Circulating tumor DNA guided de-escalation of frontline diffuse large B-cell lymphoma therapy

PROJECT SUMMARY The treatment paradigm for newly diagnosed diffuse large B-cell lymphoma (DLBCL) has largely remained unchanged: a uniformly applied regimen of 6 cycles of rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). A newer regimen for intermediate or high-risk patients replacing vincristine with polatuzumab vedotin (R-pola-CHP) has recently shown improved efficacy. The randomized FLYER study demonstrated that shortening frontline therapy for patients younger than 60 with limited-stage, low- risk DLBCL from 6 cycles of R-CHOP to 4 cycles of R-CHOP followed by 2 cycles of rituximab alone is possible with equal efficacy and decreased toxicity. However, eligibility for this approach is based on a pre-treatment risk assessment applicable to a minority of patients. Stratifying a patient with advanced stage DLBCL as low risk by their interim response could be a new way to identify patients for treatment de-escalation. Fragments of circulating tumor DNA (ctDNA) containing cancer-specific sequences can be isolated from patient peripheral blood. Next generation sequencing-based assays that detect ctDNA have shown promise as “liquid biopsies” in DLBCL, allowing for molecular response assessment and detection of minimal residual disease (MRD). However, this technology has not yet been used in real-time. All studies on ctDNA in DLBCL have been retrospective or non-interventional to this date, involving sequencing of archival samples without clinical integration. Phased variant enrichment and detection sequencing (PhasED-seq), which involves the detection of fragments of ctDNA harboring multiple somatic mutations, boasts the lowest limit of detection of available techniques. In retrospective and prospective analyses, patients who were MRD negative by PhaED- seq after 2-3 cycles of frontline DLBCL therapy experienced long term remissions in >90% of cases. Patients with advanced stage DLBCL that achieve interim MRD-negativity are likely overtreated with 6 full cycles of therapy, and a ctDNA-guided de-escalation approach could still lead to cure while sparing patients from unnecessary toxicity and treatment. Demonstrating the feasibility of real-time ctDNA sequencing during DLBCL therapy is needed before adopting said approach into routine clinical practice. Our central hypothesis is it is feasible to perform interim ctDNA sequencing during standard frontline DLBCL therapy to determine which patients should discontinue chemotherapy early. The proposed single-center pilot study will enroll 8 patients with newly diagnosed, stage II-IV DLBCL intended for 6 cycles of R-CHOP or R-pola-CHP. Patients MRD-negative by PhaED-seq after cycles 2 and 3 will de-escalate to rituximab alone for cycles 5-6, while all other patients will receive 6 full cycles. We will investigate the following aims: 1. Determine the feasibility of ctDNA assessment in real-time to guide decision making during frontline therapy for DLBCL, 2. Assess the outcomes of patients who become undetectable for ctDNA during frontline therapy for DLBCL and discontinue chemotherapy early, and 3. Evaluate the safety profile associated with a de-escalated course of frontline therapy for DLBCL.