Brain MRI Changes in Children and Young Adults with B-cell Acute Lymphoblastic Leukemia Following Chimeric Antigen Receptor T-cell Therapy

Abstract

Objective:

To evaluate brain MRI findings in children and young adults after chimeric antigen receptor (CAR) T-cell therapy for B-cell acute lymphoid leukemia (B-ALL) and associate results with clinical and neurological symptoms.

Methods:

We reviewed pre- and post-CAR-T cell therapy brain MRIs of B-ALL patients aged 25 years or younger who underwent therapy between April 2015 and October 2023 at a single institution. MRI abnormalities were categorized as no change, exacerbation of preexisting lesion, or newly developed lesion. Clinical CAR mediated toxicities including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) grades, were recorded. Patients were grouped into those with and without ‘exacerbated/new lesion,’ and clinical and neurological symptoms were compared using Fisher’s exact test.

Results:

Sixteen patients with pre- and post-CAR brain MRIs (median age 16 years [interquartile range, 11–21]; 9 males, 7 females) were included in the analysis. Post-CAR brain abnormalities were observed in 81% (13/16) of patients, including white matter (WM) signal changes (12/16), leptomeningeal enhancement (1/16), and cerebellar embolic infarction (1/16). Of the post-CAR WM lesions, 50% (6/12) were exacerbated, 33% (4/12) were newly developed, and 17% (2/12) remained unchanged compared to pre-CAR brain MRI. No difference in CRS (p= 0.079) or ICANS grades (p> 0.99) was observed between patients with and without ‘exacerbated/new lesions’.

Conclusion:

Children and young adults with B-ALL can develop brain MRI abnormalities after CAR T-cell therapy, predominantly WM signal changes. These brain abnormalities did not show association with higher CRS or ICANS grade.

Introduction

Chimeric antigen receptor (CAR) T-cell therapy is an immune-based treatment method for acute lymphoid leukemia (ALL) [1]. It involves genetic modification of a patient’s T-cells to express CARs that recognize and bind to specific tumor antigens, resulting in elimination of antigen-positive malignant cells [2]. CAR T-cell therapy achieved complete 65%–90% response rates in children with B-cell acute lymphoblastic leukemia (B-ALL) [3–7]. However, significant toxicity has been reported in up to 95% of patients, including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) [4].

ICANS is the second most common adverse effect of CAR-T cell therapy and manifests with a variety of neurological symptoms such as confusion, tremor, or aphasia in the early phase after CAR T-cell infusion [8, 9]. Gust et al. investigated the pathological characteristics of ICANS in 133 adults treated with CD19 CAR-T cells and found that serum cytokines, including Fin and TNFα, led to blood brain barrier disruption and severe neurotoxicity, with some patients presenting with T2 FLAIR hyperintense signal in the white matter (WM), global cerebral edema, or multifocal hemorrhages on brain MRI [10]. At present, there is limited literature on MRI brain findings following CAR T-cell therapy in children and young adults. In 2019, Gust et al. reported acute brain MRI abnormalities in six out of 15 children and young adults who underwent CD19 CAR T-cell therapy and brain MRI within 28 days after therapy [11], that included T2 FLAIR hyperintense signal in the WM, brainstem, or the cerebellum, and in some cases, diffusion restriction in the cerebral cortex. In 2022, Shalabi et al. also reported brain abnormalities such as hyperintense T2 FLAIR signal in the WM or diffusion restriction in nine out of 11 children and young adults who had neurotoxicity after CD19/CD28 CAR T-cell therapy [12]. However, these previous studies lacked comparison with pre-therapy scans. This limits our ability to determine which brain MRI findings are related to CAR T-cell therapy. Similarly, previous studies provided limited information on how other clinical and neurological condition such as CRS and ICANS grade can be related to brain MRI abnormality after CAR T-cell therapy. This limits our ability to determine the clinical impact of brain MRI changes after CAR T-cell therapy.

The purpose of our study was to evaluate brain MRI findings in children and young adults after chimeric antigen receptor (CAR) T-cell therapy for B-cell acute lymphoblastic leukemia (B-ALL) and associate results with clinical and neurological symptoms.

Materials and Methods

Patients

This study was approved by the institutional review board (CCT5019). All patients signed a written informed consent to participate in a prospective clinical trial with CD19 or CD19/22 CAR T-cell therapy and associated clinical and image data analyses. For this retrospective study, we identified 66 children and young adults with an age of 25 years or younger who underwent CAR T-cell therapy between April 2015 and October 2023 at our institution, Lucile Packard Children’s Hospital Stanford. We identified 23 patients who received an MRI of the brain after CAR T-cell therapy. Of these 23 patients, 16 patients received brain MRI before CAR T-cell therapy (Figure 1). The 16 patients represent those who had both pre- and post-CAR T-cell therapy brain MRI, regardless of the interval between the scans and the CAR T-cell therapy. For this group of patients, we did not define a specific interval between pre- and post-CAR T-cell therapy. There were 7 patients who received methotrexate (MTX) between pre-CAR T-cell brain MRI and CAR T-cell infusion. The timeline of the therapies and brain MRIs are summarized in Supplemental Figure 1.

Figure 1.

Flow chart of the patient selection

We investigated a subgroup of seven patients who underwent brain MRI within 30 days after CAR T-cell therapy to avoid confounding findings on brain MRI attributed to other causes. This interval was selected to capture peak neurologic toxicities, as median time to onset of ICANS is 9 days (interquartile range, 7–11 days) and median duration is 7 days (interquartile range, 4–19 days) [8].

Clinical and neurological information

The following clinical information was recorded: age, sex, prior stem cell transplantation history, prior methotrexate (MTX) treatment history, disease burden, CRS grade, and ICANS grade. Disease burden was categorized into three categories: high disease burden, low disease burden, and no detectable disease [13]. High disease burden was defined by ≥ 5% bone marrow lymphoblasts, any peripheral blood lymphoblasts, central nervous system 3 status (central nervous system 3 indicates the presence of overt central nervous system leukemia, in which the sample contains >5 white blood cells and lymphoblasts), or non-central nervous system extramedullary site of disease. Low disease burden was defined as morphologic or flow cytometry detectable bone marrow or central nervous system disease not meeting the high disease burden criteria. No detectable disease was defined as the absence of morphologic or flow cytometry-detectable disease or extramedullary disease. CRS grade and ICANS grade were based on the severity of systemic adverse effect and neurotoxicity, respectively [14]. We followed the ASTCT consensus grading criteria for both CRS and ICANS. The reason for brain MRI examination after CAR T-cell therapy was also documented. All patients received seizure prophylaxis with Levetiracetam.

We included 16 patients (median age 16 years [interquartile range, IQR, 11–21]; 9 males and 7 females) with B-ALL who underwent CAR T-cell therapy and had brain MRI before and after therapy. Demographic and clinical characteristics of 16 patients are summarized in Table 1. Prior stem cell transplantation was received in 4/16 (25%) patients and either systemic or intrathecal prior MTX treatment was received in all (16/16, 100%) patients. There were 2/16 (13%) high disease burden patients, 4/16 (25%) low disease burden patients, and 10/16 (63%) patients with no detectable disease at CAR T-cell infusion. CRS grade ≥ 3 was observed in 2 patients (2/16, 12%) and grade 0 was seen in 4 patients (4/16, 25%). ICANS grade ≥ 3 was observed in no patient (0/16, 0%) and grade 0 was seen in 13 patients (13/16, 81%). ICANS grade ≥ 1 was observed in 3 patients (3/16, 19%).

Table 1.

Clinical characteristics of patients who received a brain MRI before and after CAR T-cell therapy

Characteristics	Total patients (n = 16)
Age* (year)	16 [11–21]
Sex (M:F)	9 (56) : 7 (44)
Prior SCT
Not received	12 (75)
Received	4 (25)
Prior MTX treatment
Not received	0 (0)
Received	16 (100)
Disease burden
No detectable disease	10 (63)
Low	4 (25)
High	2 (13)
CRS grade
0	4 (25)
1	6 (38)
2	4 (25)
3	1 (6)
4	1 (6)
ICANS grade
0	13 (81)
1	1 (6)
2	2 (13)
3	0 (0)
Reasons for post-CAR brain MRI
Follow up for pre-CAR abnormality	8 (50)
Headache or orbital symptom	5 (31)
Altered mental status	3 (19)
Interval between CAR T-cell and post-CAR brain MRI (day)	79 [18–188]

Data are number of patients; data in parenthesis are %

^*Age in median; data in brackets are interquartile range

CAR, chimeric antigen receptor; CRS, cytokine release syndrome; ICANS, and immune effector cell-associated neurotoxicity syndrome; SCT, stem cell transplantation; MTX, methotrexate

Seven patients underwent brain MRI within 30 days after CAR T-cell therapy (Table 2), including 3 males and 4 females with a median [IQR] age of 16 [11–18] years and interval between CAR T-cell therapy and post-CAR T-cell therapy and brain MRI of 16 [4 – 27] days.

Table 2.

MRI findings of patients who had brain MRI within 30 days after CAR T-cell therapy

Patient No.	Age	Sex	Prior SCT	Prior MTX Tx.	Disease burden	CRS grade	ICANS grade	Interval between therapy and post-CAR MRI (day)	Reasons for post-CAR brain MRI	Pre-CAR T-cell therapy brain MRI findings	Post-CAR T-cell therapy brain MRI findings
1	22	F	No	Yes	Not detectable	3	0	4	Altered mental status*	WM signal change Subdural hemorrhage	No change of WM signal change No change subdural hemorrhage
2	17	F	No	Yes	Not detectable	1	0	3	Headache	WM signal change	Exacerbation of WM signal change
3	15	M	No	Yes	Low	2	2	23	Headache, orbital Sx.†	Normal	Normal
4	18	M	No	Yes	Not detectable	1	0	4	Headache	Normal	Newly developed WM signal change Newly developed leptomeningeal enhancement
5	1	M	Yes	Yes	Not detectable	0	0	27	Follow up imaging	Normal	Newly developed WM signal change
6	16	F	Yes	Yes	High	1	0	29	Follow up imaging	WM signal change Parenchymal hemorrhage	Exacerbation of WM signal change No change of parenchymal hemorrhage
7	11	F	No	Yes	High	4	2	16	Headache	Normal	Newly developed cerebellar diffusion restriction

^*This patient had altered mental status, but the symptom was considered irrelevant to CAR T-cell therapy by an oncologist; therefore, the ICANS grade was 0.

^†This patient’s orbital symptom was retro-orbital discomfort.

CAR, chimeric antigen receptor; SCT, stem cell transplantation; MTX Tx., methotrexate treatment; CRS, cytokine release syndrome; ICANS, and immune effector cell-associated neurotoxicity syndrome; Sx., symptom; WM, white matter

Brain MRI acquisition

Brain MRI scans were performed when clinically needed under our institution’s oncologist’s guidance. Brain MRI after CAR T-cell therapy was conducted using the following MRI scanners: DISCOVERY MR450 1.5T, OPTIMA MR450W 1.5T, SIGNA Premier 3T, DISCOVERY MR750 3T, and SIGNA PET/MR 3T (General Electric Healthcare, Chicago, IL, USA). One out-side hospital MRI scan was performed using MAGNETOM Prisma 3T scanner (Siemens Healthcare, Healthcare, Erlanger, Germany). All patients underwent precontrast imaging, including T1-weighted, T2 FLAIR, T2-weighted, and diffusion-weighted sequences, followed by the intravenous administration of a gadolinium-based contrast agent and postcontrast T1-weighted imaging. Not all technical parameters were identical and imaging protocol and scanner information for the patients are summarized in Supplemental Table 1 and 2.

Brain MRI evaluation

All brain MRI scans were reviewed independently by two experienced pediatric radiologists (a faculty pediatric radiologist with 15 years of experience and a pediatric radiology fellow with 8 years of experience). The radiologists reviewed each case together and reached a final assessment by consensus. The reviewers recorded the presence or absence of abnormal findings on MRI scans before and after CAR T-cell therapy [9–12, 15, 16]. Imaging evaluation included identification of specific pathologic features, such as presence of T2 FLAIR hyperintense WM lesions, leptomeningeal enhancement, abnormal brain diffusion signal that might indicate recent ischemia or focal necrosis/demyelination, as well as presence of T2* hypointense hemorrhage or mineralization. Examples are shown in Figure 2. Other intracranial abnormalities including ventriculomegaly or extra axial hemorrhages (e.g., subdural hemorrhage) were also recorded.

Figure 2.

Schematic images of previously reported brain MRI characteristics following CAR T-cell therapy.

Next, the reviewers compared ‘MRI findings after CAR T-cell therapy’ with ‘MRI findings before CAR T-cell therapy’ as follows: (a) no interval change of preexisting lesion (b) exacerbation of preexisting lesion, and (c) interval development of new lesion. In the case of preexisting abnormal WM signal, the WM lesions were considered exacerbated if the extent of signal abnormality had enlarged. The final MRI findings and categorization of the lesions were recorded in consensus by the two radiologists.

Statistics

We identified two groups of patients, who did (Group 1) or did not (Group 2) demonstrate ‘exacerbation of preexisting or newly developed brain lesion(s) (exacerbated/new lesions)’. Group 2 included patients with no abnormality on brain MRI or with no change of preexisting lesion(s). Age and interval between CAR T-cell and brain MRI were compared between group 1 and 2 using a Mann-Whitney test. Gender, prior SCT history, and prior MTX treatment history, disease burden, CRS grade, ICANS grade, and reason for taking post-CAR T-cell brain MRI were compared between the two groups using a Fisher’s exact test. We used Stata 18 (College Station, TX: StataCorp LLC) for statistical analysis and p values less than 0.05 were considered to indicate statistical significance.

Results

Brain MRI findings before CAR T-cell therapy

Among 16 patients who obtained both pre and post-CAR brain imaging, seven showed no abnormalities on brain MRI before CAR T-cell therapy and nine showed abnormal findings. Eight of the nine patients with pre-existing brain abnormality showed T2 FLAIR hyperintense WM signal in bilateral cerebral hemispheres; and one patient showed diffuse ventriculomegaly. The 8 patients with abnormal WM signal showed the following features: diffusely confluent abnormal WM signal (n = 5), bilateral periventricular WM involvement (n = 2), and multifocal WM signal abnormality (n = 1). Two patients with diffuse WM signal abnormality also had other coexisting lesions: subdural hemorrhage (n = 1) and multifocal parenchymal hemorrhages (n =1).

Brain MRI findings after CAR T-cell therapy

Reasons for taking post-CAR brain MRI were following imaging for pre-CAR brain imaging abnormality (n = 8), headache or orbital symptom (n = 5), and altered mental status (n = 3). Median [IQR] interval between CAR T-cell therapy and post-CAR brain MRI was 79 [18–199] days.

Post-CAR brain abnormalities were observed in 81% (13/16) of patients, including WM signal changes (12/16), leptomeningeal enhancement (1/16), and cerebellar embolic infarction (1/16). Of the WM lesions, 50% (6/12) were exacerbated, 33% (4/12) were newly developed, and 17% (2/12) remained unchanged compared to pre-CAR brain MRI (Table 3, Figure 3). White matter signal changes exacerbated on serial follow-up brain MRI scans in one patient (Supplemental Figure 2).

Table 3.

Absolute and relative number of patients with brain MRI findings after CAR T-cell therapy compared to the findings before CAR T-cell therapy

Post-therapy MRI finding	Total†	No change‡	Exacerbated‡	Newly developed‡
Abnormality on MRI	13/16 (81)
WM signal change	12/16 (75)	2 (17)	6 (50)	4 (33)
Diffusion restriction	1/16 (6)	0 (0)	0 (0)	1 (100)
Leptomeningeal enhancement*	1/16 (6)	0 (0)	0 (0)	1 (100)
Parenchymal hemorrhage*	1/16 (6)	1 (100)	0 (0)	0 (0)
Ventriculomegaly*	1/16 (6)	1 (100)	0 (0)	0 (0)
Subdural hemorrhage*	1/16 (6)	1 (100)	0 (0)	0 (0)
Normal brain MRI	3/16 (19)

^*The patients had coexisting WM signal change

^†Data are the number of patients with the percentage of the total patients in parentheses.

^‡Data are the number of patients with the percentage of the specific abnormality in parentheses.

CAR, chimeric antigen receptor; WM, white matter

Figure 3. Representative example of an 8-year-old male patient with B-ALL showing newly developed white matter high signal change on FLAIR after CAR T-cell therapy.

A pre-CAR T-cell therapy axial T2-weighed FLAIR image shows no abnormality on cerebral white matter (A). A post-CAR T-cell therapy axial T2-weighed FLAIR image shows newly developed diffuse cerebral white matter high signal change in both cerebral hemispheres (B).

One patient demonstrated a 4 mm non-enhancing lesion with restricted diffusion in the left cerebellar WM (Figure 4). The patient underwent post-CAR brain MRI 16 days after the CAR T-cell infusion due to headache. The diffusion restriction lesion was resolved on serial follow-up brain MRI (Supplemental Figure 3). One patient demonstrated newly developed leptomeningeal enhancement (Figure 5). The leptomeningeal enhancement resolved on serial follow-up brain MRI (Supplemental Figure 4). Both restricted diffusion in the left cerebellar WM and leptomeningeal enhancement were newly developed brain lesions.

Figure 4. Representative example of an 11-year-old female patient with B-ALL showing newly developed diffusion restriction lesion in cerebellum after CAR T-cell therapy.

A pre-CAR T-cell therapy axial diffusion-weighted image shows no lesions in the cerebellum (A). A post-CAR T-cell therapy axial diffusion-weighted image shows newly developed a 4 mm diffusion restriction focus in the left cerebellar hemisphere (B).

Figure 5. Representative example of an 18-year-old male patient with B-ALL showing newly developed leptomeningeal enhancement after CAR T-cell therapy.

A pre-CAR T-cell therapy axial T2-weighed FLAIR image (A) and an axial T1-weighted contrast enhancement image (B) show no abnormal finding. A post-CAR T-cell therapy axial T2-weighed FLAIR image shows a newly developed white matter high signal lesion in the left cerebral hemisphere (C, arrow). A post-CAR T-cell therapy axial T1-weighted contrast enhancement image demonstrates newly developed leptomeningeal enhancement along the right cerebral hemisphere (D, arrow).

One patient showed newly developed WM signal change on FLAIR, despite unchanged pre-existing hemorrhage (Figure 6). Since multifocal parenchymal hemorrhages were present on the pre-CAR brain MRI, the hemorrhages were not considered to be related to CAR T-cell therapy. Other preexisting brain abnormalities of ventriculomegaly (1/16, 6%) and subdural hemorrhage (1/16, 6%) did not show significant change after CAR T-cell therapy.

Figure 6. Representative example of a 16-year-old female patient with B-ALL showing exacerbated white matter signal change but with stable parenchymal hemorrhage.

A pre-CAR T-cell therapy axial T2-weighed FLAIR image shows pre-existing high signal intensity lesions in both cerebral white matter (A, arrow). A pre-CAR T-cell therapy axial susceptibility-weighted image shows a microhemorrhage in the right temporal lobe (B, arrow) along with other focal hemorrhages (not shown). A post-CAR T-cell therapy axial T2-weighed FLAIR image shows increased in extent of hyperintense white matter high signal changes in both parietal lobes (C, arrow). A post-CAR T-cell therapy susceptibility-weighted image shows no change in the microhemorrhage in the right temporal lobe (D, arrow).

Patient by patient analysis

On a patient-by-patient basis for post-CAR brain MRI, three of 16 patients (19%) showed no abnormality, two patients (13%) had no change of preexisting lesions, seven patients (58%) had exacerbation of preexisting lesions, and four patients (25%) had newly developed lesions.

Of the seven patients who received their post-CAR brain MRI scan within 30 days after therapy, five patients (71%) showed exacerbation or newly developed abnormality: four patients with WM abnormality (57%) (Figure 5, 6) and one patient with diffusion restriction lesion (Figure 4) (Table 3). Two patients (29%) showed no change in their MRI scan. Reasons for taking post-CAR brain MRI were headache or orbital symptom (n = 4), altered mental status (n =1), and follow up for pre-CAR imaging abnormality (n = 2). One patient presented with altered mental status 9 days after CAR T-cell infusion and underwent brain MRI at 14 days post-CAR, which showed no changes compared to the pre-CAR brain MRI. Another patient experienced headaches and orbital symptoms 14 days post-CAR and underwent brain MRI at 23 days post-CAR, which also showed no changes compared to the pre-CAR brain MRI. A third patient reported a headache 1 day after CAR T-cell infusion and underwent brain MRI at 4 days post-CAR, revealing newly developed WM abnormalities and leptomeningeal enhancement (Figure 5).

Association between clinical and neurological symptoms

When patients with and without ‘exacerbated/new lesions’ were grouped, there were 11 patients (69%) with ‘exacerbated/newly lesions’ and 5 patients (31%) without ‘exacerbated/newly lesions’ (Table 4).

Table 4.

Comparison between patients with and without exacerbation or newly developed lesions

Characteristics	Patients with exacerbated/new lesions (n = 11)	Patients without exacerbated/new lesions (n = 5)	p value
Age *	11 [9–17]	22 [14–24]	0.082
Gender			0.596
Male	7 (64)	2 (40)
Female	4 (36)	3 (60)
Prior SCT			0.245
Not received	7 (64)	5 (100)
Received	4 (36)	0 (0)
Prior MTX Tx.			> 0.99
Not received	0 (0)	0 (0)
Received	11 (100)	5 (100)
Disease burden			> 0.99
No detectable disease	7 (64)	3 (60)
Low	2 (18)	2 (40)
High	2 (18)	0 (0)
CRS grade			0.079
0	4 (36)	0 (0)
1	5 (45)	1 (20)
2	1 (9)	3 (60)
3	0 (0)	1 (20)
4	1 (9)	0 (0)
ICANS grade			> 0.99
0	9 (82)	4 (80)
1	1 (9)	0 (0)
2	1 (9)	1 (20)
3	0 (0)	0 (0)
Reasons for post-CAR brain MRI			0.038
Follow up for pre-CAR abnormality	7 (64)	1 (20)
Headache or orbital symptom	4 (36)	1 (20)
Altered mental status	0 (0)	3 (60)
Interval between CAR T-cell infusion and post-CAR brain MRI (day) *	40 [16–316]	117 [23–137]	0.702

Data are number of patients with percentage in parenthesis

^*Data are median with interquartile range in brackets

SCT, stem cell transplantation; MTX Tx., methotrexate treatment; CRS, cytokine release syndrome; ICANS, immune effector cell associated neurotoxicity syndrome; CAR, chimeric antigen receptor

No difference in CRS (p= 0.079) or ICANS grades (p > 0.99) was observed between patients with and without ‘exacerbated/newly lesions’. Reasons for brain MRI after CAR T-cell therapy were different between the two groups (p= 0.038) with more frequent patients with ‘follow up for pre-CAR imaging abnormality (64% vs 20%)’ in patients with exacerbated/new lesion and more frequent patients with ‘altered mental status (60% vs. 0%)’ in patients without exacerbated/new lesion. There was no difference in age, gender, prior stem cell transplantation, prior MTX treatment, disease burden, and ‘interval between CAR T-cell therapy and post-CAR brain MRI’ between the two groups (all ps> 0.05).

Discussion

Our data showed that 81% of patients showed abnormal findings on brain MRI after CAR T-cell therapy. Clinical symptoms, quantified by CRS or ICANS grades, were not associated with exacerbated or newly developed brain abnormalities.

The most common brain MRI abnormality observed in our study was WM signal abnormality. T2 FLAIR hyperintense WM signal abnormalities have been reported in adult studies [9, 10, 17] and in children [11, 12] who underwent CAR T-cell therapy. Gust et al. reported brain abnormalities in 7 of 23 adult patients (30%; [10]) and 6 of 15 children and young adult patients (40%, [11]) who obtained brain MRI. In their adult study, WM signal abnormality was observed, though the number of affected patients was not specified [10]. In their pediatric analysis, two of 15 patients with neurotoxicity who underwent brain MRI showed acute WM signal abnormalities [11]. However, both studies lacked direct comparison with baseline brain MRIs prior CAR T-cell therapy, limiting direct association between CAR T-cell therapy and development of WM lesions. Additionally, some of the diffuse WM abnormalities were considered unrelated to CAR T-cell therapy, potentially preexisting or related to other non-CAR T-cell therapies. In our study, comparison of both baseline pre- and post-therapy MRI brain scans helped identify brain abnormalities stemming directly from CAR T-cell therapies. Here, we found that diffuse WM signal abnormalities can arise or worse from CAR T-cell therapy.

The underlying mechanism of WM signal change on MRI after CAR T-cell therapy is considered multifactorial [10, 11]. CAR T-cell mediated release of cytokines can disrupt the blood brain barrier and lead to a vasogenic edema of the WM. This increased extracellular fluid accumulation can manifest as areas of high signal intensity on T2-weighted FLAIR sequences [18]. Gust et al. found an abnormal astrocyte state during neurotoxicity following CAR T-cell therapy [11]. Since astrocytes are key osmotic regulators in the brain, astrocyte abnormalities can contribute to cerebral edema [19]. In addition, CAR T-cells can exert direct cytotoxic effects on the brain [20]. While the primary target is cancer cells, Parker et al. reported that CD19-targeted CAR T-cells can directly attack CD-19 expressing pericytes and vascular smooth muscle cells (mural cells) in the central nervous system cells [20].

Rubin et al. reported foci with restricted diffusion in an adult patient following CAR T-cell therapy, suggesting an embolic event [21]. This is consistent with an observed focus of diffusion restriction in one of our patients. Patients undergoing CAR T-cell therapy have an increased risk of thromboembolic events owing to pro-inflammatory cytokines and coagulopathies [9, 10]. There is also evidence of an association between CAR T-cell therapy and cardiovascular events such as cerebral ischemia and cerebral hemorrhage [22]. Among 140 adults, 7% showed CAR T-cell therapy-related thrombotic event including myocardial infarction and pulmonary embolism [23]. Gust et al. reported diffusion restriction along the occipital cortex of brain in a pediatric patient with B-ALL following CAR T-cell therapy [11]. Santomasso et al. reported thrombotic emboli in the corpus callosum in an adult patient [9]. Therefore, diffusion-weighted images should be evaluated for embolic infarction in patients after CAR T-cell therapy.

One patient showed leptomeningeal enhancement associated with WM signal abnormality. The brain MRI abnormalities were considered post-CAR changes, as the patient had a normal baseline brain MRI and developed the abnormalities in the post-CAR MRI taken 4 days after CAR T-cell therapy, following the onset of headaches. Leptomeningeal enhancement was reported in patients with ICANS in previous studies [10, 16]. Vascular disruption and endothelial activation observed in leukemic patients with ICANS can contribute to this inflammatory process in the meninges [9, 10]. However, other causes, such as meningeal leukemia, infective meningitis chemical meningitis, or reactive changes due to irradiation should be differentiated [24, 25]. Therefore, cerebrospinal fluid study and prior treatment history should be carefully reviewed for differential diagnosis of the leptomeningeal enhancement.

We found no association between CRS or ICANS grades and exacerbated or newly developed brain abnormalities. The neurotoxicity rate in our study (19%) was lower than that reported in previous studies (40 – 62%). One suspected reason for this is the use of seizure prophylaxis with Levetiracetam in every patient, regardless of CRS or other neurological symptoms. In a previous study, seizure prophylaxis was initiated only for patients who developed severe CRS or neurological symptoms [11].

Our study has several limitations. First, the number of patients included was small. With this small number of patients group comparison between those with and without exacerbated/new lesions would have limited value. Second, our study was a cross-sectional study based on a retrospective data set. Only selected patients underwent post-CAR T-cell therapy brain MRI and underwent pre-therapy brain MRI. Given that neurotoxicity was more likely in patients who had known brain MRI abnormalities before CAR T-cell therapy [11], selected patients with both pre- and post-therapy brain MRI would have increased the overall rate of abnormalities observed in the brain MRI after the therapy. Third, not all patients underwent brain MRIs using the same scanner or protocol. Brain MRI scans were obtained in a clinical setting with multiple different scanners. Half of the patients (Supplemental Table 2) underwent MRIs with different scanners before and after therapy. Scanner and sequence variation can affect the appearance of the lesion, thereby affecting the final categorization of the brain MRI characteristics. Fourth, the intervals between CAR T-cell therapy and brain MRI scans varied. Due to these varying intervals and differing clinical scenarios, establishing a direct relationship between CAR T-cell therapy and the observed brain MRI changes, while excluding potential confounding factors, was not possible. Lastly, we did not include functional outcome data for the patients. This limits our ability to interpret and understand the brain MRI changes in relation to patients’ neurodevelopmental outcomes.

In conclusion, we found evidence of brain MRI changes after CAR T-cell therapy in children and young adults with B-ALL, but these changes were not linked to the severity of clinical-neurological symptoms. Future studies could evaluate the long-term neurological outcomes of these patients by correlating the quantifiable brain MRI abnormality to neurodevelopmental assessment scores. This would help to determine if the observed WM changes lead to long-term cognitive, motor, or sensory deficits.

Key points:

• Brain MRI findings after chimeric antigen receptor (CAR) T-cell therapy for B-cell acute lymphoblastic leukemia (B-ALL) and their association with clinical and neurological symptoms are not well understood.

• Brain MRI abnormalities, mostly white matter changes, were seen in 81% of patients but were not associated with CAR-mediated toxicities.

• Brain MRI abnormalities, commonly observed post-CAR T-cell therapy, do not correlate with the severity of CAR-related toxicities, aiding in the clinical management and monitoring of these patients.