A model for predicting effect of treatment on progression-free survival using MRD as a surrogate end point in CLL

Our objective was to evaluate minimal residual disease (MRD) at the end of induction treatment with chemoimmunotherapy as a surrogate endpoint for progression-free survival (PFS) in chronic lymphocytic leukemia (CLL) based on three randomized, phase 3 clinical trials (ClinicalTrials.gov identifiers NCT00281918, NCT00769522, NCT01010061). MRD was measured in peripheral blood (PB) from treatment-naive patients in the CLL8, CLL10, and CLL11 clinical trials, and quantified by four-color flow cytometry or allele-specific oligonucleotide real-time quantitative polymerase chain reaction. A meta-regression model was developed to predict treatment effect on PFS using treatment effect on PB-MRD. PB-MRD levels were measured in 393, 337, and 474 patients from CLL8, CLL10, and CLL11, respectively. The model demonstrated a statistically significant relationship between treatment effect on PB-MRD and treatment effect on PFS. As the difference between treatment arms in PB-MRD response rates increased, a reduction in the risk of progression or death was observed; for each unit increase in the (log) ratio of MRD-negative rates between arms, the log of PFS hazard ratio (HR) decreased by -0.188 (95% confidence interval, -0.321 to -0.055; P = .008). External model validation on the REACH trial and sensitivity analyses confirm the robustness and applicability of the surrogacy model. Our surrogacy model supports use of PB-MRD as a primary endpoint in randomized clinical trials of chemoimmunotherapy in CLL. Additional CLL trial data are required to establish a more precise quantitative relationship between MRD and PFS, and to support general applicability of MRD surrogacy for PFS across diverse patient characteristics, treatment regimens, and different treatment mechanisms of action.