Roche's Columvi win in earlier DLBCL faces questions over inconsistent patient survival data

Roche’s up-and-coming Columvi has demonstrated it could save lives when used as a second-line therapy in diffuse large B-cell lymphoma (DLBCL). But the Swiss pharma still finds itself fielding questions about the bispecific drug’s market potential thanks to an imbalance noticed in the phase 3 data.

Combining Columvi with the chemotherapy regimen GemOx significantly reduced the risk of death by 41% compared with Rituxan and GemOx in DLBCL patients who had tried at least one prior line of therapy and were not eligible for stem cell transplant. The results came from the primary analysis of the phase 3 STARGLO trial and were presented during the plenary session of the European Hematology Association 2024 Congress.

After a longer follow-up of 20.7 months at median, the overall survival benefit stood strong at 38%. Patients who took the Columvi combo lived nearly twice as long as those in the control group with median overall survival clocked in at 25.5 months and 12.9 months, respectively.

Overall survival is the gold standard for measuring a cancer drug’s efficacy, and it would have cemented Columvi’s role in an earlier treatment setting of DLBCL—the CD20xCD3 T-cell engager got its initial FDA approval in 2023 as a third- or later-line therapy. However, inconsistencies in the drug’s efficacy in different geographic regions were hard to ignore in the STARGLO trial.

Regional inconsistencies

An exploratory analysis of certain prespecified subgroups showed that Columvi’s overall survival benefit appeared to be driven entirely by patients outside the U.S. and Europe. Among the 88 European patients and 25 North American participants, the risks of survival in both cases favored Rituxan-GemOx, although the data didn’t carry statistical significance. Data in the North America subgroup looked especially alarming. There, eight (53%) patients in the Columvi-GemOx arm have passed away, versus two (20%) in the control arm.

As one analyst duly questioned during a Roche conference call on Sunday, Columvi’s lack of showing in the U.S. and EU might be explained by the availability of efficacious later-line therapies such as CAR-T therapies and bispecifics. If that were the case, questions could be raised around whether Columvi should rather be used in later lines versus the second line.

In response, Jeremy Abramson, M.D., from the Massachusetts General Hospital Cancer Center and principal investigator of STARGLO, first stressed that the number of patients in those two regions was simply too small to draw any conclusions. In a situation like this, “minor variations across baseline characteristics become massively magnified,” he said.

That’s especially true for the North America subgroup, which only had 25 patients between the two treatment arms, because “any single event is going to massively sway the numbers within a population of that small size,” Abramson said. 

Jeremy Abramson, M.D. (Massachusetts General Hospital)

In the North America cohort, the Columvi arm had a higher proportion of patients with high-risk features—including early relapse disease—and high tumor volume, Abramson noted. In the European cohort, five deaths in the Columvi arm were due to COVID, versus none in the control group; and the control group had more use of CAR-T therapies and bispecifics as subsequent treatments.

Besides, geographic location is merely one variable for any given patient, whose baseline characteristics are defined by multiple variables. The STARGLO researchers conducted a multivariable analysis and found “no impact whatsoever of geographic region on overall survival,” Abramson said.

As to whether Columvi is better off as a later-line therapy, Abramson argued that “the best treatments work best the earlier you use them.”

Although no formal studies have compared Columvi itself between second and later lines, the Columvi-GemOx regimen “produced significantly deeper rates of complete remission and a longer progression-free survival than was observed in the monotherapy study” in the later-line setting, “suggesting that earlier use of this regimen is inducing deeper and more durable responses,” Abramson said.

Going against CAR-T

Roche plans to submit the STARGLO results to the FDA. If approved, Columvi will go toe to toe against CD19 CAR-T therapies, including Bristol Myers Squibb’s Breyanzi and Gilead Sciences’ Yescarta in second-line DLBCL.

CAR-T therapies can be used for patients regardless of their eligibility for stem cell transplant, but Abramson noted that about two-thirds of newly diagnosed DLBCL patients are considered transplant-ineligible, including those who are older. The vast majority of patients who’re considered transplant-ineligible would be eligible to receive the Columvi-GemOx regimen because there’s no upper age limit and it’s generally well tolerated with a fixed duration of treatment, he added.

Nevertheless, Abramson said he would reach for a CAR-T first for patients who are candidates because these cell therapies have longer follow-up data compared with STARGLO.

However, many patients at his practice are not suitable to receive CAR-T therapies, which require weeks of manufacturing. And these individualized treatments are not accessible to most patients outside large academic centers. Indeed, Yescarta recently reached a plateau despite the second-line label expansion, and Gilead’s Kite Pharma is working to improve the infrastructure bottleneck.

“So the real merit of [Columvi-GemOx] is that a time-limited therapy can be given second line off the shelf to really patients anywhere across the world who has this treatment in their pharmacy,” Abramson said.

Columvi is one of two CD20xCD3 bispecifics in Roche’s offerings. The other one, Lunsumio, is also undergoing phase 3 testing in second-line DLBCL as part of a combination with Roche’s CD79b antibody-drug conjugate Polivy. During Sunday’s call, Roche pushed back the expected readout for that study, coded SUNMO, from this year to 2025.

Together, Roche expects second-line DLBCL could give the two bispecifics up to $2 billion in peak sales, and “the opportunity would even become significantly bigger” should they move into the first-line setting, Bruno Eschli, Roche’s head of investor relations, said on the call.

Roche last fall started the phase 3 SKYGLO trial, which is evaluating Columvi on top of the Polivy-R-CHP regimen in previously untreated DLBCL. The company reported encouraging early-phase data for the combo in December.