Hello, and welcome to Arcus Biosciences First Quarter 2024 Earnings Call. My name is Lydia, and I will be your operator today. I will now hand you over to Pia Eaves, Vice President of Investor Relations and Strategy.

Hello, everyone, and thank you for joining us on today's conference call to discuss Arcus' First Quarter 2024 Financial Results and Pipeline Updates. I'd like to remind you that on this call, management will make forward-looking statements, including statements about our cash runway and our expected clinical development milestones and timelines. All statements, other than historical facts, reflect the current beliefs and expectations of management and involve risks and uncertainties that may cause our actual results to differ from those expressed. Those risks and uncertainties are described in our annual report on Form 10-K and quarterly report on Form 10-Q, which have been filed with the SEC. We strongly encourage you to review our filings. Today, you'll hear from our CEO, Terry Rosen; COO, Jennifer Jarrett; and CFO, Bob Goeltz. We'll also be joined by our CMO, Dimitry Nuyten; and President, Juan Jaen, for questions after the prepared remarks. For ease of listening, we will refer to abbreviations of our molecule names: domvanalimab as dom; zimberelimab as zim; casdatifan as cas; quemliclustat as quemli; and etrumadenant as etruma. During today's call, we'll refer to slides in our corporate deck, which can be found on the Investors section of our website. With that, I'll turn the call over to our CEO, Terry Rosen.

Thanks very much, Pia, and thanks to all of you on the call for listening in today. As you know, 2024 is shaping up to be an incredibly catalyst-rich year for Arcus. By the end of this year, we'll have data to support all 4 of our later-stage clinical programs: dom-zim in lung and upper GI cancers; cas in clear cell RCC; quemli in pancreatic cancer and etruma in colorectal cancer. We'll spend most of the call today setting the stage for these upcoming data events. With over $1 billion of cash on hand, a runway into 2027, partnerships with Gilead, AstraZeneca, Taiho, and others, along with a very diversified pipeline, we're extremely well positioned to capitalize on these data sets and advance our potential first-in-class and best-of-class treatments towards approval and commercialization quickly and efficiently. Let me start with ASCO. ASCO is less than a month away. So we're almost there. We're thrilled and honored to have two oral presentations, both of which will provide strong support for our efforts and programs in the GI cancer field. Importantly, both data sets are in settings where there's limited competition and huge unmet need. These are genuine opportunities to make a meaningful difference for patients. First off, on Saturday, June 1, we'll have updated data from cohort A1 of the Phase II EDGE-Gastric study evaluating dom plus zim plus chemo in first-line gastric cancer. As you're aware, dom is the only Fc-silent anti-TIGIT antibody in late-stage clinical development, and we believe that the data presented to date indicate that dom may potentially have an improved safety profile, when combined with chemotherapy relative to that of the Fc-enabled anti-TIGIT antibodies, when they're combined with chemotherapy. As a reminder, we presented initial data from this cohort of EDGE-Gastric at the ASCO virtual Plenary Session in November of last year. At the time, median PFS was immature. However, we did present mature landmark 6-month PFS numbers, which you can see on Slide 16 of our corporate deck. What you can see is that 6-month landmark PFS was 77% for the overall population and 93% for the PD-L1 high population. Given that the median PFS for standard of care in this setting ranges from 7 to 8 months, these data were obviously very, very encouraging. At ASCO, we're very excited to be presenting mature median-PFS data. We expect the updated data will further support the potential for dom-zim to provide clinically meaningful benefit relative to the standard of care in gastric cancer. Importantly, EDGE-Gastric evaluated the same setting and similar patient population as our ongoing Phase III study STAR-221. Therefore, we expect these data to foreshadow our confidence in our STAR-221 study. Enrollment in STAR-221 is expected to complete by midyear. The incredibly rapid enrollment of the study is indicative of the lack of competition in the gastric cancer market and the immense need for new therapeutic options, particularly with overall survival in this patient population ranging from only 13 to 14 months in studies with anti-PD-1 antibodies in chemotherapy. We also felt that we actually achieved some tailwinds with our data presentation at the end of last year. So putting this all together, you can infer that there's a line of sight to data and that dom-zim will have a very substantial head start over potential competitors, given there are no other anti-TIGIT antibodies in Phase III development for the same. We think we're going to have a clear first-to-market advantage. This creates an exciting opportunity for us to be first in this setting with an addressable patient population of over 25,000 patients in the U.S. alone and 100,000 patients in the G7 countries. This equates to the potential worldwide market of over $3 billion. Now moving on to our second presentation at ASCO. On Sunday, June 2, we'll be presenting data from our ARC-9 study in third-line colorectal cancer. This will be the first presentation on this study. The data will be from Cohort B, which is evaluating a tumor in combination with zim, FOLFOX, and bev versus regorafenib, one of the standard of care treatments for third-line colorectal cancer. On Slide 41 of our corporate deck, you can see the study design and the conclusion criteria for this portion of the study. Patients in this cohort must have received bev unless contraindicated, a prior oxaliplatin-based regimen and additional regimen. These are the current standard of care therapies for first- and second-line CRC. 105 patients were enrolled who were randomized 2:1 between the etruma arm and the regorafenib arm. This is a relatively large data set. ARC-9 also includes 2 additional randomized cohorts, which evaluated etruma plus zim plus FOLFOX and bev versus FOLFOX and bev in second-line colorectal cancer. These data are not yet matured and will be presented at a later time. The second-line setting, as you know, has a substantially longer OS. The ASCO presentation will include mature PFS and OS data with a median follow-up of over 20 months. Our data will include patients with and without liver metastasis. This is important since patients with liver metastasis tend to have a poor prognosis, and they're not always included in late-line CRC trials. As many of you know, there are very limited options in third-line plus colorectal cancer. Patients are typically treated with regorafenib and more recently with the combination of Lonsurf and bev based on the SUNLIGHT study, which showed a PFS of 5.6 months and OS of 10.8 months in the third-line patient population. While acknowledging the limitations of cross-trial comparisons, we all do it based on our data, which will be shared with you next month, our combination regimen may represent a very substantial improvement over current options for patients in the third-line setting. I also want to point out a small data set of 35 patients that was just presented in a poster at AACR, which captured quite a bit of interest and further supports our hypothesis that adenosine blockade enhances the immune-activating benefits of chemotherapy. These data were from the MORPHEUS-PDAC study, a randomized Phase Ib/II trial operationalized by Roche, that evaluated our molecule etruma plus Roche's anti-PD-L1 atezo in chemotherapy versus chemotherapy alone in first-line metastatic pancreatic cancer. We've highlighted the design on Slide 37. On Slide 38, we've shown the spider plots for both the control arm and the etruma-based regimen, which showed durable responses. In this trial, the etruma-containing arm demonstrated a meaningful improvement in both PFS and OS. We've shown these data on Slides 39 and 40 of the corporate's deck. Specifically, the PFS hazard ratio was 0.48 and the OS hazard ratio was 0.67 with the etruma-based regimen yielding an absolute improvement in median OS survival of 4.4 months over chemotherapy alone. The median OS for the etruma-containing arm was 16.5 months, very similar to what we saw in ARC-8. Before we leave etruma, I want to highlight that now with the ARC-9 data presentation, we will have 3 data sets presented in a very short period of time for our 2 molecules that inhibit the ATP adenosine pathway, quemli and etruma. Between ARC-8, which evaluated quemli in combination with chemotherapy in first-line pancreatic cancer, MORPHEUS-PDAC, very similar, which evaluated etruma with chemotherapy in first-line pancreatic cancer. And now ARC-9, we have 3 independent but similar data sets, which together provide compelling evidence demonstrating that mitigating the immunosuppressive action of adenosine, combined with immunogenic chemotherapy, may prolong survival relative to that associated with chemotherapy alone. For those of you who've been following us for a long time, keep in mind, this was the original hypothesis that drove us to the adenosine axis in the first place. So we're getting to the point where the data are matching the hypothesis. Importantly, two of these studies showed a meaningful improvement in OS versus the standard-of-care control, and all 3 showed a substantial improvement of our historical benchmarks when adenosine blockade was combined with immunogenic chemotherapy. Now that we've covered what we'll be sharing at ASCO, I'd like to turn it over to Jen to discuss expectations for our HIF-2 alpha inhibitor program later this year.

Thanks, Terry. I'll now turn to our data events for the second half of 2024 and specifically for cas, our HIF-2 alpha inhibitor, which we are developing in clear cell renal cell carcinoma, or ccRCC. As a reminder, HIF-2 alpha is a validated mechanism, with belzutifan recently approved as monotherapy for baseline clear cell RCC patients. Cas has PK and PD advantages over bel, which should enable it to hit the target harder with the goal of achieving greater clinical efficacy than that of belzutifan. As we talked about on our last earnings call, there are multiple opportunities to improve upon the profile of belzutifan. A lower rate of primary progressive disease, a higher overall response rate, and more prolonged responses, any of which could translate into a higher PFS and ultimately longer OS for cas relative to that of belzutifan. Our Phase Ib study for cas, ARC-20, now has enrolled over 80 patients. It was designed to answer numerous questions, so I want to spend a minute describing the various components of this multi-cohort study, which is summarized on Slide 27 of our corporate deck. First, for the dose escalation, as of our last earnings call in February, we had completed enrollment of the 20 mg, 50 mg, and 100 mg dose cohorts with no dose-limiting toxicities observed. We have now completed enrollment of the 150 mg cohort. Again, we did not observe any DLTs, and we just cleared that dose. Additionally, we continue to see linear dose proportional PK for cas even at the 150 mg dose. The dose expansion portion of ARC-20 was designed to serve a few purposes. First and foremost, to generate data for a proposed Phase III dose of 100 mg of cas. These data will be used to support the initiation of our first Phase III study, which is an advanced stage treatment for cas. The monotherapy dose expansion portion is enrolling patients that have received at least 1 prior anti-PD-1 therapy and at least 1 TKI. Approximately 1/3 of the patients have received 4 or more prior lines of therapy. So the patients were relatively advanced. On our last earnings call, we disclosed that while the majority of patients in this cohort had only received 1 or 2 scans, we were already seeing a response rate, including responses pending confirmation, in line with LITESPARK-005, the Phase III study for belzutifan. We also mentioned that we have observed a lower rate of primary progressive disease where the percent of patients that progressed prior to their first scan was reported for LITESPARK-005. This is important as one weakness in the LITESPARK-005 data set, as we've heard consistently from clinicians, is the high rate of primary progressive disease. Bringing this rate down should result in more patients benefiting from treatment to prolong PFS and survival. We are very excited to present detailed data from the cohort at a medical conference in the second half of this year. We also designed the dose expansion portion to satisfy the dose optimization work required for regulatory submissions. Specifically, we included 2 additional monotherapy cohorts in ARC-20, which are evaluating a 50 mg dose and 150 mg dose. We have now completed enrollment of the 50 mg cohort, and we just initiated enrollment of the 150 mg cohort. So in addition to enabling us to complete the dose optimization work required for Project Optimus, these cohorts will also generate a lot of additional data that will further elucidate the clinical profile of cas. In addition to the 3 monotherapy expansion cohorts in ARC-20, we are about to initiate enrollment of a cohort to evaluate cas in combination with cabo, the most commonly used TKI in clear cell RCC. Data generated from this cohort, in addition to the data generated from STELLAR-009, our Phase II study being operationalized by Exelixis, will be used to support the planned initiation of at least 1 Phase III study for cas in combination with the TKI. Given the enthusiasm for cas and the rapid enrollment of ARC-20, we expect our 100 mg dose expansion cohort data presentation later this year to be followed quickly by data from the various other cohorts. We are aggressively advancing cas towards the initiation of our first Phase III trial early next year. Importantly, we are taking a strategic approach to our development plan to maximize the value of what we believe is a best-in-class HIF-2 alpha inhibitor. For example, the STELLAR-009 study is evaluating cas with a next-generation and potentially best-in-class TKI with belzutifan. We believe that both zanza and cabo could have advantages over belzutifan’s TKI partner, lenvatinib, with a better tolerated AE profile. Therefore, we expect our combination of cas and TKI to be differentiated from both an efficacy and safety and tolerability perspective relative to belzutifan plus lenvatinib. I'd like to end by spending a few moments on the RCC market, of which approximately 80% of patients have clear cell histology. The annual addressable patient population for first-line clear cell RCC includes over 12,000 incident patients in the U.S. and approximately 30,000 in the G7 countries. Around 2/3 of these patients progress on first-line treatment in our addressable second line, resulting in more than 8,000 patients in the U.S. and approximately 20,000 across the G7 countries. The ccRCC market is fragmented, and cabo is the most frequently used TKI. Cabo sales in 2023 in the U.S. alone were over $1.6 billion, driven almost entirely by RCC, with only 1/4 of market share in the first line and under 50% market share in the second line study. In terms of the treatment paradigm for first-line RCC, patients are typically treated with anti-PD-1 plus a TKI or a combination of nivolumab plus ipilimumab. When patients experience tumor progression on first-line therapy, they cycle through different TKIs frequently for a few years or more. With few options beyond TKIs, the clinician community has been eagerly waiting for the introduction of HIF-2 alpha inhibitors, particularly given the relatively benign safety profile relative to TKIs. Not surprisingly, the positive Phase III data for belzutifan resulted in an immediate and steep inflection in script trajectory. Monthly scripts for belzutifan are now at an approximately $400 million run rate in the U.S. alone and growing. As we've discussed, we believe that cas may improve clinical outcomes relative to belzutifan, and we plan on developing cas in differentiated combinations, all of which support a very meaningful market opportunity for cas, which we believe is north of $2 billion. Before we conclude, I'll now turn the call over to Bob to review our quarterly financials.

Thanks, Jen. Arcus continues to be in a strong financial position. Our cash as of March 31, 2024, was $1.1 billion as compared to $866 million as of December 31, 2023. Turning to our P&L, we recognized GAAP revenue for the first quarter of $145 million, which compares to $31 million in the fourth quarter of 2023. Our revenue is primarily driven by our collaborations with Gilead and Taiho, and in the first quarter included a cumulative catch-up of $107 million resulting from the Gilead amendment we executed in January. We expect to recognize GAAP revenue of approximately $30 million per quarter for the remainder of 2024. Our R&D expenses for the first quarter are stated net of reimbursements from Gilead and were $109 million as compared to $93 million in the fourth quarter of 2023. In the first quarter, non-cash stock compensation represented $10 million of our R&D expenses. The increase in the first quarter was related to higher clinical manufacturing, clinical trial, and headcount-related costs associated with our late-stage programs. We continue to expect modest increases in R&D expenses as our Phase III studies mature and spend will fluctuate primarily based on the timing of clinical manufacturing activities and the purchase of standard of care therapeutics for our clinical trials. G&A expenses were $32 million for the first quarter compared to $29 million in the fourth quarter of 2023. Non-cash stock compensation represents $10 million of our G&A expenses for the first quarter, and we expect G&A to remain stable for 2024. Total operating expenses in the first quarter were impacted by a $20 million non-cash impairment charge resulting from the intended sublease of a portion of our office space. Finally, we continue to expect our cash balance at the end of 2024 to be between $870 million and $920 million and to fund operations into 2027. This guidance includes a $100 million partnership continuation payment from Gilead in the third quarter and excludes potential opt-in payments and approval milestones from our partners. For more details regarding our financial results, please refer to our earnings release from earlier today and our 10-Q. I'll now turn it back over to Terry for concluding remarks.

Thanks, Bob. As you've heard from us today, we are genuinely at a significant inflection in the evolution of markets. With the completion of enrollment of STAR-221 first-line gastric cancer expected by midyear, we will return our attention to our first Phase III readout. Behind this, we'll have Phase III data from our second registrational study for dom-zim, STAR-121, in first-line non-small cell lung cancer, with enrollment for this trial also expected to complete this year. In addition, we'll be starting at least 2 additional Phase III trials by early next year for cas in clear cell RCC and quemli in first-line pancreatic cancer. As we talked about today, we have several important near-term data sets that may provide further validation and meaningful derisking of several of our programs. This will all kick off shortly at ASCO with 2 oral presentations for dom-zim and etruma-zim. Together, they'll showcase our emerging GI cancer franchise. And we're, of course, looking forward to sharing much more on our HIF-2 alpha program later this year from the dose expansion cohorts of ARC-20. I want to highlight how unusual it is for an early-stage company to be executing multiple Phase III studies for several different molecules, all targeting massive opportunities and doing so in parallel. This is all enabled by our current cash position and our partners that include Gilead, AstraZeneca, Exelixis, and Taiho. These partnerships have and will continue to provide a combination of development funding through receipt of option fees, milestones, and cost-sharing to reduce our overall development expenses. I want to conclude by thanking all of you for your continued and ongoing support of Arcus and our mission to bring innovative therapies to patients in need. We'll now open the call for questions.

Our first question comes from Kaveri Pohlman of BTIG.

For HIF-2 alpha, some of the literature suggests that HIF-2 alpha is upstream of VEGF and it leads to production, so they're basically in the same pathway. Do you think that could make cabo-responsive patients more sensitive to CAS-521 and because of the same biology, do you expect to see any overlapping toxicities with zanza, which also targets VEGF?

Yes. So they are in the same pathway. We do believe that HIF-2, on the other hand, though, since it controls probably 100 genes, there are other effects ongoing in cancer. We don't really expect, and in fact, we know what the side effects and adverse event profiles are. So we do not expect that you would see overlapping toxicities. I would just say that with HIF-2 alpha, those toxicities in adverse events have already been well defined by belzutifan; what you primarily see is very manageable anemia. We've seen similar, and even though we believe and know, in fact, that we're hitting the target harder, essentially, nature is built in a break because only so much HIF-2-mediated EPO expression is happening in the kidney, and there's other sources of EPO. We've seen a safety profile that continues to look very similar to what has been reported for belzutifan. As Jen noted, we've actually even completed a 150-milligram cohort, and again, HIFs have not seen any DLTs to date.

Got it. That's helpful. And maybe a quick one regarding the STAR-121 trial for first-line non-small cell lung cancer. It has primary endpoints of PFS and OS. Do you have to win on both? And what are the expectations from Arm C? How much variation from the control arm and/or TIGIT arm is acceptable?

Dimitry, do you want to handle that question?

Sure, Terry. The first question is relatively simple. Statistically speaking, the trial is set up with dual primary endpoints, meaning that either PFS or OS, if either one of those is statistically significant, it will be a positive trial. I think we've been very clear in previous calls and in lots of interactions that are in the public domain that from a regulatory perspective, OS is really the registrational endpoint; PFS is supportive. Your second question is can you repeat it, please? I didn't get the entire question.

Sure. I wanted to clarify that the progression-free survival and overall survival are really comparing Arm A and B. Arm B is the KEYTRUDA combination, but there is also Arm C for zim. I was curious about how much variation from that arm versus the control arm is acceptable for the KEYTRUDA arm.

Yes, that's a hard question as the FDA would say it's a review issue. That arm, I think, is clear from the public domain is a 4:4:1 randomization. So a very limited number of patients are going onto zim and chemotherapy to establish the contribution of components. Based on FDA guidance, I can share that contribution of components is mostly established by looking at the response rate. Of course, in a randomized fashion, a time-to-event endpoint could also be assessed. There's no absolute guidance on what the variability can be; that's a review issue. I can summarize and I'm very confident with the data we've generated for zim that a positive trial, meaning dom and zim combined with chemotherapy, clearly beat KEYTRUDA, that the contribution of components discussion is a modest part of the discussion, and I'm confident that we will be able to achieve that successfully with the study design.

The next question comes from Peter Lawson of Barclays.

Just around HIF-2 alpha, what defines success for you from the second half data set? Could you remind us what Gilead's potential trigger points are for opting in?

Sure. Thanks, Peter. So on what defines success, as we've articulated, there's a number of variables and opportunities for differentiation from belzutifan. Starting with response rate, rate of primary progression, depth of response, even if we have PFS by that point, so we want to do better than belzutifan. I'll remind you, in thinking about our overall program goal, based upon what we've already seen, the data looks good. We're full speed ahead to the registrational trial. Keep in mind that we are not going to be developing this, at least in the near term, as a monotherapy. We also believe, as Jen was discussing, that we'll be combining with a better TKI than lenvatinib, so that will give us another opportunity for differentiation. There are multiple places, and we think that we have an opportunity to beat those on more than one of those. In terms of the Gilead opt-in, that's something that we would expect. We've converged on what would define opt-in, and we would expect a decision either towards the end of this year or early into next year.

The next question comes from Yigal Nochomovitz of Citi.

This is Ashiq Mubarack on for Yigal. I have a few on the EDGE-Gastric update at ASCO. Based on the data you've shared, it looks like the capital narratives on PFS are trending well beyond the 7 to 8 months hurdle, but naturally with a small end. So I'm just curious how much greater do you think the PFS benefit really needs to be to derisk STAR-221? Also, how well do you think PFS correlates with OS benefit in this setting, given that the primary endpoint in STAR-221?

So thanks for the question. You're right on the sample size. I think you'll be able to make your own call on that. As you suggested, it's known. The data has been pretty tight. In addition, the CheckMate 649, there have been 2 other registrational studies with anti-PD-1 and chemo. They both come in at roughly just under 7 to 8 months PFS. That’s a pretty firm number. We think our data will be quite meaningful. One other thing to keep in mind when you compare our data sets, interestingly enough, CheckMate 649 had 60-40 high PD-L1, low PD-L1, and we're actually 40-60. So a pretty substantial difference. You saw that even our numbers in the all-comer population look pretty good. We're 3 weeks away. We're excited about the data. You'll make your own call on how much you think they've derisked, besides, interestingly enough, as we also mentioned, we're on the cusp of that registrational study being completed. You’ll see those data in a couple of weeks. In terms of the correlation, PFS and OS are certainly qualitatively correlative, but I wouldn’t necessarily think that you can plot a line for them. But given our data set, I think you'll clearly be able to form some speculation as to what might happen on the OS, just as you formed some speculation on the 6-month landmark numbers on what might happen. I think those dots are going to connect at least qualitatively in a direction that will be confidence-enhancing.

Yes. And if you look at the CheckMate 649 data set, you could say that the PFS and OS benefit was pretty correlative and pretty similar, if you look at the hazard ratios. So that's at least one data point that I think shows that their correlation to be pretty strong between PFS and OS in this disease.

Got it. I have one more. You mentioned that the data at ASCO will only be from Cohort A1. I'm curious where things stand with the other cohorts, especially Cohort A2, which is I think the zim plus FOLFOX arm. That one will be important to help me understand the contribution of dom if that cohort varies from the 7 to 8 month number you're talking about. So just curious where things stand there.

Sure. That cohort actually enrolled sequentially for the first one, so the final patient actually just was enrolled a few weeks ago. Keep in mind, not only was it sequential, but since it's essentially a single-arm study, the fact that you don't have an experimental arm is another thing that probably slowed its enrollment. It's probably trending a year behind the initial cohort. As you've probably seen on clinicaltrials.gov, we have 2 additional arms that are probably more designed to take on the contribution of components, which will be run in a randomized fashion and with a larger end. The other thing that may be important about that cohort that you just referenced is we recognize it’s another opportunity, albeit a single-arm, nonrandomized data set that will give people another look at how zim and chemotherapy compare to historical numbers for other anti-PD-1s such as KEYTRUDA or nivo chemotherapy. So we think it will be an interesting data set as well, but it’s at least a year behind this cohort.

Our next question comes from Jonathan Miller with Evercore ISI.

I have a couple of questions about HIF-2. We are very excited about the release of the 100-milligram cohort, as it will provide us with a clearer understanding of HIF-2's efficacy compared to the competitor. However, I am curious why the 50-milligram cohort is not included in that same release. I understand that the 150-milligram cohort has just been completed, but why not incorporate the 50-milligram cohort in the year-end release?

I'm going to answer that because I'll forget what you said. Your brain is younger and faster than mine, so I'll forget the first question. So let me just answer the first one, then I'll take your second question. The 50-milligram cohort was actually enrolled after the 100-milligram cohort. We added both the 50 and then we wanted to go with the higher in the context of Optimus. The 50-milligram cohort just very recently was fully enrolled, and in fact, it won't be part of the medical conference presentation because, clearly, it won't be mature. But whatever we know about it, you'll be able to get that out of me without much inducing. So we'll share whatever we know about it. In fact, when we first saw the very earliest data, even when only 15 patients or so have been scanned, the data were looking interesting. Even though it's 50 milligrams, that's 2.5 times the PD equivalent of the Merck belzutifan clinical dose. The 150-milligram cohort is just starting, as you noted, which is why 50 came after 100.

Okay. Makes sense. Last quarter and this quarter, both, you highlighted that ORR at the 100-milligram dose, I assume, is already similar to belzutifan. The implication there being that given belzutifan's time to response, that ORR could, in fact, deepen as you get more scans under your belt. Have you seen ORR increase? I noticed that in both this quarter and last quarter using similar language saying that you're hitting similar ORR levels to belzutifan. But now they've gotten presumably more scans on these patients, do you see those late responses or later responses more akin to belzutifan's time to respond?

Yes. We were using the same language because we were restating what we mentioned at the end of the last update rather than offering a new description. The data still looks promising, showing a longer trend compared to what we observed previously. We do not plan to provide any updates until the medical conference, at which point you'll see the complete picture.

The only thing to add is, you're right, this is a mechanism where the response kinetics are on the slower side, and the average time to response was about 4 months in the LITESPARK study. Yes, still early; some of our responses may take more time.

Okay. Makes sense. Lastly, as we think about the 2 different TKIs that you're exploring for combinations here, can you give us a little bit of color about what you're looking for to pick between them? You've highlighted the importance of safety versus LENVIMA as a driver here. So are there clear no-go signals that would push you one way versus the other, or is it something else that you're really looking at?

Yes. So I would say it's more strategic in timing and we're thinking about different settings. I think what you're going to see, and this is primarily for competitive reasons, not playing any games. As this year goes along, we'll be very transparent about our development strategy. Some of this is not even that we're in the decision-making process; we know, but we want to let time play out a little as we get closer to those studies coming online, and you'll see there's different lines of therapies. There are other combinations we're thinking about. In addition, you can imagine even that we might go in both directions in different things. So we'll describe that as the year goes along.

Our next question comes from Jason Zemansky with Bank of America.

Congratulations on the progress. Maybe just to take a step back on the adenosine pathway. Just starting development efforts, it looked like the pathway was upregulated across a number of tumor types. While there are certainly very encouraging signals in colorectal and pancreatic cancer, have there been some admitted signals, well, both in prostate tumors and mainly at this point, NSCLC? I'm curious, have you cracked the code at this point in terms of which tumors are likely to respond, especially as your data sets emerge and mature? Are there additional signals that give you confidence in the mechanism at this point, especially as you start to look longer term, maybe some additional indications?

Thanks, Jason. Right now, when we consider the areas where we have found success, we believe it's less about the organ and more about the biology and treatment. The three studies we are currently focused on, MORPHEUS-PDAC, our pancreatic study, and the colorectal study, all involve approaches that integrate immunogenic chemotherapy. From the beginning, we thought this was a key area because it involves killing cells through immunogenic means with chemotherapy. It's well understood that when these initial cells die, they release ATP, which leads to the production of a significant amount of adenosine. If any immune response is triggered, T cells are affected by adenosine, and research shows that when T cells encounter adenosine, they become inactive. The responses we are observing in each of these studies are quite similar. We are also considering expanding to other situations where standard care includes immunogenic chemotherapy but where we can mitigate the effects of adenosine. This is our primary understanding to date, and there's potential for further insights as we continue our work.

Sorry, I think a third element would be tumors that tend to be very high in CD73 expression. That pretty much takes you to the GI tract and long tumors as the places where you have the convergence between high-level CD73 and immunogenic chemotherapy.

Our next question comes from Daina Graybosch of Leerink Partners.

This is Jeff on for Daina. So we have one on cas and one on the HIF-2. I guess starting with cas. We noticed Novartis has an oral ASCO for their HIF-2 inhibitor. How is cas differentiated from their molecule? And in that context, what are your expectations for that ASCO presentation? Do you think Novartis' decision to discontinue has any read-through to cas and your program?

Juan, why don't you comment on what we know about the Novartis molecule? I'll start off, 0 read-through, but I'll let Juan comment on the Novartis molecule.

Yes. Our in-vitro evaluation of the compound suggests that it's a reasonably potent HIF-2 alpha inhibitor, but in our hands, it may be an order of magnitude weaker than cas is. As you know, there's much more that goes into making a high-quality drug and what it does in terms of pharmacokinetics and drug interaction. We'll have to see, but definitely going in, it was not a great contender clearly for a potency standpoint.

Okay. Great. And then how is the recent Roche PDAC data and the upcoming ARC-9 data change your view at all on etruma's potential therapeutic opportunity relative to quemli? Put another way, do you see any clinical settings where you expect etruma to outperform quemli? Do you anticipate issuing any additional randomized Phase II or Phase III trials for treatment this year?

On the first question, while some people speculated otherwise, we look at both molecules as great molecules with great potential. To be honest, if things continue along the way, we said from the beginning, once we start to get good data and the more good data that come, the more this is something one will contemplate. You might even think about combining the two at some point because there are certain situations where you may be able to generate adenosine through a mechanism that's not CD73 mediated, in which case blocking its action could be advantageous. We look at both molecules as presenting compelling data, and when we start the pancreatic trial, we expect it as late as early next year or as early as late this year. We do expect to do more with etruma; the only reason we aren't sharing specific details of what the plan will be. After you see the details of our data, there'll probably be more questions about exactly what we're going to do in the future, but we're still working through those details together with Gilead. We're preparing for discussions with the FDA. We want to think about what would be the ideal control. That control and standard of care have evolved; when we started, it was regorafenib, now it's moved more in the direction of Lonsurf plus bev. There are a number of things that we want to work through, but you can expect that we'll be doing more with the molecule.

Yes, it means a lot. I think it indicates that we achieved an oral presentation at ASCO, which is a significant accomplishment. Receiving any presentation at ASCO has become increasingly challenging. The acceptance of our oral presentation suggests that the clinical community finds the results and the study to be meaningful.

And our next question comes from Li Watsek of Cantor Fitzgerald & Co.

This is Rosemary on for Li. Just a quick one on your next-gen programs. How are you thinking about prioritization as you have all these late-stage trials going on? How does inflammation fit in with your oncology franchise?

Yes. I think the question was sort of the early-stage portfolio, given that we have all these late-stage things, how do we think about it? We were talking about the collaborators and what it enables us. In fact, it's another unusual aspect of the dynamic we created with this collaboration. When we set this up with Gilead, Day 1, people don't always take these things of its value, was designed to be an R&D engine type of collaboration. Given the number of late-stage programs, dollars are being invested correspondingly. With that said, we've continued, and this was when we discussed the most recent equity investment that Gilead made; one of the things that we discussed at that time and the $100 million fee that they'll be paying shortly maintains the relationship. There was a big emphasis on continuing to enable the discovery part of it; clearly, if you look at what we're generating, we started as a blank piece of paper. We're generating what we believe are good, high-quality molecules for targets of interest. The most notable one that's moving along nicely and we think is going to be a big deal is our AXL inhibitor, which recently has gone into patients; it performed well in healthy volunteers and generated some initial PK data. That will be the AXL inhibitor that tests the hypothesis. It's going to have the PK and selectivity that we'll be able to hit AXL hard and see how that works out. The early-stage portfolio continues with the same emphasis and investment as it has previously. To your question about the more immunological component, we haven't shared the most recent targets that we're working on, and we'll probably hold off on that for some time within a year that we'll start to talk about that. But we, along with Gilead, increased our emphasis there. As we built the company, we have a very strong immunology biology group. I think you'll slowly see the percentage of the portfolio that looks like it includes immunology inflammation increasing, but it's not like we've defined a particular percentage. We're still at a stage of our evolution where the biggest driver would be best players. So best target, we'll get the next emphasis in our drug discovery group, whether for immuno-oncology or oncology or I'll use the term du jour, I&I.

We have no further questions in the queue, so this concludes today's call. Thank you very much for joining.

Thanks, everybody. Appreciate it.