M-Easix (Better Than EASIX) Correlates with Specific Endotheliopathy Biomarkers, Predicts Severe CAR-T Cell Toxicities and Discriminates from Sepsis

Background Chimeric antigen receptor (CAR)-T cell-based immunotherapy has become a groundbreaking advance for relapsed/refractory hematologic malignancies' treatment. However, cytokine release (CRS) and immune effector cell-associated neurotoxicity (ICANS) syndromes are highly incident and life-threatening toxicities, in which endothelial damage plays a fundamental role in their pathophysiology. Higher EASIX (Endothelial Activation Stress Index) and modified-EASIX ((m-EASIX), which substitutes creatinine for C reactive protein (CRP)) values calculated peri- CAR T-cell infusion have been correlated with increased risk for CRS and ICANS. The present study aims to assess the applicability of EASIX and m-EASIX to predict severe toxicities and to discriminate from sepsis, which clinically might have overlapping symptoms. Methods Adults with relapsed/refractory hematological malignancies (CD19 positive or multiple myeloma), admitted to our center to receive immunotherapy with CAR-T cells with any construct available, were included prospectively from 2018 to 2024. Blood samples were collected: (A) before the CAR-T cell infusion (day zero); (B), 24-48h after; (C) at the suspicion of CRS; (D) 24-48h after immunomodulatory treatment for CRS; (E) at the suspicion of ICANS and (F) 24-48h after corticosteroid treatment for ICANS. The results of EASIX and m-EASIX were compared with those in septic patients from an independent cohort. Correlations between EASIX, m-EASIX and levels of biomarkers of endothelial dysfunction (sVCAM-1, sTNFRI, TM, ST2, Ang-2), innate immunity activation (NETs, sC5b-9) and hemostasis/fibrinolysis (VWF:Ag, ADAMTS-13, alpha-2-antiplasmin, PAI-1:Ag), that were previously measured in 62 of the patients, were assessed. Results One-hundred and ten patients treated with academic or commercially available CD19 or BCMA-targeted CAR T-cell product for hematological malignancies (n=87 with B-cell malignancies, n=23 with multiple myeloma) were included. Patients who developed CRS (n=64) presented a significant increase of m-EASIX at its clinical onset vs.post-infusional values (median ± interquartile range (IQR) 12.7 ± 29.7 vs. 2.8 ±7 respectively, p<0.0001) and a significant decrease four days after treatment with tocilizumab ± corticosteroids (3.9 ± 13.33, p<0.0001). Interestingly, m-EASIX presented a non-significant decrease in patients who developed ICANS (n=20, 19 of them after CRS) (point E) with respect CRS values (point C) (7.6 ± 15.06 vs.13.5±30.12, p=0.089, but a significant one after corticosteroid treatment (1.25 ±1.42 at point F with respect point E, p<0.0001). EASIX, and m-EASIX especially, showed a significant positive correlation with Ang-2 and ST2, both biomarkers of endothelial dysfunction (Correlation of m-EASIX with Ang-2 and ST2, r Spearman of 0.5, p=0.012 and r=0.73, p<0.001, respectively). m-EASIX at points A and B showed a good discriminative power to discern which patients would develop grade >3 CRS or ICANS (values of area under the curve (AUC) in a ROC analysis of 72 and 74%, respectively) or need for intensive care unit (ICU) admission (AUC of 68% and 74%, respectively). High CRP levels and low platelet counts were the variables that had individual impact on the development of severe toxicities needing ICU admission (CRP and platelet count, both at point B, with OR 1.4, p=0.004 and OR 0.98, p= 0.012, respectively). Moreover, m-EASIX was significantly higher in septic patients than in CAR T cell treated patients at the CRS onset (64.87 ± 157.95 in sepsis vs. 11.53 +/- 28.42 at point C, respectively, p<0.001) and demonstrated to discriminate between these two syndromes (AUC for the classification of sepsis of 75%), even when considering only severe ICU patients in both groups (APACHE-II >12 points), AUC 61%. Conclusions M-EASIX arises as a feasible tool that might help to predict CAR T cell toxicities at early stages of immunotherapy, severity classification and differential diagnosis with sepsis. Additionally, it could be considered a complement to assess endotheliopathy at different time-points, as it shows an acceptable correlation with specific circulating biomarkers.