Roche Holding AG to Discuss Ph III STARGLO Results in 2L+ DLBCL Presentation Transcript

The following is excerpted from the question-and-answer section of the transcript.

Question: Richard Parkes - Exane BNP Paribas - Analyst : Thanks, Bruno. Thanks for taking the questions. So I've got two questions, firstly, for Dr. Abramson. Just in terms of the geographic differences in overall survival, there's a possibility that could be caused by improved availability of second-line therapies like CAR T -- third-line therapies like CAR T and bispecifics. So it's likely could lead to a debate over whether it's better to use this treatment or kind of later lines. without the prior written consent of Thomson Reuters. 'Thomson Reuters' and the Thomson Reuters logo are registered trademarks of Thomson Reuters and its affiliated companies. JUNE 16, 2024 / 11:00AM, ROG.S - Roche Holding AG to Discuss Ph III STARGLO Results in 2L+ DLBCL Presentation So could you give us any thoughts on that? Is there any data on post-progression treatment? And maybe you could -- you mentioned your co-variant analysis. Could you just explain in layman's terms why that makes you comfortable that there isn't another explanation for that difference other than just a small land. So that was my first question. So maybe we'll start that. Jeremy Abramson - Massachusetts General Hospital Cancer Center - Director of the Lymphoma Program at the Jon and Jo Ann Hagler Chair in Lymphoma Okay, sure. No, that's an important question. We've obviously thought a lot about these regional inconsistencies that were observed. The most important thing is that the number of patients treated in North America and Europe was a small number. And that is largely related to the fact that this is during the COVID-19 pandemic, where we had really inability to conduct clinical research at most centers in the United States and Europe for an extended period of time. So when we look at this, we looked at a few things. One is we looked at the drug itself. Is there a possibility that glofitamab is differentially effective by race or region? And the answer is no. The pharmacokinetics are no different across race or region and monotherapy clinical efficacy is also no different across race and region. So we then looked, of course, within these subgroups. And what's important to recognize and we identified is that minor variations across baseline characteristics become massively magnified within a subgroup of small patient numbers. So getting to your point regarding next line of lymphoma therapy. Within the European cohort on the R-GemOx arm, there was a higher use of the most active medications, including CAR T-cells and bispecific antibodies. So that's clearly going to affect the behavior and overall survival of the control arm. Furthermore, in the European cohort, the seven deaths related to COVID, I mentioned, 5 of them. We're on the Glofit-GemOx arm of the European cohort, and that's going to massively impact a study of that size within that cohort in terms of progression-free and overall survival. In the North American cohort, which I should note was only 25 patients, we saw also some variations in baseline characteristics, particularly on the Glofit-GemOx arm, which were enriched for high-risk features, including the presence of primary refractory or early relapsed disease, which is a dominant risk factor in second-line therapy for large B-cell lymphoma, as well as high tumor volume was increased in the globe with R-GemOx arm. So ultimately, these are minor variations that become exaggerated in small patient numbers, which is why we really can't draw firm conclusions. What I can say is that these small variations don't make an impact in the overall findings of the study. And that's why randomized clinical trials like this and others are designed to be interpreted based on the intent-to-treat population as opposed to small underpowered univariable analysis. Regarding the other part of your question, which was the multivariable analysis, so because these are univariable analysis, meaning you look at whether a patient had relapsed or refractory disease, independent exclusively, you look at these other factors. Short of a multivariable analysis, very difficult to draw conclusions about the independent predictive power of a single variable, because patients that turns out are all multivariable. There's no such thing as a univariable patient. Patients come from a region, patients have relapsed or refractory disease, patients have received one or more prior lines of therapy. You have to view them all together. And what we find is that you have to adjust for risk factors in aggregate. And when we do that multivariable analysis, which is a highly refined statistical tool that's, of course, completely standard in these analyses, we find no impact whatsoever of geographic region at overall survival.

Question: Richard Parkes - Exane BNP Paribas - Analyst : Perfect. And then my second question, hopefully a bit shorter, for Daud, just on the Hemlibra autoinjector. I just wondered if you are able to comment on potential timelines. So just wondering if you're optimistic this could be available by the time in [by May] launches? And what technical risk is remaining? I'm just wondering if you've done initial formulation work. So is the technical execution risk relatively low?

Question: Richard Parkes - Exane BNP Paribas - Analyst : Yeah. It was great. Thank you.