Background
Prospective, multicenter, non-randomized cohort study (EURO-SKI, EUROpean Stop Kinase Inhibitor study). N=755 patients with CP-CML in sustained MR4 (BCR-ABL IS ≤0.01%) for ≥1 year on any TKI (imatinib, nilotinib, or dasatinib). The largest and most rigorous TFR study to date, designed to validate TKI discontinuation across all TKI classes and identify predictors of successful treatment-free remission. Published Lancet Oncology 2018. Builds on STIM (Mahon 2010) with larger N, any-TKI eligibility, and formal multivariate predictor analysis.
Interventions and follow up
Regimen: TKI discontinuation (imatinib, nilotinib, or dasatinib) in patients with sustained MR4 ≥1 year
Monitoring: Molecular monitoring monthly for 6 months, then every 6 weeks; restart TKI on loss of MMR (BCR-ABL IS >0.1%)
Primary endpoint: MR4 at 6 months (maintenance of BCR-ABL IS ≤0.01% without TKI)
Median follow-up: 27.9 months
Monitoring: Molecular monitoring monthly for 6 months, then every 6 weeks; restart TKI on loss of MMR (BCR-ABL IS >0.1%)
Primary endpoint: MR4 at 6 months (maintenance of BCR-ABL IS ≤0.01% without TKI)
Median follow-up: 27.9 months
Results
TFR (MR4 maintained at 6 months): 62% (468/755)
TFR at 24 months: 50.3%
Molecular relapse (loss of MR4): 38% by 6 months; nearly all within the first 6 months
Re-achievement of MR4 after TKI restart: >95% of relapsers
Predictor of TFR (multivariate): TKI treatment duration ≥5.8 years, strongest predictor (P<.001); longer MR4 duration before stopping
TFR by TKI type: Similar for imatinib, nilotinib, dasatinib, no significant difference
TFR at 24 months: 50.3%
Molecular relapse (loss of MR4): 38% by 6 months; nearly all within the first 6 months
Re-achievement of MR4 after TKI restart: >95% of relapsers
Predictor of TFR (multivariate): TKI treatment duration ≥5.8 years, strongest predictor (P<.001); longer MR4 duration before stopping
TFR by TKI type: Similar for imatinib, nilotinib, dasatinib, no significant difference
Adverse events
Withdrawal syndrome: Musculoskeletal pain, myalgia, arthralgia any grade 26%, most resolving within 6–12 months; grade ≥3 withdrawal symptoms 2%
Disease-related: No CML progression to accelerated or blast phase in any patient; no deaths attributed to the TFR attempt; rare fatigue and weight gain after stopping
Disease-related: No CML progression to accelerated or blast phase in any patient; no deaths attributed to the TFR attempt; rare fatigue and weight gain after stopping
Conclusions
EURO-SKI confirmed that approximately 50% of CML patients in sustained MR4 on any TKI can maintain treatment-free remission for ≥24 months after stopping, with all molecular relapsers safely restarting TKI. Longer TKI duration (≥5.8 years) is the strongest independent predictor of TFR success, providing an actionable recommendation for patient selection.
Key Limitations
Non-randomized observational cohort without a continued-TKI comparator; TFR rate is a single-arm estimate. Eligibility required sustained deep molecular response (MR4 ≥1 year), limiting generalizability to selected patients. Standardized, frequent molecular monitoring with rapid restart on loss of MMR is essential and not always feasible outside trials. The duration threshold (≥5.8 years) is data-derived and not prospectively validated as a cutoff. Late molecular relapses beyond the reported follow-up cannot be excluded; longer-term durability and quality-of-life data continue to mature.
Clinical Context
EURO-SKI is the definitive evidence base for treatment-free remission in CML, extending STIM's findings across TKI classes and quantifying predictors. It underpins current ELN recommendations that an attempt at TFR is appropriate in eligible patients with CP-CML in sustained deep molecular response (typically MR4 for ≥2 years on adequate-duration TKI therapy), with mandatory frequent molecular monitoring and prompt TKI restart on loss of MMR. ESMO-MCBS: not assigned.