Bone marrow blasts- and modified EASIX-guided risk stratification for coagulopathy and outcomes after CAR-T therapy in relapsed/refractory B-ALL.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND: Chimeric antigen receptor (CAR) T-cell therapy has revolutionized the treatment of relapsed/refractory (r/r) B-cell acute lymphoblastic leukemia (B-ALL). While CAR-T-associated coagulopathy (CARAC) remains a critical complication, significantly increasing the risk of hemorrhage and disseminated intravascular coagulation (DIC). AIMS: To develop an effective CARAC risk stratification and outcome prediction model. METHODS: This multicenter retrospective study enrolled r/r B-ALL patients who received CD19 CAR-T therapy between January 2016 and July 2025. Machine learning, logistic regression and cutoff values were utilized to select key variables and develop predictive models. Model performance and clinical applicability were evaluated by receiver operating characteristic, calibration, and clinical decision curves. Survival analyses evaluated the impact of CARAC severity on overall survival (OS) and progression-free survival (PFS) and to validate the prognostic value of the prediction model. RESULTS: Bone marrow (BM) blasts and the modified endothelial activation and stress index (mEASIX) were independent predictors for CARAC. Patients with mEASIX<4.0 were identified as low-risk CARAC. For patients with mEASIX≥4.0, CARAC was stratified by BM blast percentage (blast%). Blast% between 10 % and 44 % indicated high-risk CARAC-nonDIC, and a blast% ≥ 44 % indicated high-risk CARAC-DIC. Moreover, compared to non-CARAC patients, CARAC patients had inferior OS (hazard ratio [HR]: 2.62, p = 0.002) and PFS (HR: 2.11, p = 0.002). The stratified prediction model revealed progressively worse OS and PFS from low-risk CARAC patients to high-risk CARAC-nonDIC, with high-risk CARAC-DIC patients demonstrating the most unfavorable outcomes. CONCLUSIONS: An effective CARAC risk prediction and stratification model was established. High-risk CARAC patients, particularly CARAC-DIC, were associated with significantly worse outcomes.