Berlin-Hannover ICANS 严重程度评估——一种评估 CAR-T 细胞相关神经毒性的新型床旁检测

Chimeric antigen receptor (CAR) T-cell therapy has transformed the treatment of refractory hematological malignancies but is frequently complicated by immune effector cell-associated neurotoxicity syndrome (ICANS). Early clinical recognition remains challenging, as the commonly used Immune Effector Cell-Associated Encephalopathy (ICE) score lacks sensitivity for subtle deficits.

In this prospective bicentric study, 100 patients treated with CAR T-cells at Hannover Medical School and Charit - Universit tsmedizin Berlin underwent systematic neurological assessments using both ICE and the newly developed Berlin-Hannover ICANS Severity Assessment (BHISA). Examinations were performed at baseline prior to CAR T-cell infusion, on day 6-7 ( 1 day) post-infusion, and during ICANS episodes. Data on the clinical course, other toxicities, comorbidities, CAR T-cell products, and ICANS treatment were collected.

Thirty-seven patients (37%) developed ICANS, which was associated with preceding cytokine release syndrome and specific CAR T-cell products. While ICE scores clustered at maximum values both at baseline and follow-up, BHISA showed a broader distribution and higher sensitivity to subtle changes. Correlation analyses confirmed agreement between ICE and BHISA, but BHISA captured early cognitive decline more reliably. Receiver operating characteristic analyses demonstrated comparable diagnostic accuracy (BHISA: AUC = 0.783, ICE: AUC = 0,777), with consistently higher sensitivity of BHISA at matched specificity. (Specificity target = 0.7, BHISA sensitivity = 0.743, ICE sensitivity = 0.571; Specificity target = 0.8, BHISA sensitivity = 0.629, ICE sensitivity = 0.571).

BHISA may provide a more sensitive and more differentiated screening tool for ICANS than ICE by incorporating additional cognitive and motor domains, while remaining easy to use. This may enable earlier and more nuanced detection of CAR T related neurotoxicity, potentially improving patient monitoring across a heterogeneous population.