Frontline chemo-free therapy for follicular lymphoma (FL) and marginal zone lymphoma (MZL) raises a sequencing question: combine active agents from the start, or begin with one and add treatment only when needed.
A phase II trial of response-adapted, chemotherapy-free treatment in untreated FL and MZL produced an overall response rate of 100%, a complete response rate of 86%, and two-year progression-free survival of 89% among 42 patients (Lynch et al., 2026). Most patients reached CR on mosunetuzumab alone, without intensification or maintenance. PET and circulating tumor DNA (ctDNA) are the candidate tools for identifying who needs more.
Background: Bispecific Antibodies Move Toward Frontline Use
Chemoimmunotherapy, combining rituximab or obinutuzumab with bendamustine or CHOP-based chemotherapy, remains the established frontline option when FL or MZL requires systemic treatment. These regimens control disease durably but carry short- and long-term cytotoxicity. In the phase III RELEVANCE trial, rituximab plus lenalidomide produced a three-year PFS of about 77%, comparable to chemoimmunotherapy, but did not meet its prespecified superiority endpoints.
CD20×CD3 bispecific antibodies are the next candidate. Mosunetuzumab and epcoritamab have shown substantial activity in relapsed or refractory FL, which supports testing them earlier. Their best frontline role is undefined: they could be combined with other agents from the outset, or given alone with treatment added after an insufficient response.
Study Design
The single-center, investigator-initiated, open-label trial ran at Fred Hutchinson Cancer Center and the University of Washington. It enrolled 42 untreated adults with FL grade 1–2 or 3A (37 patients) or MZL (5 patients) who had an indication for treatment. Median age was 60 years, 93% had stage III–IV disease, and 31% had bulky disease exceeding 7 cm.
All patients received eight 21-day cycles of subcutaneous mosunetuzumab with step-up dosing in cycle 1. PET-CT after cycle 8 determined the next step: patients in CR stopped treatment, and those in partial response were eligible for six cycles of polatuzumab vedotin plus obinutuzumab.
The primary endpoint was CR rate after all protocol therapy, with a prespecified threshold of at least 34 of 42 patients (80%). Exploratory ctDNA analysis used PhasED-Seq, which detects molecular residual disease through tumor-specific phased variants.
Results: CR After Mosunetuzumab Monotherapy
All 42 patients were evaluable. After eight cycles of mosunetuzumab, every patient had responded and 30 (71%) were in CR. These 30 received no further treatment.
Twelve patients had a PR and were offered intensification. Seven received polatuzumab vedotin plus obinutuzumab: six completed treatment and reached CR, and one stopped after a single cycle because of toxicity and remained in PR. Of the five who did not receive consolidation, two converted to CR without further therapy, one had histologic transformation, and two received no further study treatment after patient or physician decisions.
At the end of protocol therapy, 36 patients (86%) were in CR, meeting the primary endpoint. That figure combines patients treated with mosunetuzumab alone and those who received consolidation. Only seven patients received the polatuzumab vedotin-obinutuzumab combination, and the two agents were given together, so the study can’t establish the contribution of either.
PFS and Relapse Patterns
At a median follow-up of 34 months, two-year PFS was 89% and two-year overall survival was 100%. Four progression events occurred: two CD20-negative FL relapses and two biopsy-confirmed histologic transformations.
Mosunetuzumab and obinutuzumab both target CD20, so antigen loss is a plausible escape mechanism and could limit the value of subsequent CD20-directed therapy. PFS appeared similar across FL grade 1-2, FL grade 3A, and MZL, but subgroups were small, and MZL included only five patients.
ctDNA: Exploratory Data on Interpreting PET
All evaluable patients had falling ctDNA by the start of cycle 3, but depth of clearance differed by eventual response. After two cycles, 62% of patients who reached CR had undetectable ctDNA, whereas 83% of patients with PR still had detectable ctDNA.
At the end of mosunetuzumab, 93% of evaluable patients in CR had undetectable ctDNA. Among patients with PR, 67% had persistent ctDNA. Four PR patients had rising ctDNA at the end of treatment; all four reached CR with polatuzumab vedotin and obinutuzumab, with undetectable ctDNA.
The reverse pattern also occurred. Some patients with PET-positive residual disease and undetectable ctDNA later reached CR without further therapy. FDG uptake after T-cell-engaging therapy could reflect immune activity instead of viable lymphoma, as described after CAR T-cell therapy and checkpoint inhibition. Escalating on PET positivity alone could expose such patients to additional treatment during an evolving response.
Safety
Cytokine release syndrome occurred in 64% of patients, all grade 1 and all resolved. No patient had grade 2 or higher CRS, needed tocilizumab, or developed immune effector cell-associated neurotoxicity syndrome.
Grade 3 neutropenia occurred in 14% and grade 3 thrombocytopenia in 5%. Treatment was delayed in 33%, and two patients discontinued because of toxicity. Infections occurred in 57%, most often upper respiratory. No deaths occurred during treatment or follow-up.
Frontline Strategy: When to Escalate
Many ongoing frontline bispecific trials test combinations or extended schedules, which may deepen responses in some patients but add exposure, toxicity, and cost. A response-adapted design starts with a single active agent and reserves intensification for insufficient responders.
Timing and criteria for escalation remain unresolved. The PET used to assign consolidation came after cycle 8, which may have missed earlier CRs. PET combined with ctDNA could separate persistent lymphoma from delayed metabolic resolution.
Durability: Could Short Therapy Produce Long Remissions?
FL is generally considered incurable, but long-term data show exceptionally durable remissions in some patients after frontline treatment. The authors cite 15-year follow-up from SWOG S0016, where the PFS curve plateaued at about 40%. Patients with early PET CR and molecular clearance after mosunetuzumab may form such a subset.
Follow-up of 34 months can’t establish cure or exclude late toxicity, since relapse, transformation, and prolonged immune suppression in indolent lymphoma can emerge years after treatment. SWOG S2308 compares mosunetuzumab with rituximab in low-tumor-burden FL, a different population. Other trials are testing bispecific combinations and response-adapted strategies in patients who require treatment.
Limitations and Clinical Implications
The single-arm design precludes comparison with chemoimmunotherapy or other chemotherapy-free regimens. Nineteen percent of patients were treated for progressive or symptomatic disease that did not strictly meet the Groupe d’Étude des Lymphomes Folliculaires (GELF) criteria used in some historical frontline trials, which further weakens cross-trial comparison.
Only four progression events occurred, and escalation was not randomized. The results support further study of finite mosunetuzumab-based therapy in untreated FL and MZL, particularly for patients who want to avoid chemotherapy.
