Interim assessment by circulating tumor DNA in primary mediastinal large B-cell lymphoma: a multicenter LYSA study.

Primary mediastinal large B-cell lymphoma (PMBL) achieves excellent outcomes with dose-dense immunochemotherapy, yet response assessment by positron emission tomography (PET) remains limited. In this prospective multicenter observational study, we evaluated the clinical relevance of circulating tumor DNA (ctDNA) minimal residual disease (MRD) in patients with newly diagnosed PMBL and assessed whether MRD enhances outcome discrimination beyond PET. Plasma and PET images were collected at baseline and after 2 and 4 cycles. ctDNA and tumor biopsy were analyzed by high-depth, error-corrected sequencing (limit of detection ∼10-3). Associations between MRD, PET response, and progression-free survival (PFS) were evaluated. Among 84 patients, baseline ctDNA was detected in 98%. After 4 cycles of treatment with R-CHOP14 (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone, administered every 14 days) or R-ACVBP (rituximab, doxorubicin, cyclophosphamide, vindesine, bleomycin, prednisone), 87.7% had undetectable MRD. Persistence of minimal residual disease after 4 cycles of therapy (MRD4) was associated with shorter PFS (hazard ratio [HR], 78.1; 95% confidence interval [CI], 9.5-641.8). The 1-year PFS was 98.4% (95% CI, 95.4%-100%) for patients with undetectable MRD4 vs 33.3% (95% CI, 13.2%-84.0%) for patients with detectable MRD4 (P < 10^-4). MRD4 showed a higher positive predictive value (89% vs 50%) for disease progression than PET4, (PET after 4 cycles of chemotherapy) maintaining a similar negative predictive value (100% vs 93%). In multivariate analysis, only PET4-/MRD4- remained associated with PFS (adjusted HR, 0.07; 95% CI, 0.01-0.90). Plasma ctDNA represents a highly abundant source of genetic markers for tumor fingerprinting and disease monitoring. MRD detection after frontline therapy strongly complements PET response criteria in predicting outcome. These findings support the use of ctDNA monitoring as a valuable tool for future risk-adapted strategies in PMBL. This trial was registered at www.clinicaltrials.gov as NCT04980222.