Clear Cell Ovarian Cancer

>> we're going to get going, and our first speaker is christine annunziata, from the georgetown medical school and came to nci to work and she's currently an investigator at nci in the medical oncology branch, her title ovarian cancer in the genomic era.

tina. >> use this mic for now? okay. thank you for that introduction. actually my branch has changed, and it is -- i can talk in this, i can see my slides. i'm actually in the women's malignanes branch, which i

moved to after being in the metabolism branch where i studied lymphoid malignancies but now i'm studying ovarian cancer. how do i advance this? like this? here we go. just some background on ovarian

cancer, as an overview and i'll talk more about more recent clinical trials with more targeted therapies. ovarian cancer is the most lethal gynecologic malignancy in the u.s., fifth most common cause of cancer deaths for women because 70% of cases are

diagnosed with advanced stage disease which is the reverse for breast cancer, much more common disease, but probably 70% of breast cancer or more is diagnosed at the early stage breast cancer, which is much more treatable. so ovarian cancer is much more

in the advance stage disease at the time of diagnosis, less than 30% of the advance stage patients will be alive at five years. the stage is defined by the federation for gynecologic oncology, stage i is confined to the ovaries.

as you can see that occurs in 20% of the situations. but those women have 90% chance of long-term survival. stage ii is confined to the pelvis, beyond the ovaries, an arbitrary distinction, it's not an anatomic distinction. as you know, the pelvis is

contiguous with the abdomen. to be confined to the pelvis is based on is it still below your hip bones or is it all over your abdomen. stage iii is beyond the pelvis, and stage iv is distant metastasis, long invasion into organs such as liver.

stage iii and iv, the overall survival, long-term survival, is much less. what are some prognostic features of ovarian cancer? the stage at diagnosis is what i just told you. the extent of cytoreduction, i'll talk about that in a couple

slides, that means how much tumor you can remove at the time of initial surgery, how much is gone, can you remove 100% of the visible tumor or is there a situation in which the doctor, the surgeon, cannot get everything, is it inside of an organ, right next to a vessel,

is it in the area that can't be resected. histology and grade, we'll talk about that again in a couple slides. so the grade is how differentiated the tumor is. histology i'll talk about. performance, how active is the

patient, how sick is the patient at the time of surgery. whether p53 is mutated has to do with histology. how well their organs are functioning, and how basically active they are, their physiologic age. the therapy is platinum and

taxane. interperitoneal is a novel approach to delivery, whether you get information from a second look surgery, whether the patient has the bcra mutation, and vegf production. the treatment for newly diagnosed ovarian cancer is

complete surgical staging, go into the abdomen, open, look where the cancer is, and whether it's involved in lymph node or distant metastasis, optimal reductive surgery and chemotherapy. these are lymph nodes, surface

of the liver, or other lining of the bowel or peritoneum. lymph node dissection, except age 1. in the abdomen, the ovaries are in the pelvis. the circulation of peritoneal fluid, not fluid in the vessels but circulating, the abdominal

structures, circulates in a clockwise direction. so in the abdomen, the first place that the tumor is going to go is sort of to the right and up, it's going to lodge on the outside of the liver. so typically, during ovarian cancer surgery, the surgeon will

take scrapings from the underside of the diaphragm, surgeon will take samples around the liver, and will remove various parts of the abdomen. removal of the ohmentum is typical, the next place the cancer cells will lodge. removal of lymph nodes, uterus

and cervix are tapeically taken out. peritoneal washings are done to give you a different, if the patient is given a stage a,b, or c within the i, ii, iii, iv. the removal of the omentut, that's what cancer cells sit.

stage iii and iv optimal is arbitrary, again, but residual disease is less than one centimeter. most of the time cancer surgeons are trying to get everything, that gives you a better prognosis if everything can come out, but if they can leave less

than 1 centimeter, that's optimal. if you look at survival, this is an old slide but if you look -- if the patient was left with greater than 2 centimeters, there's a much worse overall survive than if the patient was left with zero centimeters, no

microscopic, no macroscopic, excuse me, no macroscopic disease. this is the goal of surgery, to get out all of the disease. but even so, there can still be microscopic disease which is why we have to use the adjudant in some situations ovarian

cancer it can be given before or after of the time of surgery. classical chemotherapy regimen is platinum, either cisplatin or car bowplatin in combination with taxane. intraperitoneal administration is given if the patient has stage iii optimally reduced

ovarian cancer. just a quick background on the time line of treatment, in the 1960s, the five-year survival for advance disease was essentially zero. that was before we had effective in the 1970s, cisplatin was introduced with a 5% overall

long-term survival, the addition of chemotherapy is shown to be beneficial. in the '80s there was the addition of a combination with platinum chemotherapies but it wasn't until the 1990s that the taxanes are aimproved andim intruced, and that gave a

pretty dramatic increase in overall survival. still, only 35% but much better than zero, much better than single agent at 5%. the introduction of the intraperitoneal therapy did give up to 40%. the ovarian cancer tends to

spread within the peritoneal cavity and not through the lymph nodes, the lymph system, not through the hematologic system. this is why we see the effective intraperitoneal chemotherapy, the initial benefit is taking care of disease at the site of disease as opposed to

intravenous chemotherapy, taking care of the disease that would be metastasizing through the intravenous system. the intravenous chemotherapy does reach the peritoneum but when you give it directly you get a higher exposure in the peritoneal cavity.

this is why the spread does not go to the lungs or bone like breast cancer that spreads through blood and lodges in the bones and blood. ovarian spreads through the peritoneal cavity and the structures in the peritonenuum. gog is the cooperative group

that does multi-institutional phase iii trials. this trial tested standard intravenous chemotherapy, platinum, versus intraperitoneal with cisplatin. the reason for that, when you

give cisplatin intraperitoneal, it goes into the blood, circulates throughout the body as well as the peritoneum. the paclicaxel does not get absorbed into the circulation and it was thought we should not be, you know, robbing patients of standard therapy which with a

would be the intravenous paclitaxel. we did not want to not give them standard of care so this is standard of care, plus an additional intraperitoneal administration of the some people argue we're giving overall a higher dose to a total

patient because in the top versus the bottom, the bottom are getting an additional 60 milligrams perimeter squared of paclitaxel. another thing that's interesting, 42% of the people in the intraperitoneal arm were able to complete six cycles

because it's fairly toxic. people get a lot of side effects when you directly give chemotherapy into the abdomen. you can imagine there was a lot of abdominal pain, nausea, vomiting, diarrhea, side effects that made it impossible to give people, some people, all six

cycles. but even with that, only 42% getting the full six cycles, there was a substantial increase in overall survival and this is a slide, of course this is almost eight years old now, there's been updated data where this is sustained but there was

a dramatic increase in overall survival from 50 months to more than 65 months of overall survival when the intraperitoneal therapy was given. i'm going to switch gears and get more into the biology of in those trials we're grouping

all ovarian cancers but there's lots of different types of even at that stage, we knew there was different histologies, so what it looked like under the microscope, 80% of the time looks like serous, 10% from the end oh meet rum, 5% like a

kidney cancer, and 3% like muvinous. you can imagine these have different types of biologies. what we don't know is even where are these cancers coming from. is the ovary -- we would look in the abdomen, there's a mass on the ovary, we call it ovarian

cancer but is it really coming from the ovary? what is really the tissue of origin? so nowadays we're thinking that based on molecular studies, perhaps serous cancers are from fallopian tube. so of course we think that

increasing our understanding about the biology and biochemical events underlying the ovarian cancer will create avenues for new treatments. should we treat all ovarian cancers the same or can we find more of the targeted. we all know about the atlas

research network which published their study of ovarian cancer in 2011. cancer genome atlas, we looked at clinical annotated high grade serous, and potentially gave us new therapeutic targets. what they did was they did

microarray analyses on 489 of these, looking at messenger rna, microrna, dna copy number, dna methylation, and 316 of them did whole exome dna sequencing. they included only newly diagnosed patients, a key point, can we extrapolate this to the

recurrent setting which we tried but doesn't always necessarily apply. only serous adenoid carcinoma, any stage or grade but predominantly stage iii cancers. each frozen tumor specimen had a companion normal tumor specimen which could be either adjacent

normal tissue or peripheral blood lymphocytes with previously exacted germline dna. the clinical data collection, from cancer.gov. this is the work flow of what they did, from the publication. interestingly, what they found was that the most frequent

mutated gene was p 53 in these serous cancers, that defines them looking back at this. the next most common gene was brca 1. you can see the vast difference in the number of cases, these are not mutually exclusive. only i would say less than 10%

are the brca mutation. after that it falls off a lot, quite quickly after that. there's a lot less frequent mutations that were identified. from a copy number standpoint looking at the chromosomal copy numbers, it looks compared to gbm

which is a brain tumor, compared to gbm which has distinct copy number gains or losses, high grade serous ovarian cancer has diffuse genomic abnormalities in the genome. red is amplified, blue is deleted region, and the chromosomes are listed one

through 22, from top to bottom. you can see throughout the genome there's is a vast number of dna copy number abnormalities, however what we're finding now is that the 3q gain and 5 -- where am i looking here? the 5q loss seem to be a

recurring feature of the high grade serous. there's more on that that's being talked about right now. not a lot of definitive targets but it's an area of study. from a gene expression standpoint they were able to identify four different gene

expression subgroups, using all 489 tumors and grouping them by 1000 genes. and they found there was a differentiated and immunoreactive mesenchymal and proliferative signature that define these subgroups, which were similar to subgroups

identified in this reference 25 data set published in 2008. what are altered pathways? looking at the dna copy number, the mutations and the gene expression signatures, what they noted was that in 67% of cases, the rb signaling pathway was altered, in 45% of cases the pik

was altered, large amounts of data, you come up with pathways and not necessarily individual genes. and this will come into play with therapeutic targets down the road. not signaling pathway was another recurring theme in the

cancer genome atlas. and homologous recombination. this is related to brca and methylation. another pathway that we kind of are integrated here, dna repair with multiple mutations or aberrations found in all of

these, this is kind of trying to tie in the -- you see that p53 was mutated in almost all of the cases. so just to summarize, the cancer genome atlas was a large scale integrative view of aberrations and high grade serous subtype of ovarian cancer, and what they

found was that mutational spectrum was what they call surprisingly simple. p53 was mutated in 96% of cases, brca 1 and 2 when combined was 22%, but the others were infrequent, so 2 to 6% of cases had other significant mutated from other literature we know

this high grade serous is distinct from other histologic subtypes, there were recurrent mutations in arid. there was frequent mutations, in mucinous there was prevalent kras. in the future we won't treat

them all the same but that has yet to come. some other quotes from the paper, there was a remarkable degree of genomic disarray as i mentioned with the dna copy number, strikingly in contrast to the glio blastoma. mutations might be related to

copy number aberrations. where does that believe us? do we have new therapeutic approaches? 50% are homologous defects. commonly disregulated pathways might provide pathways, still in the clinical trial setting. inhibitors for 22 genes and

regions of recurrent amplification, but then the other issue that was introduced here, can we target the network rather than the genes themselves, in order to be an effective treatment do we need to target multiple regions of these networks, can we target a

central core of these networks, these are all questions that are raised by these data. so that bring us to the next sort of precept of treating newly diagnosed ovarian cancer. clinical trials and how can we incorporate targeted therapies? i mentioned parp inhibitors,

used in brca mutant cancers, still in the experimental stage. normal processes of the cellular lead to single strand breaks. parp is critical to the repair of single strand dna breaks. in replicating cells, if parp needs to be there in order for this single strand break to

be repaired. if not repaired single strand gets converted to a double stand break. in the normal cell, the normal cell can repair the double strand break by a pathway called homologous recombination. in brca deficient tumors

homologous recombination is knocked out and so there's no repair of the double strand dna break, that leads to catastrophic consequences and the cells die. olaparib is novel. we did a phase 1 b stud whichy

to push along the dna damaging, we used a car bowplatin to induce the damage in the brca-mutated cancers and used olaparib. our result, we enrolled 45 patients, 37 had ovarian cancer, 8 had breast cancer, the brca

gene gives you disposition. we had a phase 1 expansion cohort with 15 patients at maximum tolerated dose to look at the patient's cells and in their tumors, in their peripheral blood and tumors. maximum dose was 5. we gave that on day one and gave

olaparib 400 milligrams twice daily on days one through seven repeated every 21 days. and we did find the remarkable response rate in all these patients. these are the color coding by the dose level,ened you can see we had an expansion cohort at

the dose level six, maximum tolerated dose, more red patients than the other colors. even at the intermediate there was a reduction in disease. we did have one patient with a complete response, 100% of her disease went away.

these are patients that had been -- all had been previously treated with carboplatin, some were refractory, meaning they had disease growing while on single agent carboplatin. parp gave most a response. to summarize, one patient had a complete response, and has been

on a study for 23 months, most recently published this past june in jnci. most patients had a partial response, or stable disease, and six patients had progression of the overall response rate was 87% in the breast cancer and 44% in the ovarian cancer, if you

include response rate and stable disease, stopping it from growing, we had a clinical benefit rate of 82% in ovarian cancer, 100% in breast cancer. oral olaparib was well toll traited. greater activity was seen at the higher doses but it was positive

proof of the concept of activity and tolerablebility with o, whetheraparib in the brca-deficient cancer. move the to the gene expression how can we use the gene expression subtypes, the network pathway approach to looking for targeted therapies for subtypes

of serous? we looked at immunoreactive subtype, defined by a set of 200 genes here in this green subtype. and when you look at those, take those 200 genes and look at what are the networks that are involved in this, in this set of

genes, the nf-kappab comes out as one of the most predominant networks influencing these -- that's putting together, bringing together, these set of 200 genes. so what is the nf-kappab pathway? nf-kappab you probably all are

familiar with. this is a very brief schematic overview, nf-kappab can be triggered by cytokines such as tmf signaling through the tmf receptor activates a receptor, to then trigger phosphorylation of ikke, kinase beta, a kinase that's involved in -- that's

part of the a heterotrimeric complex. beta is the predominant that phosphorylates, that is a protein that binds the transcription factor, the nf-kappab transcription factors themselves. the need to translate to the

nucleus to activate genes of interest. however, in the inactive state, they are held in the cytoplasm by inhibitor of kappa b. the degradation releases factors to the nucleus, bind and provide transcription of nf-kappab target genes, and they are

involved in survival and proliferation of cancer cells and of normal cells as well. so we first started by looking, are any of these transcription factors expressed in ovarian cancer, was the relationship to survival? one study, we found that

nf-kappab p-50 transcription factor was associated with worse overall survival, in a group of patients we had long term ten-year follow-up on. and we looked at the expression of the ikks and asked whether any of the ikks were associated with survival.

overexpression of ikk beta was associated with worse overall survival in a different data set. so we wanted to look at what is nf-kappab, how do we study nf-kappab in ovarian cancer? we moved to a pre-clinical model to look at ikke signaling,

downstream in ovarian cancer in a plea clinical cell line model and treated cells with ikke inhibitor to define pathway in and we used an ikk beta and sh-rna and came up with a gene signature that defined nf-kappab is ovarian cancer. these are all known nf-kappab

target genes, but they seem to be specifically co-regulated in what i mean by that is shown on the next slide, in large datasets of gene expression, from ovarian cancer patients, this is one but we've done it in three sets now, if you look at the co-regulation of these

genes, the genes are very tightly coordinated with each other. if you look at genes that are other nf-kappab targeting, for example ones that are in multiple myeloma, that we looked at in my previous life, when i studied multiple myeloma, the

genes are disregulated and don't fit together, they don't define a signature. they define a signature in multiple myeloma but not a sent of patients in ovarian cancer because they are not regulated, not so regulated. these are specifically

co-regulated in ovarian cancer. so we can now use these as a tool to study ovarian cancer datasets and find out what is going on with nf-kappab in what we found in these datasets was that there was a worse survival when the signature was expressed at a level higher than

the median compared to cases where the signature was lower than the median. it might be important. we did a body of work in cell lines and looked at the proliferation, survival, inflammation, adhesion and invasion and angiogenesis and

found nf-kappab by blocking ikke beta was important in cell lines, in all these situations and that was related to the known functions of the nf-kappab target genes. if we take a closer look on proliferation, it doesn't look like -- it did define a subset

of cancers that were sensitive but it was a slow death, not a dramatic death with a broad effect on the functions but was not promoting active cell death like apoptosis. we looked in a different part of the pathway,backing up, what is apoptosis, there's the intrinsic

and extrinsic pathway. what i'll point to here is the molecule smac, it acts in the intrinsic and extrinsic pathway to promote the apoptotic cascade finally end of in cleavage of caspase 3. the importance of smac, this function is to inhibit the

inhibitors of apoptosis, or the iaps, and if we take that back to the nf-kappab pathway, smac is there, let me just take you back to the nf-kappab pathway. you note that iap is one of the molecules that i pointed out in an earlier slide that's important at the level of nnt

receptor signaling to promote activation of ikk beta. we're looking at a smac mimetic, it down regulates iaps. we thought we could down regulate the nf-kappab signaling by downregulating iap and then promote apoptosis and

another activated cell death pathway called necroptosis by this pathway. the smac blocks the iap which then releases the block on the rip kinase down extreme of the tnf receptor, to activate apoptosis. rip 1 with activate extrinsic or

intrinsic pathway, activating casp 8 and apoptosis. rip 1 can turn off nf-kappab. rip 1 can also activate another pathway, necroptosis. cancer cells typically have apoptotic pathway, another pathway promoting death via rip 1 to activate necroptosis.

we use the smam mimetic and found that there was increased cell death in ovarian cancer cells when we treated with all three together, still in vitro. we inhibited that cell death with incontributor of apoptosis, shown in the gray bars, that's inhibiting specifically casp 8,

and inhibited with nec 1 which blocks the necroptotic pathway, there was still some rescue of the cells there. we would hypothesize and state here that the treatment with tnf ikk beta inhibitor and smac mimetic were killing, this might be a good way to treat ovarian

cancer patients. so we of course -- there's much more pre-clinical and animal models. i'm skipping a lot of that. eventually we took that to phase ii using the smac mimetic in platinum-resistant patients, and this is noted by light blue

arrows, on a weekly basis for three weeks out of four. we did core needle biopsies of tumors before starting treatment and after six doses of the smac mimetic and cat scans every two cycle. we did multiple correlative studies using the frozen tumor,

fixed tumor and plasma, as well as peripheral blood mono nuclear cells, and a whole blood for what's going on with the immune cells. i'm just going to previously summarize who findings that we had here. we found that the -- in the

tumor biopsies and pbmms, the smac hit its target, down regulating in the tumor biopsies and peripheral blood mono nuclear cells at the time and after six doses of the drug. that was good. the drug actually hit its target.

what happened to the nf-kappab in the peripheral blood mono nuclear cells, there was good attend downregulation of-and-a-half signaling. the way we measured was phosphorylation and ratio of p65 in these samples. but what we didn't see was

downregulation of nf-kappab in the tumor biopsies. we're investigating what is the reason for that. why when we downregulated iap why did p65 not go down. there are multiple reasons that might happen. and we're looking at that right

now. in the lab. but we didn't see the downregulation, nor did we see actually tumor response. these patients were treated with the smac mimetic for at least two cycles, some for six, but we never saw tumor shrinkage, we

did not see the apoptosis we were looking for that we had seen in the cell culture, we did not see that in the patient. ongoing studies will investigate that. there are multiple smac mimetickings mimetics in the clinic now with similar

responses. there's not good clinical activity of the smac mimetic. there's stable disease or most of the clinical trials that have gone on right now there were complete responses in two patients in one study, going to a phase ii.

we'll see what's going on there. a lot of promise in the lab, disappointing clinical activity. i think that's the main summary of this smac mimetic story. what we're doing in the lab is trying to take this a step further, maybe we can't induce apoptosis when treating with

single agent, is that a pharmacological reason or biological reason in the tumors? we're now looking for combinations that we can do, with the smac mimetic to enhance this activity. here we did a screen with national center for the

advancement of translational sciences, and what they did first was to generate a 6 x 6 matrix, what we're doing on the y axis. we have the smac mimetic and on x we have the experimental drug with a library of 476 compounds we tested in the 6 x 6 matrix.

we then chose the most synergistic combinations and expands to 10 x 10 to confirm, with a combination of less than 1, beta parameter of less than 1.5 in the model. now coming back to the pathways that are disrupted, these agents fell into multiple -- into

categories of drugs. the ones we found that were synergic were categories. we had inhibitors, various miscellaneous drugs. what dear doing here, take this back to the clinic, we're going to focus on docetaxel. this confirmed the synergistic

activity of the combination. we do plan on going to a randomiseds phase ii clinical trial of docetaxel versus with smac mimetic. we'll see if we can categorize women by the level of nf-kappab sag signaling present before targeting and what

happens to the signal afterwards. those are two scenarios that i can present. there are many other vast clinical trials ongoing in but i just wanted to highlight some research that we're doing here, in a targeted manner in a

subtype specific manner for with that, i'd like to just highlight the women's cancer team, and translational sciences who did much of this work in the laboratory. i'll be happy to take any questions. [applause]

yes? [off mic] >> no, these are all questions. we think we have a way, that's why i wanted to show you like a successful trial and a not successful trial. we think we know what's going on in the lab, but it's much more

complex in vivo. whether that patient simply had like upregulation of mdr gene and pumped out the drug and the drug never got to the tumor, that's one reason. there's many reasons why this could happen. they are all fascinating to

study on an individual basis. that's what we're moving towards. can we do whole genome sequencing? can we profile individual patients that are sort of the outliers to find out more about who is responding and who is not

responding? >> that's an excellent question. are they deletions, different mutations, is that what you're asking? we haven't looked in detail in that, but that's an excellent point because all mutations are not the same.

some of them, some mutations completely knock out the function of p53. some mutations actually give p53 a different function. so are the mutations responsible for this aberrant cycle, dna repair mechanism for the genomic instability?

i think that is an area of research. we're not actually doing that directly in the mci, but that's ongoing research outside. people are looking at those questions. that's an excellent point.

yes? >> yeah. so, yeah, a lot of tumors have p53 mutations. this is another i think huge benefit of the cancer genome we have all this information. i don't know if you saw the speaker on friday for grand

rounds. we have i think like 17 cancers that have been profiled now, by all of these methods. so now we can try to pick up what are the similarities and differences between the cancers, and what that investigator found and others have found is that

the serous ovarian cancers look very similar to triple negative basal-like breast cancers. and the 3q and 5q implication was consistent across both basal breast cancers and serous ovarian cancers. yes, they do occur throughout those of cancers.

kras mutations seem to mimic kras mutation in colon cancers, and clear cell in kidney cancers as well. there are recurring themes. that being said, i mean, there's still something to tissue of origin, because we know like in the kras mutations, for example,

the ovarian cancers with the kras mutations don't really respond to the same chemotherapy regimens and colon cancers. we have to keep that in mind moving forward. yes, all right. thank you.

>> okay. we have one announcement, last week i spent you an e-mail about the course and tumor board, you should respond by the end of the week, if you're interested in that activity. and then we'll be visiting the course and tumor boards between

columbus day and thanksgiving day. so our next speaker is melinda merchant, she got her md, ph.d. it the miami school of medicine and did residency training at the children's national medical center in pediatrics, joined the pediatric

oncology branch in 1999, and she went to memorial sloan-kettering, and then she returned to the pediatric oncology branch, she's now a staff clinician. her title, immune checkpoint inhibitors, overviews of immunotherapy in the translation

for pediatric solid tumors. melinda? last year in science magazine, immune checkpoint inhibitors was listed as the biggest advance in cancer treatment. >> thank you. good evening. i was asked to talk specifically

about immune checkpoint inhibitors and thought i would do a couple things. is broach immune checkpoints inhibitors and teach you about the angle i'm coming at. many of you know immune checkpoint inhibitors had some exciting results in melanoma but you may not know pediatric

melanoma exists and other tumors in pediatrics can be targets for the immune system. so we're going to go through those today. science magazine named cancer immunotherapy the breakthrough of the year last year, as dr. moody mentioned.

it divided it into groups, which will we'll touch base on, imi'm stimulation, before we knew what the basic science was learned at branches and laboratories here and elsewhere, we had nonspecific activation, as we've drawn more complex we're using cytokines and antibodies and

complement figuressation. adopt a cell transfer is well studied here and the nci branch with tumor infiltrating lymphocytes and exciting news for tcr therapies against solid tumors, an ongoing trial with nye trials and car t-cells against solid tumors in surgery,

etib, leukemia, and pediatric oncology branch. and autologous. cancer tests, i mentioned nye cells, as one of those in the and the fusion proteins, many tumors have a transfusion

protein which essentially takes two genes, puts them together and you have a unique target within the cell that's not in any of the normal tumor cells. but we're still trying to understand how best to target things like that and the immune checkpoint blockade.

it was mentioned i spent time at nci and went to memorial and came back. there's a lot of memorial sloan-kettering and nci history in immunotherapy. and one of the ones sometimes referred to as the father of immunotherapy is william coley,

a bone sarcoma surgeon at memorial hospital and new york cancer hospital and later became memorial sloan-kettering. he took chiropractor of daisy dashiell. this was a teenage friend of jd rockefeller jr. and put the impetus for philanthropic giving

to cancer and the sciences and the basis of rockefeller university doing clinical trials work. i really want to highlight here the importance of philanthropy we we work within pediatric he was a 19th century surgeon, so he had 19th century tools

at his hand. he noticed that sometimes patients after he did surgery on them for their bone tumors had a better overall survival when they got an infection from surgery. and last week i opened up my twitter feed and saw atwitter

feed about the nci exceptional responder initiative. this is the 21st century of the twitterverse, in william coley's day his version was going into lower manhattan in the tenements, searching for a patient that had a sarcoma, an infection with strep afterwards,

and supposedly left the hospital without any evidence of tumor. he wanted to make sure that indeed was true. and so he searched for a long time, found this mr. stein, a german immigrant living in lower manhattan, and here is the picture of him, seven years

after that event where he was still free of tumor. and so i like this quote in one of the books about -- oops, i spelled cooley there. nature give us hints of her profoundest secrets. i encourage you as you're

looking, part of the rationale behind a course like this is because sometimes those little points, which may not have seemed so big can be the foundation of scientific path that could take centuries to come to fruition, but don't despair, graduate students and

postdocs. it can also lead to exciting things. keep your observational caps on and i think this idea of exceptional responders was also talked about, dr. annunziata, these patients, why are they responding well in clinical

trials? mr. stein was one of those. i have here in your handout just a pause to say some of the things that much this is coming from. you'll see this isn't my typical science talk or very detailed about all of our clinical

trials. this is really more an overview. it started with history, and the commotion in the blood is a great book by stephen hall that really gives a good overview, background, to immunotherapy. it will have things about the young dr. steve rosenberg,

things about coley's toxins, and goes through some of those things, it's a nice narrative and actually balances well as the emperor of all maladies as well, talking about some of the history of immunotherapy. there's a couple of recent reviews that are very

interesting, and sort of briefer to get a broad overview, drew pardol from hopkins and judd memorial sloan-kettering have recent reviews, as well as some reviews coming out of the rosenberg group, and gallewski was here for grand rounds, and gave a great overview worth

looking at as well. i'm going to take a pause and walk you into a little bit of the pediatric world. so adolescent and young adult melanoma is actually the second most common adolescent and young adult tumor. you see from a monograph listing

the tumors in patients 15-29, i highlighted melanoma in 11%, soft tissue sarcoma and bone tumors make up 11%. that's another forte of our branch, specialty of mine. melanoma, one age group where it's increasing in incidence and it actually is not very well

studied because it's pretty rare. there are low rates of trial enrollment in adolescents and young adults, more so for young adults, adolescents tend to get captured on children's ongoing group studies or pediatric oncology studies, young adults

not so often. there are risk factors in teenagers and young adults, some sun exposure issues that may play a role but there's genetic predisposition, one of the biggest is a large congenital birth mark that covers huge

parts of the skin, sometimes referred to as swimming trunk because they look like swimming trunks on a patient. patients are more likely to have a melanoma develop. there are familial, dysplastic, atypical nevi, you can see the link between skin and melanoma.

immunosuppression is also a setup, and so is a pediatric cancer survivor, probably for a lot of what we put them through to get rid of their first cancer, they are now more at risk of having a melanoma. there are lower numbers of patients in this aya group than

in the adult population. i'll show you a couple graphs about that. and then there's the referral pattern about whether they actually end up getting the pediatric oncologist or dermatology or adult oncology referral.

and there's a lot of psychosocial challenges to the age group, just on their way to independence and get this total road block of cancer, a rare and aggressive one at that. here is the list, a graph showing the incidence of malignant melanoma versus other

cancers, we tend to spend most of our time treating patients down in the very low numbers of melanoma, and low numbers of cancer overall when you're looking at the numbers of patients like ovarian cancer, lung cancer, we're going to talk about much smaller trials than

you just heard about in the talk right beforehand. overall there's about three or four hundred cases of pediatric melanoma in a year, 10% will be metastatic, making it an incidence of about 1 per million. this is 1.5 per 10,000 for

overall cancer in the pediatric age group, 4 in 100,000 for all, representing 1.3% of total melanomas but a decent number in the teenage years. when i came back about five or six years ago now, to the nci, i came back as a solid tumor sarcoma-focused immunotherapy

and molecular targeted translational clinician scientist. that's a whole bunch of adjectives but it made me a melanoma doctor too because our interest in immunotherapy left these patients an option to get treated with things that first

were not approved, in the clinical trial setting, and we actually became a referral pattern for melanoma patients and have seen more metastatic melanoma patients in the past five years than in other places. you'll see to get the sort of numbers that have this, it was

over a span of decades at other large referral centers. there's not a national registry but we're in the midst of writing a natural history trial that will help us get some more of the genomic details that we're starting to get on our current study.

so of interest, the majority of our patients that are shown up have had head and neck or scalp -- this was in a nevus on the head, this patient's melanoma developed in. and that's a way higher merge of head and neck cancer than you see in the general melanoma

population. most of these patients did not have sun exposure, to those areas. and i honestly don't know yet whether it's biology or the fact that it gets covered by hair a lot and takes longer and probably has deeper levels than

at a higher stage before it gets first found in the question of why do we have more of these in our metastatic population than if you look at overall at the entire database. they are in the typical places melanoma goes in the adult, which is to say anywhere.

and they can be very big and palpable as in this young girl. and can grow to be very disfiguring but can also go to multiple places in the body. the genetics have been a very big story too because the other thing that's been approved in the last couple years for

melanoma is braf inhibitors, and i'm looking at the amount of patients with mutations, it's about similar in our older populations to what you see in the adults which is a decent number of braf v600 mutations, it's less here, nras tend to be in younger patients more

associated with nevi. and ckit or pten we have not identified in our patients and we have identified other mutations through work with mari nohe and javet khan, her mentor, identified that helped us treat the patients. i mentioned sarcoma, i want to

stick with it but won't spend as much time on history but it's heterogeneous group. two key sarcomas are osteo sarcoma and ewing sarcoma. james ewing was a pathologist at memorial hospital, when coleie was there at the same time. and bessie dashiell likely had

ewing sarcoma. there's ration at why we think it can be targets and i want to capture a couple. i couldn't go as far back as william coley but this is one the first papers i published as a fellow that set the stage so osteo sarcoma can be a target of

t-cell toxicity. they grow a primary tumor just at the same rate, there's no difference. they look the same when they come out. we do an orthotopic tumor implant, put it down in the gastra, you'll see this guy in

the portrait, you can see it on the leg here. we amputate that leg and then wait and the majority of patients will -- majority of mice will end up developing lung metastatsis over the next couple months but if you give them t-cells, after the surgery or

before the surgery, t-cells that are able to make interferon gamma, those mice will survive and not get metastasis. we also did a study in the branch which has recently come to about a five-year survival rate that we can look at. and that's using high risk

sarcoma as target for immunotherapy that included dendritic cell base vaccine, when the patient was first diagnosed they came and we got a cell harvest that got it lymphocytes and tumor biopsy to make lysate. they got standard chemotherapy

for high risk sarcoma, months of high-dose chemotherapy, five different drugs, surgery and/or radiation, so you really do a lot but in these pediatric sarcomas we can get the patients to what we call no evidence of disease, actually in the vast majority of the patients.

unfortunately in metastatic patients have less than 20% survival rate, they are harder to treat once they recur. we brought them back with no evidence of disease on the scan and gave them an immunotherapy which had their monocytes grown into dendrites, we watched.

and this is an example of one of those patients, ewing sarcoma can end up very widely metastatic, much like melanoma. this is a pet scan showing a lot of bony disease, she had kidney, arm, she had a lot of disease, went through the standard therapy, immunotherapy, she was

one of the first patients that didn't get the il-7 and actually remains ned to this stayed, six years later. in fact i had some very exciting things happen in her family life this year too. we're actually very excited this is the first time we're sorting

moving some of these poor survival rates, this is a kaplan meyer curve, this was a 1986 trial of metastatic high risk ewing sarcoma patients. here is one where we used that fusion protein as a peptide vaccine, and we didn't move that survival curve much with that

peptide vaccine, but when you change it into a lysate vaccine on a with cytokine and dendritic cells, we have a majority of patients surviving on that study at this point. these are really to set the stage for the fact that we think sarcomas are a very good immune

target, melanomas are a very good immune target. now, what can we target in that immune tumor interface? think about it as a dynamic interface, it's not just a single interaction. alex wang, who had been a fellow here before and is at case

western as beautiful pictures of t-cells going in and out of lymph nodes, bopping around sampling, looking, very dynamic. and the bone tumor itself, bones are not stagnant. they are living and growing, so all these things and forces are going on and can we take

advantage of the things that are making t-cells get into a tumor better? or can we knock down things that are keeping a t-cell from getting in there? our sarcomas, we've got a lot of myeloid cells suppressing the environment.

tumors are secreting things that tell the t-cells to turn off. melanoma tends to get t-cells in there and you can grow it but can you keep them activated and fighting? with sarcoma, how can we get more t-cells in there? really, can we help bolster this

part of the immune system to eliminate the tumor, which flipping on its head, you can ask a different way, what keeps immune surveillance from happening? i mention the myeloid yo cells, tregs. cancer cells are not as strong

as viral antiagainst. they don't put so much mhc out so they can not engage as well for solid tumors like sarcomas, there's the physical barrier, and there's work done on how much can get in due to the onco-pressures essentially. there are definite problems

along the way, but one of the things with tumor cells has been able to do is to exploit negative checkpoints. some tumors from pdl-1, we'll get to that, but that's the ligand that helps put the brakes on the immune system. there are ways tumors themselves

are putting a brake on the immune system, can we block that, so we can now have the immune system attack the tumor. that's what we're going to focus on here with the immune checkpoints. there's a whole bunch of other areas of the immune interaction

between host and tumor, that i think are very useful and we're going to see things come out in many of these areas, but for right now we're going to focus on some of these ptla 4, and pd 1 and pdl 1 are classics. let's talk about t-cell activation.

i'm a pediatrician and tend to put things in concrete ideas so that my patients can understand as i talk to them about why we're doing a certain thing. so i figured i might have some immunologists in the room. don't be offended that i'm taking it back to very basic

steps. hopefully you'll learn something about pediatric oncology. a t-cell takes ignition, signal one where the tcr binds, mhc has a peptide in the group. this is putting the keys in the ignition. signal 2 is the gas pedal, that

cd29 binding b7 here and sending a second signal which makes il2, t-cell activation and il2 and ifn gamma. if you have a t-cell activated all the time the body says maybe that's not so good. actually after 24 hours of chronic antigen stimulation it

sends ctla4, that binds to b7 with higher affinity than the normal positive second signal, to send a negative signal and that inhibits the activation of the t-cell and it essentially blocks further activation of it. ipilimumab, or a version of this, ipilimumab is dmf,

monoclonal antibody that blocks ctla4 and inhibits the break and allows the activation to go on. so notice what we're doing in blocking an inhibitor, we're not turning the t-cell on. it had to be turned on already in its stage. we're not helping the t-cell get

to the tumor any better. it had to be able to get there. but basically we're blocking a break that's been put on for the body's own homeostasis, let this continue attacking this antigen because, yes, it's sticking around but it's important that it keep attacking.

this is from jim allison's first report. he gave colon cancer to mice and implanted it back here. it starts being palpable and all of them grow. this is something to get back to as we talk about the clinical trials themselves.

essentially, we've got some growth but the ones treated here with three doses of antictla4 have a regression of palpable tumor and it goes away. they don't when you block the other side of the co-stim. this is one of the first clinical trials done with what

became ipilimumab. a surgery branch tile, and essentially you see here a large melanoma metastasis, here smaller ones, here a brain met, and the second panel all going away, with several complete resolutions an partial resolutions reported in this

trial of melanoma tumors. so it was tested here at nih by dr. yang as well in that group on renal cell carcinoma. here is the pivotal -- oops -- the pivotal phase iii trial that led to approval using ipilimumab. it look at ipi versus ipi plus

the vaccine, and we'll get into some of these other ones later. essentially had stable disease, overall responders, and a prolonged survival rate. these are really -- instead of pulling out things from the early reports, which tended to be in new england journal and

tended to make very big splashes at apsco the year they were presented pulled the two recent year -- well, i guess this is two years old. and this one is within the past year. long-term follow-up of patients on the trials.

so here was the surgery branch trial, and you can see that essentially there's a plateau of patients who are long-term survivors to getting the initial upfront doses that blocked that checkpoint inhibitor. this did not require years of treatment, simply released the

brake at the beginning of treatment and then these are patients who have had ongoing clinical responses. in the surgery branch trial they did it with il2 which ended up being their best survival arm, versus with gp100 -- i can't read that so well here.

versus the ones with the vaccine, yeah, whereas the ipi trial, the phase iii that judd published, has an ipi plus 100 in red, ipi alone was the better responder, the vaccine alone was the poorer outcome. what you see as takehome, you can tell from the drawing on the

figure here, which is that really in the first three months they don't separate. if you're testing this drug and asking for typically what we would ask for like on the parp inhibitor trial that dr. annunziata showed you, we wanted to respond.

stable disease isn't the take-home they were hoping for. we want it to shrink. if you closed down the studies based on these not move ingapart would have missed this. if you look at patients which responded but grew and took them out of the nonresponder group,

here is a patient who is responding. who is not responding, sorry. and they basically are starting to have disease and progressing, and they follow what tends to be a more typical curve for melanoma, unfortunately most of them look about like this in

their survival curve. if you look at patients with response, they are much better. they go out on a plateau, and there is some dropping out and there's a lot of censored ones here. i take this with a grain of salt.

there's a lot of patients at the two to three-year mark. if you look at patients who initially progressed, but when you looked at them in a different sense and said i'll let you progress because if immune cells come in to the tumor and grow some that might

be a good thing, maybe i don't want to cut off and go to chemotherapy that will abrogate, maybe we can watch. sure if you have if you watch those they look like the patients who have had responses. now for clinical trials for immune checkpoint inhibitors and

other therapies, immune-related response criteria that takes that into account and let it grow on the study but not take a patient off trial. so here is a summary of the experience in the adults, essentially as we started the pediatric trial.

it improved survival in metastatic melanoma, had a dose effect, .3 they didn't have responses, at 3 milligrams per kilogram, they had 4% response, but at 10-milligrams per kilogram they had 11%. now what i don't have on here i guess is the immune related

adverse events also tend to go up with that too. what do i mean by immune related adverse events? when you block the checkpoint, i've said nothing in any of that about blocking it specifically for the tumor. if we had a way for me to block

it specifically for tumor, that would be great. but we don't at this point in time. maybe some by specific, maybe some good things coming out of the basic science labs that will help us discriminate now but now it's pulling the brake off

anything. you're likely to release what will look like autoimmunity. some of those looked commonplaces, colitis, diarrhea because of inflammation of t-cells in the gut, very much like graft-versus-host disease,

issues to the liver. rash is come in in adults, it's not so common in kids i'll show you soon. and endocrineopathy why, antibodies are something we look at and a decent number will make antibodies that create some problem with their thyroid

system. they take a synthroid replace it, it works out okay. more complex one is something is hypofacitis, knocking off the endocrine talking to adrenals, and can lead to lifelong need for replacement in hormones. breaking tolerance to self means

you're getting the mechanism to and across a couple of studies, if you had a better -- if you had an adverse event related to an immune activation you were more likely to have an antitumor effect. if unblocked tolerance to self, you're more likely to have

unblocked tolerance to cancer. which after all is pretty darn close to self. it is self, just a little altered. this was our study, treating 1-year-old to 21-year-old patients with essentially every three-week ipilimumab for

induction and went to every three months for maintenance therapy. we started at a lower dose level, not quite as low as the adult studies had shown since they had no responses there but at 1 milligram, 3 and 5 and 10-milligram, and this take

essentially shows you how we had increasing immune-related toxicities as we went up. in bold are the dose limiting and made us back up on certain things, but you'll see pretty much all the things that we talked about show up here. there wasn't any new safety

signal, there wasn't anything different happening in the kids but honestly before we embarked on this we had never used any of these agents in children and didn't know what the pediatric immune system would do in response to the immune we didn't have any home runs.

we had five patients with stable disease, i think it's six actually. greater than four months. one out of twelve melanoma patients was in that stable i mentioned the toxicity trial is similar, and this was the foundation for a promised phase

ii study in adolescent melanoma ongoing now. this is one patient with a bulky scalp melanoma, you'll see here at baseline, in a couple different sequences. and essentially after 15 months of this ipilimumab, it decreased in size.

she developed an auto-immune thyroiditis, lymphocyte count went up, treating with synthroid, the disease came down, we could tell after a while her antibody titers were coming down about the time we started seeing the tumor growing again.

we gave her intense again and stabilized out the disease, and saw increase in antibodies. she ultimately progressed, but was our longest responder. here the rates of side effects, grade through and four being the worst. a third of the patients had a

grade three or four, this is consistent with what was was seen in adult studies. again, this one was not. maybe 70% of the adult patients had a rash of some sort. and we were puzzling over this until a dermatologist pointed out that what we all pretty much

know, that adults have a lot more skin damage, uv damage in their skin, that's kids. perhaps it's that target of the mutations that happen, and dr. rosenberg came out with a paper in the past year identifying that, you could see each one of those induced changes to be a

target for a t-cell. and those can be even in nonmalignant tissue in your skin, causing a rash. whereas the kids, they don't have that so they didn't really have the t-cells going to the skin. also a hypothesis.

we didn't biopsy anybody but it makes essential. the conclusion, it was tolerated similar to adult studies, with less clinical activity than we hoped fore, one long-term sd in all the melanomas we treated. when we looked at biomarkers there wasn't clear-cut things.

you like to say who responded? how is that different from the nonresponder? we're hopefully we'll have responders on the trial upcoming. questions still to be answered, are there surrogate markers for response?

so far, nobody's been able to identify the patient up front who is definitely going to respond but i take you back to the ctr grands that he has the t-cell inflamed phenotype that's a marker of somebody more likely to respond to immunotherapy. here one of judd's papers will a

patient, you can see that he had a real increase in absolute lymphocytes. here at baseline. here after his first couple of doses, first four doses, he had a lot more tumor on his scan. dr. wolchuck said i feel better, they took the unorthodox stance

of just watching, and that patient underwent without any drug spontaneous resolution of all those. the other takehome message is listen to your patients. patient said, but i'm really feeling well, not quite sure why the scans look worse.

and turns out because there was an ongoing immune response in there that was decreasing what was happening. that doesn't hold out true for everybody that responds in the lymphocyte correlates. in the pedes trial we did answer that, you can activate t-cells

in a pediatric population. but you do not get a concomitant increase in t regulatory cells. i mentioned the t-cell tumor microenvironment, we were looking at crp, a marker of inflammation, in the adjuvant trial patients had a higher inflamed crp were less likely to

relapse. so that really was the first in-class drugof its kind, the first one you notice i had very different side effects than we talk about in chemotherapies. even in pediatrics we started it only here, we opened it at memorial, and dana farber, and

now the phase ii is out in multiple places but they are very different side effect profiles, and needs to be watched closely especially at onset of things like colitis s o you don't perf-ing your colon.

what put it with and make it better, take a 6 to 10% and make it higher? dtic is effective in melanoma. they put it together, it didn't create a home run in the response criteria. braf inhibitors came of age in the time we were starting to

learn how to inhibit braf and they have been putting those together, sequentially or same time there is overlap toxicity between a braf inhibitor and ipilimumab overlapping. those studies are underway as we learn how to do that better. one of the interesting things

from an immunology standpoint is radiation plus ipilimumab. and this again is a patient of judd wolchucks, showing the effect when you were irradiated at one spot and get decrease in the other. i apologize trying to fit a lot of things on the slide.

this is small but there's a decent amount of disease. they have a response in the other areas as well. this happened after a patient had been giving ipilimumab, so they gave radiation, gave ipilimumab after.

we have often used the two together when we're dealing with a progressing patient. now, here is from drew's review, switching gears to talk about pd1, showing us how they are a little bit the same. we talk about them both as immune checkpoint inhibitors,

but they actually function in different arenas, and are not overlapping in how they deal with putting the brake on. and so provide a platform to actually use them both together at the same time. so here you'll see dendritic cell like we talked about

working with the t-cells, signal one, signal two, ctla comes to the surface, you now have signal you can block with ipilimumab, happening at the t-cell and antigen-presenting cell interface. pd1, you still have to have an antigen presenting cell and

t-cell interface but the t-cells go to the tissue and it's in the tissue you have the pdl1 or 2 expressed, on the tumor, on the stroma, providing a ligand which sends a negative signal and exhausts the t-cell. this is my big antibody, this

sends a positive signal, cytokine and tumor-mediated migration. these are studies that were presented in the last two years, in new england journal of medicine, this is suzanne's paper looking at the activity of nivolumab, that's an antipd1

drug in clinical trials, the same people make ipilimumab, which makes it easier to put things together if they are both made from the same company. and essentially you see in this graph, a spider plot that shows a patient who progressed and came off, but those are a lot

fewer in number than if you go back and look at the original ones on ipilimumab. what you see is in the first couple of months the majority of patients have a decrease in tumor burden, persisting throughout the time they have been followed.

you can see some of them get a new lesion, that's what i was talking about when you get a new thing that you now see on a scan but normally would have meant that patient had to come off trial. since the experience with ipilimumab, it kept them on this

trial and most all of them except for two had a decrease in tumor size past that first new lesion. what's exciting though is an ipilimumab, i told you mostly about melanoma and renal cell carcinoma response but there have been responses in other

tumors. here is small cell lung cancer, a pseudoprogression at two months, and at four months it's gone. again, oftentimes if this was the only lesion we would be looking at, we would say, oh, the tumor is growing and maybe i

shouldn't stick with that drug. but in this case, by adding everything up, it fit the criteria to allow to continue watching and went away. and here you see some in the bottom panel as well. so it's been exciting responses in other diseases, and many more

tumor types are undergoing studies now that it's closer to approval. so here are going back to some of the basic science studies that drive this. i mentioned you're talking immune checkpoint, when i officers heard about them i

thought they would overlap in toxicity, and they do, but because their function is not redundant in the signaling, you can do blockade of both and get a better effect on the t-cells. and so here is in colon cancer, you see when you do antipd1 by itself or anti-ctla both, and

you have better surviving, combining with the vaccine as well. we hoped to do the ininhibitors on the back end, you see things with vaccine, that's the way to activate the t-cell and get the activation arm working before you block the brake.

here is judd's paper in the new england journal showing journal the activity of nivo plus ipi. this is crossovers, concurrent therapy and you see a lot of patients with this dramatic response and then continued response. the water fall plot starts

looking like something that's exciting like we saw with the parp inhibitor, there are more patients responding here with decreased amounts of disease below the axis than growing. and this year at asco it was presented as an update. you see here the nivo plus.

well, one milligram of nivo, three of ipi, or three nivo and one of ipi, these are exciting to combine them and get even better responses than we were getting with ipilimumab alone. better responses with nivo, perhaps even better with a combination blockade.

here is another tumor type that is being targeted with this and showing some wonderful water fall plots here, and these are swimmer spots showing ongoing responses with all these arrows sticking out. this is in lymphoma. one of the things that happens

in a lymphoma is they tend to have lots of ligands for your programmed death one. depending on which lymphoma, you have more of one or the other, it becomes a great hypothesis, blocking that tissue level is so high will allow the activated t-cell to then attack the tumor

and here they combined with treatment for b-cell lymphoma and retoximab, and tolerated and excite effective. getting back to the combination and toxicity, we talked about ipilimumab side effects. nivo is pretty close, one of the things that came up in higher

amounts in nivo versus ipi was inflammation in the lungs, pneumonitis. but most of the other things are typical. the rash, they had some colitis, but actually i don't even see it listed on here. and hypothyroidism.

when you put them together you start to see some of the things you had with ipi alone and start to see it at higher levels. here at grade iii or iv, 15% had liver issues. they had some colitis. the majority of patients having lower grade, so here again the

rash happens in 55%, itching in another group, so 70% overall have some sort of skin issue. again, getting back to the t-cells in the skin. but the safety algorithms put in place in the ipilimumab trials are keeping management on these more safe.

there weren't any new safety signals. so i get back to antipd1 and pediatrics, taking you through some pediatrics on ipilimumab because we have some data with we do not yet have any data in pediatrics being given antipd1 which is unfortunate because

they just approved pd1 inhibitor, the merck 1, at the beginning of this month. and i know lots of patients who are working to get that off it's a plea for the drugmakers in the country, in the audience as you go, don't forget to help us get things to pediatrics

before they get approved so they can actually get given earlier and get studied. here is some data coming out of my branch chief, crystal. stephen was a postdoc in her lab. and essentially was using a model of rhabdomyosarcoma, in

the class of high risk sarcoma we tend to treat and gave antictla4. antipd1 still had growth. however, in their hands with a blocking agent to keep the myeloid derived cells from getting to the tumor you did a lot better in survival outcome

and that's that green bar here. and so that's getting at the other arm of what things in the sarcomas are keeping it from responding to any sort of immune and one of them being myeloid suppressors. in summary, we're excited. these are the actual rates, 7

for the melanoma tumor, biomarkers for response like i said with ipi it would be great to know who is going to respond to pd1. i hinted to the fact that pdl-1 and 2 expressed in tumors might be hypothesized to respond better to, you can see that in

the new england journals, the 36% response rate in pdl-1 positive tumors versus those who were negative. it didn't mean that there weren't any responses in a negative one, and we know these can be upregulated and down-regulated, it's a problem

when you're dealing with tumor tissue at diagnosis and at relapse two years later, we're not setting up the trial to require that you have it but we're certainly looking at it as we go. here are all the inhibitors in

development. i haven't heard anybody say this one, the merck compound we just approved for melanoma, not responsive to ipilimumab. it's a narrow initial indication. they have ongoing phase iii trials and this was a

breakthrough drug mechanism. nivo, it's anticipated we'll hear something in the fda, and there are large trials, melanoma, renal cells, bladder, there was exciting data on bladder cancel cells to nico at asco, upwards of 50% of patients with pdl 1.

there aren't cohorts that looked at sarcoma, we're going to do lest you think that the two immune checkpoints you learned about today are all that's out there, there's a lot more. here are some of the other ones. all these have things in the pipeline working their way.

the hope is they will be able to have combination therapies much like nivo and ipi without overlapping toxicity and open another range of tumors to let's say one of these ligands is more important in a tumor that hasn't so far seen responses to immune targeting,

it might once we can release these brakes. here is another way of looking at it because i like to know what they interact with, because that helps. you can target both sides and learn about signaling either way.

and so b 7h 3 is expressed on pediatric tumors. we don't know the ligand yet. maybe it's a good place to this is a growing field. this push as wave of new discovery and interest in clinical trials, rather than the quick one or two months

progressive free survival changes. the field is working hard to define appropriate end points and biologic correlates so we can learn who is going to benefit tanned who not. in conclusion, inhibition of the brake, i hope by the end

talk we've gotten, so we inhibit the brake. pediatric trials are lagging but we hope in the next few months to have the nivo trial open and nivo plus ipi, and cohorts, evidence there's immune targets within the tumors and hopefully this is a lot of progress on the

horizon. the pediatrics department is over here, it's on the west wing. our labs are in the wing and the kids are on the first floor in one northwest, and so i'm hopeful at the end of this rainbow we find some good

so just to end with acknowledgements, you know, there weren't that many patients that i talked to about directly on here but each and every one of them did this out of altruism, out of hope, and we really -- i tried to advocate hard and thank them for

their involvement. crystal macall, katherine, the rosenberg and surgery branch who helped out our children, some have gone to get pills his branch, a special exemption. and i frenched judd a lot. i went to him for advice in

learning how this could be taken. i want to come back to something i put on the front pain but didn't mention today is the second to the last day of the month, childhood cancer awareness month. there's a whipping childhood

cancer video challenge that's gone viral. that really stresses every day 46 kids are diagnosed with cancer, every day 7 kids die of and so we put this gold ribbon, that's the ribbon color for pedes, into this as really a hopeful sense of we need to do

more for these kids. thanks. i'll up it up for some [off mic] >> when we talk about immunotherapy strategies like when we're doing cd-19 chimeric antigen receptors going after cd-19 positive leukemia there

can be an outgrowth. they figured out how to grow without it. we don't think the ctla-4 has the same role in the tumors. because we haven't sort of found a reason why they might just downregulate and do by other mechanisms, or it might not have

anything to do with it. we're targeting t-cells, not the tumor. i think the resistance mechanisms are more of the other brakes, the other suppressive environments that takes one or another side away from the immune system.

i don't think it's been worked out so we know o is positive and negative. sometimes it's focal, sometimes the stroma, sometimes in the tumor cells. i'm hopeful we get more to that answer, but right now i'm not willing to say it's pd-1

negative on the first look there's no chance you can respond. there will be ways of upregulating. we did another mechanism that feeds into the smam mimetic that tina was talking about using trail, and can induce apoptosis

in many of our sarcomas. and just changing chemo, radiation, or interferon gamma can all upregulate the receptors that send the apoptosis signals through better. so i think there's going to be a lot of modulation, melanoma is one of the things that's been

used in melanomas for many years, interferon alpha, how much much a ligand is expressed on the surface. >> like i mentioned, melanoma tends tore chock full of t-cells. rosenberg for decades has been growing out the t-cells and

giving them back to patients and they can be very active. and il2, a t-cell survival factor, is approved therapy for melanoma. it's also been used in renal cell carcinoma and has a proportion of patients, don't quote me on the numbers,

somewhere around ten, that can respond to just il-2 as a t-cell growth factor, and response in the tumor. so also been known that those are your typical immune sensitive. if you asked somebody what's an immune sensitive tumor, melanoma

is the poster child. we're starting to think about that differently with adult tumors like lung cancers that i showed you, colon cancers, ones with a lot of mutations back in the hypothesis, it's different and can i attack it. if you think about tumors that

build over life times and have mutations, they have more t-cells in there at some of the -- maybe not the driver mutations of the tumor but the passenger mutation and allow immune response if you can get it going. jump starting the response seems

to be easier in melanoma. now, i will admit it also seems to be a whole bunch easier to do in melanoma in a mouse, or colon cancer in a mouse. we're certainly -- there are differences to some of the immune mediators in mice versus humans as well.

so making that leap isn't always so easy. i think it comes back to the classic immune sensitive model. >> so they have -- they have attempted to do that on some of the trials. i don't think any of them rose to any of the -- any differences

rose to the level of statistical certainly there are anecdotes, and when they are published they are no longer anecdotes, small series of cases where people show you don't have any infiltrate really before ipi and then when you go and biopsy something you're worried is

melanoma growing again it turns out to be chock full of immune infiltrates. you can take something that doesn't have a lot in it and turn it into something that has more in it by blocking the checkpoint. >> so it's something that can

the oncogeneomics and dr. rosenberg, we see a ot fewer mutations than in adult tumors. at the look at all pediatric tumors we have an omix trial that allows genome and rna sequencing off any tumor that's resected here. and when you lo at it by age,

it tends to go up slowly with age, but the things that have the translocations like i talked about in sarcomas tend not to have a lot of other mutations to find. those cross-wires in the chromosomes take the place of years of built-up mutations in

the -- you know, like the colon cancers of the world. so i don't know. we might prove that right. we might prove that wrong. that's our hunch right now is that there's fewer.