New Concepts in Immunotherapy for Solid Tumors

[Music] this is an educational symposium called new concepts in immunotherapy for solid tumors novel immune approaches biomarkers and combination therapies my name is nya rizvi I'm from Columbia University and joining me for Eddie Guerin from UCLA and Jeff Gibney from Lombardi Comprehensive Cancer Center and David rim from Yale the following audio is based on a recent live webcast with dr. Niyazi dr. Jeffery T Gibney dr. Edward B Garen and dr. David L rim this audio is part of a certified educational activity titled new concepts in immunotherapy for solid tumors novel immune approaches biomarkers and immune based combinations to access the entire activity and complete the post-test please go online to WWE us comm /nw em a printable monograph slides practice aids and other features are also available so I'm going to just speak for the first few minutes around background information as you as you know the ctla-4 checkpoint inhibitors were the first to hit the clinic with a plume lab the two main ctla4 antibodies are hippie and tremie the pd-1 antibodies and PDL one antibodies are shown the pd-1 antibodies in general are the IgG for or IG one isotypes in general the PDL one antibodies that are RG 1 do require engineering to remove an effector function so so there's no ADC CCDC that can lead to destruction of the t-cells that you're trying to activate contrary to that that premises of elam AB which does have effect or function it is not engineered it's a full human eye g1 and interestingly when we look at the activity with a dilemma that seems comparable range as as the other pd-1 PDL 180 buddy so maybe that premise really isn't true this clearly is a partial list there are many companies developing pd-1 PDL money and ctla-4 antibodies so this is why we're all here it bolded are areas where the FDA has approved immune checkpoint inhibitors air appease with either an TPD one new volume at Pemberley Zuma or PDL one antibodies for lung cancer which I mostly do we now have three antibodies that are now approved for lung cancer including development of Pemba lism evan @s ilysm add the the list just has to be updated almost on a weekly basis just within the last couple of weeks I'm the velum AB was approved for Merkel cell carcinoma based on its a high activity within that tumor type there's activity really across multiple tumor types that you can see on the right on the left we still have some challenges around prostate cancer microsatellite stable colon cancer however there's some encouraging activity combining PDL one antibodies with mechanization which chef will will touch upon but I think that some of the tumor types such as pancreas cancer which is very dense tumor not a lot of immune cells I think continues to be a tough nut to crack but I have to think that will be moving those tumor types on the left over to the right as we understand the biology better and I think that the the activity that we see in melanoma is really a foot in the door for every tumor type we're really going to have a lot to cover because we're not focusing on one tumor type we're really focusing on multiple tumor types where we have clinical information and looking at the data that led to many of the approvals of checkpoint inhibitors around melanoma lung cancer head neck cancer as well as others a lot of experience now combining CTA forward and pd1 but I think that that again is and others just really step in the in the door foot in the door to give opportunity for new combinations and new targets and we'll discuss that as well one of the more befuddling aspects of immunotherapy is tumors and biomarkers you know PDL one is used differently with lung cancer versus bladder cancer versus melanoma and I think it makes it a little bit challenging and and then finally I'll provide some sort of clinical perspective in terms of some of the issues that that arise in the clinic and how we think about using these these antibodies and so editted speak on foundational evidence of checkpoint blockade in cancer care exciting thank you very much higher and in many respect I have the easiest of the talks in that I'm not having to surmise a lot I'm actually going through the data that led to the approval of all of these agents and so of course the idea of cancer immunotherapy is not a particularly new one but to be using it to the extent we are and our clinics clearly is quite new so in the 80s cytokines became used certainly early in my career in oncology my big experience with immunotherapy was il-2 for renal cell carcinoma which was something that at our institution we did quite frequently but what is generating of course the enthusiasm is really the checkpoint inhibitors which have been really over the last five or so years something that has come to the forefront of oncology care so for this audience I certainly don't need to belabor this slide as you can see a t-cell response is going to occur with the t-cell receptor and the peptide on an MHC molecule then you will have either co-stimulatory or Co inhibitory molecules and on the left you'll see of course that response scene when both of these occur when only one occurs a response is not seen what we are looking at and as an example for the PDL one antibodies or pd-1 antibodies are instead of co-stimulatory molecules really inhibitory molecules and the idea being to use antibodies to prevent that second signal so the foundation of these from a clinical perspective really is in the field of melanoma so as you can see if the lumen map led to a median overall survival that approached a year in patients with advanced melanoma as you can see here though there is the big dot at 3 years which shows that at that time the overall survival rate was 22% that in many respects was more exciting than what was happening at the median and what I think today we could say is even more exciting than that is even the longer-term data that shows that patients really are continuing to do well that this is a true durable response so melanoma even with the success of the flu map has of course seen many changes this is a study in which the pd-1 antibody Pember lism AB was compared to hip aloha map here you see the data for overall survival there was evaluation of 2 different schedules for Pember lism ab every two weeks in every 3 weeks and they're so similar than at many points you can't even tell there are 2 lines there however they clearly do perform superior compared to the voluma map and when one looks at progression-free survival the results are you know quite similar one thing that of course has been an area of interest in NOLA Noma is if you have two effective immune checkpoint inhibitors can they be effectively combined in order to further improve response and further improve outcomes and here you see data from checkmate oh six seven this is looking at the PD 1 inhibitor Nivola map and as you can see in this instance the progression-free survival of if a luma map is shown at the bottom the gray bar shows the person through survival for Nivola map and the combination is shown on the top now I think different people could look at this differently I treat lung cancer rather than melanoma and on one hand clearly the line does appear to be higher I think whether or not a combination therapy and associated toxicity makes it important to give these concurrently versus sequential I would leave to my colleagues in melanoma so of course you're all aware that an very active area of research in I mean checkpoint inhibitor therapy has been lung cancer most specifically non-small-cell lung cancer as niya pointed out there are now three different approved agents for this Nivola map temporal ISM AB and a tes ilysm AB and I will go through some of the data that led to those approvals so the initial data was with new Bolam AB this was was published back in June of 2012 by Susannah Polian and colleagues and as you can see the response rate overall was about 17% now for many of you this may not seem so good but you have to remember this was previously treated non-small-cell lung cancer the comparator would have been docetaxel at least that's what people would use and that is approved based off of a 7 percent response rate the encouraging thing in many respects is what you see over on the right Scott get injure updated this data in the Journal of Clinical Oncology just a few years ago at that point it was a three-year follow-up and you can see that over a quarter of the patients at least on what is the approved dose of Nivola mab continued to be alive at 3 years very similar to the response rate in fact so Nivola mab received then full approval based on two randomized studies they looked at both squamous cell carcinoma in the checkmate oh one seven study and non squamous non-small cell lung cancer in the checkmate oh five seven study and you can see here in the overall survival figures the blue Nivola map clearly a performed docetaxel one thing however that was interesting when one looks at the Nivola mat versus docetaxel particularly for non squamous non-small cell lung cancer which is the majority of what non-small cell lung cancer doctors see is that there was actually the somewhat unusual progression-free survival curves and what you can see is that the curves are in fact really crossing over and for those of us who do lung cancer this was very reminiscent of what was seen in data from tony mok in the iPass study in which patients who had demographic characteristics that made them likely to respond to an EGFR inhibitor were randomized to chemotherapy or an EGFR inhibitor and it least in that study the crossing over the lines indicated two separate populations in that case EGFR mutant patients versus EGFR wild-type patients the question was would there be a similar sort of grouping here and at least the first suggestions were in study that Nayar and I were involved with the keynote 0:01 study what you see here is actually the overall survival data and the gray line is for patients who receive Pember lism a band head staining for PDL one in at least half of their cells the keynote oh one no study was a confirmatory study randomized phase three study comparing member lizmac to docetaxel two different doses of Pember lism ab were evaluated and compared to docetaxel and as you can see in all the patients who needed to be positive for PDL one there was an advantage in terms of overall survival and that advantage was particularly market amongst patients who had high level staining staining at least half of their cells so is this really two separate groups i think that in some respects the checkmate oh five seven study which again was in the Bolam AB versus docetaxel non squamous non-small cell lung cancer indicates perhaps there our this is from a presentation by Louise Posner eze and what you can see here is that on top at the ten percent cut point for PDL one expression that there was a clear advantage for nivolumab over those attacks was a hazard ratio of 0.4 in the group of patients who had less than 10% standing there really was no difference that was seen and the hazard ratio there was 1.0 there of course is one additional checkpoint inhibitor that is approved in non small cell lung cancer obviously the combinations will be topics of later conversations here but at 10 ilysm AB is approved as a PDL 1 inhibitor and it is based off of data from poplar which was a phase 2 study comparing a teza Luther map to docetaxel and a very similar in design phase 3 study that oak study that compared a tensile ISM AB to docetaxel and showed a superiority of a tesla Lizza map one thing I will mention which will come up again when we go through some of the GU data they have looked at PDL 1 a little bit differently they have looked not only at what happens in the tumor cell but also at what happens in infiltrating immune cells so in many respects for I think the initial wave of pd-1 inhibitors this ended up being sort of the most substantial sort of thing to come out of it one question was if you were able to identify a group of patients who drive particular benefit from a PD one inhibitor could you then actually show that in that group it's better than frontline chemotherapy in this it differs non-small cell lung cancer differs somewhat from melanoma melanoma does not tend to have particularly effective chemotherapy regimens non-small-cell lung cancer are chemotherapy regimens we don't think they're spectacular necessarily but they do work they do induce responses they do extend progression-free survival and overall survival and in this case for the keynote o24 study Pember lism AB in patients with standing for PDL one in the least after their cells was compared to standard frontline chemotherapy this showed an improvement in progression-free survival which was the primary endpoint but also even though crossover was allowed as part of this study meaning that in patients who were randomized to chemotherapy when things got worse they were allowed to switch over to Pemberley smeb still there was a survival advantage for the Pemberley lab arm Nivola mab their frontline study did not necessarily restrict patients to PDL one expression at this this 50% cut point they looked at a much lower cut point of 5% and at least numerically the progression-free survival in that study was in favor of chemotherapy so of course the original checkpoint inhibitor approvals were in melanoma non-small-cell lung cancer was in many respects next but that is not the end of the story as you can see here this is something where you are looking at mutational load across different tumor types you can see that the melanoma is the most of heavily mutated followed by lung squamous carcinoma and lung adenocarcinoma and in many respects this is a figure not only of tumour mutational load but the relative benefit with PD 1 or PDL 1 inhibitors now I will say of course nothing in science is quite that easy some of you will note renal cell carcinoma which I will mention which is one of the approved indications for checkpoint inhibitors is not all the way to the far side with melanoma and lung cancer so one of these fairly heavily mutated cancers is head and neck cancer so head and neck cancer is oftentimes strikes a similar patient population to non-small-cell lung cancer and in this study member Liz Ahmad was given every three weeks to all patients and we can see there was a mid-teens response rate for Pember lism AB one thing that is important in head neck cancer there is a group of patients who has HPV as a cloud a factor of disease where other patients have more typical mutagens things like cigarette smoking and alcohol and as you can see there didn't appear to be a huge difference this study was small so differences based on PDL one are a little bit harder to interpret the responses tended to be a good duration similar to what has been seen in other malignancies and as you can see in this waterfall plot a majority of patients at least had some degree of tumor shrinkage when receiving Pember lism ab and this led to FDA approval in head and neck cancer in 2016 for recurrent disease we now have randomized data this data looks at Nivola mab as compared to investigators choice of chemotherapy and as you can see here again a survival advantage for the PD one inhibitor in this case in a bola MAV as compared to investigators choice and patients were more than twice as likely to be alive of the year if they were receiving the bola map as compared to chemotherapy this led as well to FDA approval in post progression head neck cancer so another active area of study has been in Gi and GU cancers and this slide focuses particularly on GU renal cell carcinoma is is a place where new Volvo MAV is approved bladder cancer is another area that has led to approval of this class of drugs here you see data looking at a tes ilysm AB in previously treated metastatic your athelia l-- cancer and here you can see the overall survival as i had mentioned earlier they have looked at staining for PDL one and immune infiltrating cells and felt that the survival was better amongst patients who had high staining in the immune infiltrating cells so this slide is a single arm experience looking at a tensile ISM AB in previously yura Theal cancer patients patients who are ineligible for cisplatin you will note that the x-axis of one is listed in years whereas the other is listed in month but you'll see that the response rate at least was impressive and this was something that I think was considered to be certainly encouraging in April 2017 the FDA approved it as ilysm AB for the management of patients with previously untreated locally advanced or metastatic euro filiol carcinoma who are not eligible for cisplatin containing chemotherapy the volume AB also has been evaluated and previously treated bladder cancer in this case looking at PDL one did appear to be associated with some difference in outcome and this also is now approved by the US FDA Pember lism AB to have a phase three study where it was compared to chemotherapy again an improvement in overall survival is compared to chemotherapy with patients being also more likely to be alive at a year if they were on the Pemberly zoom AB arm again an exploratory group looking at Pemberley zoom AB in frontline therapy of advanced bladder cancer showing a response rate of 24% one thing that I will also point out is that although the efficacy comparison between gemcitabine cisplatin for bladder cancer and attest ilysm AB is that the toxicity profile of those two regimens is certainly something that is quite different and does need to be taken into account when thinking about this data one of the early areas of evaluation for immune checkpoint inhibitors have been GI malignancies and in those initial evaluations colorectal cancer really was not a particularly effectively treated disease with immune checkpoint inhibitors however there's a small percentage of patients with colorectal cancer who do have this microsatellite instability hi phenotype based on deficiency of DNA mismatch repair genes this actually was published in the New England Journal of Medicine there of course a couple ways one can get such a high profile publication one is having a very large study and the other by having a very large effect and this one was clearly the latter you can see that the number of patients here is quite small but that in this is microsatellite instability high group of patients you can see that there was over 70 percent response rate that being said out of seven patients but still very impressive numbers as compared to a zero percent response rate in those who did not have this mismatch repair deficient tumor and that is on top of really a host of data that has looked at this as well so the zero percent seems quite real the 71 percent could be a little bit high but still is certainly exciting you can see here progression-free survival curves and as would be anticipated the mutational burden was more than twenty fold higher in the MSI high group as as compared to the others and progression-free survival as you can see here is quite different the vola map also has been evaluated in this population does lead to durable responses this is a larger group so maybe not quite as high numbers but still very impressive results one thing to point out is that colorectal cancer is not the only place that this phenotype is seen that there are a whole host of malignancies where this is a potential area of future investigation here I won't believe or this slide just to say that there is a tremendous amount of research that continues to go on in the gastrointestinal malignancies looking at how to bring checkpoint inhibition to that group of patients many of those looking at combination based approaches I think this slide is interesting just because we actually now have a PD one inhibitor approved in Merkel cell carcinoma this is not a particularly common tumor and is something that for those of you who are who are in biotech or Pharma I think many groups are looking for any sort of tumor they could possibly find to show a signal and potentially have a approved and in some ways having an approval for Merkel cell carcinoma is indicative that that even small tumor types can lead to approval for these class of agents here you can see a thirty one point eight percent response rate and again this did lead to the approval of the drug Pemberley the map also has data looking at Merkel cell carcinoma that also is quite favorable so where does that leave us overall we have many questions that remain how does one evaluate how these patients are doing when does one stop in a patient who is doing well when does one stop in a the therapy no patient is not doing well and I think these are things that people have struggled with this is a slide that looks at basically an unconventional response now these are fairly uncommon in melanoma I would say that they're just sort of a step above anecdotal in non-small-cell lung cancer and as far as other diseases I don't know how common they are but you can see here this is a patient with two lesions you can see that the one is is shrinking sort of early on in therapy but the other this lesion the area of the pancreas actually initially grows and subsequently improves for those clinicians in the room certainly in what I do which is non small cell lung cancer I would say that that most of the cases that look very unusual have already been presented or published and this is not in fact and something that is sweeping our clinical care but it is something to know if one does have an atypical response where some areas are getting better and some areas are getting worse particularly in a patient who's feeling good in terms of adverse events in many respects this class of drugs when given as a single agent that being PD one or PDL one inhibitors are pretty well tolerated however our colleague Julie Braemar often gives a talk in which she explains that if it has an itis at the end of it it can probably be induced by this class of drugs and in many respects that true you can see different immune related adverse events really in essentially any organ and this is something that is certainly a reason for caution and is a reason that we want people to be aware of the toxicities this is what I will leave people with the one thing that we want in terms of management is although the management issues have not necessarily always been evaluated in placebo controlled studies it has been fairly consistent from one study to the next that being that in people who have grade one toxicities generally you're able to continue the the drug give symptomatic therapy for that problem great - in general we have tried to us hold the drug steroids can be given often in outpatient setting and in general if there is improvement to those symptoms down to grade one and for those of you who don't conduct clinical trials I know that it is very difficult and sometimes the NCI does have a listing of all of these toxicities usually they are about what one would expect based on the grade that they're described as but sometimes it can be a little confusing but in general once the toxicity has largely resolved we have been willing to restart in terms of grade three and four toxicities and we have discontinued the immunotherapy we have in general tried to hospitalized the patients for IV therapy with steroids and then have not gone back to the therapy the one thing that I will say about that and I'm sure for all of us in the in the group treating patients are starting to run into this problem it is very easy when we're up at the podium to say that you would never start this patient back but it's not so easy particularly in someone who has done well if things start to grow again and they're in the situation where they're going to be dying of their state for cancer to be as firm as we are in a lecture and I think that that's something that is we're all starting to struggle with as some of these patients are now having sort of their disease takeover okay Jeff's going to tell us about some combinations and where the future is for immunotherapy as a starting place I thought this was a good slide we're going to try to cover some of the new combination strategies with checkpoint therapy and other immune modulator agents you're impossible to really cover them all but looking at this from a china doll as sort of a roadmap this is obviously adapted for this purpose we can really try to harness the immune system at different points where there's immune recognition and cell killing we are very familiar with the killing of the cells at the pd-1 PDL one to a micro environment and we're also pretty familiar with priming and activation of t-cells this typically is in drain lymph node basins such as anti sea Tillie four but there's obviously a lot of steps on the way before you have an infect a defective T cell response or immune response against a tumor and these all represent areas that we can enhance or augment to improve the responses or at least create a response in a patient with immune therapy so the example that I think most of us in the room are very familiar with is anti PD 1 plus anti ctla4 therapy actually I guess you look at as ctla-4 then PD 1 because we're trying to prime maybe in the periphery and then enhance the response once it's an in tumor microenvironment at Ballou ma'am it's an anti ctla4 monoclonal antibody that is interfering with a the inhibitory signal or inactivation of t-cells allowing for better stimulation there's also some other theories about it depleting t regs and other ways that may work but it's enhancing a t-cell response that allows cytokine release proliferation migration into the tumor and then once we're in the tumor micro environment we have immune recognition of the tumor cell we hopefully have successful engagement of it interfering gamma signal may abrogate some of this and you can have up regulation of checkpoints like PD one and this is where the second inhibitor comes in with PD one blockade in the bola map leading to hopefully a more effective strategy in treating patients and if we look across studies for involvement plus if we have seen increased activity are we seeing the synergism we hope we are but we're at least seeing out of effects especially in melanoma we know that with pd-1 drugs we see responses around 40 to 45 percent in at truven naive advanced melanoma patient with an anti PPD one drug it bloom AB some around 10% up to 15% depending which study you want to quote and that gets us close to 60% in the phase 3 trial checkmater 67 that we just saw previously and we see a pretty good response rate and we also see some very deep responses that we think are very meaningful again we're also seeing improvement we think in response rates in kidney cancer non-small-cell lung cancer there are a number of randomized trials moving forward really to prove this disease were in single arm studies largely or not with a true comparator arm but we're hopefully moving the ball forward here with this combination strategy at least in a number of different disease types what we also have move forward is the toxicity rates and most of us are familiar with involvement plus it blue map I'm not only increasing objective response rate but an equal increase in the grade 3 for toxicity rate - we're about half the patients experience something pretty significant and this tends to be largely in the GI tract the colitis in hepatic or hepatitis discontinuation rate by 40 percent and this is real we do see it in the clinic but patients get through it the side effects we think are typically manageable or at least in our clinics we can manage them very successfully even though they are more frequent now modification of the dose has been a theme that we've seen when we've moved from melanoma to lung cancer which is the o12 study the schedule was changed so rather than giving full dose it be lower dose of nivo as an induction phase four times it's now been spread out where you give a low dose of it be it be one you give it every six or 12 weeks and you just continue with in a bowl map and we see that the grade three for toxicity rate was lower more patients where they'll stay on drug herd at a lower rate of discontinuation due to AE and moving forward we may see that the regimen or the dosing schedules shift and what I don't show here in the 12 well actually I guess it's over here is that the response rates looked pretty good when you had to modify dosing schedule for the abloom em so that we're pretty familiar with but what about the newer targets these are not all checkpoints idea I guess you could call somewhere in between idea one is present in a number of cells the tumor cells dendritic cells macrophages B lymphocytes and as well as other cell populations and it's an enzyme it's catabolized and tryptophan to kinda riemann and that doesn't seem to be too important but until you realize that whether you have tryptophan depletion or kinder and that's accumulating that actually can dampen immune response and depending on which cell you're looking at this can have different effects but this ultimately can lead to tolerance and tumor cell evasion so it's a very exciting target for immunotherapy and trying to elicit anti-tumor responses we've now seen IDO inhibitors and this is a peptide try and complete cells that have idea of expression move forward from phase winding even to Phase three with a pack outer stack these generally have been relatively well tolerated no major toxicity concerns up front but unfortunately there was not a lot of activity stable disease was generally seen as the best response I in the mono therapy setting the I think the one exception if you were to look up this study with audio five peptide this isn't a lot of mainly lung cancer patients there was potentially some additional benefit but this is now move forward in combination so combining idea inhibitors with not just immune checkpoint therapy vaccine chemotherapy potentially others and there's a lot of preclinical data that's emerging word suggests this is a very effective strategy and I'm going to move on to one study that came out a doctor sprang and I asked these lab over a University of Chicago looking at combination checkpoint therapy both anti ctla-4 and anti PDL one agent in combination with ideal one Bishan and this isn't a b16 Mouse melanoma model and as you can see here this yellow curve is the combination with the disease burden was very well controlled when I was given combination and it was superior compared to PDL one alone anti ctla-4 lung the IDL alone or the control experiments so we think that in this preclinical model there's suggestion synergism between the combination with checkpoint therapy I both of these have now moved forward into clinical trials the anti CPOE for therapy I was involved with when I was Moffatt before I moved to Georgetown we present that data that ASCO few years back and it was very interesting potentially increased activity in patients we're going to look at on the attack at a stat plus pebble is map data that has been now presented at a few meetings including its it C and as no I this is a phase 1/2 study in patients with a range of tumor types given a standard dose of parallelism at 2 milligrams per kilogram and escalating dos attack out of stat this is an oral agent given continuously 25 milligrams be up to 300 milligrams be ID this led to dose expansion phase and did lead to a randomized phase 2 dose which is 300 and is now being explored in a number of tumor types so the the data that came out of this I was very exciting although we have to dampen it because the sample size was not very large that in the melanoma cohort especially in treatment naive patients the objective response rate was 58% so this is basically rivaling numbers that we've seen with no volum at posit balloon map and we saw a 26% complete response rate again that's only a few patients but we never saw a twenty six percent response rate with any other IO therapy that we've been studying also in a non small cell lung cancer cohort 42 percent objective response rate so the early science of potentially enhanced activity which i think is very exciting for a lot of us in the iowa space I looking at melanoma you can see in this waterfall plot clearly some very good complete responders these were PDL one positive tumors we have a few great responders though in this case that were either PDL one unknown these blue or they actually had it word negative two PRS and the melanoma core non-small-cell lung cancer in terms of the PDL one status in Reno so less data but again there's one patient non-small-cell lung cancer that was PDL and negative that I had a partial response to unknown so this might actually expand the number of responders to a PD 1 or PDL one based therapy that are not expressing PDL 1 on the tumors surface the other interesting point of the study is that the tolerability was I was fairly good that if you look at the grade 3 for toxicity rate it was only 19% for a combination strategy that's very exciting the two areas that we saw a fairly high increase in the toxicities for grade 3 for rates were rash which most of us feel very comfortable managing in the clinics and then lipase amylase were increased this is a phenomenon that we now see most of the time it's asymptomatic patients as long as it's not great for a lot of times these patients still get treated and the treatment related adverse events that led to discontinuation only 8% of patients and overall this was a well tolerated therapy that's now moved forward and the study that's leading the pack echo 301 familism at posit back at estat and unresectable or metastatic melanoma randomized trial getting femoral is map plus epic a toast ad versus pebble is aMAG and placebo this is accruing patients I won't be for a while before we see results but very exciting now a pack out that's not the only idea inhibitor I in Doc's mod is another one that's can study this is a phase 2 trial that had a very interesting design patients actually go from one checkpoint to the other and stay on the ideal one inhibitor the tolerability appeared fairly well there was no concern over a safety signal and the response rate when they grouped all the data together 36 percent this includes patients that were blue man plus index mob as well as anti PD 1 plus and X madman majority I'm actually we're getting parallelism map and in that case they saw a comparable response rate 53% to what we saw in the back out of stat study now the follow-up for this study was relatively short when I was presented at ASCO you can see one person I was actually followed out a little bit longer you also have some patients on having what would have been captured probably a stable disease so I another interest in combination with I do inhibitors but there's also a few other studies ongoing that are worth mentioning so I pack a dose that's being combined with other checkpoint therapy and Ebola map a pack outsides being combined with it as ilysm add primarily non-small-cell lung cancer another anti PDL one develop map and GDC zero nine one nine is another idea 1 inhibitor combined with the test Elizabeth so this may emerge as a very exciting new combination strategy where we have multiple regimens and different tumor types so you know where do we go from here just looking at list of trials that were presented at ask a very exciting day from a number of them obviously we can't cover them all but talking about the first long with Pemble is amad plus for 1vb agonist so this is a co stimulatory signal on the t-cells o Tamila map I think I got there right I am phase 1 b+ primal is map so this is a a GG 2 agonist monoclonal by that bind support 1 BB it's stimulating t-cell response if you recall in that very first slide I showed this could be in the priming activation phase but actually could be really anywhere in that at roadmap for immunotherapy and in preclinical models that has been shown to amplify or increase anti tumour responses and the first data that was presented on this for at ASCO looks fairly promising the toxicity profiles there was nothing too alarming in that data and there were some interesting responders so if you count up these number of bars it's not an exceptionally high response rates of 6 objective responses of 23 patients but if you look at who these patients were or are I think that's where it's a little bit more interesting so you can see the blue here is a small cell lung cancer patient I guess this is teal renal cell carcinoma and then there was an anaplastic thyroid carcinoma patient in this green that had a very deep partial response so are we seeing synergism enhance effect or they just lucky that it was a p21 response we don't know but it's at least very promising another strategy this is slightly different looking at an agent leery that's targeting more than natural killer cells we know that Nivola mom does improve survival there's a reasonably good objective response rate having a Volm AB in the second line area for squamous cell head and neck cancer eye you can see here in the phase 3 trials 13.3% I combining the Leary with neva lab there's preclinical data showing this make sense partly because net in case cell population infiltration is associated in these tumor types and that the combination may enhance the anti-tumor effect in this study it reports the activity rate at 24% so a bit better than what we saw from the phase 3 trial complete response rate 3 patients but this was 10% so this early data suggest that we may be seeing enhanced activity with the combination with an anti Kerr monoclonal bindable map at least in head neck cancer so transitioning to a slightly different approach with amino therapy this is t vex so I think a lot of people in the room are very familiar with this but if you're not this is a modified modified oncolytic virus so it's a herpes virus that's been attenuated a bit also has a gene with gm-csf and certain in it and it's a direct intralesional therapy for melanoma primarily that's been developed but it certainly is interesting as a development and other tumor types and it may turn what we call a non inflamed cold tumor there's not a lot of information going on into an inflamed hot tumor where you inject the tumor with the virus it infects the cells it causes local destruction affects the nearby cells it secretes gm-csf and it draws in an immune response and you can see in this example of a patient with dermal metastasis with melanoma each one of these were injected and at week 18 is a complete response this has now been combined with a blue map I 50 percent objective response rate in the phase one study do you have to keep in mind that a lot of these patients have early less advanced disease than some of the other sites that we're talking about but still very impressive and you can see in this waterfall plot that there is patients with stage 3 disease all the way to m1b stage 4 disease from melanoma the most closely most events this would be the most event that had complete responses now there's also data with T vac plus anti pd-1 therapy the preliminary data was presented at SMR and 57% of patients had an objective response in that study this has led to a randomized phase 3 study design in advanced melanoma patients must be treatment now except for B RAF inhibitor and they must have injectable lesions so it does select out a patient population that may differ from the other trials and it's going to be a randomized trial where you get tea-bagged plus pembo verses parallelism AB alone there is a placebo control to it so the other area that I want to touch on because this is then evolving other things very quickly and it does make a lot of sense that we may be able to combine targeted therapy with some of our immune checkpoint therapies I some of the initial attempts didn't go very well I don't know if you guys remember it bloom AB combined with M you're a [ __ ] eye that was stopped early there was a New England Journal letter the data was kind of really fully reported but there was toxicity that was unacceptable and really stopped in his tracks it blew my poster bathroom in place and tremendum be roughneck drugs in be rapping at melanoma that was also stopped early there were colitis issues with that but we hope that with pd-1 drugs that'll be a more successful combining with targeted therapy and there's strong rationale for using a targeted therapy to enhance and mean response within poppy d1i and these are just three examples enhanced tumor antigen release expression on the tumour cells surface modulation of immune checkpoints may actually bring down PDL one on iTunes or sell and it also can suppress cytokines that may be secreted in the tumor microenvironment you may also enhance the t-cell activity so think about b raf inhibitors they inhibit mutant b raf but they actually may stimulate the map kinase pathway and b raf wild-type cells which is pretty much every other cell in the body so you may actually enhance t cell activity and if you look at melanoma and these are tumor samples from patients treated with a b-raf inhibitor a baseline pretreatment and post you can see a dense infiltrate of cd8 positive T cells that were was seeing after treatment this tends to be a time-dependent effect that early on you see the infiltration down the road you may actually lose that in that progression but it's a very statistically significant influx of t-cells especially the cytotoxic cd8 positive T cells that you see with B wrap therapy in advanced melanoma and it's not only that you can see changes that are very favorable for a immune strategy where you see a decrease in some of the immune suppressive cytokines with bereft REO therapy including il-6 and il-8 you also see an increase in the cytotoxic activity of the T cells so that would be porphyrin in grams and B however you do see some mechanisms that are occurring that may dampen an immune response and that's with Tim three so this is MPD long these obviously can dampen or inhibit immune responses and the one that we're obviously all very familiar with is PDL one expression the PDL one expression can go low in the tumor microenvironment when it's exposed to a B ref targeted therapy so that may mitigate some of the immune effect in eradicating a tumor that's already responded to a upfront treatment with prf therapy so we now have trials that are moving forward or we have at least early data with PD wine PDL one therapies combined with targeted therapy melanoma this I study with the Tesla's mavin code we met Ned the MEK inhibitor this was presented at SMR this past year and there was actually I believe 21 patients but 20 with non ocular melanoma and half of them were bereft mutant they were allowed to have prior treatment an objective response rate of 45 percent across the board fifty percent response rate in the BRF wild-type patients so it's potentially enhanced activity although it's very hard to tease that out from this study I the grade 3 for the adverse event rate did go up which is not too surprising given that both of these agents can lead to an additive effects closest to this number some of the AES that were notable or diarrhea and rash at the presentation was thought that this was maybe not the best tolerated but acceptable to move forward I that I this was not prohibiting for their study and this has now moved forward with it says Elizabeth let melet them you're a fan of and Coby met nib phase one b-29 patients be rapid melanoma that were treated there is a run-in period the way they've evolved with them your afternoon Kobe met nib the MEK inhibitor are given for 28-day window before the anti PDL one agents given and then actually the dimer abstinence giving a slightly lower dose rather 960 milligrams twice a day at 7:20 eye and in this study a higher overall objective response rate 83% was seen so this is a very promising that we're actually getting numbers that may do a whole lot better than just a single strategies alone complete response rate 10% a grave 3/4 of adverse event rate of 40% eye is acceptable although we do see some fairly significant AES in some patients in the study including hepatitis this Myositis substance in one patient and diarrhea where are we going forward I it's not just anti PDL one but also anti PD one and the keynote zero zero two is a very similar concept parallelism I plus the browser have internet net and this data was presented as Quebec Tony Rivas is still very early data fifteen patients no prior be rapper Mac therapy there was no run in period so everyone got all three agents standard dosing from very beginning an objective response rate of 60% a little higher grade 3 4 verse event rate was reported again the liver was an area that was of concern neutropenia was observed in a patient that was dose limiting toxicity and 33 percent of all patients discontinue treatment due to adverse events so a little higher than what we see for member lives and I belong but still very promising both of these strategies are now moving forward in randomized trials so the phase 3 trial that says Elizabeth and Kobe met nib in metastatic melanoma this is a randomized trial phase 3 500 patients the PDL 1 will be the placebo so all patients will be ref targeted therapy and half the patients will be getting it as Elizabeth says was map on top of that key nodes are 0-2 actually was a phase and phase 1 beta phase 2 trial this is the expansion in the phase 2 randomized design 1 to 1 PI R and PFS and these are on in patients that are true and i'ii be reppin at melanoma again everyone's getting bereft I repeat raffle and post competitive and half the patients are getting a VIP d1 therapy with realism am so it's not just a melanoma story we're also seeing potentially combined enhanced effects with anti PDL one and targeted therapy and renal cell carcinoma and I think the the preclinical data looks kind of interesting in this study it came out of MD Anderson in patients primary tumors that were removed after sunitinib purposes now these are anti P I have got Jeff or betcha for therapies they did see a higher city of cd8 positive T cells they saw higher density of T rags and they also saw higher positivity rate for PDL 1 and interestingly these tumors that had higher PD aligned higher t-cell infiltrate actually those patients did worse that it's a little bit different than what we've seen from other biomarkers studies that you can see the overall survival in that in the patients that were in the blue of sunitinib was less same thing with deficits Matt PFS and I Wes it may be a little harder to interpret but this actually overlaps with a little with the data that was presented from the compares trial patients with renal cell carcinoma getting bizarre never seen it nib and these are baseline tumors where they showed that if you were high for PDL 1 in this trial so this would be the orange line and the light blue line you actually had a worse overall survival and if you look at the breakdown for median it's very obvious who are the patients that did better and these were the submitted patients that had low PDL one because often in patients that had low PDL one compared to the patients at high ptl's gaming so it does say suggest that maybe we should be combining with anti pd-1 or PDL one therapy and in the study of exiting Plus Penn bro advanced renal cell carcinoma this was presented ESMA a objective response rate of 70% was seen which is I think very exciting both agents have response rates ballpark say 25 percent so even more than an additive effect although they warn that this is very early data in fact PFS data was too immature to represent but still very exciting and you can see here the depth of the responses at majority patients at tumor shrinkage the other part is that the toxicity profile this is grade 3 or above any event that 65 percent did have an event but if you counted as immune related it was far less 19 percent this is always very tricky to interpret in studies that we've been doing but it does suggest that it you go from treatment merge them to immune-related these are the ones that are more related to the drugs themselves again you see liver toxicity diarrhea as well these strategies are now moving forward there are three randomized studies that are looking at veg F the GFR therapy in combination with the PD one or PDL one agent and Pemberly scientist exit man versus MIT named mono therapy in javelin 101 we have a valium ad plus exhibited and sunitinib and then in this last one king of a team is with his optimum and we did see some data with checkmate o 16 I just wanted to bring up with there was activity but also the adverse event rate was similarly high in the in that study so I know this was a very quick whirlwind through a number of combination strategies where are we going from here it's hopefully that we're not just going to combine every drug we possibly can in and throw it at a patient that we need to be smart about it needs to be based on patient's characteristics a tumours characteristic or some type of composite biomarker strategy and this is a very big oversimplification of the problem ahead of us but talking about things as inflamed versus non inflamed there are different strategies where you may be able to enhance and mean response based on PDL one expression are their immune targets and otherwise do you have problems with antigen expression that you have to try to overcome for example as MHC of loss you're not going to actually have great t-cell responses you have to approach a different way so this is a reference that might be worth reading but I came at all and certainly I think a lot of this on that at the table here very interested in this topic so with that that's my last slide and I will let it hand it over to there thank you so I think that you know combinations are you know very challenging for IO you know when we give chemotherapy it's you know a plus B and equals C in terms of efficacy whereas you know with ideal nibblers or MEK inhibitors you know we there may not be any single agent activity but the combinations are really compelling the the best example that I know of that is the the small data set with a Tesla lism AB and koba koba Magnum in MSI's stable colorectal cancer so these are not MSI high tumors is your MSI stable colorectal cancer where the response rate with IO is zero the response rate with mechanician is probably zero you had the combination in twenty patients with K rasputin colorectal cancer the response rate was 20% so we have a lot to understand around the combinations next you know Jeff's last slide looking at the paradigm of inflame versus non flame trying to parse out subgroups that we can identify with difference that genetic and immune phenotypes to create immuno target immuno tailored therapy is a great segue to David who really has dedicated his efforts recently around immunotherapy biomarkers and thanks Dave so actually what I want to talk to you about today is what's what's so exciting about immune therapy is actually the fact that most likely more than any therapy in the past this is one that's truly going to be a companion diagnostic therapy or a precision medicine or personalized medicine therapy and that that's what it's going to take to figure out which patients respond and as you saw in all the previous call the previous speakers you could see that there was clearly lots of different classifications of disease and lots of different responses and so it's the pathologist that's charged with being able to tell which is which and then pass on to our oncology colleagues which patients fall in which classes and then how we can determine which drugs or combinations they should get so I'm going to go through a few different topics today first we're going to talk about an update on that current FDA approved companion diagnostic PDL one and I'll tell you about the difference between that and a complementary diagnostic test which is actually true as well also FDA approved and we'll talk about some novel biomarkers first I'll talk about genetic biomarkers that is mutational load which was really first brought to our attention by our chair today dr. Uzi and then we'll talk about the future biomarker landscape including some new approaches to thinking about the biomarkers that might allow us to look a more personalized way at therapies specifically the T cell activation panel capacity to express PDL one and finally an 18 gene Manus Turing RNA expression signature dr. Guerin and dr. Gibney went through a bunch of these papers so I'm not going to spend a lot of time going through them but the salient feature of these was no matter which paper you looked at the PDL one hi patients always did better than the PDL one or showed greater response than the PDL one low so as much as we don't like this particular companion diagnostic test it really has signal and then even when we got to combinations in the more recent work combining nivo and it be we still see the same thing that is the higher the levels of PDL one the greater the response rate so I think that at least in the beginning we're definitely going to be using PDL one as a companion diagnostic test and there so what I'd like to do to start with is go through the different drugs and what their companion diagnostic or complementary diagnostic tests are and what the differences are and why and so these are the four main drugs that have been seen in this space we heard a little bit about merkel salt which adds a fifth but what you can see is that each of these drugs has a different test or assay which is positively distressing for pathologists and led to an fda and ACR convened meeting to talk about this problem because in the pathology world we usually test for a single biomarker with any number of tests which might correspond to any number of drugs but here we see that each test is associated with a different as not only is that this are different but each one is based on a different antibody and uses a different scoring system so how do we deal with this problem so where we are today is that we have only one fda approved test and that's 22 C 3 this is FDA approved as a companion diagnostic for Pember lism AB in the first line that is if there's greater than 50% cells expressing but in the second line if it's greater than 1% the rest of the tests are all complementary diagnostic tests that means it's useful and it could be helpful it's not required in order to prescribe the drug so that's an important distinction because as pathologists we may or may not decide to do that test depending on situations other than the drug the response to therapy that is market forces and other things play a role a lot of times we won't do a complementary diagnostic test because we don't have to do it and and similarly oncologists don't necessarily need to use it it's helpful but not necessary so nuvola man is the first example that does not require a companion diagnostic test but rather a complementary diagnostic test and the 28th 8si which is fda approved as a complementary diagnostic test uses a very low cut point greater than 1% tumor cells positive the SP 142 test is also FDA approved and that's a complementary diagnostic test for at is ilysm at it's in bladder cancer and more recently in lung cancer as well and not only that as I'll show you in a second you read it differently you read the immune cells as opposed to the tumor cells and give it a supporting system another test SP 263 has not yet been can approve Dazz a companion diagnostic test but will probably see it as either a companion or a complementary later on this year when their value map is evaluated by the FDA and finally there's an antibody and a test called a lab derived test that's been used which is neither a complementary or companion diagnostic test but it's a test pathologist invent in their own labs so to speak or at least validate in their own labs and this has actually been used in a number of places most significantly Morial Sloane chattering in Brigham and Women's Hospital use this as their standard test which they test most lung cancer and some melanoma patients the first question we want to talk about is are these all the same so these there's now three publications describing these three tests that are comparative this so-called blueprint project was the result of the FDA ACR ASCO meeting and it was led by Fred Hirsch but included people from a number of people from each of the companies including each of the diagnostic companies in contrast our study that was led by Ignacio with Steuben and I was sponsored by the NCCN although paid for by BMS and this was supposed to be a statistically powered prospective multi-institutional test without including any of the producers of either the drugs or the companion diagnostic tests and finally there's a last paper that came out that's got the most patients in it so I'll show you that as well so here's a comparison of the tests The Blueprint test at 39 cases read by three pathologists one from doc oh and two from Ventana and looked at more assays that are FDA cleared as either complementary or companion tests it was not statistically powered but it was a port of force to get these six companies to agree on this study and then to agree on publication the NCCN study was 90 cases it was done by 13 pathologists from seven different academic sites included four assays but excluded the SP 263 assay since at the time it was not if still is not FDA approved and instead used in a lab derived test which was II one l3n a cell signaling antibody on the Laika bond platform it was autistic we powered for inter class correlation coefficient between pathologists between assays and by subcellular localizations so we asked three questions instead of one in this study and this was led by the NCCN and you can see the difference right away here by looking at the assay and then the scores and you can see three of the assays in both studies look fairly similar although there is one outlier assay in the light green here and in the light aqua here and you can see that particular assay the SP 142 a site recognizes fewer patients this is what they actually look like when you look down the microscope tubes and you could see that while these assays look similar the different amplification systems gave them some different coloration this one looks like it's missing cells that is there's a lot of cells that are positive here that are negative in the serial section this is what they saw on the blueprint this is what we saw on the NCCN study however we're informed by the vendor by Ventana that this assay is designed to detect immune cells so what does that mean how does an antibody know whether an antigen is in a mutant cell or a tumor cell I've cobbled with that now I know that there they showed this as an example where you can see more immune cells here in the center compared to the tumour cells on the outside but I'm not sure that wasn't just actually a optical illusion since it seems like there's fewer tumor cells not more immune cells in this particular assay so we looked at that in both studies as well and in the blueprint study where there was only 39 patients or 39 specimens looked at you couldn't see the difference between any of the assays when reading immune cells however in the NCCN study where we had statistical power you could see that the immune cells were actually lower than the other three tests as well and so this was a the best evidence that we have that even the immune cells were also detected at a lower level than the tumor cells now the way that they looked at this is different between the different studies and specifically concordance was using was hard to measure because each assay came with a different set of cut points and so what they did in the blueprint study is they used the cut point for each assay and just called it positive negative and asked what their concordance was and unfortunately it wasn't very good it was only about 63% concordant between the assays which means you clearly couldn't cross utilize assays between drugs with only a 64 consent concordance and the ones that were concordant are the really high express errs and the really low expresses and the one and the ones in the middle were the ones that were non concordant we could look at this in another way in the NCCN study we used every scoring category that was possible which left us with six scoring category which you can see here and you can also see something interesting in the distribution of the scoring categories for each of these assays you can see what looks almost like a bimodal distribution except for SP 142 where there aren't as many positive cases and in fact in SP 142 over 50% fewer pate cake patients are considered positive by that assay when you're looking at tumor cell scoring the others were actually quite similar at either the higher cut point greater than 50% or the lower cut point greater than 1% the immune cell distribution is similar in that the SP 142 was about half what was seen in the other three assays where the other 3 assets were approximately equivalent the NCCN study because it was done by 13 pathologists gave us the opportunity to tist achill e test the levels of expression which we couldn't do in the real world setting in the real world one pathologist reads your slide and that's it but in this experimental setting we could average the readings of 13 different pathologists and then we could actually see a subtle difference that couldn't be seen in previous studies and that was that actually the 22 C 3s a is slightly in fact statistically significantly lower than the other two assays in fact the only two that are not statistically significantly different are the LD t e 1 l 3 n and the 28 8 assay the 22 C 3 is slightly lower a factor of 0.3 the SP 142 is dramatically lower 1.2 order of magnitude of the difference of the mean and so you could see that the average pathologist score could allow us to look at statistical significance the other thing you can do is look at inter class correlation coefficient this is what your level of agreement is for the assays and here you can see for each of the assays whether for the tumor there's a very high agreement and 0.97 is ridiculously good agreement as they read by pathologists this is again the averages so that we're looking at the differences of the assays not the differences of the pathologists however from the immune cell when you looked at all four assays it was pretty bad anything below 0.4 is really unacceptable for clinical use and you can see that 1.27 now we eliminate SP 142 we actually did have a reasonable concordance for immune cell staining now the x-variable we want to look at is not is the pathologist because the pathologist could also be a reason why there's variability in this assay and in fact this is really interesting because if you look at the inter class correlation coefficient for the pathologist regardless of which assay you use if you're reading tumor cells we do great pathologist can do this asset we all have a concordance ICC's above 0.8 which is really good but the bad news is and what's something that is concern to me is that at for the immune cells we are discordant regardless of which ask they will use and so that means even though you heard our previous oncology speakers talked about using immune cells in bladder cancer and then in head neck cancer and past and certainly in lung cancer for one of the drugs the foul just probably shouldn't read that assay at least according to this study pathologists can't agree on the level of immune cells with ICC's all the low point two now when we look at tumor cells we can look at two different variables and see how well we can read tumor cells as long as we're comparing pathologists and what you can see is that when we have a cut-off of greater than 50% we got this pathologists are really good at reading it greater than 50% but at greater than 1% we actually start to lose our concordance it goes down to as low as 0.53 and this might be also partly explaining why we do so poorly with immune cells because we're trying to tell the difference between zero and one percent or 1 percent and 10 percent which are more subtle than telling the difference between 50 percent or less than 50 percent so in another study we actually looked at this quantitatively and looking at this quantitatively we can actually look at each of the antibodies not the assays what I just showed you was looking at the assays but this study looked at the antibodies and we took each antibody and determine its maximal signal-to-noise and so in fact when we determine the maximal signal or noise we can look at the actual concentration via anybody not the dilution and figure out what the positive cases are what the negative cases are and then do a ratio and figure out the signal-to-noise and we do that using an array like or showing you over here we can actually compare the antibodies to each other to see if what we saw in the assays was an artifact of the way the assays were constructed and in fact that that may have been the case that one way to think about it is the antibodies are the eggs and the assays are the cake and so the cakes might be very different even if the egg that you start with are all equivalent and in fact that's the case this SP 142 anybody even though the assay was by far an outlier looks essentially identical when you use it at an optimal signal-to-noise this is another example of higher power view where we used SP 142 antibody and an optimal signal-to-noise and you can see it looks essentially no different from the other assays that work or generated so we compared all six of these antibodies in the study and what we found they're all essentially identical and so across about 25 different cases and then looking at regressions you can see that the antibodies all agree with each other themselves so that's the good news the good news is the antibodies are the same the bad news is some of the assays are constructed in a way that makes some of the assays or at least one of the assays quite a bit lower so the antibodies are similar the assays are not the Ventana sb 142 si as prepared detects only about one half of the signal compared to the other four assays as confirmed by two of the multi institutional studies the quantitative studies suggest the immune cell PDL 1 expression correlates with tumor cell expression actually didn't show that data but studies comparing the pathologist scoring of immune cells suggest that the concordance is poor while pathologist scoring of tumor cells is is great in fact better when we use the highest cut point greater than 50% so with that I want to talk about some of the novel biomarkers and then some of the future biomarkers so one of the novel biomarkers that was discovered early on by dr. Rizvi and his group was the mutational load the fact that you could see patients with higher mutation a load were more likely to respond to therapy than patients with lower mutation although this was the discovery cohort the validation cohort and then combined and this makes sense since it's the mute and the increasing number of mutations that ultimately leads to generation of neo antigens which theoretically stimulates the immune system another example this is shown in mismatch repair deficiency where mismatch repair proficient colon cancers don't generate so many neo antigens but mismatch repair deficient colon cancers and very high levels of mutational rates and may be an order of magnitude higher and these are the patients that actually do very well as shown in this work by me at all so does this follow all the way to neo antigens yes in fact if you look at neo antigen mode instead of mutational load the same thing as seen where high levels of neo antigens and colon cancer and melanoma and non-small-cell have all born out to be more likely to be associated with response to therapy this is further validated more recently in some larger studies where again you can see that the high mutational load or high number of mutations per megabase and morale are more commonly seen in responders and overall they do better so should we switch to this should we abandon the PDL one assay and just measure mutational load the problem is the sensitivity issue in the sensitivity of these assays it tends to be between about 50 and 70 percent and assay that we use in the clinic we don't want to have that many patients miss the opportunity to benefit from the therapy so we're looking for assays that are close to 90 percent the PDL one IHC assay is not that high but it's closer to 90 it's closer to 80 80 percent then the 60 percent that we see with mutation alone so what about other biomarkers or other sort of future things could we think about and is there hope for getting higher sensitivity or maintaining that high sensitivity but also bringing up the specificity so I'm going to talk about a couple of theoretical things that use some new boxes that might need to be introduced unto our labs one is the PerkinElmer vector platform in a nano string platform and both of them have been used toward the goal of coming up with a more complex question than just is it positive or negative the first was based on the observation that patients that our responders have a lot of t-cells in fact the higher the so called as you saw before the patients that have a lot of immune infiltrate as opposed to the immune desert patients do much better and that's what's shown here they also similarly have high PDL one and so we've come up with an assay where we try to look at a combination of tumor infiltrating lymphocytes and PDL one and combine them into three types the ones that are the immune desert those need some combination therapy to bring the immune cells in but what about the ones that have immune cells present would be good to be able to sub classify those into the responders and the non responders and the way we've tried to do that is take the positive ones and put them into one of two classes type three which are the activated t-cells they have high grams I'm and or high ki 67 or type two the ones that have high C D three lots of cells there but low grands I'm and low key 67 presumably those are inhibited by something well when we look at that a relatively small pilot population study we can see that in fact CD 3 bears out and there are more high fede 3s and the responders than the non responders but by itself neither granzymes nor ki-67 is actually predictive of response but when we combined them all together with that algorithm that is to combine them into three types that are activated or inactivated we study something very interesting and that is that in this population although there's only seven type twos none of them have died of the disease at this point and the progression-free survival is also statistically significantly better than the type ones in the type trees now the type twos remember are the ones that have a lot of t-cells but don't have Gran's I'm or t67 in those key t-cells so those are the ones that are dormant perhaps they're checked by a checkpoint when which may be what we're inhibiting and why we're seeing response to therapy so a potential way to bring an assay on that isn't based on the tumor cells themselves but actually based on the immune micro environment that we might see in the future as a mechanism to predict response to PDL one-axis therapies so this is still really preliminary data but I'm just sharing an that stand up to cancer cohort is underway and this is even a smaller study with only 37 patients is not statistically significant but it gives me increased confidence is why I share it with you it's a different disease it's melanoma and not only is it a different disease but it's a different platform this was done with a different software to try to distinguish the type 1 versus type 2 and 3 and in fact you can see the type 2 group once again does better than the type 1 and type 3 and hopefully in 3 to 6 months we'll have much more data on this potential test next one is the capacity to express PDL one so maybe the reason that we see some of those cases that are PDL one negative that responds to therapy is because we're not looking in the right place so PDL one down here is activated by a promoter IR f1 and NF kappa-b depending on which pathway you go down to actually stimulate PDL one and so perhaps you have to see the microenvironment right at the right time but we pathologists we don't get to choose the time the tissue comes when the tissue comes and so maybe we could look at potential or capacity to express PDL 1 instead of PDL 1 itself and that's what this study was based on focusing first on ir f1 and later on NF kappa-b and what you can see in patients who have high IR f1 do better in terms of response to therapy than patients with low IRF one compared to PDL one in this relatively small melanoma cohort but as you can see again here in about 70 patients the patients with high RF one have longer disease-free survival than those who had low IR f1 even though in this melanoma cohort and as has been seen in many melanoma cohorts PDL one by itself doesn't classify the patient's four outcome so another approach that we hope will ultimately promising and it's easier to read it's nucleus a nuclear expression and here's even an example of a case where you can see pretty strong nuclear expression of IR f1 this is not how we did the study this is just a parallel da b si simulating what would be done in a pathologist lab where you can see that even though it's IR f1 positive it's negative for PDL 1 in a serial section in the exact same region of the tune so the last thing I want to talk about in the last minute or so is the nana string 18 gene si and this is the idea behind this is that you could actually grind up the specimen and look at the immune signaling things insight - I think sometimes you can do much higher degrees of multiplexing and in fact that's what Nana stiring did they looked at many of these different markers that regulate and modulate the cd8 positive T cells and other T cells that may have interactions and in fact RNAs from the tumor cells as well ultimately starting with the close to 800 different genes they looked at a bunch of cases and then looked at ROC curves and decided these were the most selective genes some from the interferon gamma pathway some from the t-cell exhaustion pathway and some from pathways that sort of made sense but it wasn't clear what role they had but empirically these were the ones that they found were likely to be most predictive and I'm not showing the whole dataset but rather I'll just show the area under the curve compared to PDL one using their 18 gene classifier score is this area under the curve on top which actually looks quite quite promising compared to the PDL one assay and this is actually in squamous cell cancer the head and neck but in has been seen in other assays as well so this is something in the future we know from just from press releases that good progress is being made toward testing this even in the prospective setting and some retrospective work on some of the keynote and checkmates studies and we look forward to seeing the results of these kinds of assays which might be another future approach to pick the right patients for the right therapy thank you so I'm going to really go through a few scenarios of how immunotherapy could be used and in everyday practice it's not meant to be an exhaustive overview how I use immunotherapy and practice but I think this is a good case that's illustrative of some of the topics I'd like to cover this is a patient with a K rasputin lung cancer who received single agent PDL one therapy had a complete response after just six weeks of therapy and after about a year of therapy she developed fairly significant skin toxicity FSIA tiss that was quite symptomatic for her requiring her to come off and now she's been off therapy for over two years and her PET scan looks really identical to the one on the right so I think it begs the question such as you know how long we continue these therapies what do we do about toxicities how do we manage toxicity so the questions I'm really going to try to address are the duration of treatment if you have a patient receiving immune checkpoint inhibitors era P and they have a complete response do we stop treatment if patients have a toxicity to a therapy is that a reason to stop treatment and how do they do in terms of outcomes of patients have toxicities or no toxins and the standard treatment for patients who have an immune a is steroids which are immunosuppressive and I think everyone gets a little bit worried about whether that abrogate some efficacy of these therapies the other topic I'd like to touch upon is really what we see in the clinic where we don't see a beautiful pet response like I showed you and every patients and it's in the clinic we do see scenarios where some areas of tumor respond other areas don't of other areas progress and what do we do about those situations and finally what do we do about patients who do respond and ultimately have progression and the last topic is patients that are older or poor performance status certainly the poor patients weren't included in most clinical trials and how how do they do with immunotherapy so the first topic is really is complete response or reason to stop treatment I think that that we're seeing this more and more with immunotherapy where we are seeing complete responses for example in the lung experience with a first line member elisabeth and PDL and positive disease the complete response rate with Pember ilysm AB was a nearly 10% the melanoma doctors experience complete responses with their therapies even more frequently and what do we do with those patients so the short answer is we don't have a lot of data there is some data from the melanoma experience and the keynote oh one trial where patients had a complete response the day that looked at and what happened to them and keep in mind for this trial if you have a complete response in the melanoma patients you're actually allowed to come off trial if you have a complete response that's maintained for more than six months so the criteria from this trial was six months see our ongoing you can stop therapy and so in fact of these 95 patients or 15% of melanoma patients that had a CR 64% stopped for observation there were 17 percent that were discontinued to actually do the AES and very few patients discontinued you do a disease progression and the 19% at the time of this presentation remained on on Pember ilysm M and there's a lot of data here but I think it's really pretty compelling what's most compelling is that if you have a complete response with melanoma 97% of these responses were maintained the medium time on treatment was a nearly two years and off treatment was about 10 months and these complete responses don't occur overnight if you look at the blue circles that's the time to partial response and then the dark blue circles are time to complete response so the media time to CR was actually 13 months and as mentioned a 97% of these responses were maintained so I think there's some data certainly from this experience in melanoma where you have a CR they maintained a CR for six months one can say with some certainty that their benefit is fairly durable what about toxicity is that a reason to stop treatment and do steroids abrogate efficacy again we were starting seeing more and more data this was a publication from msk it would it bloom AB in melanoma it was patients that were treated off protocol 300 patients were treated 19 percent of patients discontinued it Balu map due to toxicity 35% required steroid using 10% actually required anti-tnf use and you can see on the left two panels with no toxicity or with any toxicity there is really no difference in terms of outcome and similarly on the right whether they receive steroids or did not receive steroids the outcomes were were very comparable actually in 12% of these patients a long-term benefit was experienced even off therapy so one of the kind of reflex of doctors maybe if you have a toxicity well if you can't go back to the immunotherapy for a bad toxicity you go on some other treatment I think that that may may not be true I think observation is is very appropriate my most extreme example is a lung cancer patient who had pd-1 therapy and after one dose developed type 1 diabetes with a glucose of 600 and renal failure and he actually had nearly a CR from that one dose of pd-1 and now he's out about nine months with that CR being maintained and he actually had osseous Matassa T's as well so the additional data with nivolumab again in melanoma this is for BMS trials that we're looking at Nivola med mono therapy with immune modulation or without immune modulation and if you look at the response rates the PR rate is in the in the mid-20s for both and the duration of response is again fairly comparable immune modularity reuse was in about 35% of patients and 25% did require steroid so whether you got steroids or not the outcome seemed very comparable finally in terms of a combination immunotherapy where we do see more immune adverse events with nivo and epyon melanoma you can see in the left panels that the number of patients from all randomized on this analysis 40% received more than four doses of nivo whereas 23% received more than 40 niebo if they were discontinued due to AE so 35 out of the 94 discs in 35 patients discontinued you today AES whereas a 94 were randomized and the panel's you can see the but not just the overall response rate but also the complete response rates those deep responses where we're really quite comparable whether they had a toxicity or not and this shows the PFS for whether patients discontinued due to AES no Iza and the top two curves the bottom curve was a wouldn't it be alone which is what you'd expect what about tumor heterogeneity or mixed response this is a patient that I'm treating them currently with PD one therapy you can see she has fairly bulky lymph nodes and a reduced einem and after six weeks of treatment you can see clear response to to her lymph nodes she actually couldn't turn her neck due to extensive cervical and supraclavicular lymph nodes as well which totally resolved but on that same scan she had a right lower lobe mass that seemed to increase and causing a collapse of lung her nodal disease obviously was was responding though so what do you do for a patient like this well for this case we actually continued the immunotherapy but radiated our lung primary shiri expanded that lung and on follow-up scans that we did a few months after the radiation she continues to respond to her nodal disease and she's having an ongoing response and and so isolated progression does not equal treatment failure and I think that's important to to note and local areas of progression can be managed with with local therapy and the immunotherapy can be continued and part of the reason may be that all the different areas of cancer may not represent the same disease David showed our data around member Elizabeth treated patients having a correlation of response based on mutational load so the more the more nonsynonymous mutations you have the more immuno genic a tumor that exists and there is that correlation but I think that we we scratched our heads around some of the data that we had where in the arrows you can see there are three patients who who had high mutation burden but those were those were non responders to immunotherapy so that the the patients on the left of the middle are the responders and the patients on the right were the non responders and those three patients actually had a significant subclone fraction there were fairly heterogeneous tumors and so despite having a high mutation burden they were heterogeneous in nature and ultimately had treatment failure so I think that one needs to factor heterogeneity in terms of clinical application acquired resistance what do we do about acquired resistance again we're seeing that more and more data from Tony rebuses group and UCLA has has begun to study this more systematically and in these melanoma cases there are very specific alterations that have been detected around cases where acquired resistance has developed loss of function jack-jack mutations loss of function beta 2 microglobulin mutations have been increasingly being described in the scenario of of acquired risk as resistance and I think going forward well as we understand this biology we'll be able to develop ways to overcome these mechanisms of resistance we see this in non small cell lung cancer as well this is a patient of mine who in the first panel was pretreatment and the second panel is post treating and showing a treatment response and then 15 months later developed an isolated nodal area of progression we actually resected that that node and again as we discussed earlier we should manage patients who have local areas of progression with local therapy and he remains off of therapy with an ongoing complete response about 18 months off of after that nodal reception we to analyze that lymph node to look for a smoking gun in terms of mechanisms of resistance in terms of up regulation of pd-1 or Tim or lag development of resistance mutations such as beta 2 m or Jack and other mutations but we really found no no smoking gun in this case and I think the challenge for understanding resistance to immunotherapy is it's not going to be as straightforward as easier for mutant lung cancer where you develop a t7 90m mutation I think that the resistance mechanism is going to be very diverse in nature and I think will be part of the challenge can we rescue patients right that's the question that you know comes up every day which is why our phase 1 clinics are very busy with immunotherapy combinations you know so if we treat a second-line lung patient with single agent P D 1 and the response rate is 20% and the durable response rate is 10% well you're clearly not helping most of the patients and what do you do after that is there an opportunity to rescue those patients and the short answer is we don't know this was an interesting case that I've been treating though it was a actually an MSI hi colorectal cancer patient who had progression of disease after standard chemotherapy and was treated on a protocol with us with single single agent PDL one he received 14 cycles of treatment he actually had a minor response he had a big pre sacral mass he had improvement in his pain and neuropathic symptoms but then did progress with escalation of his pain and and and neuropathy and we at that point we rebuy op seed him and we started him on a combination of pd-1 and ctla-4 and you can see on these images that his pre sacral mass going from August to December has clearly decreased in size in both views and he says essentially resolution of pain and resolution of neuropathic symptoms and is continuing on pd-1 and ctla-4 combination therapy so I think there is an opportunity to rescue patients but I think we need to biopsy these patients and try to test them on protocol when we when we can to inform combination strategies finally age and performance status as certainly for performance status as you as you know you know most patients were limited to e cod zero to one performance status for many of our clinical trials the other the other exclusion criteria was autoimmune disease I didn't put up a slide for that but what will we see now clinically is patients you know in the real world have autoimmune conditions and what do you do for patients that come in with a history of Crohn's disease or rheumatoid arthritis and I think that in general what we've been doing is sort of weighing the pros and cons for for immunotherapy in general we do see flares of toxicities in terms of psoriasis which is clearly manageable we get more nervous around patients out of active all sort of colitis or Crohn's disease because then those can flare but I've had a patient with active MS who we've treated who didn't flare so I think it's a bit of a learning curve and we're seeing more and more data emerging around treating patient with a pre-existing autoimmune condition so we do have some data from the bladder experience around poorer performance status or patients with comorbidities the attesa Lizza mapped while looking at patients that were cisplatin ineligible they could be ineligible for their renal impairment history of a great - hearing loss or peripheral neuropathy or ACOG ps2 in this court there was a little over 100 patients that were treated with otezla lism ab and if you can see by the baseline characteristics the overall population the response was in the in the mid-20s even for the older patients the response rates was 28% in more than 80 years of age and as you go down if you look at the impaired renal function ACOG ps2 and even renal impairment in decock ps2 the response rate remained in the in the mid 20s so suggesting that even in those patients who are more compromised activity as seen with with pd-1 PDL one therapy a similar patient population was studied with with Pember lism AB and this was in patients again who ineligible for cisplatin based on renal failure performance status neuropathy hearing loss under this case CHF as well and in these patients the characteristics are shown on the on the left and prior treatments are shown on the right and you can see in terms of the the response rates with a ps/2 patients the response rate was a little lower than what we saw with the tesla's map but clearly was still very active one of the interesting observations in this trial is of the patients who had you know visceral disease or paddock metastasis those those tumor types seem to respond less well to immunotherapy so visceral disease of pata capacities the response rates are clearly lower and that's really across tumor types that where we're seeing that observation but ps/2 i think is is not a preclusive to to IO therapy or elderly patients so to summarize the duration of treatment for these therapies is not yet defined it's open-ended and indefinite for the majority of clinical trials and approvals there are some trials which are capping at two years the I think that what one can say in terms of duration of therapy is those patients those who have a complete response certainly in the melanoma experience it's not unreasonable to stop them if they've had a CR that's been maintained for six months immune a es are not do not preclude benefit if they're giving corticosteroids so I think for the for the people in the practice don't be afraid to use corticosteroids if you need to - to treat their immune adverse event mixed responses can be seen and in practice and may not represent systemic failures don't I think that if it's a patients receiving pd-1 therapy and they they progress in every lesion then you know then I think that that's a different scenario but isolated areas of progression should be managed with local therapy multiple multiple mechanisms of acquired resistance can occur with PD one therapy we potentially have the benefit to overcome that with combinations as we go forward and finally PS to patients do benefit from IO and I will stop there thank you we are now a during our question and answer period feel free to come up to to the microphone we're not bad timewise we were allowed to go a little bit over nine and I will read the questions here we have a few here already and I think that the first one is I think everyone can weigh in and in the group David could start with this one probably is there really any purpose in measuring the levels of pd-1 PDL one whether they are low or high if PD want PD 1 inhibitor therapy is used irrespective dr. rims presentation shows that their their assay seemed to be meaningful but how does it matter clinically well I'm clinic the question clinically maybe you should address but how and that I mean it does matter and we certainly see responders and I think more than even to pick you know respond if it's a complementary diagnostic test obviously don't have to do it if you want to use it in the first line in lung cancer you do have to do it and the patient has to be greater than 50% to be on label but I think even more important than that is that when you see a patient that's completely negative that's a good patient to try to put on a clinical trial as opposed to think about just giving them Nebo or something like that Jeff do you do melanoma with Mary what which is fed by David Graham I know in melanoma we really aren't doing a lot of biomarkers to select patients upfront it's more for research purposes at this point so in some of your data in melanoma suggests that if your PDL one if your PDL one positive if in evil performs comparably to Niva whereas if you're negative the EPI Nebo does better so to spare toxicity would you do a PDL one test in melanoma so that's been brought up a few times and the data that we currently have I think is still very controversial that we may get more data soon survival data coming up right exactly so um um but I I think it all depends on how you want to look at it the original presentation by Jed wallchuck basically if your PDL one positive the PFS curves overlapped if you got the balm a born evolve mat plus if you live in heaven people use as an argument that you really shouldn't give people the ad aggressive toxicity with the combo if you're going to end up at the same place but that was despite a higher objective response rate in that same comparison so you kind of had to ignore that part of it the second time I was presented it did look like there was a separation in the PFS between the two groups that were PDL 1 positive tree with Neela p versus needle um so I I think it's really still unknown based on the data that we have I personally haven't used PDL one testing to decide between vivo or Neve Oakland City and my patient population but I'm sure there are people out there that have Eddie would you offer immunotherapy to a non-small-cell lung cancer patient who is a fast progressor so interestingly so there there are two things on that one there is data emerging that if there really is a fear that the patient is progressing too quickly that one could use combinations of chemotherapy and immunotherapy those are not currently approved therapies but but there is some data supportive of that interestingly in the patients who do respond when we initially were conducting these studies our concern was that it would take months for these patients to respond and clinically that isn't necessarily what we have seen we have generally not seen people who are the responders or for instance they're the the patients who are frontline who had high standing for PDL one did quite well even if they were presumably some of those people were progressing and did have very high disease burden so yes I would be willing to do that if their biomarkers otherwise were were suggestive and that is I think an excellent example where having that biomarker data might make one feel quite different about it all right so from keynote 24 long experience if you are PDL one positive the response rate is numerically higher than with with PEM bro than with chemotherapy and also the time to response is really identical so you're not going to risk your not going to you know lose out in terms of rapidity of response with IO versus chemo you have some questions there we were given an original questions a bunch of questions I can do a one really quickly do you foresee flow cytometry to play a role in future companion diagnostic assays and I would say the answer to that is probably not so far ever there's really not been much signal in peripheral blood at least at a sale level and taking the tumors and dissociating them for flow cytometry analysis has historically not worked because in order to dissociate them you cut off the interesting molecules that you want measure so I would say I wouldn't bet on that technology sure there is a question I think was geared towards maybe the talk that I gave our preclinical data in vivo predictive of human clinical trial IO inhibitor combination results that's a very tough one to answer obviously we've seen combinations in vivo that probably don't play out in clinical trials but then of course most of our successful clinical trials had positive preclinical in vivo data so I don't think it's a guarantee if you have a positive animal model showing I owe therapy is it's necessarily going to play out in the clinical trial that it's it's the starting place we also struggle with what is the best in vivo model so melanoma b16 has been used over and over again and that people heavily criticized it yet it's still used again and they're now gem models that might be a little bit better like Marcus Bozeman Berg's model although that may not quite replicate what we're seeing in humans so the biology of that melanoma may be slightly different so I think it's something that we still struggle with a lot so could you speak to how response correlates with immune toxicity and this is a somewhat controversial area and I will say that Erin Liz Berg from our group will be joining our faculty as a paper out for review on this topic Jeff whoever does have an article jco which indicated a correlation I would say that in general there you can it is very easy to show a correlation between a response and immune toxicity but that is heavily confounded by the idea that people who are on therapy longer are more likely to experience toxicity so I think that there is some suggestion that that may be true but it is controversial and when we look at our data from from the keynote oh one study patients we had nearly hundred patients at our institution we think that it is suggestive of that when we remove the the sort of caveat but I think it is a hard question to study in question on top of your question so is the toxicity more of a dose finding experiment that you basically proved in that patient you've rubbed up the immune system and you achieved what you're trying to do unfortunately the patient toxicity but hopefully will have benefit yeah as a as a follow-on to that there was a question was whether either steroids can be given prophylactically to prevent development of ATS I think that I think that that is something that we don't want people to do you know I think that the most patients don't have immunity s and we actually don't know whether it's detrimental to give somebody prophylactic corticosteroids we know that if you develop a toxicity they seem to do okay if you give them steroids but to hopefully give it prophylactically we that actually could be a negative effect and wouldn't be something that we'd recommend okay have you tried comparing an airway someone oh I'm sorry Raven uh fake as a microphone yeah I have a two-part question actually and maybe a little bit too basic but the first question is have pd-1 and PDL one been documented in the tumor and the tumor infiltrating lymphocytes in biopsy specimens and the second question is has anyone studied the immune repertoire of the tumor infiltrating lymphocytes and treated patients so PDL one and pd1 clearly are present in the PDL one is clearly present in tumor cells PDL one is present in some immune cells it's been shown that they're present and macrophages and some t-cells and PD one is a much harder marker to use since it's expressed in a lot of cells there's a couple of studies that have looked at the I think what you're asking the co localization of pd-1 and PDL one yeah same cell or interaction and there was one study presented last year at a CR where the interaction that is the co-localization was more predictive than either molecule by itself more weight and see if we get confirming data on that concept later this year sorry I probably asked the question wrongly but that do the antibodies have they actually been documented to penetrate into the tumor and coat the surfaces of the respective cells therapeutic antibody yes not to my knowledge if you want to talk about the t-cell repertory the second question first was want to repeat the second question yes has anyone studied that the immunoglobulin hour or the sorry that t-cell receptor gene rearrangements in the tumor infiltrating lymphocytes and can we gain any insight from from those rearrangements yes so arguably that's a potential prognostic factor that I omitted in my genomics part of this best and there are there's a company called adaptive and there are a number of other competitors in the space where people are looking at the T cell receptor repertoire and with the hope that they would either find a specific clonal clonal alteration that would lead them to a prediction of a therapeutic response date on that is I think still forthcoming thank you okay Eddie our PD 1 and PDL one treatments overlapping for / interchangeable would you use an anti PDL one in a patient who failed an TPD one so I would say this question is one that I get asked by colleagues very frequently um I would say that my my bias is that I think that it is unlikely that someone will respond to one when they have progressed on the other although I would also say we don't have definitive evidence and I think that we're getting of that empirically because some of those people are doing it no one has ever called me after I've told them that I would that I would find it unlikely that they would do well to tell me that the patient did great but but that doesn't rule it out okay couple more Jeff can you discuss pros and cons of intramural drug development delivery versus IV drug devout problematic is it to give T back or some of the newer agents that were that are entering the clinic such as sting agonist I mean how tough is it to give it can you give it you know with a through IR and start injecting liver lesions and everything else so the easy answer is yes but it does oppose new challenges so you need to have a patient that has injectable lesions and injectable could be sort of in the eye of the beholder we're normally talking about in melanoma someone who has dermal metastases either in transit or distant those are the ideal patients that in the clinic could be injected but in theory it can be done in interventional radiology and that certainly people are thinking about that at the moment I the some of the nuances though can vary whether you're talking about an oncolytic virus versus a compound or an antibody for example T vectors require special handling I require special preparation in the pharmacy so it's not like you could see non-small-cell lung cancer patient I was getting Pam Braille and the next person that comes in you're going to set up and treat them with T vac and then flop back and forth so there's a lot of logistical issues with that for the injectable lesions I would say the surgical oncologists have played a major role in administering that develop that treatment I certainly met onks are doing it but I think the surgeons have been more integrated and then obviously the IV antibody therapy we've dominated our clinic so I think that may be a feasibility issue in certain clinics certain patient populations so that might also play well in how we utilize the treatments okay thank you for coming out thank you the speaker thank you for listening to this activity to view the rest of the CME activity download materials and complete the post-test for instant please go online to wtiu press comm /m WM this activity has been jointly provided by medical learning institute incorporated and pvi peer-review institute for medical education this activity is supported by educational grants from Genentech insite Corporation and Merck & Company Incorporated

Channel: PeerView Oncology

Views: 8,308

Rating: 4.7468352 out of 5

Keywords: CME, Bladder Cancer, Clinical Pharmacology, Colorectal Cancer, Dermatology, Genitourinary Disorders, Geriatrics, Immunology, Lung Cancer, Melanoma, Oncology, Pathology, Pulmonary, Renal Cancer

Id: iR5X_bLDOX0

Length: 117min 9sec (7029 seconds)

Published: Wed Aug 16 2017