IL-6 and IL-2 Plasma Levels at the Time of Discontinuation Identify CML Patients Who Fail to Sustain Treatment-Free Remission

VASCONCELOS CORDOBA, BIANCA; SÁNCHEZ, MARÍA BELÉN; PAVLOVSKY, CAROLINA; MOIRAGHI, BEATRIZ ; VARELA, ANA INÉS; CUSTIDIANO, MA DEL ROSARIO; FERNÁNDEZ, ISOLDA; FREITAS, JOSEFINA; VENTRIGLIA, VERÓNICA; BENDEK, CAROLINA; OSORIO, M JOSÉ; MORDOH, JOSÉ; SÁNCHEZ AVALOS, JULIO CÉSAR; LEVY, ESTRELLA MARIEL; CUETO, GERARDO RUBÉN; BIANCHINI, MICHELE

Congreso; 24° ESH - iCMLf International Conference on CML - Biology and Therapy; 2022

European School of Haematology (ESH)

Introduction: Recent studies suggest that a proportion of chronic myeloid leukemia (CML) patients in deep molecular remission can discontinue the tyrosine kinase inhibitor treatment without disease relapse. However, there is still need for better prediction tools for choosing the right patients for stopping attempts. Easily accessible biomarkers such as plasma proteins would be a valuable complement for guiding decision-making in the treatment-free remission (TFR)scenario. This study aimed to identify plasma cytokine biomarkers in CML patients at discontinuation to predict subsequent TFR failure.Methods: This study was conducted as an exploratory sub-study of the Argentinean Stop Trial (AST), which recruited 46 CML patients in chronic phase. Plasma samples were taken at the time ofdiscontinuation and cryopreserved at -80°C until analysis with Luminex MAGPIX® platform. The levels of 20 cytokines were measured in duplicate plasma samples using three different panels(Merck-Millipore): EOTAXIN/CCL11, GMCSF, IFNa2, IL-1a, IL-1b, IL1RA, IL-2, IL-4, IL-6, IL-7, IL-8/CXCL8, IL-9, IL-15, MCP-1/CCL2, TGFa, LIF, SCF, TGFb1, TGFb2, TGFb3. For statistical analysis, nonparametric Mann–Whitney test was used for comparing differences between molecular relapsed (R) vs. no-relapsed (NR) patients. Molecular recurrence-free survival was estimated by the Kaplan–Meier method and compared within groups by the log-rank test. Each cytokine was dichotomized according to receiver operating characteristics curves to optimize their cutoff points. Logistic regression was used to create predictive models, considering molecular relapse as the outcome variable. Differences were considered statistically significant when p