Replimune Group, Inc. (REPL) Earnings Call Transcript & Summary

Good morning, and thank you for joining our midyear data update event. And before we begin, I need to highlight our forward-looking statement on Slide 2. I'm Philip Astley-Sparke, CEO of Replimune. And I have with me today, Dr. Robert Coffin, our President and Chief R&D Officer; we also have Dr. Mark Middleton, Professor of Experimental Cancer Medicine and Consultant Medical Oncologist at the Oxford Cancer Center and the Head of the Department of Oncology at the University of Oxford. Slide 3 shows the agenda for the call, which will start with a brief [ berth ] introduction from me and a brief summary of the high level updated results by Rob, followed by discussion of the results in number of the individual patients by Dr. Middleton. Rob will then provide an overview of what will be one strand of our future development efforts with RP2, 3. Following the presentation, we'll be holding a live Q&A. To ask a question, please type your question and click submit in the Q&A section on the bottom of your webcast. All questions will be held until the Q&A portion of the events. So moving to Slide 4. We are continuing to make great progress towards our ambition of making our products a cornerstone of immuno-oncology treatment regimens. That's the most practical and effective way to initiate a systemic antitumor immune response. The studies that have been ongoing the longest are [indiscernible] RP1 in our Phase II skin cancer cohorts, which support the 2 studies we are running, in cutaneous squamous cell carcinoma, an anti-PD-1 failed melanoma with registrational intent. We have always said from the get-go, our modality and products will result in higher rate of complete response. It is these responses that transform lives, give patients the potential to cure and relieve them of disabilities and/or the [ segments ] caused by disease. 7 out of 9 of our cutaneous squamous cell carcinoma responses are now complete ones with a further patient to be assessed as a probable eighth complete response pending biopsy. Further, in melanoma, a number of our past responses be shaped metabolic complete responses by PET scan. While this further supports our expectation for a positive outcome in our registration-directed studies, we also believe that if our well-tolerated products are pushed earlier into disease courses, the rates of complete response could go higher still and result in many patients never developing the type of end-stage disease we are currently treating. With RP2, the signal in patients who have failed anti-PD-1 has really picked up from where RP1 left off and reconfirms the ability of our platform, treat anti-PD-1 failed disease. We look forward to presenting the updated data in this regard today. We also look forward to explaining our decision to advance RP2 or 3 into later-stage development to treat patients with liver metastases with very poor prognosis where we have seen very encouraging reproducible activity. We've also made solid progress putting in place the building blocks to build an entity capable of transforming the immuno-oncology landscape. Our own manufacturing facility is fully operational, Berlin GMP batches, and we agreed to pass forward recharacterization release assays with the FDA. In addition, we have hired a Chief Commercial Officer who is starting the planning process to ensure, if approved, our products are widely adopted in the marketplace. Finally, we have a strong balance sheet to execute on our vision. As a reminder of our technology in MOA on Slide 5, oncolytic immunotherapy is the use of viruses that when injected into tumors, partially or comply destroys them through virus replication. First, when tumors cells open and exposing all the release cancer antigens to the immune system in an environment and necrotic cell death. This leads to activation of a patient-specific systemic immune response and the disruption of uninjected deposits. We believe that Herpes Simplex bars is an optimal specie to bars for oncolytic use. They're being both highly mitigant inflammatory and having the ability to carry multiple immune stimulating proteins into the tumor microenvironment to further amplify the immune response. Our next-generation constructs have been specifically designed to maximize each neurological signals, 1, 2 and 3 in harmony to provide full activation of adaptive immunity, combined with the potent activation of the innate immune system as well. Our strain of HSVs has been deliberately selected for its lytic properties in human tumor cells. And from all our products in our base platform, we express a fusogenic protein that greatly increases direct tumor killing, immunogenicity of cell death and systemic immune activation. These design features ensure maximum presentation antigen on the MHC antigen presentive cells or so-called Signal 1. We then expressed various proteins from the virus intended to maximize cost imagery signals of the antigen presenting cell T cell interface or so-called Signal 2 to ensure optimal T cell priming. Our lead product RP1 additionally expresses GM-CSF, RP2 additionally expresses an anti CD34 antibody, stop a negative feedback loop of the CD28 [indiscernible]. And RP3, too further in the [indiscernible] pathway activating proteins targeting CD40 and 4-1BB which also lead to downstream inflammatory [indiscernible] release or so-called Signal 3. Moving to Slide 6. We believe our products are the most practical and effective way to ensure a tumor is recognized as foreign, where there is an absence of an effective pre-existing immune response. Practical, as all our products are off the shelf, are coming as simply vial, manufacturing is relatively straightforward and cheap, and our products are well-tolerated, effective, while it approaches off the shelf, it is also patient-specific and acts as a pan-universal [indiscernible] vaccine as the tumorous lives open and all the cancer antigens within are exposed to the immune system in an environment of necrotic cell death, a major immune danger signal. Our approach has markedly further MOAs packaged into 1 product through the expression of immune stimulating proteins, which will carry into the tumor microenvironment to further amplify the immune response, as I described on the previous slide. Our development plan is shown on Slide 7. The data providing an update on today pertains to the top 2 bars, where we fully enrolled 30-patient melanoma cohort, containing both PD-1 failed and PD-1 naive patients. We are closed to full enrollment in the non-melanoma screen cancer cohort as it pertains to the first 30 patients who are anti-PD-1 naive, of which 15 patients have cutaneous squamous cell carcinoma. However, this cohort has now been expanded to include non-melanoma skin cancer patients who have failed anti PD-1 where enrollment is now underway. As a reminder, we have 2 registration studies ongoing in cutaneous squamous cell carcinoma and anti-PD-1 fail melanoma as depicted. We're also providing an update on RP2 single agent and releasing initial days at the first time at RP2 in combination with Opdivo. However, it should be borne in mind that this data set is immature and as much as they include many patients who remain on therapy and may therefore further respond or progress. Finally, we intend to expand the [indiscernible] RP2, 3 beyond Phase I, we'll be giving our initial thoughts on where we plan to focus our efforts during 2022 during the course of this presentation. I will now hand over to Rob to summarize today's data sets.

Thanks, Philip. So I'll just now briefly summarize the data update today, which is for RP1 in skin cancers, as Philip said, and for RP2, before I hand over to Professor Middleton, who'll present the data in much more detail. So first, cutaneous squamous cell carcinoma. As Philip said, we've now enrolled 15 patients into the IGNITE study with CSCC. And this shows that while we recently enrolled a number of patients with a particularly high tumor burden, who quickly went off study, which Mark will discuss a little bit as well. The theme of achieving a high rate of response and particularly complete responses continues. The complete response rate in this group is now 46.6%, with 2 additional complete responses since October, and one further response awaiting biopsy confirmation, which would bring that to 54.4% if that biopsy shows to be negative. The response rate in CSCC is now 60% with continued very good durability of those responses and multiple responses now out over 600 days. So in our fully enrolled melanoma cohort of 30 patients, the registrational directed indication we're also pursuing. The objective response rate in anti-PD-1 naive patients is now 62.5%, and for anti-PD-1 failed patients, it remains at 31.25%. Remember that cohort fully enrolled early last year, and so it's now really quite mature. With again very good durability being seen and multiple patients now having converted from originally stable disease or partial response to now complete response. We now have 11 patients with other non-melanoma skin cancers enrolled and clinical activity continues to be demonstrated across the breadth of these different cancers. And we've also found interestingly that giving a second course of RP1, the initial first quarter is only 8 doses over a couple of months. So a small number of patients may benefit from additional doses of RP1. Those being patients who achieved an incomplete response or progressed after initial response, and we're in a very small handful of patients. We've now done that. This has been seemed to be well-tolerated. And also each of the patients have shown evidence of clinical activity, which is something which we think is quite unique in oncology drug development. Overall, the data with RP1 in skin cancers continues to show that the combination with Opdivo is well-tolerated. And that the combination drives deep and durable responses, which we do believe strongly supports our registrational directed activities in both cutaneous squamous cell carcinoma and in anti-PD-1 failed melanoma. So if we now move to RP2, a brief summary there. So with regard to the single-agent part of that study, that was fully enrolled last year, and we presented initial data in October. And there were 3 out of the 9 patients enrolled who had responded. These patients with further follow-up have continued to show good durability with the patient with mucoepidermoid carcinoma, who has a complete response in October, maintaining that out now to 15 months. The patient with esophageal cancer has a maintained very good partial response now out at 18 months. And the uveal melanoma patient progressed at 15 months, although obviously showing quite impressive durability before that. So relevant to our liver metastasis strategy, which Philip mentioned and which will be discussed later, 3 of those monotherapy patients had injections into the liver of RP1, and all 3 of those patients showed evidence of activity following those injections. If we move on to RP2 in combination with Opdivo, the new data, which will be presented today. There are 27 patients who have been enrolled into that combination portion of the study so far with a range of different tumor types typical for a Phase I trial. And we're now actually currently intend to expand the number of patients in that study with liver metastases to be treated with RP2 also to provide further support to the liver met strategy, which we'll come to later. So while this data is still rather early and immature, so far, out of the 27 patients, many of which still ongoing treatment, 6 have achieved an ongoing response in patients with uveal melanoma, cutaneous melanoma and head and neck cancer. All of those patients having had prior anti-PD-1, and which at this early and immature stage of the study, we think, is very promising. And also has reinforced the signal we've previously seen in some of those tumor types with in the past and which is supporting our registrational development at the moment. So with that very brief summary, I'll now hand over to Professor Middleton, who will go through the data in somewhat more detail. So Professor Middleton, as Philip said, is our lead investigator in Oxford, who's been extremely important in our enrollment to date and really has recruited the most patients of any investigator at any site globally. So it has, by far, the greatest experience with both RP1 and RP2 of anybody, so he really is a perfect person to describe the data in detail. So over to Mark.

Thanks, Rob. So to start with RP1 and the IGNITE study where it's given in combination with nivolumab, just to review that for melanoma, we're now concentrated on patients who failed PD-1 directed therapy and also enrolling non-melanoma skin cancer cohort. Nivolumab is given in the standard way for up to 2 years, but the RP1 is given every fortnight as an intratumoral injection. The 8 doses additional show that completes treatment well within 3 months of starting. And although we can inject the RP1 directly into superficial or helpful nodal tumors. We also use image-guided injection quite a lot so we can access peak tumors, including industrial and [indiscernible]. And since the start of this year, we've had the option to give up to 8 further doses with RP1, according to clear criteria specifying of protocol. So for single progressing lesions or where we can go response and store. We would benefit from further treatment. And we've got some very early data of this, which is a pretty unusual approach for oncology job. Because generally, the principle has been that once you've tried it, if it's not working, then you move that off and try something different. So moving on to talk about safety, broadly speaking, the safety information is unchanged from the previous presentation with treatment pretty well tolerated and with the toxicities that you might expect for patients who are receiving nivolumab. There has been one significant toxicity in March of this year of patients who have already completed their RP1 portions on the nivolumab maintenance stage of treatment. [ Mitral ] [indiscernible] think associated this year and succumbed to that. The investigator assessment business is clearly related to the nivolumab [indiscernible] than the RP1. Otherwise, as the table shows, it's a very well-tolerated regimen. Moving on now to talk about efficacy in the situation and focus initially on non-melanoma skin cancer. This table summarizes the response data. And you'll see in the red columns, focuses on the changes that there have been since October. As Rob touched upon briefly, we've enrolled 4 more patients with cutaneous squamous cell cancer. I think it's probably fair to say that I and my fellow investigators gave up a pretty hard challenge in rolling some very advanced patients who have not benefited from treatment. In fact, all of them progressed moving very, very quickly. And a batch patients but in retrospect didn't precisely meet the bill with their rapidly progressive disease. Having said that, as Rob's already touched upon we've seen some of the responses that we've previously reported deeper so that now 7 of the responding patients have got complete responses and has the potential for one more to join them in the near future. We've also seen additional responses in other tumor types, basal cell cancer, merkel cell cancer and angiosarcomas, all of which are listed in the subsequent columns, show a broad inspection of activity across an important set of patients. It's going to a little bit more detail and apologize for the slightly gruesome slide, but it does speak to the issue of the severity of the disease of patients that we enrolled since the last briefing, both pictorially here, I'm also looking at the scan, very, very extensive tumors that defines treatment led to rapid progression and then leaving the study within a couple of months to move on to parity of care. If you look at the water board box describing the extent of response than what you can see here. Color coded according to histology is how deep these responses are with a very significant proportion of patients experiencing complete response, both [indiscernible] and with the potential for more to join the number of patients still on treatment. And looking at that in [indiscernible] bottom -- sorry, [indiscernible] to show the duration of that. Again, Rob gave the headlines earlier. We're now starting to see patient approach to [ Genmark ] in complete response and several big partial responses and the durability of the spots is really very impressive across this coding. And indeed, with other patients who have been treated with RP1 and 2. And to look at this another way, individual by individual, the spider block, what you can see here is the depth of that response, the speed over which it developed, which is often quite quick in this patient population. But there's merit in persevering with treatment because there are some examples of patients who have a modest response initially, which then deepens significantly over time. I draw your attention also to the right-hand part of this, both way you see a patient who has reinitiated. There's a small uptick in the line when you see it going down again as a result of reinitiation of therapy. And we'll talk a little bit more about that in a few more slides' time. So updates on individual patients now for patients -- since what we talked about back in October, if we look at this patient -- the average of new complete response. It's [ subtle ], but what you can see in the baseline scanning in June is deformity of the Maxilla. You can see how it pushed back where the tumor is and with ongoing treatment at 6 months at the end of last year and now early this year, where we've got a complete response. What you can see is normalization of the architecture around that. And that, obviously, has very, very profound consequences for the patient in terms of how they feel, how they look as well as being impressive on the radiology. Then a second patient who we reported previously was a partial responder on the basis of the pictures you've seen in May and in October of last year, has had further treatment, which has now converted this to a complete spot. And what's really important is that as you might expect, looking at these pictures of the patients go to regain full mobility, having previously been unable to walk on that foot. And so it's not just pictures on a scan, but a very important and practical difference to the patient's life as well. You've got -- in this slide said now described a partial response to patient with basal cell cancer. And we think that this may that be a complete response because as you pull other scans from October through to February, you can see that they are approximating normality which part and clear whether this tumor which incidentally as a patient who failed does not [indiscernible] standard care in this situation. We're increasingly optimistic that this represents scar tissue and so the patient will have a biopsy to see whether this is the case and may end up being a complete responder. And then finally, to look at a patient with merkel cell cancer what we see here is a very impressive response in the injected lesion, in the top panels, where they are very significant mass, 5 centimeters in diameter. But also an obstacle response in non-injected lesion, which again is tending towards modality, but we'll need to get biopsy of foot whether this is a complete scan. Well, I'm now going to turn to the data for melanoma in -- with the combination of RP1 and nivolumab. What you see here is a summary of the disposition of the patients treated so far in the initial cohort, we're principally patient with cutaneous disease, although we do have some of the [indiscernible] uveal melanoma. I think it's important to note that this is a relatively advanced population with the majority of patients having prior exposure to PD-1 agents and being of a relatively high end-stage where over 3/4 of the patients had a M1b or c disease. So this is, by no means, a highly selective population. So again, same format is with the non-melanoma skin cancer in the red column, we've got the update in February -- in October of last year. And this is relatively material cohort because enrollment completed early last year. And what you can see here again, is a similar story with increasing responses across the [ piece ] and we're seeing deep responses in the anti-PD-1 naive [indiscernible], now 3 complete responders. And overall response rate is roughly 2 in 3 and then a response rate of that 1 in 3 in the PD-1 failed cohort with a complete responder as well as 4 partial responders. One of whom we have assessed as a metabolic [indiscernible]. And then looking across at the new casual data, for the complete and casual responder. That patient who we previously reported is having a partial response. And stable diseases but still with the potential to respond albeit in the late-stage in the newbie October. And we'll talk a little bit more about new growth then we discuss in RP2. Looking at the waterfall clots. Again, what you can see is a high proportion of patients with complete responses, which, I think, is significant and marks out this approach to treatment from some of the other IV drugs and development, and then we sure clots, again, a similar story to what we've presented already in the non-melanoma skin cancer with lots of patients still on treatment, starting to approach that 2-year mark, and with enduring response whether partial or complete. And then the -- despite of that, perhaps a slightly slower trajectory towards the best response, and you see the non-melanoma skin cancer. But again, I draw your attention to a patient with stable disease. Who didn't progress. And then responded to reinitiation of PD-1, which, in fact, is one of my patients who have failed anti-PD-1 before joining the study at a prolonged period of stable disease. And then we responded -- and is firstly achieving the PR based upon new initial treatment. So just to talk a little bit more about reinitiation. As I said before, this is a relatively unusual approach but it spends to the ability based upon the mechanism of action to get a subtly different response from injecting a new lesion or a lesion that's progressing in the face of prior successful or reasonably successful treatment. And it's an opportunity to deepen responses that we've already achieved or to achieve response at more where it's been lost often locally in an isolated lesion. So we've got evidence of activity in all 4 of the patients that we've treated so far. So we're still at a very early stage. And as an example, here of the patient not treated at my center, but at Liverpool by Dr. [ Gary Sackett ]. So what you can see here is the baseline scan on the left, the initial response to therapy through until March of last year. Local regrowth of the tumor at the beginning of the year and was at this time that we've gained to approachable amendments, the ability to reinitiate treatment. And that was done with the resulting complete response evidence on the far right picture at the nimble of last month. So to talk in a bit more detail about RP2 now. Next slide, please, Rob. Then the fee eligibility criteria here, I get a broad set of tumors, eligible, RP2, given again, every couple of weeks, the 8 doses, with the nivolumab started with the second large dose and being able to continue for up to 2 years. And again, in exactly the same way as a thing with RP1, we didn't specify, but we had to go for superficial [indiscernible], particularly [indiscernible] enthusiastic use with image injection to broaden the scope of patients that we can treat. And slightly earlier than the RP1, we had the opportunity to reinitiate the RP2 then in the fall of last year. So this table summarizes the patient disposition. As Rob highlighted at the beginning, we had 9 patients in the RP2, only dose that has previously been reported. And we've got 27 of the 30 patients enrolled in the combination. It's a broad speculative tumors that you'll see that we have melanoma very well represented into the uveal melanoma, some squamous cell cancer that hold neck [indiscernible]. Safety has been very tolerable. The majority of patients experienced only grade 1 and 2 , and this is very similar to the story that we've seen with RP1 in the larger patient set. And in surprise, we have seen higher grade [indiscernible] very consistent with either injection. And therefore, lot of disease [indiscernible] chooses to deliver it or within the valid markets that give the RP2. Just to summarize briefly, the monotherapy service [ clots ]. We'll talk about these responding patients. The new real melanoma patients will be treated with [indiscernible]. Sadly, we lost control of that excellent response after 14 months. And patients progressed and died very rapidly. And I suspect that COVID was a factor in them to seek medical attention, obviously, came from some way away and have that felt response. But I'd also draw your attention to the top patients. Again, Rob had mentioned previously, another patient of mine with esophageal cancer and liver metastatic disease, who's had an excellent partial response. And incidentally, this pressure is being treated in the clinical trial, and therefore, had been exposed to anti PDL-1 agent as well as to chemotherapy before joining the study, and we were able to confirm with a PET scan last month there was no metabolically active disease remaining that we can't frame a complete response because there are visible lesions from cross section needed. But I think this points to the exciting potential for RP2 to form deep and durable response. We move on now to the patients who received RP2 in combination with nivolumab. We've got across the whole cohort. The response rate of just over 20%. But there are still a significant number of patients -- too early really to determine that ultimate disposition with [indiscernible] treatment concerned because these data really aren't very sure. I would draw your attention though to the patients who have cutaneous [indiscernible] melanoma, where we've got ongoing responses in patients who've had significant fire exposure during the therapy. And then looking at the [indiscernible]. You'll be familiar with this story now. Obviously, we're a little bit earlier in the piece. Therefore, we don't quite get out for the 700 days that we're able to look at the RP1. But what you can see here is a number of patients ongoing the treatment now getting on to the 6-month mark with good responses. Interestingly, as per RP1, a number of patients who show initial progression including new lesions have had responses thereafter and raises the question of whether these are inflammatory reactions to the therapy and consistent with the mechanism of action. And I think one of the learnings that we've taken from the whole program is that what started it pays to persist with viral injection and the nivolumab out until at least the completion of the viral injection part of the treatment. But before making -- taking a view as to whether this patient benefits. And just to talk through this with some examples. So this is a patient of Professor Harrington in the Royal Marsden Hospital with uveal melanoma. What you can see here is a local tumor occurrence. The patient also had bone metastases, which are not shown here. So this patient had a nucleation. And this is a tumor in the sharp pic, which you can see in the picture as well, dated October of last year. And then as the scans and the picture shows this very good resolution with the tumor, and again, will have a significant impact on quality of life as well as longevity. Then just a further example from Dr. Sacco. Again, this is a patient with cutaneous melanoma who've had prior exposure from nivolumab above the disease in the neck as well as small lung lesions and there've been substantial initial progression. If you look here at the baseline scan and focus on this lesion in the upper neck, circled in red, you can see that over the course of the first 3 months we've respected the progressive disease. But we've worked to grit our teeth, to carry on with treatment, as you can see here that about 3 months down the line, this has achieved a partial response as the impact of treatment starts to take hold. This was associated with stability in the more decelerated smaller regions and significant functional improvement that allows the patient to return to work after a considerable amount of time, unable to do so. And then finally, a patient with squamous cell cancer in the head and neck, have had prior exposure to anti-PD-1 including nivolumab as well as standard fluoropyrimidine and platinum chemotherapy as well as regular therapy. And what you can see here is on injected and uninjected lesions, both of whom are responding to treatment in perhaps a more conventional way within a couple of months of initiation treatment. So in summary, RP1, combined with nivolumab continues to provide deep and durable responses in the range of stim [indiscernible] and remains pretty well tolerated. What's new compared to previous presentations is that we've got some experience now of re-initiation of RP1 which remains very well tolerated and for which we've got early indications of clinical activity where 4 of the patients treated so far. You recall the biomarker data we mentioned at ACR continues to support the mechanism of action of the agent and broadly speaking, turning cold tumors hot. As far as RP2 is concerned, it's very well tolerated whether given alone or in combination with nivolumab, and we continue to see exciting initial activity in patients who failed anti-PD-1 and also in the range of patients with cancers that perhaps haven't tethered -- to be considered to be particularly a meaningless immunotherapy. So I'm going to hand back to Rob now who's going to talk about some of the development plans that we have.

Great. Thank you very much, indeed, Mark. That was wonderful. So we're now going to move to a bit of a discussion as to how we're intending to move beyond skin cancers, particularly with RP2 and RP3 and in particular, our plans to initiate a relatively broad development program in patients with liver metastases specifically. And so as everyone will be aware, the liver is one of the most common sites of metastatic disease, including for many of the larger tumor types like lung cancer, breast cancer, colorectal cancer, et cetera. The prognosis for these patients is particularly poor, and they also respond particularly badly to immune checkpoint blockade. This may be due to the fact that liver metastases appear to selectively remove tumor reactive T-cells from the circulation, which I'll come back to in a minute. But as oncolytic immunotherapy aims to directly kill tumors and also unleash a systemic onslaught of tumor reactive T-Cells, the hope is that with RP2 and RP3, particularly, we'll be able to reverse that, which our initial data suggests may indeed be the case, some of which Mark alluded to and showed. So this slide shows the numbers of patients with liver metastases in some of the larger indications. Shown on the left that the potential commercial opportunity in just these 3 highlighted indications is around 80,000 patients per year in the U.S. alone, and on the right, the proportion of patients with liver metastases from a broader range of tumor types is shown, which altogether, as you can see, adds up to a very large number, indeed, such that the potential market is clearly very substantial indeed, if one had a therapy, which was able to provide activity in patients with liver metastases. Sorry, I'm experiencing a bit of a delay. So on Slide 48, this shows on the left that patients with liver metastases treated with immunotherapy tend to fail at distant sites rather than only in the liver. So if they only had liver mets in the first place, they don't actually tend to fail locally, they tend to fail systemically, which really does indicate a systemic deficiency in immune activation. And that on the right, a patient with liver metastases across tumor types also achieved less benefit following immunotherapy than patients who don't have liver metastases. Whereas interestingly, this difference isn't seen with things like chemotherapy or targeted agents. The difference between liver and non-liver mets is actually specific to immunotherapy, if the immunotherapy which is particularly impacted negatively by the presence of liver metastases. And if 1 then looks specifically at response rate and survival across tumor types,to immunotherapy, these are both reduced in patients with liver mets, as you can see, where survival is considerably impaired in patients with liver mets as is the response rates achieved across these tumor types mentioned in the slide. So a recent publication on Slide 50 looked into the mechanism by which liver met patients do particularly poorly on immunotherapy. This is still somewhat hypothetical at the moment, but the data does seem to back it up. So this indicated that macrophages, which accumulate in liver metastases present antigens derived from tumor, which can then bind antigen reactive T-Cells, which are then killed by apoptosis induced by the macrophage by defense pathway. And therefore, T-Cells which react to the tumor are selectively depleted in the systemic circulation, resulting in poor systemic efficacy of immunotherapy. If, however, we're able to kill off tumors in the liver, for example, with RP2 or RP3, that would reduce not only the tumor in the liver, but also the number of macrophages associated with that tumor, which should effectively remove for T-cell sync if you want to think about it in that way, so while at the same time as inducing the army of new tumor reactive T-cells to increase overall systemic efficacy. With RP3, in particular, the expression of CD40 ligand and 4-1BB ligand should also reduce the degree of T-Cell apoptosis and also increased T-Cell activation, further enhancing the therapeutic effect. All of which suggests to us that there really is a particular opportunity for us in patients with liver mets, which we're hoping to exploit. So this next slide -- which there may be a little lag. So this next slide does show a number of patients so far we've treated with liver mets, which really we do think is pretty impressive, bearing in mind the nature of the disease these patients had. There are 6 patients here, all of which with -- had substantial tumors in the liver. Some of which were injected in the liver and a couple of which were injected at other sites and some of which were treated with RP1 plus nivolumab and others with RP2 but what in aggregate it shows are the patients with liver metastases cannot only respond but respond systemically and also have very good durability of effect as well, which one of these patients out now to nearly 2 years following initiating therapy. So this body of data, we really do think supports that we really can potentially do useful things in patients with liver metastases, which has inspired us to begin this development pathway for those patients. So our development strategy in patients with liver metastases described and outlined here is still early and work in progress. But the plan at the moment is to initially expand enrollment into the RP2 study to include patients with specific tumor types with liver mets, including from GI cancers, lung cancer and breast cancer. Together with a few additional patients in uveal melanoma to further confirm the signal we're already seeing in uveal melanoma with RP2. And then in parallel, we'll be continuing to assess the safety and clinical activity of RP3 which includes patients with liver metastases. We're currently in the Phase I dose escalation, single agent part of that trial and will move to combination therapy later in the year. Then depending on the data we see with each of those, RP2 and RP3, we'll then initiate a multi-cohort Phase II program in liver met patients with specific tumor types, which will provide us with further signal confirmation and tell us which tumor types we have the strongest signal in and where, therefore, to enter a registrational development. So it's a step-wise approach, which should be able to be relatively rapidly executed. And will be very much stated -- about dependent. And the exact details will depend on that data as it will -- whether it's RP1 or RP -- or sorry, RP2 or RP3 which is used in a particular circumstance. But all of that will kick off later in the year with the expansion of RP2 into further patients with liver mets. So Slide 54 sums up the prior comments I've just been making and reiterates that based on the theoretical considerations, the huge unmet need in patients with liver mets and on our emerging data that we believe liver mets to be a very real opportunity for us, which is worthy of aggressive development as we so far mainly apply to patients with skin cancers. So we're now aiming to take a similarly aggressive approach and proceed, hopefully, rapidly through development in patients with liver mets as well, particularly with RP2 and RP3 which was specifically designed to try and treat patients with less immune responsive tumor types than we're currently targeting with RP1. So with that summary of our expanding strategy, I'll hand back to Philip, who's going to do some summing up.

So Rob, and thanks, Mark. So looking ahead over the next 6 to 12 months, we look forward to being able to share further data across our programs, support our ability to treat patients who have failed anti-PD-1 across the broader swath of tumor types, including initial data in anti-PD-1 failed cutaneous squamous cell carcinoma and anti-PD-1 failed non-small cell lung cancer. We also look forward to presenting around the year-end single agent data with RP3 and providing more details on the Phase II program we plan to run with RP2/3 in patients with liver metastases from prevalent tumor types. The responses we have highlighted today in the liver are profound. They include single agent responses. They include PD-1 failed responses. They are deep, durable and provide a very large value opportunity beyond our skin cancer franchise. In skin cancer, we're maintaining guidance. We expect to release top line data from our 2 ongoing studies with registration in turn in CSCC and anti-PD-1 failed melanoma in 2022. And we'll provide further details on timing expectations later in the year. We'll now turn over to Q&A. [Operator Instructions] We will get to as many questions as we can. Thanks, again, for everyone's time this morning. And we now look forward to answering your questions.

So if we have a question, the patients recently enrolled with high tumor burden CSCC, would these patients have been eligible for CERPASS based on inclusion, excluding criteria.

Rob?

We would have to look at the patients in detail to determine that. It's quite likely based on other inclusion criteria not relating to tumor burden relating to comorbidities and performance status that they wouldn't have been eligible, but I can't formally answer the question. But we do have to remember in CERPASS, it's a randomized controlled trial. And therefore, there would be expected to be a balance of such advanced patients between the arms, if they were to be enrolled, which would mean that the trial would not be negatively impacted one way or another by enrolling such patients.

We've got a question for Dr. Middleton. Dr. Middleton, how confident are you that you are seeing abscopal effects that is not oncolytic virus spreading by, for example, needing vascular change?

I think very confident indeed. The majority of patients who've been treated have had prior exposure to herpes simplex virus have carry antibodies that can neutralize any virus that escapes in the systemic separation. And if we look across the range of data that we presented previously and today, what you can see is clearing distant tumors which are responding. So it's not possible to break that down to viral leakage in the direct healing effect. This is a new mediated effect by tutoring the immune system to act systemically on the basis of local self-heal. And the fact that you can then see that again, the re-initiation further supports that contention.

Great. And we've got another one for you, Dr. Middleton. What are the disadvantages of using oncolytic virus versus antibody for small molecule based therapies? Do oncolytic virus limit the potential uptake in academic centers?

The obvious disadvantage is that current systems of care are very well set up or intravenous or oral drug administration and you have to learn a new way of doing things in order administer things intratumorally. And you have to have a cooperation of skilled interventional radiologists to access in to access visceral tumors. But one of the reasons we've been able to contribute so significantly to this program is because we have those capabilities in placing oxygen and, thereby no means, specialty experts. Interventional radiology is available in almost all major hospitals in the western hemispheres. It's used well beyond cancer as well. And where we have effective treatment, then experience tells us that uptake will be enthusiastic. And I can tell you that already that our radiologists extremely enthusiastic, having had to be persuaded that this was a program we wanted to pursue. They can see the fruits of their labor from scan to scan, from cycle to cycle, and were therefore very in vested in the program. So although it undoubtedly adds a layer of complexity, and you have to be prepared to manage the small number of side effects that go with deep injections into viscera. It's by no means insurmountable. And if you compare it with the complexities associated with intravenous adoptive cell therapies or bone marrow transplantation, they're trivial. This is something that couldn't be sorted out if you've got the efficacies.

Great. We have a question on dosing, Philip. Are you planning to extend RP1/2 dosing beyond 8 doses for all patients? Do you think this could extend the durability for response?

Rob?

No, we're not. We do think that the 8 doses, we got it about right. The vast majority of patients have up to 8 doses. In some cases, as in quite a few cases, nothing left to inject well before that full 8 doses and then achieve good durable responses, which in nearly all cases has remained ongoing to date. It's only been a very small handful of patients who it has been thought could benefit from a further course of up to 8 doses. So we are now comfortable that having the initial dosing regimen of the first course of up to 8. And then the allowance for the small number of patients who may benefit from the second quarter -- a further quarter up to 8 really is right, and we have no plans to amend that further.

If I could just add from a clinician's perspective, the ability to re-treat or reinitiate RP at makes a massive difference. With all immunotherapies, whether it's levo, epi, when we first started using that, and so forth, there's always been this question how long we'll go on for. And we all strongly suspect that we overtreat patients. So we were very pleased with the initial 8 injections only approach because it draws a line, and we saw patients who continue to respond and have deeper responses after the end of the injection, which gave us -- we've got the confidence that we have triggered an enduring change in the immune system. The fact that you can say to a patient, "Look, we go with 8, we know we can get good results of that. But if it happens not to be right for you, we know we can come back and give more," is, I think, a hugely better approach that we're going to sign you up to having a rejection delivered every 2 weeks over time.

I think if we go back to this slide, it is worth pointing out here in relation to these patients who've got single agent RP2, the treatment course for single-agent RP2 in the Phase I part was actually just 5 doses, to really just sort of get the Phase I part done more quickly. So highlights in the green. But as you can see, the patients didn't actually fully respond -- the responding patients until after the period of treatment. And then that those responses continue to deepen. And none of these patients had any other therapy with anything else since these first 5 doses of RP2. So we really are achieving deep, durable effects following a short initial course and just having the opportunity to give a second course in the relatively rare cases where that might be needed, we really do think is quite novel and also about right, as Mark was highlighting.

Maybe moving on to liver mets. When treating liver metastases, would the clinical approach be to inject only tumors in the liver or also inject the primary tumor organ?

So I'll answer the question and Mark can further comment. So the rules for dosing are of patients -- are really the dose, the largest easily injectable tumors. And you have up to 10 mills of injectate to use on each injection day. So the largest easily injectable tumor is in the liver. You inject the appropriate volume into that tumor. And then if there's any left over the 10 mills, you would inject into other injectable tumors, which may or may not be in the liver. So in many cases and in most cases, so far, when patients have been given injections into the liver, they just get the full dose into liver. But there may be some occasions where they get injections into both liver and elsewhere.

Yes. So we've been slightly changing our approach a notch based upon the results that we've seen today. So I'm -- consented the patient to this program yesterday, who has relatively small subcutaneous lesion on the back and has liver disease. And perhaps 2 or 3 years ago, we said, "Well, we'll go for the low-hanging fruit. We'll inject a small amount of virus into the surface lesion and see what that does." But based upon what we've been seeing to date, we're keen actually to inject the full 10 mills. So we'll be injecting a right side with liver lesion in this patient as well as a subcutaneous disease. And that's not driven so much by the nature -- mentioned paper that Rob referenced. By the observation and the 2 patients that were called out in the RP2 only slide of good examples of that. So we've seen some really impressive results in patients with predominantly liver disease by injecting one lesion there and in a way that is out of line with what we've come to expect with IO, where it's off in a site that does less well. So we're interested, obviously, when we've discussed with the patient, what's involved in participation, we're frank about the slightly increased risk by introducing the needle into a vascular organ, might deliver. But I'm happy to say that I think that the risk-benefit falls on the side of benefit to this because of what we've seen, acknowledging that the data are glowing so far.

We can also look at these examples just to further illustrate that. We go through these patients. So this patient just had injections into the liver. This patient, likewise, just an injection into the largest liver tumor. They did have disease outside of the liver, a much smaller disease. This patient only had disease in the liver, but quite a number of tumors in the liver, as you can see. And just the largest one was injected. Likewise, this patient, I believe, is one of marks that only had the disease in the liver and was injected in the liver. This was an MSI-high patient, again, disease in the liver, injected the largest tumor in the liver. This patient was interesting. This patient had quite extensive disease in the liver, but actually, it wasn't the liver, which is injected, the tumor which was injected was in the thigh, but both the thigh and the liver responded to achieve a complete response, which is ongoing. And then this final patient had quite a lot of disease in the liver, where just one lesion, the largest was injected in the liver. But the patient also had quite a lot of disease outside of the liver, none of which was injected and all of which responded and this patient has an ongoing partial response now, and an actual fact, the PET scan was done a day or 2 ago, and none of the sites of disease are avid any longer, including large tumors in the lung.

Dr. Middleton, this is especially for you, does the LAG-3 plus PD-1 data from Bristol, used a way -- or delay the use of potential oncolytic virus therapy?

No.

I would add that we at Replimune are very keen to combine with things which aren't anti-PD-1 including additional promising immune modalities. So that includes things like anti LAG-3, anti-TIGIT. Also molecules which impact the macrophages in the tumor microenvironment as well or regulatory T-cells. And we're also interested in combining with things which aren't immunotherapy at all, combining with standard chemotherapy agents and targeted therapies where we think there could be a lot of potential value. There, particularly to combine with standard of care in tumor types, which are not currently treated with immunotherapy.

Yes. And to expand on my answer, I think where we start to see the potential for changes in standard of care in earlier lines of treatment, that might certainly change the patient population, for example, in melanoma, that comes to us as a PD-1 progressor. But I think the bottom line is that if you look at the therapeutic landscape at the moment, I don't foresee any significant changes, which mean that we end up with a fundamentally different patient population that might have a different response to this approach. If we look beyond melanoma, I'm a big fan of combining immunotherapy with chemotherapy. I think that the data from non-small cell lung cancer, the early data from esophageal cancer, which is my other tumor area of interest away from drug development, point to the potential there. And this program is in its early stages. It's asking a very particular question. I think it's asking the right questions to its next stage of development. But with infinite time and resources, then one could see adding it in much earlier in the patient pathway and in combination with chemoimmunotherapy, for example, as being a reasonable way to go. But as you'll know, these are significantly more expensive and therefore, potential risky studies and well, I leave it to the company to discern when they have the runway and the data to support that growth.

Yes. I mean, we haven't gone into that level of detail yet, but the RP3/2 liver match program isn't intended to be limited to only combining with anti-PD-1, but also potentially combined with actual standard of care in tumor types where anti-PD-1 isn't active or approved.

Next question. It seems that persistence pace where continued treatment often results in PRs after PD-1 therapy. How is this built into the study protocols and what is the plan to treat patients through pseudoprogression in future clinical trials?

So I'll comment first and Mark can further comment. So our protocol has already allowed for pseudoprogression to the extent that progression needs to be confirmed on 2 scans for the patient to go off therapy. If they see progression on a first scan and it's felt to be clinically warranted to continue, the patient gets then consented to continue therapy past progression and carry on treatment. And as you'll have seen, a number of patients where that has occurred, patients have then achieved response. And our response criteria require that for progression to count is progression has to be confirmed, but if it isn't confirmed then they achieve a response, then their best response is still that response. The protocol also allows for treatment for up to 2 years with the anti-PD-1. So they get the short course of RP1 or RP2 or RP3, then they carry on with the anti-PD-1 for up to a standard 2 years and are monitored throughout that process or that period for response and can have a best response documented anytime during that period. So we already account for pseudoprogression in a way which we believe the FDA is fully happy with.

Yes. I think that the principal difference is the rules that we have to see the progression, treatment beyond progression, are pretty much the industry standard rules that you see across that large number of trials we run here at Oxford. So I think the big difference is that it is one of experience. We've treated a larger number of patients in this program. And if you can compare it to true progression versus tumor progression per se, combination with Opdivo in melanoma, we talk about it a lot. We see it rather less often, and for most patients whose runts get bigger on if we move out, the reason they're getting bigger is they're not responding to immunotherapy. And it's the occasion of patient who is a late responder, and is a true pseudoprogression. What's been striking throughout this program has been right from the get-go, the number of patients who had their best response after completing viral therapy and also the proportion of patients with an initial flair, who then go on to get clinical benefit, which does mean that we are able to be much more aggressive than perhaps I would choose to be with the nivo because I think the chances this turns out to be a pesudoprogression are really very much higher. I can't put a quantity on that, and I apologize for that, but it's not something that you track formally within the data. But it's -- our fellows on the music rotate every 9 months. They start slightly cynical that this notion of tumor progression may leave us with the nivo. And I think that's compelling.

This is also why it's sort of important why investigators get experience with our approach because an investigator who, for example, had only treated -- this was their first patient, if it was. So this sort of increase may not be noted to continue, whereas an experienced investigator has treated a number of patients and have seen good responses in them, knows that it's worthwhile to persevere, and then this lady has ended up in a very good place. So experience is sort of very important to learn about what to really expect for your patients with this type of therapy.

We've got another question for Dr. Middleton. For the rapidly progressing non-melanoma skin cancer patients, is RP1 still the most promising approach you have? And would you consider adding some form of chemo to slow the tumor down and give IO a chance to get going.

I think that's a very sensible suggestion. It's a shame Professor Harrington is not here. Because he's really the expert in non-melanoma skin cancer and has treated the majority of patients. I think, to answer the first part of the question, it's very hard to get past it. Very impressive complete response rate from this cohort. So yes, I do think this is the most promising approach we have. What the last 4 patients of -- who have tested it, it's not a panacea. There are limits to what can be achieved, and we will need to think more creatively about how we create -- how we set the circumstances in which RP1 can do its best work. And I think chemotherapy is certainly an optional although it's not terrifically tumor-sensitive, too.

So I mean, those patients were patients or at least a number have already failed prior chemotherapy. They're not patients who haven't already had everything thrown at them. So obviously, we did talk about earlier, combining RPX with chemotherapy as well, which we think is a potentially promising approach. But those patients already had chemotherapy and not benefited from it. At least 2 of them had.

To stay in the obvious, then obviously, what really ought to happen is that those patients are treated earlier in their disease course with the type of regimen that involves an oncolytic and a check-logged [R8] drug, a fair number of those patients will never get to that stage.

Yes. I mean, obviously, as Philip was saying, it takes quite some time for tumors to get to be this sort of size. And if there were a known to be effective therapy available for those patients, they should not really have waited to get to tumors of this size before they were treated. So one would hope in a real-world population post-approval, the patients wouldn't have patient tumors of this size but would be treated far earlier and hopefully, cure them much earlier and never -- I mean, they never get to this rather nasty spot in their life.

On RP2. For RP2 plus Opdivo, were you surprised to see a lack of PRs with the PD-1 naive patients?

Can I just quickly answer that? I mean, so for any patient who has be -- eligible for anti-PD-1 in the combination with RP2, there weren't -- they have to have had anti-PD-1 first. It's a Phase I population who failed all available standard of care. So the only patients who wouldn't have had anti-PD-1 are tumor types where anti-PD-1 is not approved. We didn't enroll, for example, any melanoma patients who haven't had prior anti-PD-1 or any other tumor types where anti-PD-1 is approved, and they haven't had it.

] Another RP2 question. How do you view the response dynamic from RP2/nivo combo, for example, post outright response, post-stable disease or progression?

I think it's too early to say, but based on the data that we've presented, today, if you look at the [ 5 o'clock, the 2 o'clock ] who outlines the individual patient experience, which is all that you can really talk about at the moment, the dynamics do not look significantly different to Mark 1 where we have more mature data. And the key issue there is I don't know of anything else out there at the moment, if you look at RP1, where 1/3 of patients can progress with CB1, you can get a response -- an objective response. Again, that's better than Tebbe and some of the other agents I've been involved in developing in the space over the last 5, 6, 7 years. So to come back to the RP2 question, the early indications are not of anything that's substantially different. More what I expect it to be. If you think about what's different between RP1 and RP2, it's essentially the CTO in force in -- deficient, and we know that the connector to that are relatively slow and -- compared with chemotherapy or childhood therapy. And therefore, it something of a surprise if extreme news comes up.

So this slide here just reminds you of the tumor types which were involved in the RP2 plus nivo combination part. As you can see, they're a mixture of quite rare tumor types. Generally, they really are a Phase I population. They failed everything which they could have under standard of care. So it is a salvage population. And those with melanoma, uveal melanoma, head and neck, have already had anti-PD-1 before coming on to the trial. And obviously, the other tumor types, anti-PD-1 is not approved because it's either not being tested or known to not be effective.

We've a question. What are the considerations for deciding whether you will be advancing RP2 or RP3 in any setting?

So we definitely think that the RP2 data, both as a monotherapy and in combination is extremely promising. And if we didn't have RP3 following along rapidly behind, I think we'd be very motivated to enter broad development with RP2. However, as we do have RP3 following quite rapidly behind, we think it is sensible to gather additional data with RP2 and the initial data with RP3 before deciding which to push the button on for broad future development. There is a sort of interim potential approach, which is progressing RP2 in 1 indication, for example, uveal melanoma, where we've clearly got a signal and then reserving everything else for RP3. But exactly what we do will depend on the data as it accumulates over the rest of this year and early next, which will tell us whether we should do everything with RP3 or do most things with RP3 and just a smaller number of things with RP2, for example, or in any combination thereof. So it's all going to be data-driven, and we'll follow the data. But if RP3 really has properties we really like, the intention would be to do most things going forward with RP3.

Another question here. Can you elaborate on how you're thinking about a registrational development path for patients with liver metastasis?

That really is too early at the moment to further comment on. We do have some internal things we've been thinking about, but it's too early to comment on publicly at the moment.

A question. It appears like persisting through pseudoprogression is providing good outcomes. How consistent are the protocols in allowing for this persistence when there seems to be initial progressive disease?

All of our protocols are exactly the same wording. So it's entirely standard across all our -- everything we're doing.

Yes. Just to reiterate the point made earlier. What you write in approach goal is very different than what you train the investigator and the value of experience there. So all protocols carry roughly the same language amongst investigations these days. So it's very much about the experience with the agents. And I think the power of anecdotes, either somebody else's or when an investigator first sees for themselves what happens by persisting, that drives this. And I don't think there's a way around that. It's about choosing investors of a site carefully. It's about picking people with the relevant experience and I think also the right patient in patient population will be greatly helped because the nature of this early part of the program is that patients don't have huge numbers of options when they come off the program. And therefore, it's not that they're going to come off because they're attractive to some other effect. They are likely willing to persist.

I think it's also about communication, and when I say communication, I mean, with regard to running the trial, such that we do have investigator meetings, such that the investigators are hopefully properly collaborating together and all as a group, aware of the group experience. So even if they themselves have only recruited 1 or 2 patients and haven't necessarily seen this sort of thing occurring as yet or indeed haven't recruited a patient at all, the group experience is communicated such that the group motivation is there.

Another question coming in. Given the desirable investigator experience with RP, what are the company's plans to get U.S. physicians and KOLs involved in trials?

The protocols are active. So the RP1 protocols are active globally. In the U.K., in Australia, in Europe and in the U.S., we have a substantial number of U.S. sites participating in the CERPASS study and participating in the IGNYTE RP1 trial, including the melanoma 125 patient cohort, and they are delivering substantial numbers of patients into those trials. With RP2 and RP3, we've always taken the approach at Replimune due to our relationships with investigators and history that we've conducted the initial Phase I part of trials in the U.K., in collaboration with particularly Mark, the Institute of Cancer Research, Royal Marsden in London and now Joe Sacco in Liverpool and then moved out more globally, in particular in the U.S., as we get past Phase I, and the same would be intended to be the case with RP2 and RP3. So the liver mets program we've been describing or in early outline would be conducted globally, particularly including the U.S.

Rob, can you talk about the advantages in CD40 and 4-1BB over other co-stimulatory agents?

Yes. I mean, I think the greatest advantage of 4-1BB and CD40 over the other costims, which include things like OX40, GITR, ICOS, are that while those other pathways are clearly very active in mice, the human experience with those other pathways has been really rather disappointing. CD40 and 4-1BB are also active in mice, but not any particularly more than OX40, but there is human data which suggests that targeting CD40 or 4-1BB in humans really is a workable thing to do in humans. So the reason we chose them was based on not only the fact that we could get very good outcomes in mice, but also in contrast to the other things, we also tested, which went well in mice, antibodies agonistic antibodies developed by others have shown promising early activity, although in some cases, compromised by toxicity caused by the systemic administration approach of the antibody. So we obviously aim to retain the activity but get rid of the toxicity by delivering the activating ligand directly into the tumor where it's really needed and limiting the systemic exposure to a very low amount.

What is the likelihood of generating CRs in visceral tumors by RPS, the abscopal effect?

So we have demonstrated CRs in visceral tumors following administration of RP1 into superficial tumors. So we do see that.

Those 6 patient examples, as Rob previously presented, including examples where the liver has been injected, when you've seen abscopal effects in nodes in the abdomen, you've seen on abscopal effects and in the spleen after injection in the liver with complete resolution, disease in the spleen. And as Rob also referred to, resolution of disease as far away as the lump, where a metabolic complete response has recently been declared in patient 6.

Yes. So this patient #5 is specifically a patient who had a lesion in the thigh, which was injected and also these lesions in the liver, which completely resolved.

So these are systemic responses.

We have a question here. What data do you need to be able to move into earlier lines of treatment in larger cohorts and when you anticipate having this?

Well, with -- in CSCC, we're already conducting a registrational study in the first-line setting or at least in the patients that are eligible who haven't had any prior therapy. They just need to be not eligible for radiation or surgery. So already in the first-line setting there. But for us to get earlier in disease courses in general, it will be starting to combine with whatever the standard of care is in the first-line setting, whether or not that first-line standard of care happens to be an anti-PD-1 agent. One can also think further down the line to particularly neoadjuvant approaches in various diseases, which is something we're also extremely interested in. So a combination of the 2, really.

What are your thoughts on evaluating RP2 for second-line PD build, head and neck squamous cell carcinoma?

We certainly think that, that's a potentially attractive indication for us. However, also upfront combination with chemo radiation in a neoadjuvant approach is also a potentially attractive indication in head and neck cancer. We've certainly got interest from investigators to test RP2, particularly in head and neck cancer. And we'll see where that interest leads.

Next question. Do the costims built into RP2 and 3 play a role in the abscopal effect or do they only play a role in the injected tumor?

The aim of expressing those is to increase the systemic immune activation and therefore, the abscopal effect, we would expect we have more limited impact on the injected tumor. The idea is they're expressed in the tumor and provide the co-stimulatory signals to maximally activate T-Cells to provide a systemic army of T-Cells to systemically treat disease. So the aim is to increase systemic effect, not local effect. An actual fact, if you look at our mouse data, you see that adding in the CD40 ligand, 4-1BB ligand or indeed, anti-CTLA-4 has a much greater impact on uninjected tumors than it does on the increased activity in injected tumors.

We have another question on RP2. With RP2, are you seeing deepening of responses over time in either the monotherapy or in the early PD-1 combination data.

Mark can also further comment, but if one looked back at this slide again, absolutely, we are. As I've said earlier, they get the short cohort -- they got the short cohort course of 5 doses with single-agent RP2. At the first scan, this patient was a stable release which then deepened into a partial response, which has been maintained out to the present day. And as Mark said, has recently had a PET scan indicating no metabolic evidence of active disease. Similarly, this patient at the first scan after completing the course of RP2, the patient had a stable disease, which then deepened to a good partial response. Unfortunately, as Mark indicated, he succumbed to his disease at roughly 15 months. And then this third patient had a PR at their first scan which deepened to a complete response, which is still ongoing in a very stunning fashion. I was talking to Kevin Harrington yesterday who's the investigator. And this patient really is a true miracle as compared to what would be expected for his prognosis otherwise. So in all these 3 patients, depth responses deepened over time and after completing the course of RP2 treatment.

Great. Maybe time for a couple more. One coming on RP1. Does RP1 re-initiation always involve a new lesion? When it worked, was the lesion a deeper lesion than the original and how many lesions are generally injected?

So we only have a very small number of patients so far who have had re-initiation. As I've said, in most cases, just the first quarter of 8 is about right. I believe it's 4 patients who've been reinitiated with RP1. And as I think about, I haven't got it off the top of my head, but they have so far been 1 patient who had a local recurrence in the area of the original tumor in the nose, which is responding to that re-initiation of that local recurrence. The second was the local recurrence in the thigh, the example we showed, which is the site of the originally injected tumor. The patient additionally had tumors in the groin and in the lung, which responded in the first place and have not further progressed. So the only site of relapse actually was the initial tumor in the thigh, which was reinjected and responded. So that's 2 of them. The third patient is Mark's patient who had disease in the shoulder, as far as I remember, who initially achieved a very durable stable disease. And then that begun to progress and re-initiation occurred. And now they're now down to nearly a partial response, better than their first response. So it's again at the local site of the initial disease. And then the last patients is a patient, had lots of lesions in their leg who achieved a very good and durable partial response for whom re-initiation was pursued not because any sort of relapse, but to try and convert that patient into a CR and that patient is -- don't seem to have further benefit from that second course. So there hasn't been a case where there's been a distant relapse, which has been reinitiated. It's actually only been very small local relapses, which so far have been kept under control by the re-initiation.

Let me -- to max in new regions, or lesions that have arisen where previously one gone and then come back. And then a couple of patients, there has always been something to speak to the original question.

Yes. So we have -- I was talking about RP2. So with RP2, we've had a number of patients who've had some new lesions appearing after the initially injected lesion had gone away. Actually before the full 8 have been used, and then those new lesions were injected and response had occurred.

And a final question. When do you expect to provide more specific time lines for CERPASS and IGNYTE pivotal relapse?

Philip can answer that. Philip?

Yes. My current guidance is quite broad with the primary readout within the '22 time frame, obviously, as we move to second half, is appropriate given more granularity on that, and we will do so within the next 3 months or so in terms of giving guidance and when we expect full enrollment and when we expect the primary readout probably at the next quarter's release.

So just before turning it back to you, Philip, if we didn't answer any questions, we'll try to get to those future in the future in future calls, but we'll turn it back over to you Philip.

Yes. I just want to thank everyone again for their time this morning. We look forward towards the end of the year, giving more granularity on this -- particularly this Phase II program in liver. This will involve prevalent tumor types, like colon, breast and lung. And obviously, look forward to future updates as well on the data side around the year-end. And I'd just like to thank Mark for his participation and again, everyone, for their time and time to close out the meeting. So thank you.