Despite substantial progress in chronic myeloid leukaemia (CML) treatment, long-term survival and progression-free survival (PFS) remain influenced by baseline risk scores and early molecular response. Stratification by scores such as EUTOS, ELTS, and BCR::ABL1 levels may improve prediction of outcomes and guide treatment.

To assess overall survival (OS) and progression-free survival in a CML cohort, stratified by prognostic scores and molecular response, and to identify predictive markers associated with favourable outcomes.

Kaplan-Meier analysis was performed for OS and PFS, followed by log-rank tests for group comparisons. Stratification included EUTOS (>87), ELTS (>2.21), and BCR::ABL1 at baseline (>100%) and month 3 (<10%). Survival curves were interpreted alongside clinical variables and response kinetics.

The overall survival probability remained above 75% at the end of follow-up, with a gradual decline from month 20. Stratified analysis showed significantly worse OS for patients with EUTOS >87 (p = 0.014), while ELTS >2.21 showed a non-significant trend toward poorer survival (p = 0.084). Baseline BCR::ABL1 >100% was not associated with OS differences. For PFS, approximately half the cohort remained progression-free at the final follow-up. EUTOS >87 was significantly associated with lower PFS (p = 0.023), while ELTS >2.21 again showed a non-significant trend (p = 0.082). Baseline BCR::ABL1 >100% did not predict PFS (p = 0.693). In contrast, achieving molecular response <10% at month 3 (MR3) was significantly associated with superior PFS (p = 0.005).

The findings highlight the prognostic importance of the EUTOS score for both OS and PFS. While ELTS showed a consistent trend towards worse outcomes, it did not reach statistical significance. The most robust predictor of long-term disease control was early molecular response (MR3 <10%). Baseline BCR::ABL1 >100% did not predict survival outcomes, potentially due to early therapeutic adjustments in this subgroup.

In this real-world cohort, high EUTOS scores and failure to achieve MR3 <10% were key markers of poor prognosis. BCR::ABL1 baseline level was the strongest predictor of early molecular response, while early response itself emerged as a relevant factor for long-term disease control. The integration of molecular and clinical risk parameters remains essential for optimising CML management.