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

Ladies and gentlemen, thank you for standing by and welcome to the Replimune RP1 Data Update from its Phase II Cohorts in Melanoma and Non-Melanoma Skin Cancer Investor Event. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, to Philip Astley-Sparke, CEO. Thank you. Please go ahead, sir.

Good morning, and thank you for joining our midyear data update event. And before we begin, I need to highlight our forward-looking statements. I am 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. Kevin Harrington, Professor in Biological Cancer Therapies at The Institute of Cancer Research, London, and Consultant Clinical Oncologist at The Royal Marsden. Dr. Harrington will describe some of his experiences of treating his patients with cutaneous squamous-cell carcinoma with RP1 combined with Opdivo. Also with us is Dr. Mark Middleton, Professor of Experimental Cancer Medicine and Consultant Medical Oncologist at the Oxford Cancer Centre, U.K. and Head of the Department of Oncology at The University of Oxford. This is the agenda for the call, which will start with a brief further introduction from me and a brief summary of the high-level results by Rob, followed by a discussion of a number of the individual patients by Dr. Harrington and Dr. Middleton. Rob will then provide a further pipeline update, followed by some closing comments by me, after which we'll go into Q&A. Dr. Harrington's biography is described here, where, as you can see, he has a wealth of relevant experience in oncology and with novel therapies, and in particular, oncolytic viruses in addition to treating patients on our trial. And likewise, Dr. Middleton has extensive experience in treating late-stage cancer patients and in the development of novel anti-cancer therapeutics. I want to start with a brief summary of Replimune's purpose. Despite all the advances in immuno-oncology, particularly antibodies targeting anti-PD-1, it still remains the case that the majority of patients see little or no benefit, most probably due to a lack of a pre-existing immune response to their cancer. At Replimune, we believe that oncolytic immunotherapy has the potential to be the most effective and practical way to initiate an anticancer immune response. Replimune was started to realize the full potential of oncolytic immunotherapy and we believe our products will become the second cornerstone of immuno-oncology. In order to achieve this ambition, we are taking a stepwise approach. RP1's development is focused on immune responsive tumors where we're conducting a 240-patient registration directed, randomized controlled clinical trial in our lead indication of cutaneous squamous cell carcinoma, or CSCC, and have also recently initiated a 125-patient single-arm study in anti-PD-1 refractory melanoma. We will present these data today from our ongoing Phase II clinical trial, RP1 in combination with Opdivo, focusing on the patients enrolled with CSCC and PD-1 refractory melanoma, which we believe support a positive outcome in the larger registration-directed studies underway. As a reminder of our technology and mechanism of action, oncolytic immunotherapy is the use of viruses that, when injected into tumors, partially or completely destroy them through virus replication, bursting the tumor cells open and exposing all the released cancer antigens to the immune system in an environment of 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 virus is the best species of virus for oncolytic use, it being both highly lytic and inflammatory and having the ability to carry multiple immune stimulating proteins into the tumor to further amplify the immune response generated. Our strain of HSV has been deliberately selected for its lytic properties in human tumor cells. And from all our products as our base platform, we express a fusogenic protein from the virus that greatly increases the direct tumor killing, immunogenic cell death and the systemic immune-mediated antitumor activity of the construct. Our lead product RP1 additionally expresses GM-CSF, RP2 additionally expresses an anti-CTLA-4 antibody and RP3 to immune-costimulatory pathway activated proteins targeting CD40 and 4-1BB. The Phase II development plan for RP1 is shown on this slide. The data providing an update on today pertains to the top 2 red bars, where we're fully enrolled the melanoma cohort and enrolled approximately half the non-melanoma skin cancer cohort. The non-melanoma skin cancer cohort contains predominantly patients with CSCC, all of whom are anti-PD-1 naive. The melanoma cohort consists of anti-PD-1 refractory cutaneous melanoma patients in addition to anti-PD-1 naive, together with patients with uveal and mucosal melanoma. The CSCC data and the anti-PD-1 refractory cutaneous melanoma data are of primary importance today as they provide a guide to the likelihood of future success in our larger and registration-directed studies in these 2 settings, depicted in the top yellow bars. However, it should be borne in mind that both data sets are still immature inasmuch as they include patients who remain on therapy and may therefore further respond or progress. Summarizing the data, we are very pleased to be able to report, with the first 7 CSCC patients we've treated and who have follow-up scans, 6 are currently in response, all ongoing at out to 16 months post the start of treatment. Four of the 6 responses are complete responses, each in patients with prior advanced aggressive disease and who now, depending on durability, have the potential for cure. The anti-PD-1 drug Opdivo used in our study reported no complete responses in 24 patients this ASCO. Merck's Keytruda achieved only 4 complete responses out of 105-patient Phase II dataset and Regeneron's Libtayo has an on-label initial complete response rate not far or above. We believe this high rate in complete response provides clear differentiation of our combination as compared to anti-PD-1 alone. Switching to melanoma. Of 16 cutaneous melanoma patients that have failed prior anti-PD-1, 5 have been declared responsive, including patients with extensive visceral disease, such that current response rate is 31%. Two patients remain on study with the opportunity for response, which could increase the rate further. The response rate to a second line of anti-PD-1 alone in this population would be expected to be in the mid-single digits. In that, given 4 out of 5 of these responding patients have previously failed not just anti-PD-1, but also anti-CTLA-4. We believe that both of these data sets point to a high probability of success in our lead indications where we have registration-directed clinical trials ongoing and provide a solid foundation to build towards our goal of making our products a universal second cornerstone of immuno-oncology. I will now hand over to Rob to talk about the commercial opportunity and the data some more.

Thanks, Philip. I'll first turn to cutaneous squamous cell carcinoma and talk initially about the disease and the market opportunity. Cutaneous squamous cell carcinoma is the second most common skin cancer with approximately 700,000 cases annually in the U.S. and as for melanoma, largely results from sun exposure to the skin. While the majority of these 700,000 patients are cured by initial surgery, the disease recurs in about 10%, after which it's high-risk disease with no better and quite possibly a worse prognosis than for many patients with melanoma. Advanced cutaneous squamous cell carcinoma can either be metastatic, that being it's spread beyond local draining lymph nodes or local regionally advanced. And in fact, unusually for cancer, it's actually local regional disease, which results in 80% of the mortality for patients, usually through invasion of critical structures often in the head and neck area. Therefore, treating not only metastatic CSCC, but also locally advanced disease represents a significant medical challenge and unmet need. Deaths in the U.S. from cutaneous squamous cell carcinoma are estimated at between 7,000 and 15,000 per year, representing the most conservative population which could be addressed, which rises to up to 28,000 patients, including a higher proportion of the patients with recurrent disease. So far, there's only one therapy specifically approved for the treatment of CSCC, that being Regeneron's Libtayo, which is in the process of superseding the use of the prior chemotherapy-based approaches. As shown on Slide 12, Regeneron made great strides in the commercialization of Libtayo for CSCC, with the first-year sales of around $200 million and steadily increasing market share, including, compared to other anti-PD-1 agents which were all selling these off-label. We do, therefore, believe that CSCC has a significant market potential for Regeneron -- for Replimune, which is one of the reasons why it was chosen for the lead indication for RP1. This slide highlights the efficacy data with anti-PD-1 agents in CSCC, where, as you can see, with Libtayo in something over 100 patients, an approximately 50% response rate was seen both in locally advanced and metastatic disease, the vast majority of which were partial responses. While the complete response rate has risen over time with longer follow-up now to over 2 years, at the primary analysis, this was, as Philip said, in the single-digit range. For Keytruda, where a 105-patient Phase II trial was conducted, the response rates and CR rates were lower with a 34% objective response rate and 4% CR rate, which probably reflects a particular patient population enrolled. And for Opdivo where a 24-patient trial was reported at ASCO, the response rate was around 55% with no CRs having been seen. From this, you can see that the numbers we likely need to beat with RP1 are a response rate of somewhere between -- around 35% and 55%, depending on the population you enroll and where we believe we can also achieve a CR rate two to threefold higher than with anti-PD-1 alone. As Philip described, we have a registration-directed Phase II trial of RP1 combined with Libtayo versus Libtayo alone in 240 patients enrolling at the moment, where to achieve a statistical win, we need to see an approximately 15% absolute improvement in ORR in the combination arm but where, as I said, we would also hope to see at least a two to threefold improvement in CR rate too. As a result, we're particularly interested in how our current data with Opdivo supports our expectation, which, as you will see, we believe is solidly the case based on the data so far. We do also have a further clinical trial open for enrollment in CSCC using single agent RP1 rather than combination with anti-PD-1, this being solid organ transplant recipients with CSCC. CSCC is a particular problem for all these patients due to their immune compromised status with up to 70% of patients developing CSCC within 20 years. And so while it's a relatively niche indication, it does provide a significant unmet medical need and the use of anti-PD-1 therapy isn't indicated due to the high-risk of rejection of a transplanted organ, which can be up to 40%. This, therefore, gives us the opportunity to generate clear single agent efficacy data with RP1 as well as a potential for a niche approval and where we hope to have some initial data to share by the end of the year. Finally, we're also considering a clinical trial in the neoadjuvant setting for RP1 combined with anti-PD-1 in CSCC, which we're planning at the moment, which may be an investigator-initiated study. So with that summary of the CSCC landscape, I'll move on to the current data on CSCC, which is summarized in the next slide. So far, we've enrolled 9 patients with CSCC in the nonmelanoma skin cancer cohort in the clinical trial of RP1 combined with Opdivo, together with 5 patients with a number of other skin cancer tumor types, as indicated on the right. Of the 9 patients with CSCC, 7 have follow-up scans and it will be these patients we're talking about today. Of the patients enrolled, roughly half had a locally advanced disease and half had metastatic disease, with roughly half also having had prior systemic therapy in addition to surgery or radiation. As a reminder, 80% of patient mortality is driven by local regional progression in this disease. Of the 7 patients with follow-up scans, we're very pleased, indeed, to report that all but 1 of these has good -- responded to therapy, with 4 out of the 7 having ongoing complete responses and 2 having ongoing partial responses with all 6 of those patients currently continuing on treatment. We're, therefore, very pleased with both the reproducibility of the effects we've seen and also with the depth of the response we're seeing, which we do think is clearly distinguishing from anti-PD-1 therapy alone. As you'll see when we get to the patient examples, these responses are in patients with extensive and difficult-to-treat disease with complete responses not only at injected sites, but also in distant uninjected visceral sites, demonstrating the systemic benefit the patients are receiving. The depth and kinetics of these responses are shown on the next slide. The waterfall plot on the left shows the depth of response and the swimmers plot on the right, the timing of response, including that all are ongoing to date between approximately 6 and 16 months from the patients having started the trial. And as I said, we do believe that the data, albeit in a relatively modest number of patients so far, is clearly distinguishing from the effects of which you might see with anti-PD-1 alone, particularly bearing in mind the profile of the patients we've enrolled. With that, I'll hand over to Dr. Harrington to describe some of the patients that he's treated with CSCC in the trial so far. So Kevin?

Thank you, Dr. Coffin. I'm very grateful for this chance to present some of the cases that Rob has already outlined to you in the summary data. And so I'm going to show you examples, focusing on really difficult-to-treat cutaneous squamous cell carcinomas of the head and neck region, which, as again, Dr. Coffin has already explained to you, are associated with significant morbidity and indeed mortality for our patients. And I'm going to show you these patients sequentially, and these are patients who I have personally treated. So this is a gentleman who presented with a recurrent cutaneous squamous cell carcinoma low in the right neck. You can see the lesion there in the initial baseline photograph. And I guess on first inspection, it doesn't look as if it's too extensive. I'll show you the scan in a moment. And you'll see that this is indeed quite an extensive lesion. The patient had contralateral neck disease, also had disease in the retroperitoneal, in peritoneal cavity and also had extensive bone metastasis. He had previously been treated with platinum-based chemo radiation and had indeed also received first-line systemic chemotherapy with combination platinum 5-FU. So both the large injected tumor and the contralateral neck disease, which I will show you on the CT scan, showed a reduction after the first injection with RP1 before we initiated treatment with Opdivo. The next slide shows us data from this patient's CT scans, going only as far as 24 weeks. This patient has been followed for significantly longer than that now and is in a maintained complete remission. But you can see, as I hinted from the photograph, at the baseline on the scans on the left-hand side, at the top left, you see the status of disease at the outset of treatment, high in the neck and then lower down adjacent to the thyroid gland. And you can see this is really a very unpleasant and highly symptomatic recurrence of disease. And on the right-hand side of the slide, you can see a smaller but nonetheless, visible disease on the contralateral neck on the baseline scan. And what I hope you can appreciate is really the very rapid resolution of this disease. And indeed also a resolution of disease in the retroperitoneum. You can see that both before and after in treatment at 16 weeks, this patient achieving an ongoing CR at these sites. But I think even more impressively on the next slide, I show you data from this man's CT scan, and here is a parasaggital image looking at the extent of bone disease that this man had extensive lytic deposits. And as we go through treatment, you can see increased brightness, whiteness indeed of the vertebral bodies, which have been arrowed for you showing healing sclerosis of that disease that was causing him symptoms. And of course, we wondered whether or not this is active disease. Regrettably, we did not have a baseline pretreatment PET scan, but I can show you on the next slide, a PET scan performed for the patient when he did have active disease previously in 2018. Indeed, that volume of disease underestimates the volume of disease he had on study entry. But on the right-hand side of the slide, you can see a parasagittal and then a coronal image taken in February 2020, indicating a metabolic response, a significant response amounting to a complete remission of his disease. Biopsies of this man's lesions demonstrated no active disease in both injected and uninjected sites. And when we look at both the baseline and then the on treatment week 6 samples, you can see on the left-hand side, the CD8-positive staining, and on the right-hand side, the PD-L1. Very little T cell infiltration at baseline, a smattering of PD-L1 staining at baseline and dramatic and impressive changes at week 6 of treatment showing both CD8 infiltration and PD-L1 staining. Now the next slide demonstrates a patient and I'm hoping it's coming soon, a patient with, again, what I hope you will appreciate is an extremely unpleasant recurrence of disease at the supraclavicular fossa on the left-hand side. This is low in the neck on the left-hand side. A patient who had been treated previously with surgery and then with radiotherapy with immediate relapse of his disease and the development of systemic metastasis with disease in the lung, which I will show you on a scan. You can see the time course of response of this patient. So with flattening of the lesion after the first injection of RP1. And again, this is illustrated for you. Thank you, Rob. And you can see flattening of the lesion after the first dose before the start of Opdivo. And you can see that this patient has an ongoing partial remission of his disease. And on the next slide, what we show you is the evidence of that on the CT scans. And again, you can see the before and after scans, at multiple levels in this patient's neck, showing a dramatic response in the exophytic tumor, albeit the fact that the ulcerated area has been relatively slow to heal. The active disease around the margins of it has shown a significant reduction. And you can see also a response in systemic disease with a reduction in lung metastasis. On the next slide, I show you data from another patient. So again, an extremely unpleasant recurrence of a squamous cell carcinoma of the skin that had arisen around this lady's left eye that had required multiple surgical procedures, none of which was able to achieve local control. And at the time that I saw this patient, she had become inoperable with a deeply infiltrating cutaneous deposit that was eroding into the left maxillary antrum and essentially untreatable with surgical approaches. She entered on to study. And after the first injection of RP1 before we had started nivolumab, she noticed an improvement, not only in the extent of the disease but also in her pain symptoms. And as you can see over the sequential photographs, remarkably, this lesion has essentially healed and biopsy from this lesion has confirmed complete remission of the disease. In the next slide, I show you a patient again with a very unpleasant recurrence of a squamous cell carcinoma arising in the skin of the nasal vestibule. This patient who had been recommended by surgical colleagues to undergo a complete rhinectomy, removing the whole nose and replacing it with a prosthetic nose had understandably declined surgery and had gone forward instead for combination chemotherapy and radiation therapy. Regrettably, her disease relapsed almost immediately after completion of radiation therapy, and she entered into the study receiving an initial injection of the RP1 agent. And you can see a dramatic response. Indeed, an apparent worsening of the appearances of the lesion, but very rapidly, as she began the nivolumab treatment, you can see a deep response and then indeed a complete remission, again, confirmed by biopsy with no evidence of active disease at the biopsy site, this patient is achieving a response. And indeed, in a further photograph on the next slide, we see the differences in the appearance of this patient -- I beg your pardon. This is the -- thank you, Rob. This is the appearance that you can see for the patient, really a dramatic change from baseline to Day 57. And if you go back, please, to the CT scan, you can see that at the level of where this disease was most bulky on the CT scan demonstrated in the December 2019 scan compared to the subsequent February and then again in the April scan, a complete radiological remission of this disease. The scans appear, the disease -- the amount of brain you see on the scans is slightly different, but they have been windowed in order to see the same area of the nose. And again, from this patient, the sequential biopsy show at screening versus Day 43 a significant increase in the number of CD8-positive T cells within the tumor and an increase, less dramatic than for the previous patients, but nonetheless, a significant increase in PD-L1 staining within the tumor.

Okay. Thank you very much, Kevin, for those thorough description of your very interesting patients, which I certainly do believe show the significant clinical benefit is being achieved by the patients on the combination therapy. And obviously, in particular, the frequency and depth of response, particularly complete responses, really does seem to be very different to what you might expect with Opdivo alone. So we'll now talk about the data generated in melanoma with a focus on cutaneous melanoma patients with anti PD-1 refractory disease. So in melanoma, while great strides have been made in the development of new drugs, including with immune checkpoint blockade, it does remain the case that roughly half of advanced melanoma patients will still die of their disease. And bearing in mind the standard first-line option is now treating with anti-PD-1 or anti-PD-1 combined with anti-CTLA-4, there's obviously now a significant unmet need for patients who've failed those approaches, after which there's no approved or known-to-be-effective therapy. Data shows that the response rate to a further course of anti-PD-1 following clinically significant progression on the first course of single-agent anti-PD-1 is in the 6% to 7% range and where Yervoy may be used as a second line option after failing single agent anti-PD-1, the response rate to Yervoy is just 13% as was actually reported at this ASCO in 162 patients. So clearly, where prior combined anti-PD-1 and anti-CTLA-4 have both been failed, these responses to a second line of therapy would be lower still. Each of these numbers putting into context the significant unmet need for patients who failed anti-PD-1 or combined anti-PD-1, anti-CTLA-4. Hence, based on the early signals of activity we've seen in this population and the great unmet need and therefore, market potential, we announced in January our intention to enroll a 125-patient cohort of patients with anti PD-1 refractory melanoma to be treated with RP1 combined with Opdivo, for which we're currently recruiting patients. We intend to be discussing the details of exactly what we need to achieve for potential approval with the FDA later this year within this clinical trial, us targeting patients who would not be expected to respond to a second line of anti-PD-1 therapy as all will have received the first-line for at least 12 weeks and have had progression documented on at least 2 successive scans. Therefore, in that background, of particular interest to us is how our current data in anti-PD-1 refractory melanoma supports the potential future outcome of this ongoing 125-patient trial, which brings us to the data we currently have in melanoma, which is summarized on this slide. So, so far, 36 melanoma patients have been enrolled into the clinical trial, including the Phase I expansion cohort of RP1 combined with Opdivo and in the Phase II melanoma cohort. Of these, 16 had anti-PD-1 refractory cutaneous melanoma. So 9 of these 16 patients showed initial clinical benefit, which was defined as stable disease or better with evidence of antitumor activity, which is tumor shrinkage and/or biopsy evidence of antitumor activity with 7 patients currently still ongoing on treatment, 5 of whom have achieved an actual response and 4 of those 5 having failed anti-CTLA-4 in addition to anti-PD-1. That means that with 2 further patients still on treatment and with the opportunity for response, we have a minimum final response rate for this group of 31%, which may rise to up to 44%. And bearing in mind what's been seen with other approaches in this setting, we do think that, that range of 31% to 44% is very promising indeed. So we do, therefore, think that this look at what's still an immature data set is highly supportive of our approach and of our expectations for a positive outcome for our 125-patient trial, which is currently recruiting. So we've also enrolled 8 anti-PD-1 naive cutaneous melanoma patients, of whom half have currently achieved a response with 2 further ongoing with the opportunity for response, and therefore, the final response in that group of patients will, therefore, be between 50% and 75%. For the difficult-to-treat melanoma subtypes of mucosal and uveal melanoma, we've also seen activity with 6 patients having been enrolled with each of those tumor types. And in mucosal melanoma, 2 of the 6 patients have responded, 1 having failed prior anti-PD-1 and 1 treatment naive. And in uveal melanoma, where all the patients enrolled have failed prior combined Yervoy and Opdivo, 2 patients are ongoing with 1 of those patients having achieved a 27% disease reduction by RECIST of extensive liver disease, which is currently a 61% reduction when assessed by WHO bi-dimensional measurements, which better assess volume. So clearly, that patient's doing very well. Overall, therefore, we've seen a very promising activity across the different melanoma settings, including data which we think is very supportive of our currently enrolling cohort with anti-PD-1 refractory disease. So this slide describes the baseline demographics of the melanoma patients enrolled, which in particular makes the point that we have a very late-stage population, more advanced than most melanoma clinical trials. Within the anti-PD-1 refractory cutaneous melanoma group, of particular interest here, 87.5% of those patients having stage 4 M1b or stage 4 M1c disease, that being having visceral metastases of liver, the lungs or elsewhere, so very advanced patients. We'll now move on to the example of melanoma patients with Dr. Middleton from the University of Oxford, who's actually treated most of the patients -- most patients with RP1 of any investigator so far, who will take you through. Dr. Middleton will also summarize the safety data with RP1 combined with nivolumab together with some of his observations indicative of clinical activity in some of other tumor types as well. So over to you, Mark, to talk about those patients and also any other general observations you may have.

Thanks, Rob. So on the next slide, we'll start with one of the first melanoma patients. I should add that not all of these are my patients. We have a very diverse group of investigative sites, all of whom have contributed to this part of the slide presentation. So this is a patient treated in the U.S., who is refractory to ipilimumab and nivolumab with baseline disease in the liver and lung and only the largest liver lesion best seen in the top line of images second from the left was injected. And what you can see here, clearly, over the course of 5 months on treatment, is an impressive shrinkage in the number of liver lesions and then a similar evidence of an abscopal effect on the next slide when we look at the lung CT scan, with the lesion adjacent to the mediastinum on the right-hand side showing particularly impressive, diminishing in size and well away from the original area of injection. The next patient is also refractory to ipilimumab and nivolumab. Their baseline disease was mainly subcutaneous in the first, the leg and the groin with injections into the lesions in the heel and the groin. And as with some of the patients that Kevin has shown initially, if you look at the second column of pictures, we saw some evidence of disease progression, which then started to show much more impressive shrinkage going from left to right over the course of 9 months, and response in all areas of disease. And it may well be that this is now just melanosis, but we require a biopsy in order to understand whether this is a complete response or an ongoing partial response. And looking at that on the CT scan, we can see what's happening in the groin lesion and also in some of the thigh lesions in the top row and the bottom row there, with very similar radiological evidence of response to that seen on the photographs. The third patient treated in the U.K. in Liverpool by Joe Sacco is another patient with extensive subcutaneous disease in the thigh, lymph nodes in the groin and also lung lesions. And what you can see here is evidence of a very rapid effect of injection of RP1 on its own, which is given briefly before the Opdivo starts. So you can see that in the 2 left-hand pictures with an improvement in appearance and flattening of the lesions before the Opdivo starts and then even more impressive response with combination treatment. And if we look at this on the scan, it's perhaps more impressive than with the photographs because you can get some appreciation of the depth of the lesion and how it improves, now going from top to bottom during the course of therapy, and this patient is still on treatment nearly a year out now. Further images show what's happening in the lung and also in the lymph nodes, with more evidence of an impressive response globally to treatment. Then looking at this from the biological point of view, what we see here on the baseline pictures is one of the classical mechanisms of resistance to checkpoint inhibition, in particular, in immunotherapy in general, which is T cell exclusion. So what you can see here in the brown rim around the tumor, if we look at the top left, is that there are CD8-positive cells in the biopsy, but they're all held at the periphery of the tumor, and that's also recapitulated with the PD-L1 expression. But when a biopsy is repeated during the course of treatment, so 7 weeks in, you can see evidence of T cell infiltration into the middle of the tumor and associated PD-L1 staining as well, which this being in an ipi/nivo refractory patient, I think, gives a very impressive demonstration of a reversal of a well-understood resistance mechanism to the standard of care. I'd like to turn now to mucosal melanoma patients. They do respond to checkpoint inhibition in the same way as cutaneous melanoma but not nearly so frequently. I mean, in a combined analysis of the early pembrolizumab data from the KEYNOTE-001, 002, 006 studies, for example, we know the response rate to PD-1 is about 19% compared with 33% in cutaneous melanoma patients. So though it is checkpoint-inhibitor-sensitive, it's not nearly as sensitive as is the case for cutaneous disease. So here's a patient with mucosal disease. You can see a very impressive sinus lesion on the top left and more extensive areas of disease in the abdomen as well, lymph node and peritoneal deposits with shrinkage of all of the areas of tumor, not just that, which was injected, which is the primary lesion on the left-hand side. Moving on to one of my own patients. This is a patient who had proved refractory to pembrolizumab in the first-line and came to us for treatment of recurrence -- or a reactivation of the anorectal primary shown here, which although relatively small, would need to be treated surgically with loss of sphincter and devastating consequences for the patients in the longer term. So we started treating her last summer and by the beginning of this year, on the bottom, you can see a small tag is all that remains, which we had removed in a minor surgical procedure rather than the extensive operation that would have been required last year and was shown to be tumor-free. So a terrific result for the patient. And then finally in this melanoma section to turn to those patients with uveal primaries. This is a very different story. The activity of single-agent PD-1 in this setting is very, very low, in single-digit percentage-wise. And even the combination of ipilimumab and nivolumab only results in objective responses in around 1/5 of patients. So this is a patient of Joe Sacco who's up in Clatterbridge in Liverpool in the U.K. with very extensive disease, who is refractory to ipilimumab and nivolumab being the standard approach in this situation. And in whom normally we would revert to any experimental treatment or perhaps even cytotoxic chemotherapy. And what you can see here is a terrific response over the course of only 3 months with a very large scalp lesion, which was injected shrinking to subcentimeter proportions over the course of 3 months. Now this patient also had far more extensive disease at multiple sites, including lung and liver, which also responded, although he did ultimately have to discontinue treatment due to the development of a new brain lesion nearly a year into treatment. The next patient is also a patient with uveal melanoma who was refractory to ipilimumab and nivolumab. And as is common for this disease, extensive hepatic involvement and the vast majority, nearly 100% of patients have liver metastases. And often, this is the only site of disease. And what you can see here is a significant shrinkage not quite partial response by RECIST because we only consider this in one dimension, but using WHO bidimensional measurements, we can see that the volume has reduced substantially and this patient continues on treatment and therefore does have the potential to develop into an objective response in time. So turning now to the safety aspects. I think next slide, please. What you can see here is that treatment is really very, very well tolerated. But this is a combined data set from patients treated with both RP1 and with nivolumab from the nonmelanoma skin cancer and melanoma cohorts. And you can see the low-level reactions in the form of fever, chills and influenza type symptoms are very common, but the higher grade side effect are relatively rare with only the fatigue being experienced by more than 1 patient across the whole of this cohort of over 40 patients. And all of these are readily manageable and some of them not even noticed by the patients, such as the lipase increase in the seroconversion test, with very few patients being discontinued due to treatment-emergent adverse events at 10%. And this compares very favorably with what we see with PD-1 alone in cutaneous disease. So moving on to other tumor types. First, that we've got a patient with angiosarcoma, another skin cancer tumor type, in which you can see very, very extensive disease on this photograph from late last year, which has visibly improved by February of this year -- oh we seem to have moved on, by February of this year. But, unfortunately, the patient did have to withdraw from treatment due to side effects from the Opdivo side of the combination. The next patient is a patient of mine, a young woman with microsatellite unstable rectal cancer who had progressed through neoadjuvant disease -- neoadjuvant chemotherapy with a fluoropyrimidine, irinotecan and oxaliplatin in combination and came to us for treatment with the combination of RP1 and Opdivo. And similar to some of the patients that Dr. Harrington has shown you, you can see here that there was extensively progressive disease soon after starting treatment with significant enlargements of the exophytic portion of her disease relapse at the stoma site. But this then very quickly started to improve so that by the end of March, we had a near-normal appearance for a stoma, as shown on the picture on the far right. And we have taken biopsies from the area which show no evidence of tumor. And we see a similar picture if we look at the CT scan, where we just got the picture here from cycle 5 at the end of last year, where the disease was perhaps at its most exuberant and then shown significant improvement with the last scan on the 25th of March being at the same time as the picture I showed you. And obviously, it doesn't return to normality because the stoma is present there. The eagle-eyed amongst you will be able to see the subtle improvement in the peritoneal disease, which doesn't lend itself to objective measurement if you look at the relative thickness in and around the bowel on the left-hand side compared with the picture on the right-hand side at the end of March. And overall, in terms of measurability, we've seen an 80% reduction in disease burden, which is impressive by any standard. We've also treated patients with esophageal cancer. This patient is heavily pretreated with large number of prior therapies, almost all chemotherapy with paraesophageal and lung lesions, and we injected a lesion in the lung which is shown here, but we also saw responses in uninjected lesions as well, so that this patient has an ongoing partial response at 10 months, which speaking as somebody who treats the GI cancer as well as melanoma regularly, is a relatively unusual occurrence in this group of patients who often don't tolerate multiple lines of treatment terribly well and certainly don't respond other than to first-line chemotherapy. And this is backed up by the biology shown here from the biopsies taken again at baseline and again at day 43, where we do see some evidence of CD8 infiltration at baseline, but this is clearly much more involved at the time of the later biopsy and consistent with the fantastic clinical effect that we've seen to date. Next slide.

Thank you, Mark. I think we've exhausted your patient examples.

Yes, I think we'll move to the bladder category.

Yes, yes. No, so I'm going to take over again now. So thanks very much for that very thorough summary of some of the patients treated so far, which I really do think are providing compelling evidence that RP1 is adding significantly to the anti-PD-1, not only in skin cancers, but also more anecdotally, maybe in the other tumor types mentioned as well. So now we'll turn to some other news which we announced today as well, that being our intention to initiate trials in anti-PD-1 refractory non-small cell lung cancer as well as we've decided to terminate enrollment of the cohort we were enrolling in metastatic bladder cancer patients, which is based on the changes in the competitive landscape in particular, with the good progress, which is being made with Seattle Genetics' drug antibody conjugate drug. So initiating development in non-small cell lung cancer was driven not only by the great unmet need, there currently being no standard care or viable option for patients who fail anti-PD-1 therapy, but also based on data with RP1. So with RP1, both a single agent and in combination, we have seen reductions in lung metastases of other tumor types in both injected and uninjected lesions and including an anti-PD-1 refractory disease, a couple of whom you've seen in the prior slides. So RP1 is also found to be able to be safely administered to tumors in the lung where quite a few injections into quite a few patients have been performed so far. And the lung is also known to be immune responsive site. So overall, we think this background is very supportive of testing RP1 in non-small cell lung cancer, including in anti-PD1 refractory disease. We're, therefore, very pleased that BMS have agreed to us, including this new cohort in the study and look forward to beginning enrollment of the protocol amendments made a bit later in the year. So now we'll very briefly move to our pipeline products of RP2 and RP3. So while RP1 was designed with the general intention of treating tumor types which have a level of immune responsiveness, including in anti-PD-1 refractory disease, RP2 and RP3 were designed with the intention of addressing less or nonimmune responsive disease from the outset. And respectively, as well as the genes in RP1, additionally encode an anti-CTLA-4 antibody like molecule in RP2 and for RP3, also CD40 ligand and 4-1BB ligand as well as anti-CTLA-4. So heavily armed oncolytic immuno gene therapies, which will really get, we believe, the immune response going to a patient's cancer. So we believe that these have the potential to increase activity and reduce toxicity as compared to systemic antibody based approaches to addressing the same targets. And with RP2, we're looking forward to presenting safety and efficacy data from the single-agent portion of our ongoing Phase I clinical trial later in the year, together with initial data also in combination with Opdivo. So for RP3, we're also pleased to report that we're on track to initiate clinical development in a Phase I trial of RP3 alone and combined with anti-PD-1 later in the year, for which the first results will be presented in due course which will be expected sometime next year. So with that summary of where we are in the clinical data to date, particularly in melanoma and CSCC, I'll hand back over to Philip for some concluding remarks.

Thanks, Rob. I would like to close out our presentation by summarizing our data once more and the opportunity ahead. Our lead indication of CSCC, where there is a clear path to market, represents a significant commercial opportunity which alone has the potential to drive substantial value. The number of complete responses we've generated is highly suggestive that our combination approach with RP1 can drive better patient outcomes than anti-PD-1 alone. We are also very encouraged by the high rate of response in difficult-to-treat melanomas, refractory, to anti-PD-1 and anti-CTLA-4, where minimal to no activity to a further line of anti-PD-1 would be expected. The clinical responses are supported by compelling [ biomarker data, including reversal of T cell ] exclusion. We are also excited to be moving to test RP1 in non-small cell lung cancer given the large unmet need particularly in the anti-PD-1 refractory setting. Looking ahead, by the end of the year, we expect, COVID-19 notwithstanding, to be able to share further data across our programs to support our potential registration studies in CSCC and melanoma. Further, we plan to expand our CSCC franchise into the transplant setting where anti-PD-1 is contraindicated as well as into the neoadjuvant setting. In anti-PD-1 refractory melanoma, we believe we have a strong signal and are optimistic that our 125-patient study could generate data able to form the basis of a product approval. We look forward to discussing with the FDA their thoughts in this regard. We will also present single-agent data with RP2 from our Phase I study and expect to progress RP3 into the clinic by the end of the year. We believe we have established clinical proof of principle with RP1 in immune-responsive tumor types and now have a solid foundation upon which to build and further to establish our products more broadly as the second cornerstone of immuno-oncology. On the finances front, given the uncertainties on achieving time lines due to COVID-19, we also announced today an expansion of our debt facility with Hercules to $40 million, of which only $10 million has been drawn to date, to ensure we are funded through 2022, and we'll be able to achieve multiple value inflection points with our current financial resources at hand. I would now like to thank you once again for joining our call and hand back to the operator for Q&A.

[Operator Instructions] I show our first question comes from the line of Anupam Rama from JPMorgan.

Congrats on the update. Maybe a question for both the company and the physicians on the line. Maybe you could talk a little bit about the onset of CRs here with the RP1 combo. Rob, in your comments, you talked about the ASCO data here suggesting CR data -- CRs with cemiplimab actually improve with time. So just thinking about your onset of CRs, what that means for depth of response over time with RP1 and any potential implication for duration of response.

So I'll hand over to Kevin in a second. But really, the kinetics that we're observing is quite variable from quite rapid to taking quite a number of months to fully mature. And often, that is because one has to take biopsies to confirm a response and it often takes quite some time before it's decided to be appropriate to take that biopsy. Obviously, to-date, all of our responses are ongoing, and we're seeing good durability with our first patient treated with CSCC, who did achieve a complete response, being an ongoing complete response at 16 months. But I'll also hand over to Kevin to talk about his observations regarding kinetics of complete response in CSCC patients.

Yes. Thanks, Rob. And thanks for the question. I think it's a really good and pertinent question. I think what we've shown you in the summary data, of course, is that the onset of CR is somewhere usually between 3 months and 6 months. But I think you will have appreciated having seen the photographs that I showed you and some of the images that I showed you in patients, many of whom had been heavily pretreated including with multiple rounds of surgery and indeed with radiation at the local site. We really were at great pains to confirm that a complete remission really was a complete remission. The radiology took time to evolve in terms of resolution of inflammatory changes and our being able to be confident that any residual scarring was not active disease. And Rob is absolutely right that in order to confirm that with biopsies, we really wanted the patients to be in as good a shape and as good a response both visually and radiologically as we could. So what I'm sort of saying, I guess, is that these complete remissions may indeed have occurred earlier, but our declaration of a complete remission may have been delayed by the care that we took over ascertaining that that's indeed what we've seen. So I think the headline is somewhere between 3 and 6 months but possibly a bit quicker than that, were it not for the fact that we took such pains to confirm them.

Our next question comes from the line of Jonathan Chang from SVB.

Congrats on the update. First question, maybe just a follow-up on the previous one. Can you talk about the importance and implications of achieving CRs in CSCC?

Well, actually -- thanks, Jonathan. I actually think the best person again is to hand back to Kevin, the expert on the matter to -- for him to comment on the clinical relevance of that.

Yes. Thanks again, and I think another really relevant question. I hope again, you will have seen that for the patients that we recruited, many of whom really were very late-stage and very recurrent diseases, the partial remission that we did achieve and the partial remissions achieved en route to complete remission were meaningful clinically for the patient. They were often associated with significant symptom benefit and also in terms of just the unpleasantness of the wound, the unsightliness, the problems with repetitive dressings of those wounds, major, major implications on quality of life, although we didn't necessarily capture those data by formal quality-of-life assessment. But the question you ask is actually about the importance of complete remission. And for the patient themselves, many of these patients, I think, have more or less given up on any hope that they might have a chance of achieving a durable complete remission that would convert potentially into a cure. But for me as a clinician, of course, the ability to obtain a complete remission is the absolute -- the necessary step on the route to long-term durable remission and potentially even cure. So I think the fact that we're achieving CR means that we can begin to hope that in some of those patients, the disease may not ever return. And I think that's the most exciting thing for me, albeit in small numbers, we're seeing very deep and complete responses of nasty disease. And my hope is that we actually may have some patients where when we look longer term at bigger data sets, we see a true tail on the curve that when it goes out to 2, 3, 4 and 5 years, we can begin to start talking about curing these patients.

Got it. And then just a second clarifying question from me. What -- should we expect another update of the data presented today in the second half of the year?

So yes, Jonathan, we are planning to present further updates from these cohorts towards the end of the year.

Our next question comes from Don Kim (sic) [ Do Kim ] from BMO Capital Markets.

Congrats on the data. My first question is for Dr. Harrington. When you look at this data set from CSCC and all the compelling complete responses, could you talk about your level of confidence that RP1 is the underlying driver of the benefit in comparison to the single-agent studies that this company has talked about?

Yes. You asked the really important question. I guess with the first 1, 2 and 3 responses, of course, it could be that this was just a group of patients that we struck it lucky and we had patients who had disease that was primed to respond to single-agent PD-1 inhibition. I guess some of the biopsies that I showed you that demonstrated relatively low levels of PD-L1 staining might suggest that, that wasn't the case. And indeed, maybe these tumors were not particularly immune-hot; were not particularly likely to do well with single-agent immunotherapy. But of course, as we've gone on and we've treated patients with increasingly unpleasant disease and have seen really dramatic responses, my level of confidence and my level of belief in terms of the fact that the virus is adding something to the package and indeed may be the dominant component of this package and is essentially allowing or licensing the PD-1 to have a proper effect against the disease, has become increasingly great. And of course, I have to issue a disclaimer in that I have been working in the field of oncolytic viral therapy for the best part of 20 years. And so I am a believer in this technology, and that's why I work on it. But I think [ the data here ] really are beginning to become more and more compelling, and I believe that the RP1 really is bringing something to this combination over and above what you might expect from a PD-1 agent alone.

Great. And for the company, since you've seen this much higher CR rate for CSCC with the combo plus anti-PD-1, could you possibly see that kind of benefit in frontline melanoma where CR rates are still fairly low? And thoughts on moving to the naive melanoma setting after the refractory study.

I mean certainly, the concept of frontline melanoma is not something we wouldn't consider. We do not have any specific plans in frontline melanoma at the moment because we do think the nearer-term opportunity is in anti-PD-1 refractory disease for which there's a clearer path to registration due to the great unmet need and therefore greater -- less complexity in development and where smaller and single-arm studies could be appropriate. However, we certainly wouldn't in any way discount in the future moving into the PD-1-naive population as well as the data further evolves.

I show our next question comes from online.

Rob, we have a question from the webcast. Does the deepening of responses seen with Libtayo over time confirm the relevance of your mechanism of action? Or is the read-through not as clear as that?

So we certainly anticipate that responses in any tumor types we are treating will deepen over time with variable kinetics, depending on both the patients and the tumor type. And we're certainly seeing that to be the case in CSCC, where obviously, patients start off as PRs and then take time to evolve into CRs. So we're hopeful that other patients we have who are PRs will evolve into CRs. And we're seeing the same with -- in melanoma as well. So I do think that the Libtayo data demonstrating deepening of response slowly over time indicates that our initial looks at data should also further mature into further improvements, these being, in most cases, relatively immature patients at the earlier stages of their therapy-straight response and certainly having had the 2-year follow-up that now is the average in the Libtayo data.

[Operator Instructions] I show our next question comes from the phone line from Robert Driscoll from Wedbush.

Just so we know that T-VEC doesn't work as well in stage IV melanoma patients compared to earlier-stage patients, here, you've mostly enrolled stage IV. Do you expect the makeup of patients in the registration-directed study will be similar?

In our registration-directed study, the 125-patient cohort, there's no reason to particularly expect that it will be different. However, obviously, until more of it is enrolled, we can't further comment. It is the case obviously that we were enrolling the melanoma -- the Phase I part of the trial and the Phase II cohort in a sort of Phase-I-y-type setting where -- at which point investigators didn't have huge experience with RP1. And I do think in that setting, there's often a tendency to enroll the more advanced patients who really haven't got any other options, whereas once there becomes a bit more experience and a bit more data, the tendency may be to enroll patients who are a little less advanced. But other than that, there's no particular reason to expect that we won't have a similar number of M1b and M1c, for example, patients as we're currently enrolling.

And then -- and with just one follow-up maybe. And I know it wasn't the point of today's update, but have you been able to look at expression levels of RP1's expressed proteins such as GALV or GM-CSF?

We haven't done that in patient samples although we do have data from animals.

Our next question comes from Peter Lawson from Barclays.

I've got 2 -- I guess questions for Dr. Harrington and Dr. Middleton, just on what other trials are you putting patients on in CSCC just to give us an idea of where RP1 sort of ranks in your opinion amongst the other CSCC trials?

So probably that's a question for Kevin. Bear in mind he's our CSCC man on the phone.

Yes. I was just waiting to see whether or not Mark would have first dibs at that question, Rob. So in terms of CSCC, we have -- in the United Kingdom, we have approval to use cemiplimab and Libtayo through the National Health Service. And so that tends to be the treatment that is offered to patients. But for patients in clinical trials, the RP1 and indeed the RP2 study options are what we were -- we go to first. We have activity with STING agonists with and without anti-PD-1 in addition. And other than that, our only options for patients are cytotoxic chemotherapy, and I'm pleased to say that we virtually never use that because most of these patients tend to be rather frail and don't withstand that treatment very well. You will notice that the patients that we treated and enrolled in this study, a number of those had, had previous chemotherapy. So they had been heavily pretreated. But we have relatively limited options in cutaneous squamous cell carcinoma at least in the United Kingdom.

I don't really have a special interest in CSCC here in Oxford. So this is the only trial that recruits that particular patient population.

Got you. And then how easy is it to use RP1? Is there any reason why you couldn't see this moving into more of a community setting, kind of outside academic sites?

I could answer that.

You've used more of this than anyone else, Mark.

Yes. It's relatively straightforward to use. There are some logistic complexities and some regulatory complexities around using a GMO that you have to get your head around, but they're relatively straightforward to overcome. And certainly, in terms of injecting cutaneous lesions, it's pretty straightforward. It does require interventional radiology to access deeper lesions. And certainly, as far as the U.K. is concerned, the infrastructure isn't there to do that at scale at the moment. But clearly, if the preliminary evidence of activity is borne out, then the cost effectiveness of that approach will ultimately win through.

So I believe it's a case, Mark, that you've treated a few tens of patients by imaging-guided injection to liver, particularly in lung tumors?

Yes. And certainly, it's -- as long as you can see it on ultrasound, it's really very simple. It's a 30-minute outpatient procedure, which compares very well with the sort of chair time that you have to use to give intravenous checkpoint inhibitors. For lung or for lesions that require CT image guidance, it is a little bit more complex, but still, we only need about an hour of machine time in order to do that. And the complication rate is very low, much lower than for image-guided biopsies, for example, because we're using very much smaller needle sizes. So we've had the occasional patient who get a pneumothorax, and usually, that will settle spontaneously. I think we've had one this year who's required a drain. But even then, 2 weeks later, we're able to come back and reinject without ill effect. So as with all things in medicine, it's about experience and ascending that learning curve. But there is nothing that hasn't been done before in an oncological or interventional radiology setting that means that this is a particularly special challenge compared with some of the other things that we're thinking about introducing, like cell-based therapies and so forth.

And if I could just add a couple of comments in relation to the cutaneous lesions. And I'd echo everything that Mark has said about the additional regulatory and logistical issues. But in terms of injecting these cutaneous deposits, essentially, patients are in and out, treated in a normal clinic slot in a normal outpatient facility. I mean they're in and out within less than 30 minutes. Once they're established on treatment, it's extremely straightforward once people are comfortable with doing this. And so it's actually something that I could see -- certainly with the levels of response that we're seeing, if those are continued and in larger numbers of patients, it is something that could be done as a routine without a great deal of trouble, I believe.

And just finally, on the safety, you had like a 10% discontinuation rate approximately. What was the reason there? And does that cause any concern?

Mark, do you want to comment on that?

Sorry, I didn't -- I couldn't hear that question very well. So could you repeat it, please?

Yes. Just on the safety, you had a 10% discontinuation rate. What was the main reason? And is that a concern in any sense? Or is that a typical discontinuation rates you see in clinical trials?

No. I think that's a relatively low discontinuation rate for a PD-1-based therapy. So from that perspective, no, absolutely not a cause for concern. Although PD-1 -- for example, nivolumab treatment enjoys a good reputation. To a certain extent, that's because it's much less toxic than ipi/nivo, where about 45% of the patients will withdraw due to toxicity within the first 3 months of treatment. But it's far from 0 for single-agent PD-1, and 10% is about the going rate. It's also a function of time. So the longer that patients stay on, the more likely it is that they will develop a toxicity that requires the treatment to stop. So from that point of view, I'm very comfortable with the tolerability of the combination.

Our next question comes from Tony Betley (sic) [ Tony Butler ] from ROTH Capital.

Tony Butler. There is some extraordinarily strong abscopal responses observed in CSCC. I think there was one that Dr. Harrington showed there as well as Dr. Middleton in cutaneous and mucosal immunity, respectively. These were, I believe, PRs. The question is, is there any evidence that you have to date that abscopal effects, that is distant -- distal lesions, have actually demonstrated CRs? Because it strikes me -- and I'd be interested in your thoughts on this, that this, by definition, would unequivocally demonstrate that it's in -- it's the RP1 that's really doing the work. And I think perhaps Dr. Harrington alluded to the notion that the allowance and licensing of the PD-1 to have an effect, which I thought was quite interesting. So I'm curious, most importantly, on those abscopal effects.

So this is Rob. Thanks very much, and I'll further hand over to Kevin again, too. But those CRs, which we're showing, are CRs obviously of all disease, including in patients with visceral, distant, uninjected disease which has completely gone away. So in the patient examples you saw, there was a retroperitoneal lymph node which completely went away, which is obviously nowhere near the neck where it was injected; lung metastasis and bone metastasis, I think most impressively, completely resolving following injections into only the neck. So I do agree with you that the best data which distinguishes the abscopal effect we're having from that, which might be achieved with anti-PD-1 alone, is the complete response abscopal effect, which we certainly have seen in a number of patients, including at visceral sites and indeed in bone lesions, too. Kevin, I don't know if you have any further comments.

We -- well, Rob, I'm obliged to object to the use of abscopal when there's a systemic therapy on board, but that's a long-standing -- almost a joke we have between us. I think when there's a systemic therapy on board, of course, the lesion that is at a non-injected distant site might be responding to that systemic therapy. So I think you have to go back and you do have to look at what the historical data show, for instance -- mostly, for instance, with the cemiplimab data, demonstrating that complete remissions at distant sites will indeed -- in the cemiplimab data, there is no locally injected lesion -- are relatively rare. So I think the data that I have shown you, and I would make the point and reiterate what Rob has said, we truly have seen biopsy-confirmed complete remissions at sites that were not injected with RP1. Now of course, they may be responding completely to the systemic therapy, the nivolumab that was being delivered at the same time, but that's not what the experience with single-agent nivolumab tells us we should expect. So we are seeing complete remissions. We're also seeing some partial remissions, but I am convinced that we are seeing deeper and more impressive responses even at the uninjected sites, which I do believe are being conditioned, licensed and being promoted by the local effects that we have at the injected site. So I think there is evidence for both of those effects, both the local and the distant systemic non-injected response.

If I may -- this is Tony again. Just what I neglected to perhaps state is especially in patients who have been found to be refractory to prior anti-PD-1 therapy because then, again, I would argue in the -- maybe the word abscopal could be overused here because you would have a subsequent systemic agent on board. But let's be fair, it didn't work the first time, as you alluded to before. The only thing that's different is really the RP1 at least in those refractory patients.

Yes. So there are a number of examples of tumors very significantly reducing or having apparently gone away in patients who failed, for example, ipi/nivo in melanoma. So for example, the patient Mark described from Liverpool who had the disease in the -- particularly in the thigh but also in the groin and the lungs, actually, the suspicion of the investigators is that, that patient is in complete response, including in the lungs, but they're not going to go in and take biopsies for various reasons to actually prove that. The belief is there's actually no active disease. So I think we really are seeing exactly what you're alluding to.

Our next question comes from Jim Birchenough from Wells Fargo.

So a few questions from me. Just first, for the 16 melanoma patients, are they all primary progressive on PD-1, i.e., did they flow right through PD-1 therapy? Or did some previously respond and then subsequently relapse? Just wanted to understand that first.

It is a -- we haven't fully documented that information in this data set because that wasn't part of the inclusion criteria to do so, and that is being fully documented obviously in the 125-patient cohort enrolling because it is part of the inclusion criteria. So it is a mixture of patients who have -- who responded to some extent and didn't respond at all. However, they have all fully failed anti-PD-1 rather than just having had some sort of marginal response -- sorry, marginal progression before coming on to our trial. So it's not as important a group in that regard.

So of the 5 responders, could you maybe break out which were primary progressing on PD-1 and which had previously responded and then relapsed?

I can't, off the top of my head, provide you exactly accurate data. But from best recollection, they -- all but 3 of them had not responded. One of them had, had a minor response -- or a mixed response, sorry, with clear progression in the [ liver ], I believe, before coming on to our trial. And then it was a fifth patient who failed actually adjuvant PD-1 therapy before coming on to our trial, which is also allowed in our 125-patient cohort.

So only 1 of -- out of 5 of those patients had evidence of prior activity to anti-PD-1.

Got it. And maybe for Dr. Harrington, just as you think of the 125-patient study, is there any sense of what proportion of patients would be accessible with the cutaneous lesion or where you might have to inject the liver? And then the second part is just, do you see any role for potential combination with a TIL therapeutic? If you're in injecting a lesion, could you take a biopsy to support TIL recruitment? And does that make any sense at all?

Before I hand to Kevin and/or Mark, and Mark's probably recruited more melanoma patients on this trial. The melanoma patients treated to-date are a good mixture of patients who've had cutaneous or subcutaneous disease injected and visceral disease, including lung and liver metastases, injected. So we're not in any way seeking patients with cutaneous disease. We're very happy to enroll patients who don't have cutaneous disease but only have, for example, lumps in the liver, which could be injected, which, as Mark said, is actually relatively straightforward because ultrasound would be the preferred approach there, which is particularly easy for largish liver tumors. Kevin and/or Mark?

I think Mark is probably better able to comment on this, but I would just make one brief observation about the idea about harvesting for TIL therapy. I think it's a really interesting idea and, of course, begs the question, when do you access those TILs? Do you do it before you start treating or more likely when you're already underway with treatment and you've excited an inflammatory response within the lesion? But I'd be interested to hear what Mark had to say as the melanoma expert.

So I think it's a really interesting idea. And my bias not supported by data because obviously, our biopsies are taken at prespecified time points within the protocol would suggest that somewhere around 4 to 6 weeks in tends to be where we see the flare, which is profoundly inflammatory in those patients who do flare and suggests that, that may be a good time to think about harvesting. I would see that as an entirely separate activity though from injecting the tumor because the sorts of approach that we take to injecting doesn't give us a needle track or point of access that would get nearly enough tissue in order to get an adequate harvest for developing a bespoke therapeutic in a TIL system after that. And of course, those patients in whom we've seen the most exuberant flare have tended to go on and do very well, which likely begs a question of whether you would end up getting the best yields from those patients who least need the boost in therapy that TIL might offer. But I think it's certainly something to consider and, in larger cohorts and as we learn more, understand whether that's a way of driving patients into complete responses that are the hallmark of really long-term benefit in melanoma as much as in CSCC.

Just maybe a final question from me. And that is in the biopsy data, have you looked at particulars of response? And are you looking at things like high mutational burden versus low mutational burden, whether that makes a difference? And then I guess associated with that, are you looking at T cell repertoire and how broad a repertoire and if there's a certain pattern that's supporting a response?

So to date, in the clinical trials, we haven't been doing that other than we do have -- generating some Nanostring data from the biopsies as well. And obviously, to date, the patient numbers have been relatively modest and also relatively heterogenous groups of patients. And therefore, it does make trying to interpret that sort of data rather challenging. However, in the cemiplimab CERPASS, randomized controlled Phase II trial, we are collecting biopsies and other samples to enable that to be done a lot more thoroughly, including tumor mutation burden and any other analyses we choose to perform on the samples when we have them.

[Operator Instructions] I show our next question comes from the line of Mark Breidenbach from Oppenheimer.

Maybe this one is directed at Rob. Since I don't think we saw swimmer plots in this presentation, I was hoping you could give us a sense for median follow-up time both from the CSCC and melanoma trials and the ranges of response durability you've seen so far. I'm just wondering if you've hit median duration of response for any of these cohorts.

So there is a swimmers plot in the deck for CSCC. And bearing in mind that all responses are still ongoing, we have not obviously hit median duration of response. All we can say is response is up to 16 months from initiating therapy with the minimum at the moment, I believe, of 2 or 3 months. But obviously, we can't hit a median until we've had some number no longer responding. At the moment, the melanoma data is a bit immature. The last patients in the melanoma cohort were only enrolled in January and therefore -- and the shape of recruitment was enrolling more towards the end. And therefore, they've often had a rather short follow-up, but everybody -- we're seeing good durability to date, but I can't comment more than that.

I show no further questions in the queue. This concludes our Q&A session. Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.