The second-line setting in relapsed/refractory diffuse large B-cell lymphoma (DLBCL) has become one of the more actively contested spaces in hematologic oncology, and on June 29 AbbVie and Genmab added a notable data point. The Phase 3 EPCORE DLBCL-4 trial met its primary endpoint: epcoritamab plus lenalidomide reduced the risk of disease progression or death by 60% versus R-GemOx under U.S. censoring conventions (HR 0.40; 95% CI, 0.30–0.55) and by 56% under the censoring rules applied outside the U.S. (HR 0.44; 95% CI, 0.33–0.60), both with p<0.0001. The dual hazard ratios are not a discrepancy in the effect; they reflect the different progression-free survival (PFS) censoring conventions U.S. and ex-U.S. regulators expect, and both land in the same range.
A hazard ratio near 0.40 against an active comparator in a randomized Phase 3 is a strong result, and it was generated with a fixed-duration regimen that contains no cytotoxic chemotherapy. But this is a topline readout: overall survival (OS), response depth, durability, and the full safety dataset were not disclosed and are pending a medical meeting. The size of the PFS effect is the story; whether it becomes a category-defining one depends on data not yet in hand.
The competitive backdrop
Epcoritamab (Epkinly), a subcutaneous IgG1 CD3xCD20 bispecific built on Genmab‘s DuoBody platform, first reached the U.S. market in May 2023 under accelerated approval for third-line-and-later DLBCL, based on single-arm response data. It competes most directly with Roche‘s glofitamab (Columvi), an IV CD20xCD3 bispecific that holds an analogous third-line accelerated approval. Both companies have worked to move their T-cell engagers earlier in the disease course, and the two programs took visibly different design paths.
Roche’s second-line bid, STARGLO, paired glofitamab with gemcitabine-oxaliplatin and reported a 41% reduction in the risk of death versus R-GemOx (OS HR 0.59; 95% CI, 0.40–0.89), roughly doubling median OS (25.5 vs. 12.9 months). Despite that, the FDA’s Oncologic Drugs Advisory Committee voted 8-1 in May 2025 against the applicability of the trial’s population and results to U.S. patients, roughly 9% of enrollment was from the U.S. versus about 48% from Asian regions, and a complete response letter followed in July 2025.
Higher toxicity in the glofitamab arm (serious adverse events in 52.3% vs. 17%) was part of the regulatory concern. Glofitamab’s third-line accelerated approval remained intact, and Roche has since pointed to a frontline study (SKYGLO) as a path forward.
It’s tempting to read EPCORE DLBCL-4 as a reply to that setback, but the timeline doesn’t support that framing. EPCORE DLBCL-4 began enrolling in August 2024, before the STARGLO advisory-committee vote and complete response letter, at a point when glofitamab looked headed toward full approval. The chemo-free, broadly-enrolled design was a strategic choice made on its own terms. What can be said is that, with hindsight, that design happens to avoid the two issues that sank STARGLO in the U.S.: it did not depend on a heavily region-concentrated population, and it did not layer chemotherapy toxicity onto a bispecific. Whether U.S. enrollment in DLBCL-4 was in fact more representative is a question the full dataset will have to answer.
The rationale for pairing an IMiD with a T-cell engager
The mechanistic logic behind the combination is coherent, though the topline readout does not by itself prove it. Epcoritamab bridges CD3 on a patient’s T cell to CD20 on a malignant B cell, forming an immune synapse that redirects polyclonal T-cell cytotoxicity onto the tumor independent of T-cell receptor specificity. Its activity therefore depends on the fitness of the patient’s own T cells, a real constraint in a heavily pretreated, immunologically exhausted population.
Lenalidomide, a cereblon-binding immunomodulatory drug (IMiD), targets that constraint on paper. By driving degradation of the transcription factors Ikaros and Aiolos, it lowers the T-cell activation threshold, increases IL-2 and interferon-γ production, and expands effector T- and NK-cell populations in preclinical and clinical models. In principle, that conditions and amplifies the very effector cells epcoritamab recruits, which could translate into deeper or more durable responses. EPCORE DLBCL-4’s PFS result is consistent with that hypothesis but does not establish that this mechanism, rather than additive single-agent activity, produced the benefit; response-depth and durability data will speak to that.
The toxicity argument is more concrete. Chemotherapy is lymphodepleting, which can blunt the effector compartment a bispecific relies on, so a chemo-free backbone is one reasonable way to preserve T-cell fitness while limiting toxicity overlap, lenalidomide’s dose-limiting toxicities (myelosuppression, thromboembolism, rash) sit largely outside epcoritamab’s profile of cytokine release syndrome (CRS), neurotoxicity, and B-cell aplasia. That said, STARGLO is a reminder that chemo-bispecific combinations are not mechanistically doomed; glofitamab plus chemotherapy produced a strong OS signal.
The chemo-free approach is best read as a differentiated, tolerability-driven bet rather than the only valid design. AbbVie and Genmab report the topline safety profile was consistent with each agent’s known profile, with no new signal flagged, but the detail clinicians will scrutinize is the full safety dataset, particularly CRS when an immune activator is added.
The population and the comparator
R-GemOx is an accepted but modestly effective salvage regimen, used largely for patients who cannot tolerate intensive therapy, with responses that are often not durable. EPCORE DLBCL-4 enrolled patients who had relapsed after or were ineligible for autologous stem cell transplant and were also ineligible for or unable to receive CAR-T, a deliberately hard-to-treat group, and included several aggressive large B-cell histologies beyond DLBCL NOS (double/triple-hit high-grade B-cell lymphoma, follicular lymphoma grade 3B, T-cell/histiocyte-rich large B-cell lymphoma, and EBV-positive DLBCL). Patients received fixed-duration subcutaneous epcoritamab (weekly in cycles 1–3, then every four weeks through cycle 12) plus oral lenalidomide, versus up to four cycles of R-GemOx.
That target population matters commercially and clinically. Second-line CAR-T (axi-cel, liso-cel) is now standard for transplant-eligible patients with early relapse, but a large share of DLBCL patients are too old, too frail, or too geographically removed from CAR-T centers to receive it, and CAR-T’s manufacturing lead time excludes patients who need treatment quickly. For that group, an off-the-shelf, time-limited regimen delivering this magnitude of PFS benefit is a meaningful option. CRS management remains the practical hurdle for any bispecific; epcoritamab’s subcutaneous step-up dosing is designed to keep CRS predominantly low-grade and concentrated in cycle 1, and phase 2 experience (EPCORE NHL-6) supports outpatient monitoring, a real logistical and cost advantage against IV CAR-T and IV glofitamab.
The business stakes
A second-line label would move epcoritamab from a later-line option into a substantially larger commercial pool, and it fits a broader platform strategy rather than a one-off. Epcoritamab is co-developed under a 2020 AbbVie–Genmab collaboration in which AbbVie paid $750 million upfront with up to $3.15 billion in potential milestones; the companies share U.S. and Japan commercialization (Genmab books sales and shares profits in those markets), with AbbVie handling the rest of the world and paying Genmab tiered royalties of 22–26%. A second-line expansion therefore feeds a revenue line both partners depend on, at a moment when Roche is regrouping in the same setting.
The platform angle is more advanced than a single readout suggests. In November 2025 the FDA approved epcoritamab plus rituximab and lenalidomide, the “R²” backbone with a bispecific added, in relapsed/refractory follicular lymphoma. EPCORE DLBCL-4 is a leaner doublet (epcoritamab plus lenalidomide, dropping rituximab), but it extends the same idea into aggressive lymphoma: substitute a CD20 T-cell engager for the anti-CD20 antibody in an IMiD combination. If the DLBCL data hold up through OS and safety, the approach is a credible template that could extend to other CD20 bispecifics and other histologies, and it strengthens the case for epcoritamab as a combination backbone rather than a salvage monotherapy.
What still has to land
The restraint is straightforward. This is a topline PFS readout; OS was not reported, and OS is precisely the endpoint that earned glofitamab its clinical credibility and that regulators weigh most heavily. Cross-trial comparison to STARGLO is unreliable: different experimental arms, different endpoints as the basis of the win (PFS here, OS there), and different follow-up. Full safety, response depth, and durability remain undisclosed pending a medical meeting, and the degree of U.S. enrollment, the specific issue that sank STARGLO, has not been detailed. The companies plan to discuss the data with global regulators and present full results at a future meeting.
A hazard ratio near 0.40 is a strong hand. It is not yet the whole hand. The survival data, the full safety profile, and the enrollment breakdown are what stand between a strong PFS result and a second-line label AbbVie and Genmab can defend where Roche could not.
Read more biotech insights on OncoDaily Biotech.
Written by: Semiramida Nina Markosyan, Editor, OncoDaily Canada