B11 | TREATMENT FREE REMISSION IN CHRONIC PHASE-CHRONIC MYELOID LEUKEMIA: ANALYSIS OF PREDICTIVE FACTORS AND NOVEL SCORING SYSTEMS FOR MOLECULAR RECURRENCE

BACKGROUD: Treatment free remission (TFR) is a key treatment goal in chronic-phase chronic myeloid leukemia (CP-CML). Identifying predictors for TFR success is crucial for better patient selection in real-world. METHODS: Our single-center retrospective study, on patients with CP-CML who discontinued TKIs from 2009 to 2024, aimed to evaluate the rate and duration of successful TFR in real-life clinical practice and to identify factors predicting molecular recurrence (MRec). Eligibility required prior TKI treatment ≥ 3 years, a minimum molecular response (MR) of MR4.0 (IS) before discontinuation, and subsequent serial MR monitoring. MR was assessed using RT-qPCR at predefined intervals. MRec was defined as loss of major molecular response (MMR). RESULTS: Of the 118 consecutive CP-CML patients, 60.2% were on imatinib and 39.8% were on 2G-TKIs. Median TKI treatment duration was 10.6 years, with a median stable DMR duration of 6.1 years. After a median follow-up of 72.5 months, the cumulative incidence of MRec was 20.3% at 6 months and 28.6% at 48 months. The estimated MRec-free survival was 79.7% at 6 months, 71.2% at 48 months, and 69.9% for the entire follow-up. Multivariate analysis identified three factors associated with shorter TFR duration: high-risk Sokal score (HR 2.93, 95% CI, 1.20-7.12, p=0.018), stable DMR duration before TKI suspension < 5 years (HR 3.63, 95% CI, 1.60-8.23, p=0.002), prior unstable DMR history (HR 2.47, 95% CI, 1.16-5.23, p=0.019). A score called BASE-TFR was developed by assigning one point for each of these factors, and patients were classified into low-risk (0), intermediate-risk (1) HR 4.19 (95% CI, 1.43-12.29, p=0.009) and high-risk (≥2) HR 14.06 (95% CI, 5.18-38.18, p<0.001) for MRec (Fig. A). Univariate analysis showed that a one-month post-discontinuation MR > 0.0032% (IS) (HR 5.82, 95% CI, 2.80-12.10, p<0.001) and a difference of ≥ 0.0017% between the last available MR before TKI-stop and first-month MR (HR 9.66, 95% CI 4.62-20.20, p<0.001) were associated with a shorter TFR period. A score called F1RST-TFR was developed by assigning one point for each of these parameters, and patients were classified into low-risk (0), intermediate-risk (1) HR 4.11 (95% CI, 1.49-11.31, p=0.006) and high-risk (2) HR 8.90 (95% CI, 4.08-19.38, p<0.001) for MRec (Figure B). CONCLUSION: TFR is an achievable goal in real-life. Predictive scores like BASE-TFR may help to better identify candidate for TKI discontinuation in real-life settings.