FDA ODAC: STARGLO Glofitamab DLBCL Trial Inapplicable for US Populations

The FDA’s Oncologic Drugs Advisory Committee (ODAC) regards the STARGLO international clinical trial of glofitamab plus gemcitabine and oxaliplatin (GemOx) for diffuse large B-cell lymphoma (DLBCL), in its enrolled cohort and outcome results, to be inapplicable to patient populations in the United States.¹ It heard arguments for and against applicability in a meeting at the Center for Drug Evaluation and Research (CDER) on the FDA White Oak Campus in Silver Spring, Maryland, on May 20, 2025.²

Glofitamab is a CD20xCD3 bispecific antibody, developed and marketed by Genentech, Inc. under the brand name Columvi.³ The FDA in June 2023 granted accelerated approval to monotherapy of this agent for relapsed or refractory DLBCL not otherwise specified (DLBCL NOS) or large B-cell lymphoma developed from follicular lymphoma in adults who have undergone at least two lines of systemic therapy.⁴

Currently before the FDA is a supplemental Biologics License Application (sBLA) submitted by Genentech for a combination of glofitamab injection with GemOx to treat relapsed or refractory DLBCL NOS in adults ineligible for autologous stem cell transplant (ASCT).² This sBLA involves STARGLO, a global phase 3 open-label randomized clinical trial in which glofitamab plus GemOx was compared against rituximab plus GemOx in patients with relapsed or refractory DLBCL NOS who had received at least one line of systemic therapy and were ASCT-ineligible. The trial met its primary endpoint of statistically significant improvement in overall survival (OS) and two of its three secondary endpoints in progression-free survival (PFS) and complete response (CR).³

Before the ODAC, the FDA raised several concerns about the generalizability of STARGLO to patient populations and medical practice specifically within the US. It argued that the number of US patients enrolled in STARGLO was insufficient, regionally inconsistent, and lacked representativeness of the US population.^1,4

Moreover, the FDA’s subgroup statistical analysis of STARGLO’s results found differences in OS and other statistical measures for outcomes between Asian and non-Asian patients in the trial, with a reduced benefit from the study treatment in North American and European patients compared with Asian patients. Disparities became apparent over the course of the trial, and Genentech requested an additional 9 months to analyze and address them.^1,4

“With an additional follow-up of approximately 11 months, the inconsistent treatment effect in overall survival persisted. While a positive result was observed in the ITT [intention-to-treat] analysis, with a hazard ratio of 0.62, the inconsistent results between the regions persisted with similar discordant treatment effects observed for patients treated in the Asian region, with a hazard ratio of 0.39, compared to the Non-Asian region, with a hazard ratio of 1.06,” stated Margret Merino, MD, of the FDA’s Division of Hematologic Malignancies II and the cross-disciplinary team lead for the related sBLA.^1,4

Other concerns that the FDA raised regarding the Asian-versus-Non-Asian-region dimension within STARGLO were that the patient cohorts that statistically drove the study, specifically the Asian regional subgroup, differed in key aspects from the US patient population. This included patients’ baseline disease characteristics, disease cell of origin, reasons for transplant ineligibility, treatment exposures, schedule of study regimen administration, and reasons for treatment discontinuation.^1,4

The responses Genentech presented at the meeting to the FDA’s critiques were led by the company’s senior vice president and global head of oncology and hematology drug development, Charles Fuchs, MD, MPH. The company insisted that the enrolled cohort and findings in STARGLO applied to a US patient population, that the subgroup-analysis-based concerns raised by the FDA did not affect the results, and that the low enrollment at US centers in the study was an effect of the COVID-19 pandemic. It repeatedly argued that the STARGLO results were impressive and that the treatment approach under investigation in the trial would meet an unmet need in the US population.^1,3

Among the points of disagreement between the company and the FDA was whether imbalances in new anti-lymphoma therapy (NALT) use across the study centers affected the study’s outcome statistical results. The company maintained that there was an effect, while the FDA maintained that there was not.^1,3,4 Genentech also objected to the subgroup assessment applied by the FDA to STARGLO, holding that the trial had not been designed for that type of analysis.

The FDA noted that it was not seeking the opinion of the ODAC as to whether the STARGLO trial could be used to convert the accelerated approval of glofitamab to a traditional approval, but only on whether the trial’s enrolled population and results applied to US patient populations.^1,4 To this question, eight ODAC members voted “No” and one “Yes.”¹

Several members of the Committee, before and after the vote, expressed concern about the small number of US patients enrolled in STARGLO.

“If there’s a low percentage of patients enrolled in a trial and we see this directionality in multiple endpoints, we have to call into question, and we can’t know if this therapy may be effective in this specific population,” stated Neil Vasan, MD, PhD, of NYU Langone Health in New York City, who voted “No” on the discussion question.¹

Linked to this concern of the small number of US patients enrolled in STARGLO, Heidi McKean, MD, from Avera Cancer Institute in Sioux Falls, South Dakota, who voted “No” on the discussion question, spoke from her own experience in community practice of a need to be conscious of the high cytokine release syndrome (CRS) risk faced by patients who may receive glofitamab.¹

“We as community oncologists need to make sure this is safe and really effective, and 25 patients is really tough to make that call,” Dr. McKean said.¹

Committee member Daniel Spratt, MD, of University Hospitals Seidman Cancer Center and Case Western Reserve University in Cleveland, Ohio, who voted “No” on the discussion question, said that the subgroup analysis applied by the FDA to STARGLO was appropriate given the study design’s established organizations of patient region and race.¹

“There is a significant difference in the differential benefit of this therapy by the subgroups that were prespecified. It’s not simply that all of this is just post hoc subset-of-subset analysis. There’s actually a differential in treatment effect,” Dr. Spratt said.¹

The FDA is not legally bound to follow the recommendations it receives from ODAC but typically has done so.²