Standardization Protocol to Mitigate Progression of Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS) in Elderly Patients Treated with CD28 Co-Stimulatory Anti-CD19 CAR-T Cells – a Pilot Study

Background

Early access to CAR-T has resulted in a growing population of older patients previously considered ineligible for CAR-T to receive this therapy. Among the elderly population, specifically those with pre-existing neurologic comorbidities, the incidence and severity of ICANS are higher. ICANS treatment guidelines are mainly derived from data excluding older aged patients. As such, we designed a protocol to prevent escalation of ICANS in the older age population receiving a CD28 co-stimulatory constructs.

Methods

Eligibility included all patients treated with CD28-based anti CD19 CAR-T cells aged ≥75 years and patients ≥65 years with at least one risk factor for ICANS (pre-existing neurologic comorbidity, ECOG ≥2, or high disease burden).

All patients received prophylaxis levetiracetam and thiamine from day 0. Concomitant dexamethasone was added to tocilizumab when administered for CRS. Grade 1 ICANS was treated with daily dexamethasone 20mg, Grade 2 – dexamethasone 20mgx4/d, Grade 3-4 – methylprednisolone 1-2gr/day, respectively, for 3 days with concomitant anakinra 100mgx2/d. Refractory ICANS was defined as lack of improvement within 24 hours of treatment or grade escalation and managed according to the subsequent grade. Diagnostic workup including neuroimaging and electroencephalogram were performed for all grade ≥2 ICANS. ICANS was defined by the ASTCT consensus (2019), response by the Lugano criteria (2016), and CAR-T expansion was measured by immunophenotyping.

Results

Between 04/2023 and 06/2024, 32 patients treated were eligible (axicabtagene ciloleucel for LBCL, n=24; brexucabtagene autoleucel for MCL, n=7 and for B-ALL, n=1). Median age was 75 (range 65-86) years, 34% had an ECOG score of ≥2, and 41% had pre-existing neurologic comorbidity (19% CNS involvement). Among patients with LBCL 56% received CAR-T as second line. Median LDH prior to LD was 424 (IQR 380-574) U/L and 36% of patients achieved CR after bridging and prior to LD.

Twenty-three patients (72%) experienced ICANS of any grade with 28% grade ≥3. Median days to ICANS onset of any grade and grade 3-4 was 5 (range, 0-12) and 5 (range, 4-8), respectively. Median ICANS duration was 5 (range, 1-27) days. ICANS progression was observed in 4 patients (15% of patients with grade 1-2), and time to resolution to grade ≤1 was 2 (range 1-6) days.

CRS was observed in 96.9% of patients, 3 (9.4%) grade ≥3. Time to CRS onset and CRS duration was 2 (range 0-8) and 4 (range 1-10) days respectively. 32% of patients received tocilizumab for CRS.

Mean cumulative steroid dose (by dexamethasone equivalent) was 368 (IQR 25-647) mg, and 28% required treatment with anakinra. Median day +7 expansion of CAR-T was 65 (range 0-1725) cells/microL. Median inpatient days was 23 (IQR 19-37). Three patients (9.3%) were discharged to a rehabilitation center.

Prior neurologic comorbidity was associated with ICANS development (p=.003) and day +7 expansion was numerically higher in patients who developed ICANS compared with those who did not (median 90 (IQR 44-306) vs 27 (IQR 1-52) cells/microL, respectively, p=.056).

In a univariate analysis, performance status, age ≥75, disease status at LD, CNS involvement, number of prior lines, modified EASIX score, CRS incidence and grade, and use of tocilizumab did not predict ICANS development, grade, or duration.

One and 3-months ORR were 82% (71% in CR) and 75% (all in CR), respectively. Non-relapse mortality at 1 month was 13% (95% CI 0-22%), majority due to infections (n=3) and 1 patient due to pulmonary embolism.

Conclusions

In this pilot study, we utilized a novel standard and goal-directed treatment in elderly patients with ultra-high risk characteristics for development of ICANS. Incidence of ICANS was similar, duration was shorter and recovery was faster compared to pivotal trials. Previously reported risk factors for neurotoxicity which are commonly considered when evaluating older patients as candidates for CAR-T, were not discriminatory. If reproduced in a larger cohort, these findings suggests that risk for ICANS in older adults can be ameliorated using an individualized assessment and treatment strategy.