Long-term 3-year follow-up of mosunetuzumab in relapsed or refractory follicular lymphoma after ≥2 prior therapies

Follicular lymphoma (FL), an indolent B-cell non-Hodgkin lymphoma (NHL), is the second most common lymphoma globally.^1,2 Although the overall survival (OS) in FL continues to improve because of improvements in diagnostic tools, supportive care, and novel therapies, many patients experience multiple relapses, decreasing progression-free survival (PFS) rates with successive lines of therapies, and face the risk of disease transformation to aggressive lymphoma.^3-7 Furthermore, clinical prognosis can vary, with a subset of patients experiencing progression of disease within 24 months of initial therapy (POD24), affecting OS.⁸ Treatment options are limited for late-line relapsed or refractory (R/R) FL, thus newer therapies are required to combat increasing refractoriness and to improve outcomes.

Mosunetuzumab is a first-in-class, humanized immunoglobulin G1–based CD20×CD3 T-cell engaging bispecific antibody that redirects T cells to engage and eliminate malignant B cells.^9-11 Mosunetuzumab has been granted conditional marketing authorization by the European Medicines Agency and accelerated approval by the US Food and Drug Administration for the treatment of R/R FL after ≥2 prior systemic therapies.^12,13 A phase 1/2 dose-escalation study (ClinicalTrials.gov identifier: NCT02500407) of mosunetuzumab established that increasing IV doses, using a step-up dosing schedule in cycle (C) 1, were well tolerated and induced durable complete responses (CRs) in patients with R/R NHL.^11,14,15 In the phase 2 portion of the study, fixed-duration mosunetuzumab demonstrated a favorable safety profile with a high rate of CRs in patients with R/R FL after ≥2 prior systemic therapies.¹⁶ At a median follow-up of 18.3 months, this study demonstrated an independent review committee (IRC)-assessed overall response rate (ORR) of 80%, and an IRC- and investigator (INV)-assessed CR rate of 60%, which was significantly higher than a historical control CR rate of 14% (P < .0001) with copanlisib at the prespecified primary analysis.^16,17 Here, we report 3-year follow-up data from the phase 2 portion of the NCT02500407 study in heavily pretreated patients with R/R FL, assessing the durability of response (DOR), along with additional exploratory analyses.

The study design has been described in detail previously.^11,16 Briefly, the NCT02500407 trial is an ongoing phase 1/2, pivotal, multicenter, open-label study in patients with R/R NHL.

This analysis focused on patients with R/R FL (grade 1-3a) after ≥2 prior therapies, who were treated with the approved dose of mosunetuzumab^12,13 (the recommended phase 2 dose).

Mosunetuzumab was administered for a fixed duration by IV infusion in 21-day cycles with step-up dosing in C1 (C1, day 1: 1 mg; C1, day 8: 2 mg; C1, day 15/C2, day 1: 60 mg; C3, day 1 and onward: 30 mg). Hospitalization for treatment was not required. Patients achieving CR by C8 completed therapy with no further treatment cycles; patients with a partial response (PR) or stable disease continued treatment at 30 mg for up to 9 further cycles (17 cycles in total), unless progressive disease (PD) or unacceptable toxicity occurred. Premedication with corticosteroids (IV dexamethasone, 20 mg; or IV methylprednisolone, 80 mg) was administered 1 hour before each dose of mosunetuzumab in C1 and C2 and was optional from C3 onward. Retreatment with mosunetuzumab (1/2/60/30 mg) was permitted for patients who relapsed after CR.

The study protocol was approved by institutional review boards at each center. The trial was performed in accordance with the Declaration of Helsinki, International Conference on Harmonization Guidelines for good clinical practice, and applicable laws and regulations. Written informed consent was obtained from all eligible patients.

The primary efficacy end point was IRC-determined CR rate (as best response), according to Cheson et al¹⁸ criteria. Safety and tolerability were assessed by describing the incidence and severity of adverse events (AEs), and by changes in laboratory parameters. Secondary efficacy end points were INV-assessed CR rate; and IRC- and INV-assessed ORR, DOR, duration of CR (DOCR), PFS, and OS. Updated efficacy outcomes were analyzed based on INV assessment after a median of 37.4 months follow-up.

Post-hoc analyses evaluated the time to next treatment (TTNT; defined as the time from the first dose of mosunetuzumab treatment [C1, day 1] to the start of new antilymphoma therapy or death from any cause). Additional exploratory analyses examined: (1) the clinical outcomes of patients with and without a CR at the end of treatment (EOT; defined as CR at the last assessment before EOT or within 30 days after EOT), (2) the association between efficacy outcomes and the time required to achieve a CR (early CR was defined as patients with first CR achieved by the first mandatory tumor assessment at 3 months [±0.5 months], and late CR was defined as patients with first CR achieved after the first mandatory tumor assessment at 3 months [±0.5 months]), (3) efficacy outcomes with mosunetuzumab compared with those with last prior therapy, (4) efficacy of mosunetuzumab retreatment, (5) efficacy according to mutational status, and (6) the relationship between cytokine release syndrome (CRS) and tumor response.

Computed tomography (CT) and positron emission tomography-CT (PET-CT) scans were performed at screening, at 6 weeks (optional), 3 months, and once every 3 months thereafter during treatment, as previously reported.¹⁶ During posttreatment follow-up, CT (with or without PET) scans were performed once every 3 months during the first 18 months, at 24 months, and then once every 12 months thereafter until disease progression, start of new anticancer therapy, or study discontinuation. All CRs were confirmed by PET and bone marrow examination (if bone marrow was involved at baseline).

AEs were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 4.0). CRS was graded using American Society for Transplantation and Cellular Therapy criteria.¹⁹ 

Mutation profiling was retrospectively performed by whole-exome sequencing to assess the activity of mosunetuzumab in patients with known prognostic variants.^20-24 Tumor biopsy samples (n = 51) were collected as archival or fresh specimens at screening, for baseline biomarker evaluation. Further details of the methodology of biomarker analyses are described in the supplemental Materials, available on the Blood website.

CD19⁺ B-cell exploratory analyses were performed to determine B-cell depletion and B-cell recovery levels in response to mosunetuzumab treatment. B-cell counts were quantified by a standard T- and B-lymphocyte and natural killer cell flow cytometry panel at LabCorp (LabCorp, Inc, Burlington, NC). The assay had a lower limit of quantitation of <5 cells per μL. B-cell counts were evaluated at C1 (pretreatment), C2, C4, C6, C8, and every 3 months during follow-up or until PD, or until the next lymphoma treatment. The biomarker-evaluable population for monitoring B-cell depletion and B-cell recovery included patients with a known pretreatment sample, at least 1 on-treatment sample, and at least 1 follow-up sample. The analyses were restricted to patients with a CR, because circulating tumor B cells may also be present in patients who had not maintained a CR. B-cell depletion and recovery were defined as 5 cells per μL (lower limit of quantification) and as 70 cells per μL (lower limit of normal), respectively. A time-to-event analysis was performed to assess time to B-cell recovery in patients with a CR who had an EOT sample and at least 1 follow-up sample.

The 95% confidence interval (CI) of CR- and ORR-based end points were calculated using the Clopper-Pearson method. The Kaplan-Meier (KM) method was used to estimate the medians and event-free rates at 30 months and 36 months for DOR (in responders and complete responders), DOCR, PFS, OS, and TTNT. The Brookmeyer-Crowley method was used to calculate 95% CIs for the medians of DOR, DOCR, PFS, and OS. The Greenwood formula was used to calculate 95% CIs for PFS and OS event-free rates. FACSDiva software (BD Biosciences, version 6.1.3) and SAS (version 9.4) were used for data analysis. The clinical cutoff date (CCOD) for this analysis was 2 May 2023.

Between 2 May 2019 and 25 September 2020, 90 patients with R/R FL were enrolled and received IV mosunetuzumab at the approved dose (intent-to-treat population, defined as all enrolled patients assigned to mosunetuzumab; supplemental Figure 1). As of the CCOD (2 May 2023), the median follow-up was 37.4 months (range, 2-48). All patients completed or discontinued from the initial treatment: 56 patients (62.2%) completed the initial treatment; 24 patients (26.7%) discontinued initial treatment because of PD; and other reasons for discontinuation included AEs (n = 4), physician decision (n = 3), use of another anticancer therapy (n = 2), and withdrawal by the patient (n = 1).

Five patients had received retreatment; 2 patients completed retreatment; 2 patients discontinued retreatment because of PD; and 1 patient was active on retreatment. Pretreatment biopsies for all 5 patients who received retreatment were CD20 positive. At the CCOD, 68 patients (75.6%) remained in the study; 15 deaths (16.7%) were reported.

Patient baseline characteristics in the overall population have previously been published.¹⁶ Briefly, the median age was 60 years (range, 29-90), 28 patients (31.1%) were aged >65 years; 55 patients (61.1%) were male, 69 patients (76.7%) had Ann Arbor stage III/IV disease, and 31 patients (34.4%) had bulky disease (Table 1). The median number of prior lines of therapy was 3 (range, 2-10). All patients had received anti-CD20 therapies and alkylating agents before study entry, including 59 patients (65.6%) who had received prior bendamustine therapy. In total, 62 patients (68.9%) were refractory to their last prior therapy (Table 1), 71 (78.9%) were refractory to any prior anti-CD20 therapy, and 48 (53.3%) were refractory to both prior anti-CD20 therapy and a prior alkylating agent. Overall, 47 patients (52.2%) had a history of POD24.

As of the CCOD, the INV-assessed ORR and CR rates were 77.8% (95% CI, 67.8-85.9) and 60.0% (95% CI, 49.1-70.2), respectively. Median time to first response was 1.4 months (range, 1-11) and time to first CR was 3.0 months (range, 1-19; Table 2). Among patients with a response (n = 70), the KM-estimated median DOR was 35.9 months (95% CI, 20.7 to not estimable [NE]; Figure 1A) but was based on 1 event that occurred in the last patient at risk; the KM-estimated 30-month remission rate was 56.6% (95% CI, 44.2-68.9).

Among patients who achieved CR (n = 54), 49 remained in CR at the EOT; median DOCR was not reached (NR; 95% CI, 33.0 to NE; Table 2; Figure 1B). The KM-estimated 30-month remission rate in the 54 patients with a CR was 72.4% (95% CI, 59.2-85.6).

Among all 90 patients, median INV-assessed PFS was 24.0 months (95% CI, 12.0 to NE) and the KM-estimated 36-month PFS rate was 43.2% (95% CI, 31.3-55.2; Table 2; Figure 1C). Median TTNT was 37.3 months (95% CI, 18.0 to NE) and median OS was NR (95% CI, NE to NE; Table 2; Figure 1D). The estimated 36-month TTNT and OS rates were 51.8% (95% CI, 40.8-62.8) and 82.4% (95% CI, 73.8-91.0), respectively.

Five patients received mosunetuzumab retreatment after disease relapse after CR. Of these 5 patients, 2 had stable disease and 3 achieved a second CR as best response. At the CCOD, 2 of these 3 patients remained in CR, with ongoing responses at 20.6 and 8.8 months after initiation of retreatment, respectively; the third patient had disease progression 7.7 months after achieving the second CR during the retreatment period. One patient remained active on retreatment at the CCOD.

Of 60.0% (n/N, 54/90) of patients who achieved a CR as best response, 90.7% (n/N, 49/54) maintained a CR until the EOT, 1.9% of patients (n/N, 1/54) with a CR experienced PD in C8, and 7.4% of patients (n/N, 4/54) achieved a CR after EOT because of delayed bone marrow confirmation. Demographic and baseline disease characteristics in the subgroups of patients with (n = 49) or without (n = 41) a CR at EOT were broadly similar to those in the overall patient population (Table 1). Of note, 55.1% of patients with a CR at EOT were refractory to their last prior therapy compared with 85.4% without a CR at EOT. In patients with a CR at EOT, 40.8% were double refractory to prior anti-CD20 and alkylator therapy, compared with 68.3% without a CR at EOT. In the 49 patients with a CR at EOT, the median DOCR was NR (95% CI, 33.0 to NE; Figure 2A); 30-month DOCR event-free rate was 74.1% (95% CI, 60.6-87.6). The median PFS in patients with a CR at EOT was 37.3 months (95% CI, NE to NE; Figure 2B) based on 1 event in the last patient at risk; 36-month PFS event-free rate was 68.6% (95% CI, 52.2-84.9). The median OS in patients with a CR at EOT was NR (Figure 2C) and the 36-month OS event-free rate was 94.5% (95% CI, 86.9-100.0).

Of 54 patients that achieved a CR, 33 patients achieved an early CR (median DOR was NR [95% CI, 34.5 to NE]) and 21 patients achieved a late CR (median DOR was 35.9 months [95% CI, 22.8 to NE]; supplemental Table 1). Sixteen patients (17.8%) achieved a PR as best response (median DOR was 4.0 months [95% CI, 2.5-6.7]). PFS and OS outcomes in patients with an early or late CR, or PR as best response are summarized in supplemental Table 1. DOR was demonstrated for complete responders regardless of the timing when CR was achieved; in contrast, PRs were less durable with no PRs lasting >12 months.

All patients received alkylator and anti-CD20 therapy, and the majority received anthracyclines (82.2%) and bendamustine (65.6%) as prior cancer therapies (Table 1). The majority of patients received chemoimmunotherapy (63.3%) as their last prior therapy (supplemental Table 2). The estimated median time from the end of last prior therapy to first dose of mosunetuzumab was 6.7 months (range, 1-89). Efficacy outcomes of patients after receiving their last prior therapy are summarized in supplemental Table 3. Of note, the CR rate and ORR with mosunetuzumab were greater than those reported with the last prior therapy (CR: 60.0% vs 35.6%; ORR: 77.8% vs 55.6%). Median DOCR was NR (95% CI, 33.0 to NE) with mosunetuzumab compared with 15.1 months (95% CI, 11.2-26.3) with last prior therapy (supplemental Table 3; supplemental Figure 2A). The median PFS was 24.0 months (95% CI, 12.0 to NE) with mosunetuzumab and 12.1 months (95% CI, 10.3-16.1) with last prior therapy (supplemental Table 3; supplemental Figure 2B). Median TTNT was longer with mosunetuzumab vs last prior therapy (37.3 vs 16.8 months; supplemental Table 3).

Single-nucleotide variants were identified at baseline in tumor samples at a similar frequency to reported prevalence rates,²⁴ with the most common mutations found in histone-lysine N-methyltransferase 2D (KMT2D: 31/51; 60.8%), cyclic adenosine monophosphate response element–binding protein (CREBBP: 26/51; 51.0%), B-cell lymphoma 2 (BCL2: 17/51; 33.3%), and tumor necrosis factor receptor superfamily 14 (TNFRSF14: 12/51; 23.5%). Less frequent mutations were found in tumor protein P53 (TP53: 10/51; 19.6%) and enhancer of zeste homolog 2 (EZH2: 8/51; 15.7%). Clinically meaningful outcomes were observed in patients with common mutations, and none of the alterations were significantly associated with response (P > .05 for all; Figure 3).

Peripheral blood CD19⁺ B-cell levels were available for 74 patients (82.2%; biomarker-evaluable population) at the time of study onset. CD19⁺ B-cell depletion, defined as <5 cells per μL after treatment with mosunetuzumab, occurred rapidly by the initiation of C2 dosing in patients with a CR and at least 1 follow-up sample (38/54 patients; Figure 4A). The median time to B-cell recovery in patients with a CR and at least 1 follow-up sample, defined as ≥5 cells per μL as determined by time-to-event analysis (Figure 4B), was 18.4 months (95% CI, 12.8-25.0) from completion of C8. The median time to recover to the lower level of normal (≥70 cells per μL) in the same group of patients was 25.1 months (95% CI, 19.0 to NE; Figure 4C) from completion of C8. Eight patients (8.9%) received immunoglobulin therapy in this study.

No new CRS events; or fatal, serious, or grade ≥3 AEs were reported during the extended follow-up period, and no evidence of chronic toxicity was observed. A summary of AEs is presented in supplemental Table 4. As previously reported, 2 patients experienced AEs related to mosunetuzumab that led to treatment discontinuation and both recovered (both due to CRS; supplemental Table 4).¹⁶ No treatment-related grade 5 AEs were observed. Two grade 5 AEs, determined by the INVs to be unrelated to mosunetuzumab treatment, were previously reported; 1 due to malignant neoplasm progression and 1 was an unexplained death.

CRS by American Society for Transplantation and Cellular Therapy criteria¹⁹ was reported in 40 patients (44.4%) and events were mostly low grade (grade 1-2: n = 38 [42.2%], grade 3-4: n = 2 [2.2%]). CRS primarily occurred during C1; any-grade CRS was reported in 23% of patients after the C1, day 1 dose; 6% after C1 day 8; 36% after C1, day 15; 10% after C2 day 1; and 2% in any subsequent cycle (supplemental Figure 3). The median time to CRS onset at C1 day 1 was 5 hours (range, 1-24), and at C1 day 15 was 27 hours (range, 0-391). The median CRS duration was 3 days (range, 1-29). Four patients (4.4%) received both corticosteroids and tocilizumab for CRS management; 3 patients (3.3%) received tocilizumab only. Commonly reported any-grade AEs included neutropenia in 26 patients (28.9%) and pyrexia in 26 patients (28.9%; supplemental Table 4). No febrile neutropenia events were reported.

No relationship was observed between the occurrence of CRS and tumor response; ORR was 77.5% and 78.0% in patients with or without CRS events, respectively (supplemental Table 5).

Long-term data from this analysis show that fixed-duration IV mosunetuzumab can achieve long-lasting remissions in heavily pretreated patients with R/R FL after 3 years of follow-up. At a median follow-up of 37.4 months, outcomes demonstrated a high CR rate of 60% and durable benefit with fixed-duration treatment, with most patients receiving ≤8 cycles. The median DOR was 35.9 months; however, this estimate was based on 1 event that occurred in the last patient at risk and, thus, may further improve with longer follow-up. In 54 patients who achieved a CR, the median DOCR was NR and 72% of these patients are estimated to remain alive and progression-free 30 months after their first response. Additionally, an estimated 36-month OS rate of 82% and a 36-month PFS rate of 43% were observed in this heavily pretreated highly refractory patient population. The treatment benefit of mosunetuzumab occurred early, with the median time to first response being 1.4 months, and the median time to first CR being 3 months. Exploratory analyses indicate that a similar DOR benefit was observed regardless of whether patients achieved an early or late CR. No relationship between DOR, PFS, or OS and the timing of achieving CR was noted. Consistent with previous reports, these extended follow-up data also continued to demonstrate a manageable safety profile with no new CRS events, and no late-onset or chronic toxicities.¹⁶ The fixed-duration dosing regimen of mosunetuzumab was supported by CD19⁺ B-cell pharmacodynamic evaluation, demonstrating a median time to B-cell recovery of 18.4 months in patients who had a CR after completion of 8 cycles of mosunetuzumab. Within the same cohort of patients, the majority of serious infections were reported during the first 4 cycles of mosunetuzumab treatment.²⁵ Response to retreatment with mosunetuzumab was also encouraging, with 3 of 5 patients achieving a CR. Together, these findings demonstrate the benefit and safety of fixed-duration therapy and continue to support the administration of mosunetuzumab as an outpatient regimen.

The activity of mosunetuzumab was assessed in patients with known prognostic variants and demonstrated that clinically meaningful response rates were also observed in patients with common mutations, including those associated with poor prognosis such as TP53.²⁴ Exploratory analyses demonstrated that response and DOR were greater with mosunetuzumab vs the patients’ last prior therapy. As outcomes typically worsen with each subsequent line of therapy, these results demonstrate a paradigm shift from what has been historically reported in R/R FL.^3-7 This finding supports further assessment of mosunetuzumab’s activity in earlier lines of FL treatment, which is currently underway in the phase 3 CELESTIMO and MorningLyte trials.^26,27 

Although not directly comparable, the long-term response data for mosunetuzumab in patients with R/R FL are numerically superior than those with other therapies approved in this setting, such as the EZH2 inhibitor tazemetostat (patients with EZH2 mutation: CR, 13%; ORR, 69%).²⁸ Longer durable responses and higher PFS rates were observed in the current 3-year follow-up of mosunetuzumab in an all-comer R/R FL patient population compared with tazemetostat in patients with an EZH2 mutation (median DOR, 35.9 vs 10.9 months; PFS, 24.0 vs 13.8 months).²⁸ The combination of the Bruton’s tyrosine kinase inhibitor, zanubrutinib plus the anti-CD20 monoclonal antibody obinutuzumab in the ROSEWOOD phase 1b study demonstrated encouraging efficacy (CR, 39%; ORR, 69%; 18-month DOR, 69%) in patients with R/R FL. However, mosunetuzumab offers a finite duration treatment regimen allowing B cell recovery to occur, compared to the zanubrutinib and obinutuzumab regimen which included twice daily administration of zanubrutinib until disease progression or unacceptable toxicity and a 2-year maintenance of obinutuzumab after induction therapy.²⁹ 

While recognizing the limitations of cross-trial comparisons, we note that the early responses and high response rates achieved with fixed-duration IV mosunetuzumab were similar to those achieved with the CD20-directed bispecific antibodies odronextamab (phase 2 ELM-2 study)^30,31 and epcoritamab (phase 1/2 EPCORE NHL-1 study).^32,33 Patients with R/R FL in both the ELM-2 and EPCORE NHL-1 studies had similar baseline characteristics such as age, POD24 status, and refractoriness to last therapy, to those in our study. In patients with R/R FL treated with IV odronextamab monotherapy, clinical responses (CR, 73%; ORR, 80%) have been reported with a median efficacy follow-up time of 20.1 months.^30,31 Responses were durable, median DOR was 22.6 months, and median DOCR was 25.1 months. Median PFS was 20.7 months, and the 24-month OS rate was 70.1%. Subcutaneous epcoritamab induced high response rates (CR, 63%; ORR, 82%) with a median follow-up of 17.4 months in patients with R/R FL.^32,33 The median PFS was 15.4 months, and the DOR, DOCR, and OS were NR. With longer follow-up we have also reported a greater median PFS (24.0 months) and a high 36-month OS rate of 82% with mosunetuzumab. Odronextamab and epcoritamab are administered until disease progression, however our data suggest that mosunetuzumab can achieve high response rates and durable remissions as a fixed-duration treatment.^30-33 Treatment-emergent grade 5 AE rates in our study (2.2%, unrelated to mosunetuzumab) appeared lower than those reported in patients with R/R FL in the ELM-2 (all fatal AEs: 14.8%; related to odronextamab: 3.1%) and the EPCORE NHL-1 (10.0%, relation to epcoritamab not specified) studies.^30-33 Minimizing corticosteroid use is desirable in patients with R/R NHL because of an already compromised immune response from prior therapies and disease progression. Mosunetuzumab administration requires limited corticosteroid premedication (in C1 and C2). In contrast, other bispecific antibodies such as epcoritamab and odronextamab require a higher number of doses of corticosteroid premedication.^30-33 

Studies examining chimeric antigen receptor (CAR) T-cell therapies in patients with FL have reported high CR rates,^34-38 and durable remissions that are similar to those observed in the current study. Although CAR T-cell therapies are approved and have high response rates in patients with R/R FL, because of the intensity of therapy, strict inclusion criteria are implemented in CAR T-cell pivotal studies, thereby limiting accessibility.³⁹ However, more manageable toxicity profiles, such as that of lisocabtagene maraleucel (liso-cel) in patients with R/R FL, may increase the number of patients eligible for CAR T-cell therapy.³⁷ In the current analysis, the median DOCR was NR and 30-month DOCR rate was 72% with mosunetuzumab; this outcome is similar to that observed with axicabtagene ciloleucel in the ZUMA-5 study of patients with R/R FL after >3 years of follow-up (median DOCR, NR; 36-month DOCR, 62%).⁴⁰ 

Potential AEs that are associated with CAR T-cell therapy, including immune effector cell–associated neurotoxicity syndrome,^31,41 are uncommon in patients treated with mosunetuzumab.^16,42 In the TRANSCEND FL study, neurological AEs reported in patients with R/R FL treated with liso-cel included aphasia (7%) and tremor (7%), whereas aphasia was not reported with mosunetuzumab in this study, and tremor was reported in 1 patient (grade 1, 1.1%).³⁷ The CRS events observed with liso-cel and mosunetuzumab were low grade and manageable.³⁷ The off-the-shelf formulation of mosunetuzumab does however avoid logistical challenges associated with CAR T-cell administration including leukapheresis and lymphodepleting chemotherapy and may be more cost effective.^43-45 

In conclusion, this 3-year updated analysis represents the longest reported follow-up evaluating the efficacy and safety of bispecific antibody therapy in R/R FL. Results confirm and extend findings of the primary analysis results, providing further evidence of clinically meaningful and sustained benefit of mosunetuzumab, with a manageable safety profile. B-cell counts progressively recovered to normal levels after fixed-duration mosunetuzumab therapy and evidence of potential benefit from retreatment was demonstrated. This long-term analysis demonstrates that fixed-duration treatment with mosunetuzumab yields a high rate of durable responses, with a high proportion of patients remaining free from progression at 3 years.

Third-party editorial assistance, under the direction of all authors, was provided by Deirdre Kelly and Louise Profit at Ashfield MedComms, an Inizio company, and was funded by F. Hoffmann-La Roche Ltd. The NCT02500407 trial is sponsored by Genentech, Inc.

Contribution: L.H.S., N.L.B., L.E.B., M.C.W., S.Y., I.T., A.K., and E.P. contributed to the study design; and all authors contributed to the acquisition, analysis, and interpretation of study data, critically reviewed the manuscript, and provided final approval for publication.

Conflict-of-interest disclosure: L.H.S. reports research grants from F. Hoffmann-La Roche Ltd/Genentech, Inc, and Teva; consulting fees from AbbVie, Acerta, Amgen, AstraZeneca, Celgene/Bristol Myers Squibb (BMS), Gilead/Kite/Incyte, F. Hoffmann-La Roche Ltd/Genentech Inc, Janssen, MorphoSys, Sandoz, and TG Therapeutics; and reports honoraria from AbbVie, Acerta, Amgen, AstraZeneca, Celgene/BMS, Gilead/Kite/Incyte, F. Hoffmann-La Roche Ltd/Genentech Inc, Janssen, MorphoSys, Sandoz, and TG Therapeutics. N.L.B. reports research funding from ADC Therapeutics, Autolus, BMS, Celgene, Forty Seven, Gilead/Kite Pharma, F. Hoffmann-La Roche Ltd/Genentech Inc, Janssen, Merck, Millenium, Pharmacyclics, and Seattle Genetics; and reports membership on advisory committee for ADC Therapeutics, Foresight Diagnostics, F. Hoffmann-La Roche Ltd/Genentech Inc, Kite, and Seattle Genetics. M.J.M. reports research grants from AstraZeneca, Bayer, F. Hoffmann-La Roche Ltd/Genentech Inc, IGM Biosciences, Janssen, Pharmacyclics, and Seattle Genetics; reports honoraria for ADC Therapeutics, Bayer, Daiichi Sankyo, Epizyme, F. Hoffmann-La Roche Ltd/Genentech Inc, IMV Therapeutics, Janssen, MEI Pharma, Pharmacyclics, and Seattle Genetics; reports payment for expert testimony from Bayer; and owns stock or stock options in Merck. S.J.S. reports consultancy for Acerta, Celgene, F. Hoffmann-La Roche Ltd/Genentech Inc, Novartis, and Pharmacyclics; reports research funding from Celgene, Gilead, Janssen Research and Development, Merck, Novartis, and Pharmacyclics; and reports membership on scientific advisory committee for Nordic Nanovector. S.E.A. reports research grants from AbbVie, F. Hoffmann-La Roche Ltd/Genentech Inc, Lilly, Merck, and Takeda; reports speakers bureau membership with Pfizer; and is the chair of hematology group (unpaid) for Canadian Cancer Trials Group. P.G. reports current employment with Royal Adelaide Hospital. J.K. reports research grants for AstraZeneca, Kite, Merck, and Novartis; received consulting fees from AbbVie, BMS, F. Hoffmann-La Roche Ltd, Gilead/Kite, Merck, and Seattle Genetics; reports honoraria from AbbVie, Amgen, AstraZeneca, BMS, BeiGene, F. Hoffmann-La Roche Ltd, Genmab, Gilead, Incyte, Janssen, Karyopharm, Merck, Novartis, Pfizer, and Seattle Genetics; and reports membership on a data safety monitoring board for Karyopharm. M.S. reports employment with BMS; reports stock or other ownership with Koi Therapeutics; reports honoraria from AbbVie, AstraZeneca, BeiGene, BMS, Eli Lilly, F. Hoffmann-La Roche Ltd/Genentech Inc, Fate therapeutics, Genmab, Janssen, Kite Pharma, Merck, MorphoSys/Incyte, Mustang Bio, and Nurix; reports consulting or advisory role with AbbVie, AstraZeneca, BeiGene, BMS, Eli Lilly, Fate Therapeutics, Genentech Inc, Genmab, Janssen, Kite Pharma, Merck, MorphoSys/Incyte, Mustang Bio, and Nurix; and reports research funding from, AbbVie, AstraZeneca, BeiGene, Genentech, Inc, Genmab, MorphoSys/Incyte, Mustang Bio, and Vincerx. C.Y.C. reports honoraria for AbbVie, AstraZeneca, BeiGene, BMS, Dizal, F. Hoffmann-La Roche Ltd, Genmab, Gilead, Janssen, Lilly, and Menarini; reports consulting or advisory role for AbbVie, AstraZeneca, BeiGene, BMS, Dizal, F. Hoffmann-La Roche Ltd, Genmab, Gilead, Janssen, Lilly, and Menarini; and reports research funding from AbbVie, BMS, F. Hoffmann-La Roche Ltd, Merck Sharpe & Dohme, and Lilly. S.D. reports current employment with University Hospital of Duesseldorf and ended employment in past 24 months with University Hospital of Duesseldorf; and reports honoraria from BeiGene, F. Hoffmann-La Roche Ltd, Kite, and Gilead. K.F. reports current employment with Royal North Shore Hospital and St Vincent’s Hospital. M.K. reports consulting or advisory role for F. Hoffmann-La Roche Ltd; and reports research funding from F. Hoffmann-La Roche Ltd. L.J.N. reports grants from BMS, Caribou Biosciences, Epizyme, Gilead/Kite, Janssen, IGM Biosciences, Takeda, and TG Therapeutics; reports honoraria for AbbVie, AstraZeneca, BMS, Caribou Biosciences, Daiichi Sankyo, F. Hoffmann-La Roche Ltd/Genentech Inc, Genmab, Gilead/Kite, Janssen, Incyte, Ipsen, Merck, Novartis, Regeneron, and Takeda; received research support from BMS, Caribou Biosciences, Daiichi Sankyo, F. Hoffmann-La Roche Ltd/Genentech Inc, Genmab, Gilead/Kite, IGM Biosciences, Ipsen, Janssen, Merck, Novartis, and Takeda; and reports participation in data safety monitoring board or advisory board for ADC Therapeutics, Bayer, DeNovo, Epizyme, Genentech Inc, Janssen, MEI Pharma, MorphoSys, Novartis, Takeda, and TG Therapeutics. D.K. and A.K. report employment with F. Hoffmann-La Roche Ltd; and hold stock or stock options in F. Hoffmann-La Roche Ltd. M.C.W., S.Y., I.T., C.R.B., and E.P. are employees of Genentech Inc; and hold equity/stock in F. Hoffmann-La Roche Ltd/Genentech, Inc. L.E.B. reports consulting fees from ADC Therapeutics, Amgen, AstraZeneca, Genentech Inc, and Merck; and participation on a data safety monitoring board for Ziopharm Oncology.

Correspondence: Lihua E. Budde, Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, 1500 E Duarte Rd, Duarte, CA 91010; email: ebudde@coh.org.