Chimeric antigen receptor (CAR) T-cell therapy has revolutionized the treatment armamentarium of relapsed/refractory (R/R) large B-cell lymphoma (LBCL). Besides its outstanding efficacy, CAR T-cells carry unique immune-mediated adverse events, including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS); it has a wide spectrum of presentation ranging from aphasia, headache to seizure, and coma.¹ Advanced age is identified as a potential predisposing factor of ICANS.² However, whether age has a differential effect on ICANS manifestation is not well described. Herein, we report detailed neurologic presentations in R/R LBCL patients treated with axicabtagene ciloleucel (axi-cel) and the similarity/difference between older and younger patients to better characterize ICANS in older patients. We conducted a retrospective study including patients with R/R LBCL treated with axi-cel at Mayo Clinic, Rochester, MN, between June 2016, and October 2020. The axi-cel treatment process was in accordance with the productʼs package insert. We described the clinical course of ICANS, including neurologic manifestations, treatments, and outcomes. We divided patients into two age-groups according to their age at the time of infusion; younger (< 60 years) and older (≥ 60 years) groups, and compared their neurological outcomes. Neurologic presentations were stratified into different clusters using the Immune Effector Cell Encephalopathy (ICE) Score as the primary reference. Neurological symptoms which could not be characterized into ICE-based categories were classified as the independent entity (supplementary material). All patients underwent baseline brain magnetic resonance imaging (MRI) and neurological assessment by neurooncologists. Patients were assessed and managed by CAR T-providers using the standardized protocol. The severity of ICANS was reported according to the Common Terminology Criteria for Adverse Events version 4.03 (before January, 2019) and the American Society of Transplant Cellular Therapy (ASTCT) Consensus Criteria (after January, 2019). Incidence of ICANS was defined as the time from CAR T-cell infusion to the onset of ICANS with death from any causes being the competing event. The study was approved by the institutional review board of Mayo clinic. Continuous variables were described as median and range with comparison between groups by the Kruskal-Wallis test. Categorical variables were shown in percentage with comparison between groups by the chi-square or Fisher exact test. The cumulative incidence was analyzed by the competing-risk analysis. Cox proportional hazard analysis was performed to analyze for potential factors associated with ICANS and was reported as Hazard ratio with 95% confidence interval (95%CI). p-values < 0.05 were considered statistically significant.

Seventy-eight patients with DLBCL received axi-cel during the study period. The median age at infusion was 58.8 years (26.8–76.5 years) with 32 patients (41%) being ≥ 60 years old. Baseline characteristics were comparable between age-groups except for a higher proportion of high international prognostic index, prior autologous stem cell transplant and underlying cerebral microvascular disease in older patients (Tables S1 and S2). Note, ICANS was observed in 40 patients (51.3%), including 16 and 24 in the older and younger age-groups, respectively. The 30-day incidence of ICANS was comparable between older and younger patients (50%, 95% CI 29.3%–64.6% vs. 52.2%, 95% CI 35.3%–64.6%, p = 0.73) (Figure 1). Also, ICANS grade 3 or higher were observed in 13 patients (n = 8; 17.4% in younger and n = 5; 15.6% in older group). Of 40 ICANS patients, 39 (97.5%) developed CRS preceding the onset of ICANS. The median time from CAR T-cell infusion to the onset of ICANS was similar between two age-groups (p = 0.42). The median time from the onset of ICANS to maximal ICANS severity was within 24 hours and was not statistically different between younger and older cohorts (p = 0.60). The distribution of neurological symptoms and maximal ICANS severity was similar between younger and older patients (Figure 1, Table S3). The three most common initial neurological abnormalities included aphasia, dysgraphia and confusion in both age-groups (Table S3). Speech and writing abnormalities were the most common symptoms in both age-groups. Headache and tremors were more frequent in older patients but not statistically different from younger patients. Of 40 patients with ICANS, two (5.0%) developed seizures (one subclinical non-convulsive seizure and one convulsive seizure).

Radiologic investigations were performed after ICANS occurrence in 26 patients, including brain MRI in nine patients (five in the younger cohort and four in the older cohort). Abnormal MRI finding associated with ICANS (diffuse T2 enhancement in subcortical white matter) was noted in one patient from the older age-group who developed ICANS ASTCT grade 3 (Figure S1). Of all patients who developed ICANS, one patient (2.5%), from the younger age-group, required intubation due to severe altered mental status. Twenty-four patients (60.0%) received systemic steroids for the management of ICANS, including seven (17.5%) requiring pulse methylprednisolone for dexamethasone-resistant ICANS. The median total dose of corticosteroids (prednisone-equivalent) was 1.80 mg/m²/day, not different between younger and older patients (1.94 vs. 1.31 mg/m²/day, p = 0.65). Both younger and older patients received systemic corticosteroid for a median duration of 4 and 2 days, respectively (p = 0.54). Overall, there was no difference in ICANS treatment, and inflammatory markers (lactate dehydrogenase, c-reactive protein, ferritin) between age-groups (Tables S2 and S3). All ICANS resolved with the median duration of 5 days (1–20 days) comparable between the younger and older cohort cohorts (5 vs. 6 days, p = 0.18). The interval change of ICE score during the ICANS course was not different between age-groups (Figure S2). In the univariate Cox proportional hazard analysis (Supplementary Table S4), severe CRS grade ≥ 2 was the only factor associated with ICANS in the entire cohort and in the younger age-group but not in the older cohort. Poor performance status was also associated with higher risk of ICANS in younger patients. Otherwise, other factors including age, history of central nervous system (CNS) involvement, and underlying cerebral microvascular disease were not associated with ICANS. Relapse incidence, non-relapse mortality, event free survival and overall survival were similar irrespective of ICANS occurrence or age-groups (Figures S3 and S4).

Our study explored the age effect on characteristics and outcomes of CAR T-cell associated ICANS. The incidence of ICANS in our cohort was 51.3%, comparable to previous studies.³ Recently, Lin et al. reported similar safety and tolerability of older patients with lymphoma treated with commercial CD19 CAR T-cells to younger patients aged 65 years old or younger.⁴ Since there were only 16 patients aged older than 65 years in our cohort (eight had ICANS), we performed the primary analysis using the age cut-off of 60 years old. Secondary analysis comparing < 65 versus ≥ 65 years-old patients showed similar findings (Tables S5–S8 and Figures S5–S7). Like previous studies, we observed no difference in the incidence and manifestations between two cohorts. So, ICANS had a highly variable clinical spectrum with aphasia, altered consciousness and agraphia being common neurological dysfunctions in both older and younger patients similar to previous reports.² We also observed a high percentage of patients experiencing headaches and other extrapyramidal manifestations like the recent study.⁵ Although some neurologic presentations are not included in ICANS diagnostic criteria, they may result in considerable morbidity and be used to track ICANS progression. We did not see the correlation between ICANS and short-term outcomes although previous studies demonstrated the association between severe ICANS and poor outcomes.⁶ Besides the potential effect on neurological and survival outcomes, older patients might also be at a higher-risk of developing toxicities from treatments of ICANS. Currently, data on impact of ICANS including its treatments on long-term neurological status, and physical function are limited. There was no difference in treatment patterns between age-groups. However, due to the retrospective design, we did not have data on delayed neurological function, physical recovery, and quality of life of our patients. The long-term aspects of ICANS including predictive biomarkers and prophylactic strategies have been actively explored by several groups.

In our cohort, the occurrence of CRS especially grade ≥2 was the only factor associated with the development of ICANS similar to other studies.^{1, 2} However, we did not observe the difference in the incidence of ICANS by the receipt of tocilizumab and corticosteroids for CRS management. Other than CRS, we did not observe the correlation between ICANS and other clinical factors.

The limitations of our study include the retrospective design and small sample-size with limited implication to axi-cel. The retrospective abstraction of neurological symptoms could affect the distribution of ICANS presentations. However, the main study objective was to describe ICANS characteristics based on the cliniciansʼ observation but not to re-define it. The strengths of our study are that it is one of the largest cohorts that describes real-world data of ICANS and the first to highlight the effect of age on ICANS.

In conclusion, ICANS was a common immune-mediated event that had a wide range of manifestation. The characteristics and short-term outcomes of ICANS were not affected by patientʼs age. Our findings could serve as a complimentary resource to develop patient-oriented management strategies of ICANS. Future studies integrating prospective outcome measurement, symptom monitoring and risk/biomarker-adapted intervention are warranted to improve treatment experience.

嵌合抗原受体 (CAR) T 细胞疗法彻底改变了复发/难治性 (R/R) 大 B 细胞淋巴瘤 (LBCL) 的治疗方法。除了其出色的功效外，CAR T 细胞还具有独特的免疫介导的不良事件，包括细胞因子释放综合征 (CRS) 和免疫效应细胞相关神经毒性综合征 (ICANS)；它的表现范围很广，从失语症、头痛到癫痫发作和昏迷。¹高龄被确定为 ICANS 的潜在诱发因素。²然而，年龄是否对 ICANS 表现有不同的影响并没有得到很好的描述。在此，我们报告了用 axicabtagene ciloleucel (axi-cel) 治疗的 R/R LBCL 患者的详细神经学表现，以及老年和年轻患者之间的相似性/差异，以更好地表征老年患者的 ICANS。我们进行了一项回顾性研究，纳入了 2016 年 6 月至 2020 年 10 月在明尼苏达州罗彻斯特梅奥诊所接受 axi-cel 治疗的 R/R LBCL 患者。 axi-cel 治疗过程与产品说明书一致。我们描述了 ICANS 的临床过程，包括神经系统表现、治疗和结果。我们根据输注时的年龄将患者分为两个年龄组；年轻（< 60 岁）和年长（≥ 60 岁）组，并比较他们的神经系统结果。使用免疫效应细胞脑病 (ICE) 评分作为主要参考将神经系统表现分层为不同的组。无法归入基于 ICE 的类别的神经系统症状被归类为独立实体（补充材料）。所有患者都接受了基线脑磁共振成像 (MRI) 和神经肿瘤学家的神经学评估。CAR T 提供商使用标准化协议对患者进行评估和管理。根据不良事件通用术语标准 4.03 版（2019 年 1 月之前）和美国移植细胞治疗协会 (ASTCT) 共识标准（2019 年 1 月之后）报告 ICANS 的严重程度。ICANS 的发生率定义为从 CAR T 细胞输注到 ICANS 发病的时间，任何原因导致的死亡都是竞争事件。该研究得到了梅奥诊所机构审查委员会的批准。连续变量被描述为中位数和范围，并通过 Kruskal-Wallis 检验进行组间比较。分类变量以百分比显示，通过卡方或费舍尔精确检验进行组间比较。累积发生率通过竞争风险分析进行分析。进行 Cox 比例风险分析以分析与 ICANS 相关的潜在因素，并报告为具有 95% 置信区间 (95%CI) 的风险比。连续变量被描述为中位数和范围，并通过 Kruskal-Wallis 检验进行组间比较。分类变量以百分比显示，通过卡方或费舍尔精确检验进行组间比较。累积发生率通过竞争风险分析进行分析。进行 Cox 比例风险分析以分析与 ICANS 相关的潜在因素，并报告为具有 95% 置信区间 (95%CI) 的风险比。连续变量被描述为中位数和范围，并通过 Kruskal-Wallis 检验进行组间比较。分类变量以百分比显示，通过卡方或费舍尔精确检验进行组间比较。累积发生率通过竞争风险分析进行分析。进行 Cox 比例风险分析以分析与 ICANS 相关的潜在因素，并报告为具有 95% 置信区间 (95%CI) 的风险比。p 值< 0.05 被认为具有统计学意义。

在研究期间，78 名 DLBCL 患者接受了 axi-cel。输注时的中位年龄为 58.8 岁（26.8-76.5 岁），其中 32 名患者 (41%) ≥ 60 岁。除了高国际预后指数、既往自体干细胞移植和老年患者潜在脑微血管疾病的比例较高外，各年龄组之间的基线特征具有可比性（表 S1 和 S2）。请注意，在 40 名患者 (51.3%) 中观察到 ICANS，分别包括 16 名和 24 名年龄较大和较年轻的患者。老年和年轻患者的 30 天 ICANS 发病率相当（50%、95% CI 29.3%–64.6% vs. 52.2%、95% CI 35.3%–64.6%，p  = 0.73）（图 1）。此外，在 13 名患者中观察到 ICANS 3 级或更高（n = 8; 17.4% 年轻人，n  = 5；15.6% 在老年组）。在 40 名 ICANS 患者中，39 名 (97.5%) 在 ICANS 发病前出现 CRS。从 CAR T 细胞输注到 ICANS 发作的中位时间在两个年龄组之间相似（p  = 0.42）。从 ICANS 发病到 ICANS 最严重程度的中位时间在 24 小时内，并且在年轻和年长队列之间没有统计学差异（p = 0.60）。年轻和年长患者的神经系统症状和最大 ICANS 严重程度的分布相似（图 1，表 S3）。三个最常见的初始神经系统异常包括两个年龄组的失语症、书写困难和意识模糊（表 S3）。言语和书写异常是两个年龄组最常见的症状。头痛和震颤在老年患者中更常见，但与年轻患者没有统计学差异。在 40 名 ICANS 患者中，2 名 (5.0%) 发生癫痫发作（1 名亚临床非惊厥性癫痫发作和 1 名惊厥性癫痫发作）。

26 名患者发生 ICANS 后进行了放射学检查，包括 9 名患者的脑部 MRI（年轻队列中的 5 名和年长队列中的 4 名）。一名年龄较大的患者出现了与 ICANS 相关的异常 MRI 发现（皮质下白质中的弥漫性 T2 增强），该患者出现了 ICANS ASTCT 3 级（图 S1）。在所有发生 ICANS 的患者中，一名年龄较小的患者 (2.5%) 由于精神状态严重改变而需要插管。24 名患者 (60.0%) 接受全身类固醇治疗 ICANS，其中 7 名 (17.5%) 需要脉冲甲基强的松龙治疗地塞米松耐药 ICANS。皮质类固醇（强的松等效物）的中位总剂量为 1.80 mg/m ²/天，年轻和年长患者之间没有差异（1.94 与 1.31 mg/m ² /天，p  = 0.65）。年轻和年长患者接受全身皮质类固醇治疗的中位持续时间分别为 4 天和 2 天（p  = 0.54）。总体而言，ICANS 治疗和炎症标志物（乳酸脱氢酶、c 反应蛋白、铁蛋白）在各年龄组之间没有差异（表 S2 和 S3）。所有 ICANS 解决的中位持续时间为 5 天（1-20 天），在较年轻和较年长的队列中具有可比性（5 天对 6 天，p = 0.18）。ICANS 课程期间 ICE 分数的间隔变化在年龄组之间没有差异（图 S2）。在单变量 Cox 比例风险分析（补充表 S4）中，严重 CRS ≥ 2 级是整个队列和年轻年龄组中与 ICANS 相关的唯一因素，但在老年队列中则不然。在年轻患者中，较差的体能状态也与较高的 ICANS 风险相关。此外，其他因素包括年龄、中枢神经系统 (CNS) 受累史和潜在的脑微血管疾病与 ICANS 无关。复发率、非复发死亡率、无事件生存期和总生存期与 ICANS 发生率或年龄组无关（图 S3 和 S4）。

我们的研究探讨了年龄对 CAR T 细胞相关 ICANS 的特征和结果的影响。我们队列中 ICANS 的发生率为 51.3%，与之前的研究相当。³最近，林等人。报道了用商业 CD19 CAR T 细胞治疗的老年淋巴瘤患者与 65 岁或更年轻的年轻患者相似的安全性和耐受性。⁴由于我们的队列中只有 16 名年龄超过 65 岁的患者（8 名患有 ICANS），因此我们使用 60 岁的年龄截断值进行了主要分析。比较< 65 岁与≥ 65 岁患者的二级分析显示出相似的结果（表S5-S8 和图S5-S7）。与之前的研究一样，我们观察到两个队列之间的发病率和表现没有差异。因此，ICANS 具有高度可变的临床谱，失语症、意识改变和失语症是老年和年轻患者常见的神经功能障碍，与之前的报告类似。²我们还观察到很大比例的患者出现头痛和其他锥体外系表现，如最近的研究。⁵尽管某些神经系统表现不包括在 ICANS 诊断标准中，但它们可能会导致相当大的发病率并用于跟踪 ICANS 进展。我们没有看到 ICANS 与短期结果之间的相关性，尽管之前的研究表明严重 ICANS 与不良结果之间存在关联。⁶除了对神经系统和生存结果的潜在影响外，老年患者也可能因 ICANS 治疗而产生毒性的风险更高。目前，关于 ICANS 影响的数据有限，包括其对长期神经系统状态和身体功能的治疗。年龄组之间的治疗模式没有差异。然而，由于回顾性设计，我们没有关于患者神经功能延迟、身体恢复和生活质量的数据。包括预测性生物标志物和预防策略在内的 ICANS 的长期方面已被多个团体积极探索。

在我们的队列中，与其他研究类似，CRS 的发生，尤其是≥2 级是与 ICANS 发展相关的唯一因素。^{1, 2}然而，我们没有观察到接受托珠单抗和皮质类固醇治疗 CRS 后 ICANS 发生率的差异。除了 CRS，我们没有观察到 ICANS 与其他临床因素之间的相关性。

我们研究的局限性包括回顾性设计和小样本量，对 axi-cel 的影响有限。神经症状的回顾性抽象可能会影响 ICANS 表现的分布。然而，主要研究目标是根据临床医生的观察来描述 ICANS 特征，而不是重新定义它。我们研究的优势在于，它是描述 ICANS 真实世界数据的最大队列之一，也是第一个强调年龄对 ICANS 影响的队列。

总之，ICANS是一种常见的免疫介导事件，具有广泛的表现形式。ICANS 的特征和短期结果不受患者年龄的影响。我们的研究结果可以作为开发 ICANS 以患者为导向的管理策略的补充资源。未来的研究需要整合前瞻性结果测量、症状监测和风险/生物标志物适应干预，以改善治疗体验。