Meta-analysis of survival by phased-variant ctDNA and PET response in large B-cell lymphoma

Positron emission tomography (PET) is used to assess response in large B cell lymphoma (LBCL) but 20% to 30% of responding patients will relapse. Circulating tumor DNA detected by phased-variants is a highly sensitive means of identifying measurable residual disease (MRD) which may improve PET response assessment. We conducted a meta-analysis of end of treatment MRD and PET in LBCL treated with curative intent frontline therapy. Three studies and 367 patients were included. In the pooled analysis, the hazard ratio (HR) for progression-free survival (PFS) in undetectable vs. detectable MRD was 14.02 (95% CI, 7.35 - 26.74) and 5.09 (95% CI, 3.29 - 7.88) in PET complete metabolic response (CMR) vs. no CMR (p=0.01 for group difference). Two-year PFS rates were 95.7% (95%, CI 89.7% - 98.3%) with undetectable MRD and CMR, 91.3% (95% CI, 84.5% - 95.3%) in undetectable MRD and no CMR, 40.2% (95% CI, 20.9% - 63.2%) in detectable MRD and CMR and 23.4% (95% CI, 11.7% - 41.3%) in detectable MRD and no CMR (p<0.0001 for group differences). Phased-variant detected MRD provides clinically meaningful prognostic information when assessed at end of frontline curative intent chemoimmunotherapy in patients with LBCL independent of PET response assessment.