Arcus Biosciences, Inc. (RCUS) Earnings Call Transcript & Summary

Good afternoon. I'm Pia Banerjee, Head of Investor Relations at Arcus Biosciences. Thank you for joining us to discuss the results from the fourth interim analysis of our and Gilead Sciences' ongoing Phase II ARC-7 trial. We and Gilead issued a joint press release last night, highlighting the data from the ASCO Plenary Session abstracts, which can be accessed in the Investors section of our websites. In addition, the full results were presented today at 3 p.m. Eastern Time at the Virtual ASCO Monthly Plenary Series by our lead investigator, Dr. Melissa Johnson, Director of Lung Cancer Research at the Sarah Cannon Institute at Tennessee Oncology. That presentation and discussion are available for viewing on the ASCO website. For this afternoon's call, we'll be hearing from Arcus' CEO, Dr. Terry Rosen; and Chief Medical Officer, Dr. Dimitry Nuyten; as well as Dr. Bilal Piperdi, Vice President of Gilead Oncology Lung Cancer -- Lung Franchise. The Q&A portion will be moderated by Jacquie Ross, Vice President of Investor Relations at Gilead; and we'll also be joined by Arcus' Chief Operating Officer, Jennifer Jarrett, along with Dr. Bill Grossman, Senior Vice President, Therapeutic Area Head of Gilead Oncology. Our discussion today will contain forward-looking statements, including statements regarding future data disclosures and presentations, the development of current and future programs and the efficacy and safety of domvanalimab, zimberelimab and etrumadenant. All forward-looking statements involve known and unknown risks and uncertainties and other factors that may cause our actual results, performance or achievements to differ significantly from those expressed or implied by the forward-looking statements. A description of these risks can be found on Slide 3 and in the latest SEC filings filed by Arcus and Gilead, which we encourage you to review. I'll now turn the call over to our CEO, Terry Rosen.

Thank you very much, Pia, and thank you all for joining us today. We're really very excited to discuss the data presented earlier today from the fourth interim analysis of our randomized Phase II study, ARC-7 in first-line PD-L1-high non-small cell lung cancer. These data represent the largest and first prospective randomized data set to be released for an anti-TIGIT combination in this setting and reinforce our confidence in our late-stage domvanalimab development program. So starting with Slide 5 and a quick reminder of our collaboration with Gilead, our partner for the molecules in our ARC-7 study. We're fortunate to have a unique 10-year partnership. This is hugely enabling and allows us to pursue significantly more molecules, studies and opportunities than we'd be able to do on our own. With Gilead support, which has included $1.3 billion of capital to date and 50% cost sharing for 5 of our clinical stage molecules, we now have an extensive clinical pipeline, including 4 Phase III studies, all targeting billion-dollar opportunities. On Slide 6, I'd like to remind you how domvanalimab or dom works. We engineered dom with a silenced Fc region, optimizing its ability to fully leverage the fundamental mechanism of action for this target, blocking the TIGIT receptor on immune cells and reversing TIGIT-induced immunosuppression in tumors. In contrast, Fc-enabled TIGIT antibodies also tag certain immune cells, bearing TIGIT on their surface for destruction, which can potentially result in negative consequences. These include reducing the peripheral T reg cell population, which could contribute to immune-related events. As Dimitry will discuss, the emerging clinical safety data from ARC-7, including low rates of immune-related events and infusion-related reactions, suggest, in fact, that Dom may have a differentiated safety profile. Briefly, Slide 7 provides the key takeaways from today's data presentation. We saw improvement for the dom-containing arms over monotherapy for every outcome measure we evaluated. Today, we'll focus on ORR and, importantly, PFS, where we saw an impressive 12 months and 10.9 months median PFS for the doublet and triplet arms, respectively. We're going to review all of these points in detail today. On Slide 8, I'll start with an overview of the study. ARC-7 is a randomized, controlled Phase II trial that enrolled patients with first-line metastatic PD-L1-high non-small cell lung cancer and included 3 arms: zimberelimab or zim, which is our PD-1 antibody in active comparator arm; dom plus zim, which we refer to as the doublet; etrumadenant or Etruma, our adenosine-2 receptor antagonist; plus dom and zim, and we refer to as the triplet. Each arm enrolled 50 patients and study recruitment completed in August. PD-L1 testing was performed locally using either the 22C3 or SP263 assay. The primary outcome measures for the study are objective response rate, ORR, and progression-free survival, or PFS. Before I hand things over to Dimitry, I really want to make 3 essential points. First, we've seen an impressive efficacy signal with domvanalimab with median PFS for both the doublet and triplet, each surpassing anything seen historically for anti-PD-1 monotherapy and in fact, surpassing even anti-PD-1 plus chemotherapy. That's KEYTRUDA plus chemotherapy in KEYNOTE-189, and in fact, more recent studies Looking at anti-PD-1 plus chemo and PD-L1 greater than 50 patients, GSK's PERLA study, which also included KEYTRUDA plus chemo. Secondly, this is the first and only prospectively randomized data set, and it's the largest data set for an anti-TIGIT, anti-PDX combination in the setting. Lastly, as you heard during Dr. Johnson's presentation today, we believe that the Fc-silent configuration could be an important differentiator for domvanalimab. I'm now going to turn the call over to Dimitry to review the data.

Thanks, Terry. I'm excited to review today's efficacy and safety data. I spent a significant part of my career in immuno-oncology drug development and it's unusual to conduct a randomized controlled Phase II study of this size. Oftentimes, decisions are based on single-arm data. It's a pleasure to review the data set of this nature, and the data gives me strong confidence in our anti-TIGIT program. I'll begin on Slide #9 with a review of the study population. At this interim analysis, our efficacy population included patients randomized as of June 2022, who are therefore eligible for at least 2 tumor assessments as of the data cutoff date of August 31, 2022. A minimum of 2 scans provided at least an opportunity for patients to achieve a response and to confirm the response. This ITT-13 population had approximately 44 patients per arm. The ITT 13 population included both efficacy evaluable patients and any non-evaluable patients who discontinued prior to their first scan. This latter point is important. The analysis includes all patients who were randomized as of June 2022, including 1 patient who did not receive any study treatment and patients who did not make it to the first scan independent of the reason why. The baseline characteristics are shown on Slide #10. The 3 arms are generally well balanced, including with respect to PD-L1 status. However, we noted a greater population of patients over 65 as well as a higher prevalence of patients with squamous cell histology in the combination arms relative to zim monotherapy, both of these represent prognostic characteristics that might have negatively impacted the combination arms relative to zim monotherapy. A few key points I want to make on our patient population relative to earlier PD-1 studies. First, patient selection in ARC-7 may have been impacted by contemporary treatment landscape with an anti-PD-1 as standard of care. This differs from historical anti-PD-1 trials, that included chemotherapy as active control, which may have resulted in an overall healthier patient population in the older studies. Also, as Terry mentioned, as a Phase II study, local PD-L1 testing was used in ARC-7, and this is in contrast to central testing where the longer turnaround times may also have impacted the nature of randomized populations in other studies. Secondly, the study population was slightly older than historical anti-PD-1 monotherapy trials in the first-line non-small cell lung cancer setting with most studies having a median age around 63. Third, ARC-7 also had a higher percentage of patients from East Asia as compared to most other global PDX studies. Overall, we believe the population heterogeneity observed in the study is highly representative of the real-world practice today. In fact, ARC-7 included more than 70 sites across 7 countries in larger community settings. Now turning to the efficacy data on Slide 11. I want to start by reemphasizing that every outcome measure we evaluated for the ITT-13 population favor the combination arms over the zim monotherapy arm. Confirmed overall response rates in the ITT-13 population were 41% for the doublet versus 27% for zim monotherapy for a 14% differential. Overall response rate for the triplet was similar to that of the doublet arm, and it reinforces the results observed in the doublet arm. The separation on ORR across subgroups is consistent which gives us even greater confidence in the potential for anti-TIGIT therapy to provide benefit for a broad patient population. While the doublet and triplet arms appear similar as of this analysis, we do remain excited about the adenosine pathway, and we will need to continue to monitor progression-free survival and overall survival as the data matures as well as additional data emerging from other proof-of-concept studies that are underway, evaluating the adenosine pathway. Based on our analysis to date, what is known about immunotherapy in general, response rates for some patients may continue to improve with greater maturity of data. Approximately 30% of patients who responded in our study did not respond until the third scan or later. Therefore, some patients ongoing with stable disease right now might become responders as the data continues to mature. As of the data cut off in the doublet arm, there were 6 patients in the ITT-13 population on therapy with stable disease. Moving now to Slide #12 for the PFS results. We observed a median PFS of approximately 12 months for the doublet arm relative to 5.4 for zim monotherapy. For the doublet arm, this translates to an encouraging and clinically relevant hazard ratio of 0.55. As you can see in the Kaplan-Meier curve shown here, the doublet separated from monotherapy before month 2 and the separation was maintained. If we move to next slide, Slide 13, we've overlaid the triplet arm here which, again, is consistent with the results observed in the doublet arm and further strengthens our results in the overall study. Although early, these data were very encouraging with the doublet and the triplet median PFS surpassing the historical benchmarks, both for anti-PD-1 monotherapy and for anti-PD-L1 chemo combinations. In this analysis, the event rate is 61% for monotherapy, 43% and 51% for the doublet and the triplet arm, respectively. While the PFS analysis and the median PFS are still immature, our confidence is driven by the consistency of the signal across both dom-containing arms and across efficacy metrics, including landmark 6-month PFS. Importantly, this latter metric, landmark 6-month PFS is relatively mature, given the median follow-up of 12 months, and we observed 65% and 63% for the doublet and triplet arms, respectively, versus 43% for the monotherapy arm at the 6-month time point. As demonstrated by the overlay on the Kepler-Meier curves, these numbers put the doublet and triplet regimen in the same range as the landmark PFS reported in cityscape. The Swimlane plots on Slide 14 illustrates that 11 patients for 25% had progressive disease as their best overall response on the zim monotherapy arm. In contrast, only 2 patients or 5% and 6 patients or 13% had progressive disease as their best overall response in the doublet and the triplet arm, respectively. This is important, suggesting that the addition of a TIGIT antibody to anti-PD-1 brings the disease under control more quickly. The swimlanes also illustrate a much longer duration of treatment observed to date or the combination arms relative to monotherapy. To summarize our efficacy results, the overall response rate and PFS data showed substantial improvements for the dom containing arms over anti-PD-1 alone. ARC-7 represents the largest data set reported for a randomized trial of any TIGIT plus anti-PDX combination, and we believe the efficacy data clearly supports our investment in a broad dom-containing clinical development program. Turning to Slide 15 for safety now. In general, no unexpected safety signals were observed and both dom-containing arms were generally well tolerated. The incidence of Grade 3 and higher treatment-emergent events was well balanced across the arms, despite longer time on treatment for the doublet and the triplet arm. As we noted earlier, unlike dom, Fc-enable TIGIT antibodies can deplete TIGIT-bearing cells, which includes T regulatory cells, as well as other immune cells contributing to immune-related events. As shown on Slide #16, immune-related AEs were similar in both of the combination arms as compared to zim monotherapy. Specifically in the doublet arm, we saw no increase in the incidence of rash, pruritus or pneumonitis. For the triplet arm, we saw an increase in incidence of Grade 1 and 2 rash and pruritus, which tended to occur early in treatment and resolve with treatment of topical steroids and generally did not lead to dose reductions, interruptions or discontinuations. Additionally, a 4% incidence of infusion-related reactions observed for the doublet arm was noted on the prior slide. This compares favorably to rates observed for competitive anti-TIGIT plus anti-PDX combinations. Infusion-related reactions are uncomfortable for patients and may result in more intense patient monitoring and treatment. If these observed rates continue to stay low, this could be an important differentiator for our regimen. To conclude, I couldn't be more excited about the data and the potential impact of these findings for the field of lung cancer medicine and for patients in need. We are thrilled to have had the opportunity to present these results in a medical forum today, and I'm grateful to Dr. Johnson for all her support. And for now, I'll hand it back over to Terry.

Thanks very much, Dimitry. With the current ARC-7 data set that Dimitry shared in hand, we, along with our advisers, are extremely excited about realizing the opportunity to further advance a potentially novel IO backbone treatment for patients with PD-L1-high non-small cell lung cancer. And while we did review the data on this call, we're also encouraged by the intriguing responses that Dr. Johnson presented today in the ASCO Plenary Session for patients who had progressed, confirmed progressors on zim monotherapy and crossed over to the triplet arm. As Bilal will address, we and Gilead are progressing dom plus zim across 4 Phase III trials, including 3 in lung cancer settings as well as multiple Phase II platform studies, all of which are addressing large patient populations with multibillion-dollar opportunities. Finally, I want to thank the employees as well as the physicians and patients participating in our studies. We're looking forward to advancing our portfolio as efficiently as possible into bringing improved treatment options to cancer patients worldwide. I'd now like to ask our collaborator, Bilal, to say a few words.

Thanks, Terry and Dimitry. First, I want to acknowledge and extend our sincere appreciation and thanks to all the colleagues at Arcus and Gilead, who worked tirelessly to get this important data set up. We are very pleased and excited to have this opportunity to share this important randomized data set and finish the year on a high note for our program for TIGIT as a mechanism and most importantly, for patients. Now moving to Slide 18, I want to take a moment to put ARC-7 trial into context. As Dimitry mentioned previously, ARC-7 is a prospectively randomized, large Phase II study evaluating 2 dom-containing regimen against anti-PD-1 monotherapy in PD-L1 high metastatic non-small cell lung cancer. The study was designed as a Phase II proof-of-a-concept study to estimate safety and potential treatment benefit rather than to conduct a formal statistical testing and is not powered to show statistical significance. ARC-7 IA4 data give us full confidence in our joint dom-zim clinical development program and the TIGIT pathway. We are excited by the efficacy results seen in both dom-containing arm. Notably, both doublet and triplet combinations demonstrated clinically meaningful improvement in all efficacy measures compared to zim monotherapy. Additionally, we are encouraged by the consistency of safety and efficacy data for dom-containing arm across multiple analysis in the ARC-7 study. Anti-PD-1 monotherapy is the current preferred treatment option in first-line PD-L1-high metastatic non-small cell lung cancer. Although the point estimates of efficacy measure vary from trial to trial even for the same anti-PD-1 antibody, the median PFS ranges from 6 to 8 months. Basically, it means that more than 50% of the patient would have a disease progression or death within that time window. The observed median for mPFS of 11 to 12 months with dom-containing doublet and triplet is quite remarkable and have not been seen with any anti-PD-1 monotherapy or anti-PD-1 in combination with chemotherapy in this patient population. As an oncologist, being able to tell advanced non-small cell lung cancer patients that they could potentially have more than 50% chance that they would not need a second-line treatment like chemotherapy for about a year is a remarkable advancement. Of course, these data need to be confirmed with longer follow-up and in the ongoing Phase III registration studies that we are planning. With respect to safety, we have seen lower rate of infusion-related reaction in both dom-containing arms. This compared favorably to rates observed with competitive TIGIT molecules. Based on the totality of the data seen to date, we are very confident that domvanalimab with the Fc-silent design has the potential to be differentiated compared to other anti-TIGIT molecules in this space. We are moving very quickly with our partners in both proof-of-concept studies as well as late-stage trials, including 3 ongoing Phase III trials in non-small cell lung cancer. Our goal is to establish dom-zim regimen as the IO backbone for combination with chemotherapy and other novel agents. Before we open up for Q&A, I'll move to Slide 19 and briefly highlight our broad collaboration portfolio of Gilead and Arcus across the 4 assets. Notably, we're rapidly advancing and executing a lot late-stage clinical program for dom. We have 3 ongoing Phase III studies targeting different indication and setting in advanced non-small-cell lung cancer. ARC-10 will be a confirmatory study of ARC-7 data in the first-line PD-L1-high setting, where zim and dom will be compared to single-agent pembro. STAR-121 is the first-line PD-L1 all-comer patient population where dom plus zim plus chemo will be compared with pembro plus chemo. And in PACIFIC-8 study, dom in combination with durvalumab is being compared with durva alone in Stage III PD-L1-positive non-small cell lung cancer in post-chemo RT setting. This trial is being operationalized by AstraZeneca. The proof-of-concept data from ARC 7 IO 4 that you saw today reassure our expectation and excitement from this important investment, and these Phase III trials are important pillars of our overall lung cancer strategy. In addition, we have initiated STAR-221, a Phase III trial evaluating dom plus zim plus chemo in first-line gastric, GI junction and esophageal adenocarcinoma. This study represents the first Phase III trial of TIGIT combination in upper GI adenocarcinoma, an area of huge unmet medical need. Dom, zim and chemo will be compared to nivo plus chemo in STAR-221. All 4 of the Phase III studies have standard-of-care anti-PD-1 and PD-L1 antibody in the active comparator arm to maximize the competitiveness of our data sets. And thanks to our partnership, we are maximizing the speed of execution to get these promising combination to patients. To summarize, ARC-7 IA4 results show a substantial and clinically meaningful ORR and PFS improvement for our anti-TIGIT and anti-PD-1 combination over anti-PD-1 monotherapy and PD-L1 high advanced non-small cell lung cancer. We're excited to be working alongside Arcus to accelerate the development of this robust portfolio of molecule with the goal of substantially improving cancer outcome for patients. With that, I'll turn it over to Jacquie Ross, Gilead's Vice President of Investor Relations, to open our Q&A session.

Thank you, Bilal and our Arcus colleagues. As we move to Q&A, I'm pleased to welcome Jen Jarrett from Arcus and Bill Grossman from Gilead to the panel, ready to address your questions. [Operator Instructions] So our first question will come from Robyn Karnauskas from Truist.

So for some of us asking about how could ORR and PFS change over time, given the medium follow-up was around 11.8 months, how do you see that changing over time, especially given less frequent scans? And lastly, you mentioned specifically, you're excited about the triplet, you remain excited by the presenter, so the 2 arms will not look different. Are you seeing deeper responses with the triplet?

Yes. Maybe I'll take the first question. What you see here with the median PFS, we can follow up about 10 months is a clear separation of this curve. I mean, we are really encouraged by the medium PFS that we're seeing with both doublet and triplet. I mean I think median PFS in 12 to 11 months that you're seeing here is not something that you've seen with any IO monotherapy trial in the past. I mean the data is maturing. It's hard to predict and speculate where the numbers are going to go. But generally speaking, with the IO therapy, the responses are durable, and these numbers tend to improve over a period of time. So we'll present the data when it matures.

Maybe I'll take the question about the triplet, take first shot at that. I think this is something we talked about actually at IA3. Clearly, both the doublet and the triplet individually are performing well beyond what we see with the monotherapy. But definitely, at this point, it's difficult to distinguish them. As you saw, particularly when we talk about that PFS on the order of 11 to 12 months, and what we've recognized is that we're looking to wait to see more mature and OS data sometime in 2023. With that said, today, Dr. Johnson shared, for the first time, our data set on an initial 12 patients that were confirmed progressors on the monotherapy and then went on to the triplet. We actually saw -- well, this is an exploratory data set. We saw an intriguing a couple of confirmed responses. And that, again, is something that we'll both continue to monitor as those data mature as well as we have more patients that progress that go on to that therapy. That's something that we'll also update on at ASCO in June.

Maybe I'll just add on the only other thing to maybe comment on around the adenosine pathway is just one of several proof-of-concept path studies that we have ongoing, evaluating the adenosine pathway. So obviously, a tumor for this one, and we have also several proof-of-concept trials, evaluating crumbly the upstream CD73 small molecule inhibitors. So we're looking forward to those data sets emerging over the next year to two.

We'll go next to Umer Raffat from Evercore ISI.

Actually Jon on for Umer. Time zone difference here. I'd like to ask about the appropriate benchmark here because KEYTRUDA trials have really a wide range of ORR and PFS benchmarks depending on where you look. So in your mind, what's the appropriate comp for zim monotherapy? And at this point, how confident are you that zim has the strongest possible backbone for your TIGIT combos? Would you consider running combination studies and talk about other PD-1s?

Umer, thanks for the question. And maybe [ I sort of answered ], you really hit it on the nail as you said it. I mean, that's why we do the randomized studies, right? I mean, that's why we randomized particularly in registration studies to actually look at what the contemporary numbers are going to look at. So if you look at across the trial of KEYTRUDA from KEYNOTE-024 to KEYNOTE-042 to KEYNOTE-598 to even the recent data that GSK presented with the pembro plus chemo in combination, the number varies quite significantly from that aspect. So we're very confident with the date -- with zim in a combination. As you know, zim is approved for Hodgkin's disease in China. And the data that we are seeing with monotherapy across the program, we have treated hundreds of patients with zim. It's pretty compatible with any other PD-1 in this setting. And very importantly, I mean, I think what's really important here to note is that zim is in both doublet and triplet arm. And the number that you saw from zim plus dom is pretty impressive, and we are very confident. And our Phase III registration study in ARC-10 will be comparing zim and dom directly to pembro. So hopefully, that will definitely answer the question as we move into it.

And just maybe add on to Bilal's comments again, as both Melissa Johnson and Solange Peters mentioned to you that this zim is performing right in line. There's no discrepancy with what has recently been presented by Solange [indiscernible] from interval data set, too. And as highlighted, the majority of these are community-based centers for ARC-7 as well. So we have no doubt, full confidence in zim as part of the backbone, as Bilal mentioned.

We'll go next to Brian Abrahams at RBC.

It's Lee on for Brian, actually. So congrats on the data. I guess my question is ARC-7 was run largely ex-U.S. So could you elaborate a little more on the geographic differences in patients and the diagnosis or the treatment paradigms or the patient help that could account for the differences between your results and what some other trials have shown with TIGITs and with zim's performance? And I guess, with the population shifting somewhat away from Asia Pacific for ARC-10, I guess, what are your expectations for how the population differences might affect the likelihood of success in that trial? And are there any adjustments you might make to either powering or comparator arms?

Yes. Maybe I'll start it off, and Dimitry, I guess, you feel free to chime in on it. As you know, ARC-7 is a global trial. And I think just to answer to your question directly, I think ARC-10, again, is going to be a global trial. You're right that we have a little higher percentage of patients with -- from APAC region in the ARC-7 compared to the other studies. But again, when we look at the subgroup analysis, I think the benefit is seen across all the groups. So I'm not really concerned about that. I mean, to be very honest, as Dr. Johnson and Dr. Peters has discussed today, the -- ARC-7 really reflects what's really going on in the real world with what you expect to see from IO monotherapy. So I think we are very confident that what we see here in ARC-7 is going to be capitulated in ARC-10. And also to note, I mean, this is like a 2:1 randomization. I mean, you see the similar effect on both doublet and triplet, which actually re-enforces a belief in the dom moving forward. Dimitry, do you want to add anything on that?

Sure, I can add a few things, right? To emphasize again, this is the important reason why we do the randomized trial, why we really focus on the comparison between the different arms in the different trials. I think, again, like we've made multiple comments about the comparability to the real-world data sets. And also if you look at other trials that move through different regions, there's difference in performances. We can speculate on what it is. There's many reasons behind it between smoking history, perhaps histology, there could be genetic factors. But most importantly, again, within the trial where you randomize, you really compare over the active standard of care. In ARC-10, we are comparing to pembrolizumab as active standard of care and with the, let's say, the performance of the doublet both zim, dom-containing doublets, we really feel that the absolute readouts there for specifically the median PFS, let's say, that gives us a lot of confidence in the chances of success for ARC-10.

Thank you. I'll go to the Q&A chat. We had a question from Salveen Richer of Goldman Sachs. There's clearly impressive added benefit from TIGIT here, but the ORR for the doublet is lower than the 50% you previously guided to. What do you attribute this to? Did zim perform below expectations? Or did you expect even more benefit from TIGIT?

Yes, I mean, again, we like to just mention that this is a randomized Phase II study. And I think it's a proof-of-a-concept study, and what we're really looking for is the delta that we see between the arms. And I think the delta that we saw from both doublet and triplet is pretty much in line with what you'd expect from adding TIGIT over the anti-PD-1. And more importantly, I mean, the PFS results that we are seeing here are fairly aggressive. So we're pretty happy with where we are. I mean I'm going to [indiscernible] to add anything on in terms of prior expectations and stuff on it. But this data is proof of a concept for us to move into the registration studies.

Yes. I think you covered it pretty well, Bilal. I mean the gold standard here is again a randomization study, which we did for a proof of concept. And then the consistencies with both dom-containing arms is what's leading to our conviction here. And you can see, as mentioned, but it's worth mentioning it again, it's both the doublet and triplet are performing well above expectations for what you'd expect for a PD-1 monotherapy or for even PD-1 plus chemo in this population when you look at recent data sets from both historical as well as contemporary as mentioned.

Yes. The one other thing I would add, Salveen, when you talk about sort of response rate, the way we look at all these is they're surrogate. So you've got response rate, then you have PFS. Ultimately, what you're looking for is that gold standard for approval, which is OS. Clearly, now we've moved into the realm of PFS being the big driver. But what we would also note when you think about response rate and how that translates, what you'll see in certain cases and we'll see this in our study as well is that you'll actually have patients that have stable disease that benefit enormously and in fact, you might have someone that had a 25% reduction in tumor volume that's classified as stable disease. You might have someone who's 35%, and they're classified as a partial response. But now as you move into the realm of on PFS, they both may be contributing and in fact, benefiting greatly from the therapy independent of just whether they were classified as a responder or not. Hence, the importance of the survival endpoints as the trial moves along.

And then maybe one other thing that I'd point out, going back to the patient populations, and there was 3 areas of difference, particularly between the doublet arm in the monotherapy arm. First was age where we had that median age of 69, which is higher than mono arm, probably higher than any PD-1 mono study you could find. The second thing was a percent of patients with squamous cell carcinoma. So I don't know if you guys remember the data that Solange put up, but it showed that clearly, patients with squamous cell carcinoma don't do as well as patients with that are non-squamous. So it's a different amount of month in terms of PFS of 5.5 versus 4.5 based on the data that she put up there. And then the third thing that I would highlight is percent of patients that had greater than 75 PD-L1 expression was also lower in that doublet arm. So less of the super high PD-L1 expressors in that arm.

Thank you. All right. Let's find out who is playing the role of Peter Lawson at Barclays this afternoon. Peter?

Just as we think about ARC-10, when do we get the initial readout around ARC-10? Does that give us a better idea of kind of the comparability of zim and pembro?

Yes. Maybe I'll take that and maybe have others coming on. We announced recently that ARC-10 is redesigned and trying to compare zim plus dom with pembro based on treating regulatory landscape and a standard of care. So I think we'll provide some updates on the trial as we move along on this. The trial is up and running and accruing right now. So I don't have a specific date for you yet at this point.

I'll just comment. I think we -- again, we've seen a lot of data as was I think Dimitry highlighted in the discussion that we -- whether it's at the molecular level, the characterization of the antibodies, whether it's in the clinic, zim is approved as a monotherapy in China in CHL compared very favorable to the data that you would see with pembro in that setting. And in fact, just recently, a study described in cervical cancer where it even exceed. So we simply have no doubts about zim as an anti-PD-1. And I think Solange also made that point in her presentation based upon everything that's already known, zim is performing just as you would expect an anti-PD-1 to perform.

Yes. Sorry, I just want to add one point here too. I mean the fine estimate that we are seeing here from zim monotherapy arm is within the 95% confidence interval of what you expect from anti-PD1s to perform in this setting. So I think we are pretty overall confident with the data that we've seen yet.

Great. We'll go next to Evan Seigerman at BMO.

Evan playing Evan today. Two questions for me. Do you need monotherapy data from the zim to really get a dom plus in doublet approved. And I can't help but commit the comparison and kind of square with KEYNOTE-024. How should I square away these response rates? I know I'm probably the fourth person asking the question a different way. But when I look at a 27% response rate versus high 40s from KEYNOTE-024, that's a big difference for me. I'm just trying to make sure I'm not interpreting this wrong and kind of figure out what to expect as you go forward.

I'm not sure I hear the exact both questions, but I think the later part, I think you're trying to reconcile that 27% response rate a little bit. Again, just to -- probably just repeating myself a little bit here, but the point estimates of what we saw from zim monotherapy in ARC-7 is like within the range. I mean there are a couple of differences here. I think, as Jen mentioned a few of these, and I am just going to outline. ARC-7 is done, again -- in the -- predominantly in the community setting. The enrollment of the trial is based on local PD-L1 results, which is a little different than the initial pembro studies or other studies where you require central PD-1 confirmation for patients who get on the trial. So I think there's certain level of natural selection to trying to get those patients on the study. And also really to be noted, I mean, the initial registration trial, whether you look at tazo, cemiplimab, or pembro, those comparing IO to chemotherapy. So I think there's a natural patient selection there, the patient will have to be fit enough to get chemotherapy to get on some of those trials. And with IO monotherapy approved now in first-line setting, many physicians in the community are putting a wider range of patients on these trials. And so I think some of these numbers erode and evolve over a period of time. And I think that's something that Dr. Peters was alluding to today. If you actually look at the real-world data, I mean, what you're seeing here with zim is what you expect for a PD-1 monotherapy to perform in that setting. So it can help answer some of your concern on it. I think for me, the most important thing is this is a randomized data set. And what we are really seeing here from zim and dom in terms of the progression-free survival that we're seeing here is not something that we've seen with IO monotherapy in that setting. And we are really excited about it. And obviously, these need to be confirmed in the Phase III study.

I can add a few things. It's something we also mentioned before. Let's say, the recent publication of the PERLA study, it's a GSK- study where pembrolizumab plus chemotherapy is to control on just like we like it was the experimental arm in KEYNOTE-189. And that's a more contemporary trial that ran years later in different parts of the world, and the performance of pembro is shifting in that trial closer to the real-world data as we also have alluded to. So that gives us a lot of confidence when you look at trials that were run 5, 6, 7 years ago, patient selection was very different because the standard of care in those days was chemotherapy. And as Bilal already mentioned, patients getting onto those trials had to be fit for chemotherapy. In our trial, one of the things we noted, and I mentioned in my part of the presentation, is the median age in our trial is older and older. It's probably a [indiscernible] of the fact the patients who might not have gotten on to original trials earlier on because they were not eligible for chemotherapy and now they are able to get on to the trial because there's no longer chemotherapy control. The first part of your question was hard to understand. I think you were asking about zim as a combination of the regimen and the approval. We would refer to that as the contribution of components discussion. We've generated a very large data set of zim, not just in this trial, of course, specifically in this population, but we treated more than 600 patients with zim. We have large data sets in cervical cancer, in classical Hodgkin's lymphoma. And all of that data is used to, let's say, show the contribution of components when we have discussions with the regulators. And all our trial designs are discussed with the regulators before we start them to make sure they agree with our approach for approval.

Add on to what Dimitry said about the age of the patient population is shifting, that's exactly what Solange Peters also highlighted in the real-world data set from ESMO IO as well for the monotherapy PD-L1-high population in treatment. And just to reiterate, as most people get back from the health authorities, the primary endpoint for frontline non-small cell cancer is still OS at the end of the day, and we're taking on standard of care in all of our studies.

We'll go next to the line of [ Zee Shem Shu ] from Berenberg.

Great. I have a 2-part question. The first part of the question is around the -- one of the points that the discussing raised around the need to change the standard when they can change the standard of care is to have a definition of CD226 pathway biomarker. I just want to hear from you some of the ongoing biomarker work that has been ongoing on ARC-7. And then the second part of the question is around -- for median PFS of 12 months in a doublet arm is quite impressive relative to other PD-1 monotherapy. I guess we've been hearing some comparison of this number to a PD-1 chemo performance. I was wondering your thoughts on that comparison. Is it relevant at all in today's world?

Yes. Maybe I will tackle the second question first, and then maybe we get to a biomarker question a little bit, too. If you remember, I think last ASCO, I think FDA presented the data set looking at IO plus chemo versus IO and TPS, more than 50% [indiscernible] setting. And basically, what it shows is it's both are pretty much the same when they look at things together. And I think Dr. Peters today also showed in her discussion that point. So yes, there's use of both IO monotherapy as well as is IO plus chemo in that setting. But majority of the use in this setting is IO monotherapy in that setting. And that's an acceptable standard of care in that setting to design the trials. In terms of the biomarker question, maybe I'll hand it over to Dimitry and Terry to jump on it. I think there's an extensive plan in ARC-7 to look at the biomarkers and we'll report some of those as the data emerges in the future.

Yes, I can say a few words. So we are looking at TIGIT CD226, CD155, CD73. And there's a number of other things, but those are the first things we'll be looking at and that will be part of the future presentation.

And just to go back to your point about looking at IO plus chemo as a comparison, one of the things that we're so interested about our data is, at least based on the data that we've generated here, there's a possibility that we're even better than PD-1 plus chemo. And there is a percentage of patients as Bilal was referring to, they do receive PD-1 plus chemo, even in that PD-L1-high patient population because they've got really aggressive disease. We didn't show those spider plots, but when investigators see the spider plot, that's one thing that catches their eye, there seems to be fewer rapid progressors. And so something that we've heard repeatedly from them is maybe there's a potential to capture that percentage of the patient population that has really aggressive disease, and today, would receive PD-1 plus chemo based on what we're seeing so far in our results.

Great. We'll go next to Mohit Bansal from Wells Fargo, please.

This is James on for Mohit. During Dr. Peters' discussion, she showed ARC-7 sharing some of Cityscape's positives and negatives. That said, is there anything from ARC-7 that gives you confidence that ARC-10 will not end up sharing Skyscraper-01's outcome? And then secondly, I think it's mentioned that as dom moves into a larger Phase III trial, there's likely to be some PFS erosion. Can you share how much PFS erosion you anticipate? And I only ask this because Cityscape seems to have a large PFS cushion, perhaps larger than ARC-7, yet it's still missed in Sky-01.

That's like to seem to be taking the use -- thanks for the question, and maybe I'll take it. By design, ARC-7 and Cityscape are different, right? I mean, Cityscape was designed with TPS more than 1%, and TPS more than 50% was a subset analysis. And that -- I mean that's 1 select distinction here. I mean, ARC-7, this is a prospective randomization in TPS, more than 50% patient population. The other one, I mean, obviously, the sample size, which makes us a lot more confident. I mean, we are much larger compared to what you saw with the Cityscape. And the third line for me is we have 2 arms here, doublet and triplet which are independently evaluated. And I think the data looks pretty consistent on this. So that give us a lot more confidence in that setting. Again, I've been in oncology long enough. I mean, this erosion do happen in Phase II to Phase III, but also if you appropriately conduct the Phase II, some of the data from Phase II can be recapsulated in the Phase III. And you saw that with some of the pembro data. A lot of the stuff in the Phase II get into Phase III and carried. So I'm not going to completely count on the fact that there's always going to be that erosion. I think it really depends on the sample size, how you conduct these trials. And we are pretty confident in what we saw in ARC-7. Bill, do you want to add anything?

I would just add that Skyscraper-01 is still not done, right? So it's marching on towards OS endpoint, which is the primary endpoint required for approval in the setting. So -- and I know that the PFS in the first interim was a potential disappointment to some. As mentioned, they have already -- they didn't mention that they have a clear numerical enhancement with their addition of TIGIT. But we'll have to wait and see how the rest of the study plays out.

I'd just like to get a comment on the delta because what you're talking about is comparing Cityscape to Sky-01. You're talking about comparing ARC-7 to ARC-10. And I think at the highest level and the most simple answer is what was described, the fact that Cityscape was a post-hoc analysis, and then going to Sky-01, whatever that outcome, to Bill's point, we don't really know the answer. But the one thing that if we just take everything that Genentech says at face value, probably a quantitative difference. And I think one of the key things to note is that they're such different experiments. So Cityscape clearly pointed to the qualitative advantage of adding anti-TIGIT to anti-PDx. And I think that's going to be borne out time and time again. But in going from Cityscape to Sky-01, it's a very different experiment. So to expect them to be quantitatively in the same range is probably stretched. Whereas ARC-7, to the points that were made here, in fact, even the patient population very Phase III like. And then it is a randomized prospective trial, and looking very much like the design of our [ plan ]. So we feel very confident that you really couldn't have had a better progenitor to a Phase III trial when you think across the landscape, not only with the 2 arms that had the dom in them, but just the overall design about as good as you can have in going from a Phase II to a Phase III setting.

I'd like to add one more sort of commentary on that is we also have different TIGIT antibodies, right? I think Jen covered nicely in the presentation that depending upon the target, how you want to engineer the Fc component of the antibodies. And so we'll see in our Phase III trials and how that plays out, too. We know that TIGIT expression is upregulated upon activation in T factor cells. As a reminder, we had 2 antibodies that we've been watching the data emerge. One is an Fc competent IgG1 wild-type antibody, and dom is the FC silent. We -- as we've seen the dom data continue to progress, we've taken that into the pivotal trials for a reason. So we'll see how they play out.

We'll go to Matthew Harrison at Morgan Stanley.

Great. I guess, just following sort of on the theme of some of the last questions, how does the outcome of Skyscraper impact your view on the program? If it doesn't hit on OS, does that change your outlook at all just given these data and how you think about investment in the program?

Yes, I'll answer this from the clinical development perspective and have others chime in on it. I mean we designed a lot of these trials at risk because we believe in the pathway. And I think the data that you're seeing today from IO 7 for ARC-7 really reassure us that the investment that we are making is the right one, and that's going to be important for the patients. So I think for me, from the clinical development perspective, really moving forward is really the execution to get these trials executed. Obviously, external landscape changed. I mean I've been in IO business for a while. I mean we all waited for CheckMate, [indiscernible] to read out and change the world, that didn't happen. So each compounding program is a little different. They are subtilities to and how things are conducted. At least for me from a clinical development perspective, I don't think those results will change. And obviously, hope that those are positive, so that would be good for the patients and believe in a pathway, but we'll see how it play out.

Great. We'll go to Yigal Nokamobitz from Citi.

So just on ARC-7 trial design. So obviously, it was an open-label trial. Just curious your thoughts on that. Some people are highlighting that to discount some of the findings. But actually, I would argue the opposite that since the triplet didn't behave better than the doublet, that, that result would seem to substantially limit criticism around inherent bias. So we're doing agree with that interpretation and any further elaboration with respect to the fact that ARC-7 was open label?

Yes, I can take this question. I think that there's always discussions about open-label versus placebo controlled. I think you really have to think about what is the ultimate thing you try to achieve with that. If there's a specific, let's say, idea about it, let's say, a tolerability profile, safety issues, that is probably a very good reason to blind the trial. We did not have that. And as we presented in the safety analysis, we're very confident about our safety profile with limited added toxicity. If you look at, let's say, a registrational trial perspective, if you have an overall survival endpoint, then there's really no, let's say, no, let's say, reason for efficacy bias to have a blinded trial. The other thing that's a consideration is the patient perspective is you actually have, let's say, multiple agents, you have to give the patients a placebo infusion, placebo tablets. So it does put a significant burden on the patient. So overall, I believe, for every single trial, you need to make, let's say, an assessment. Is it really beneficial with the burden you put on the patients? Additionally, do you get more of an unbiased answer out there? And I think for the ARC-7 trial, that was definitely not the case. And something -- a point we made before the fact that we have 2 experimental arms comparing to zim monotherapy gives us more confidence in the overall assessment. So I hope that answers your question.

Great. We have just a few minutes left. So we're going to try and squeeze in a few more. We'll go next to Tyler Van Buren, please, at Cowen. Tyler?

Great. So for the STAR-121 study, can you discuss what gives you guys confidence in the study design, which tests the dom zim chemo triplet versus pembro chemo? And particularly as we think about the contribution of dom on top of PD-1 chemo in the lower PD-L1 expressing subgroups? And I guess, specifically, how much data do you have in the lower expressing patients? And did you do dose ranging with chemo?

Yes, maybe I'll try to check on this. I mean I think the general belief now, I think you hear from Dr. Peters today too, is the way that TIGIT is behaving is very, very similar to what the anti-PD-1s behave in the early days. So I think one of the main reason why we are trying to combine dom and zim together with the chemotherapy is really to broaden that patient population in the larger data set. The -- we will generate some data. There are some ongoing platform study that we are looking at different combinations in this setting. But based on what we have seen so far, we are pretty confident that STAR-121 was designed to definitely answer the question in that setting. And particularly, I think the IO plus chemo is -- have a place in TPS less than 50% population. And it's an all-comer study. So I mean I'd love to see what others want to add on it, but I think we are pretty confident in the design. And we actually interact with the regulatory authorities to get to where we are.

I think that what's interesting is if you think about proof of concept, so the biggest proof of concept that we have as a field is what we're talking about here in the frontline patient population. And what the biology points to is really the way you want to think about this is anti-TIGIT is turning anti-PDX into super anti-PDX. So when you start to think about your proof-of-concept and what's the most logical biologically and hence, clinically, going right down the line and going at that -- pushing the PD-L1 lower in doing that with chemo, but then bringing on what essentially becomes a single entity. That super anti-PD-1, that's anti-TIGIT plus anti-PD-1, is probably right down the middle of the fairway when you think about where you want to go next. And so we love that trial. It's a great trial. It's going right directly at KEYTRUDA. We've got an arm in there where we'll have direct side-by-side data of zim chemo that we can look at qualitatively versus KEYTRUDA chemo. It's a great trial in terms of enrollment. It's a great option for patients. So we think it's as exciting of a trial as we can really conceptualize on all fronts, clinically, market need, et cetera. So it's a trial that we think is really well supported by the data that are out there.

Great. We're at the top of the hour, unfortunately. We'll take 1 last question from the chat, it is can you discuss the depth of response over time from both the dom combo arms?

Yes. I can say a few things about that. I think looking at, let's say, the swimlane block, we provided, you can see, I would say, a more area under the curve when you go from single to double to triplet, meaning more tumor shrinkage for, let's say, treatment -- more intense treatment. One important thing that we've made before is about the dynamics of response. We know that for PD-1, responses tend to occur relatively fast, but it can take some time. We have seen in our study, as I highlighted, that 30% of our responders doesn't respond until the third scan. So that gives us confidence that part of what we see is not only deepening of response, but also an additional, let's say, time line of response. But it goes both ways because the other thing I mentioned is that we see fewer primary progressions, something that is often seen with other treatments. When you intensify treatment, you have multiple mechanisms acting on the tumor. You tend to see fewer primary progressions, which is a pretty, let's say, important finding as well. That means also, I'd say, a faster response for some patients. But more importantly, primary progression is a very poor prognostic sign. When patients primarily progress on a treatment, it's really hard to give them any benefit from a subsequent treatment. So I think these are all important observations in the trial and hopeful it will, let's say, pan out, let's say, provide even more benefits in the long run.

Super. Well, clearly, we have more questions than we had time to accommodate. So please don't hesitate to reach out to Pia at Arcus or to the Gilead IR team if you have additional questions. Thank you all for taking the time, and we wish you a very happy holiday season. Good afternoon.

Thank you, everybody.

Thank you. Bye-bye.