Utilization of Investigations for Neurotoxicity in CD19 and BCMA CART Recipients

Background:

Chimeric antigen receptor T-cell (CART) therapy is utilized for patients with refractory/relapsed non-Hodgkin lymphoma (NHL), multiple myeloma (MM), and acute lymphoblastic leukemia (ALL). Immune effector cell-associated neurotoxicity syndrome (ICANS) is a potentially life-threatening toxicity of CART therapy graded according to ASTCT guidelines. The clinical manifestations of ICANS are scored using immune effector cell-associated encephalopathy (ICE) scores, degree of unconsciousness, presence of seizures, motor findings, and signs of raised intracranial pressure/cerebral edema. While these clinical signs and symptoms are mostly attributed to ICANS if they develop within the first-month post-CART, it is prudent to consider alternate causes such as central nervous system (CNS) infections or CNS disease progression. Therefore, neurodiagnostic testing, including lumbar puncture (LP) with cerebrospinal fluid (CSF) analysis, magnetic resonance imaging (MRI) of the brain, and electroencephalogram (EEG), is often performed along with, or prior to, treatment of ICANS. The purpose of these investigations is to exclude other etiologies: CSF analysis to evaluate for possible CNS infections and disease, MRI brain to assess for acute cerebral changes secondary to ICANS, or other etiologies, such as stroke and hemorrhages. EEG studies were obtained to detect subclinical seizures associated with ICANS. The utility of these tests has not been previously reported.

Methods:

This study is a retrospective review of 347 patients with NHL, MM, and ALL who received CART targeting CD19 or B-cell maturation antigen (BCMA) at the University of Kansas Medical Center from December 2017 to November 2023. Per institutional protocol, an LP, MRI, and EEG are performed at the onset of Grade 1 ICANS or based on physician discretion. All patients received levetiracetam for ICANS prophylaxis from Day 0-30. The frequencies of post-CART therapy LPs, MRIs, and EEGs and the incidence of abnormal findings from each test were measured.

Results:

Of the 347 patients who received CART therapy, 77% of patients received CD-19-directed CART (n=266), and 23% received BCMA-directed CART (n=81). Patients with NHL comprised 72% of the study population (n=250) and had a median age of 64 years (range: 24.7-86.0). Patients with MM comprised 23% of the study population (n=81) and had a median age of 66.2 years (range: 42.3-80.8). Patients with ALL comprised 5% of the study population (n=17) and had a median age of 33.1 years (range: 19.9-68.3).

142 patients (41%) developed any grade of ICANS and 254 (73%) developed any grade of cytokine release syndrome (CRS). 42% had ICANS Grade 1 (n=59), 15% had Grade 2 (n=22), 23% had Grade 3 (n=32), 16% had Grade 4 (n=23), and 4% died of ICANS (Grade 5) (n=6). 92% (n=131) of the patients who developed ICANS experienced concurrent CRS. Of the 142 ICANS patients, 7% had CRS Grade 0 (n=11), 49% had CRS Grade 1 (n=70), 39% had CRS Grade 2 (n=55), 4% had CRS Grade 3 (n=5), and 1% had CRS Grade 4 (n=1).

141 ICANS patients (99%) underwent brain MRI. 13 (9%) of these patients had abnormal findings, which included cerebral edema, new contrast-enhancing lesions, new ischemic stroke, and acute hemorrhages. 112 (79%) patients underwent an EEG. 93 patients (83%) receiving an EEG developed at least mild encephalopathy or cerebral dysfunction. However, only three patients (3%) who underwent an EEG demonstrated seizure activity. 102 patients (72%) had an LP within 4 weeks post-infusion. The median number of LPs performed in the first four weeks post-infusion was 1 (range: 1-4). Only 1 patient (1%) had abnormal CSF – a bacterial infection.

Conclusion:

A low yield with LPs and EEGs in CD19 and BCMA CART recipients with ICANS was observed. Only one patient (1%) who had an LP was determined to have a CNS infection, suggesting that CNS infection as an alternate cause of neurological change is rare. While most patients with ICANS had non-specific encephalopathy or cerebral dysfunction on EEG, seizures were detected in only 3% of all EEGs, suggesting that subclinical seizures are rare. Brain MRI identified the highest yield of abnormal findings, with 9% exhibiting abnormalities, including alternate explanations for altered mental status. These results demonstrate a low utility of LP and EEG, which suggests that these tests may be reserved for higher grade or persistent ICANS