DNA-based MRD monitoring enhances risk stratification during TKI dose reduction in CML: Evidence from the clinical trial

Abstract Background: In chronic myeloid leukemia (CML), DNA-based measurable residual disease (MRD) analysis—either combined with RNA MRD (Machova Polakova et al. Leukemia 2020) or with characterization of DNA-positive cell subtypes (Pagani et al. Blood 2023)—has demonstrated predictive value for treatment-free remission (TFR). These strategies form the basis of the TFR traffic light stratification model, previously described in patients undergoing direct tyrosine kinase inhibitor (TKI) cessation. Aim: This study evaluated the applicability of the DNA-/RNA-based TFR traffic light model in a structured, two-step TKI dose de-escalation protocol prior to cessation, as implemented in the HALF trial (NCT04147533). Methods: Between 2020 and 2023, 95 of 207 patients enrolled in the HALF trial underwent genomic breakpoint characterization and optimization of BCR::ABL1 DNA digital PCR assays. To date, DNA and RNA MRD assessments were performed in 83 patients, generating 1,222 paired samples. MRD was monitored throughout the two de-escalation phases: (1) halving the TKI dose for six months, followed by (2) alternate-day dosing for an additional six months. Patients maintaining major molecular response (MMR) proceeded to TKI cessation. Results: At the end of phase 1, 83 patients were stratified as follows: 29 double-negative (green), 26 DNA⁺/RNA⁻ (yellow), and 28 double-positive (red). After phase 2, MRD status shifted to 31 green, 19 yellow, and 33 red. Only red-group patients relapsed during phase 2 (n = 7). By 24 months post-enrolment, MRD dynamics continued to evolve. In the yellow group, 11/19 progressed to double-positive (4 relapsed); 1 reverted to green. Among green patients, 6 converted to red (3 relapsed), and 2 to yellow. In the red group, 11 remained unchanged; 14 relapsed; and 1 improved to yellow. Stratification by MRD status at 12 months significantly predicted molecular recurrence-free survival (MRFS) at 18 months (i.e., 6 months post-TKI cessation): red group, 39% MRFS (HR: 9.71; 95% CI: 2.87–32.78; p = 0.00025); yellow group, 87% MRFS (HR: 1.68; 95% CI: 0.34–8.34; p = 0.56); green group, 90% MRFS. This stratification remained consistent at 36 months (i.e., 24 months post-TKI cessation): red group, 36% MRFS (HR: 8.10; 95% CI: 2.77–23.69; p = 0.00013); yellow group, 74% MRFS (HR: 2.12; 95% CI: 0.57–7.91; p = 0.26); green group, 87% MRFS. Conclusion: Persistent RNA positivity (double-positive MRD) is significantly associated with molecular relapse both during phase 2 of TKI dose reduction and after TKI cessation. DNA MRD analysis is essential to distinguish truly MRD-negative (green) patients from those with DNA⁺/RNA⁻ status (yellow), who carry an intermediate risk of relapse. Importantly, in yellow-group patients, the emergence of RNA expression during either phase of dose reduction indicates an increased risk of relapse. In such cases, we recommend returning to the previous TKI dose rather than proceeding to cessation. For green-group patients, the appearance of DNA positivity should prompt close monitoring, and any subsequent RNA positivity should be considered a warning sign to resume reduced-dose or full-dose TKI therapy. Overall, the two-step TKI de-escalation strategy, when combined with integrated DNA/RNA MRD monitoring in RNA-negative patients through regular follow-up, may reduce the risk of molecular relapse and improve patient selection for safe TKI discontinuation. Supported by Ministry of Health of the Czech Republic NU22-03-00136, MH CZ – DRO (IHBT 0002373), National Institute for Cancer Research Project (Programme EXCELES, ID Project No. LX22NPO5102) - Funded by the EuropeanUnion - Next Generation EU