EASIX 在 CAR-T 细胞治疗中未能在乳酸脱氢酶和 ECOG 体能状态之外增加预后价值：一项 GETH-TC 研究

The Endothelial Activation and Stress Index (EASIX) has been proposed as a predictor of endothelial complications such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) in patients undergoing chimeric antigen receptor (CAR) T-cell therapy.

However, its prognostic role for long-term survival after commercial anti-CD19 CAR T-cell therapy remains uncertain. To evaluate the prognostic value of EASIX in predicting survival outcomes and toxicity in patients treated with CAR T-cell therapy, and to compare its performance with commonly available clinical biomarkers. This retrospective multicenter study included 126 patients with aggressive B-cell lymphomas treated with commercially available CAR T-cell products across multiple centers in Spain. EASIX-d0 was calculated prior to CAR-T infusion (EASIX-d0).

Cox proportional hazards models assessed associations with overall survival (OS) and progression-free survival (PFS), while logistic regression was used for toxicity outcomes. Receiver operating characteristic (ROC) analysis compared the predictive performance of EASIX-d0 versus LDH.

A combined LDH-ECOG PS status risk model was also evaluated. Higher EASIX-d0 values were associated with inferior OS (HR: 1. 52, P < . 001) and PFS (HR: 1. 30, P < . 001). Patients in the highest EASIX-d0 quartile showed significantly worse OS (HR: 4. 40, P = . 002) and PFS (HR: 2. 50, P = . 03).

However, predictive performance did not differ between EASIX-d0 and LDH alone for OS (AUC 71. 6% vs 71. 3%, P = . 935) or PFS (68. 5% vs 66. 6%, P = . 618). A combined LDH-ECOG model identified 3 risk groups with superior discrimination of OS and PFS compared to EASIX-d0. EASIX-d0 was associated with ICANS grades 2 to 4 and 3 to 4 (OR 1. 66, P = . 001; OR 2. 10, P = . 001), but showed no association with CRS or nonrelapse mortality.

While EASIX-d0 predicts survival and ICANS in CAR T-cell recipients, however its predictive capacity was largely driven by LDH. it does not outperform more accessible markers such as LDH and ECOG PS. While recent studies suggested associations between EASIX and overall survival, our results highlight that simpler parameters such as LDH and ECOG Performance Status may provide equal or superior predictive power in real-world cohorts.