。

经机构审查委员会批准后，对 2017 年至 2021 年间接受 tisagenlecleucel (tisa-cel) 或 axicabtagene ciloleucel (axi-cel) CAR T 细胞治疗的连续 LBCL 患者进行了回顾性研究。补充数据中提供了方法学详细信息。

患者特征总结于表 1S 中。 125 名患者接受了桥接治疗：28 名患者接受了单独的放疗，11 名患者接受了联合治疗 (CMT)，86 名患者接受了单独的全身治疗 (ST)——表 2S 提供了全身治疗方案的总结。 RT 组中高龄患者（≥60 岁）比例较高 ( p = .02)。

CAR T 细胞疗法输注后，最佳总体缓解率 (ORR) 为 77% ( n = 96)，其中完全缓解 (CR) 为 51% ( n = 64)，部分缓解 (PR) 为 26% （ n = 32）。 axi-cel 与 tisa-cel 患者之间的 ORR 没有检测到显着差异，但接受 axi-cel 的患者 CR 更好（OR = 2.3， p = .02）。

细胞因子释放综合征 (CRS) 和免疫效应细胞神经毒性综合征 (ICANS) 的发生率分别为 80% ( n = 100) 和 51% ( n = 64)，其中 3 级或以上 CRS 和 ICANS 的发生率为 5% ( n = 6) 和 24% ( n = 30)。根据桥接治疗的类型，没有观察到统计学上的显着差异。

CAR T 细胞疗法输注后的中位随访时间为 9.2 个月（IQR：2.7-26.8 个月）。 axi-cel 组的中位无事件生存期 (EFS) 为 5.4 个月，而接受 tisa-cel 组的中位无事件生存期 (EFS) 为 2.9 个月 ( p = .03)。 axi-cel 和 tiso-cel 组之间的中位总生存期 (OS) 无显着差异（20.6 个月 vs. 42.2 个月； p = .9）（图 1SA、B）。

根据桥接方案的类型，RT 组的中位 OS 为 14 个月，CMT 组未达到，ST 组也未达到 ( p = 0.01)，而 RT 组的中位 EFS 为 3.5 个月组，CMT 组 3.0 个月，ST 组 4.0 个月 ( p = .9)（图 2SA、B）。根据桥接治疗的类型，在缓解持续时间 (DOR) 方面没有检测到统计学上的显着差异 ( p = .6)。

此外，对于 I 期疾病患者，根据桥接治疗的类型，OS 或 EFS 没有观察到显着差异。然后，我们根据 CAR T 细胞输注前桥接治疗后进行的 PET 扫描确定了 37 名患者 (30%) 的缓解数据。应答者中的中位 OS ( n = 15) 尚未达到，而无应答者 ( n = 22) 的中位 OS 为 18.8 个月 ( p = .3)，而应答者中的中位 EFS 尚未达到，而无应答者为 4.0 个月-响应者（ p = .03）（图3SA，B）。根据桥接治疗的类别，对治疗的反应没有差异 ( p = 1.0)。

在接受桥接 RT 的 39 名患者（包括 RT 和 CMT 队列）中，共有 45 个部位受到照射。中位剂量/分次为 24 Gy（范围，10-37.5 Gy）和 8 次分次（范围，4-15 次）。 37 个部位 (82%) 接受照射以缓解症状，8 个部位 (18%) 接受照射以控制无症状疾病。 RT 部位包括头颈 ( n = 13, 29%)、中枢神经系统 ( n = 10, 20%)、胸部 ( n = 7, 16%)、骨盆 ( n = 6, 13%)、腹部 ( n = 3, 6%)、四肢 ( n = 3, 6%) 和椎旁区域 ( n = 3, 6%)（表 3S）。

在评估的 35 个部位中，21 个部位 (60%) 在 RT 时体积较大 (≥5 cm)。 1 年现场 PFS 率为 82%（图 1A）。 7 个部位出现局部复发，中位进展时间为 5.3 个月（范围为 1.8-12.1 个月），其中 4 个部位在放疗时体积较大。 axi-cel 和 tisa-cel 接受者之间的现场反应没有显着差异。 1 年外场 PFS 率为 35%（图 1B）。 23 名患者经历了场外进展，中位进展时间为 3.5 个月（范围为 1.4-14.7 个月）；其中 9 个站点在 RT 时体积庞大。大体积部位和非大体积部位之间的场内或场外复发没有显着差异。对于仅接受 ST 方案的患者 ( n = 40) 和接受单独 RT 或 CMT 的患者 ( n = 21)，在大面积病变患者中，ORR、CR 或 DOR 输注后没有检测到显着差异。

由于临床紧急，16 名患者在血浆分离术前接受了桥接 RT。在单变量分析中，与单采后接受桥接 RT 的患者相比，在单采之前因临床紧急情况而接受 RT 的患者在 CAR T 后实现总体缓解的可能性较低 (OR = 0.2， p = .04)。这一发现很可能与高肿瘤负荷和 CAR T 输注前不利的基线特征相混淆，导致疾病迅速进展，需要在血浆分离术前进行紧急桥接 RT。在单采术前接受桥接 RT 的患者与单采术后接受桥接 RT 的患者之间的现场反应没有统计学上的显着差异。 RT 前 ALC 的平均值±SD 为 0.46 ± 0.51/μL，RT/单采后为 0.39 ± 0.41 K/μL。绝对 ALC Δ RT/单采术为 0.44 ± 0.51 K/μL。我们发现 RT 前和 RT/单采后 ALC 计数没有统计学显着差异 ( p = .8)。 RT/单采后 ALC 与 CAR T ORR 后或 RT 后现场反应没有显着相关性。 ALC Δ RT/单采术下降与 CAR T 细胞治疗后 ORR 之间也没有发现显着关联。

在接受桥接放疗的 39 名患者中，21 名患者接受了 25 个部位的综合放疗，中位剂量为 24.5 Gy（范围为 18-36 Gy），而 18 名患者接受了 20 个部位的局灶性放疗，中位剂量为剂量为 22 Gy（范围，10–37.5 Gy）（ p = .5）。高剂量放疗（BED_{​​ 10} >30 Gy）和低剂量放疗（BED_{​​ 10} ≤30 Gy）组之间 OS 没有显着差异。在经历局部故障的站点和仍保持局部控制的站点中，BED ₁₀也没有统计学上的显着差异。

接受局灶性放疗的患者更有可能出现 IPI ≥3 ( p = .006) 和晚期疾病 ( p < .001) 的情况。 LDH升高、ECOG PS不良、CNS疾病、年龄（≥60岁）、双重/三重打击或表达状态、CAR T产品类型和大体积疾病的患者之间没有检测到显着差异。接受全面放疗的患者的中位生存期为 21.5 个月，接受局部放疗的患者的中位生存期为 13.1 个月 ( p = .1)（图 4S）。

我们发现，与 CMT 和 ST 组相比，接受桥接 RT 的患者的中位 OS 较低 ( p = .01)，尽管在血浆分离术时具有相似的基线特征，包括国际预后指数 (IPI) 危险因素和大块疾病，高龄除外（≥60 岁）（ p = .02）。之前报道的一系列接受桥接 RT 的患者也具有类似的不利基线特征，例如，与 Pinnix 等人相比，我们的研究在 18/28 中具有较高的 IPI。 ^{3 名}报告 6/11 的 IPI 较高。在我们的队列中，大多数接受桥接放疗的患者主要出于姑息治疗的目的而转诊接受放疗，以缓解症状性疾病，其中 CMT 组中有 36% 的患者因全身治疗进展而接受放疗。另一方面，罗迪等人。 ⁴报道称，仅接受 RT 组中有 44% (25/62) 的患者具有高 IPI，这意味着与 CMT 和 ST 组相比，CAR T 后的总体缓解率和生存率均优于 CMT 组。自从关于桥接放疗的初步报告以来，全身治疗桥接选择已经有了显着改善，这可能有助于解释结果，因为放疗不能期望改善晚期疾病患者的 OS，因为它在这些患者中的作用仅限于局部控制。事实上，根据桥接治疗的类型，I 期疾病患者的 OS 或 EFS 没有检测到显着差异，这表明 RT 对于疾病负担有限的患者来说是一种较差的桥接方式。 在 RT 组中，与局灶性桥接 RT 患者相比，全面桥接 RT 患者的长期结局趋于改善，特别是对于局限性疾病。我们表明，RT 在这一高度难治性人群中继续提供高现场响应率。然而，就 OS 结果而言，患有多个疾病部位的患者可能更适合接受 CMT 或 ST 桥接，而不是单独的 RT 桥接。

目前，对于桥接 RT 的剂量/分次尚未达成共识。因此，我们正在进行一项试点试验，这是此类试验中的第一个，研究使用人工智能驱动的自适应技术每周一次的 RT，以在物流、时间、成本和毒性方面优化桥接 RT 剂量和领域 (NCT06004167)。此外，我们的敏感性分析结果表明，RT 在 CNS 和非 CNS 部位都是一种合理的桥接策略（参见补充材料）。

由于理论上担心 RT 可能会影响足够的 T 细胞收集和 CAR T 细胞的功效，因此建议在单采后进行桥接 RT。然而，在本研究中，我们报告了 16 名患者由于临床紧急情况在血浆分离术之前接受了桥接 RT。据我们所知，这项研究首次调查了血浆分离术之前桥接 RT 的影响，这强调了这些患者中 T 细胞的成功收集；然而，关于桥接 RT 是否影响 CAR T 水平和 T 细胞适应性，尚无法得出结论性的结论。

我们承认这项研究的潜在局限性，包括回顾性方面；放疗剂量/分次的变异性；缺乏PET定量参数；桥接疗法的选择，由医生酌情决定；以及患者群体的异质性，其中包括中枢神经系统淋巴瘤患者和接受两种不同 CAR T 细胞产品的患者。此外，与我们在 CAR T 失败后挽救环境中的发现相反， ⁵我们没有观察到有利于提高 OS 的剂量反应关系，类似于之前报道的系列， ⁶可能是由于样本量较小且局部故障数量较少。

总之，我们的研究结果支持，对于绝大多数晚期疾病患者来说，ST 和 CMT 桥接可能比放疗更好。虽然放疗在这一高度难治性人群中表现出较高的现场缓解率，但对于改善 OS 结局，患有多个疾病部位的患者可能从 CMT 或 ST 桥接策略中获益更多，而不是仅从放疗中受益。尽管如此，桥接放疗，无论是单独使用还是与 ST 联合使用，对于 CAR T 细胞治疗前局部进展性 LBCL 患者来说都是安全有效的。与局部放疗相比，适合综合放疗的患者往往会获得更好的 OS，而那些不适合的患者可能会通过增加全身治疗得到更好的服务。探索在临床紧急情况下在血浆分离术之前使用桥接 RT 及其对 CAR T 细胞功效的影响需要进一步研究，考虑到受照射的骨髓体积和受照射的肿瘤大小等因素。尽管我们的回顾性研究中系统性桥接方案改善了结果，但仍需要进一步研究来优化放疗剂量、分割、射野大小和并发治疗，以提高其对更广泛患者群体的有效性。