Chemotherapy Treatment For Ovarian Cancer

today our first speaker is kieron dunlevey. he graduated from the university college of dublin medical school in 1994. he completed his residency in fellowship training in dublin. he came to nci in 2002 and became a staff clinician in

2006. he's the title of his talk is lymphomas. >> hi, terry. thanks for the introduction. my name is karen dunleavy. i'm the clinical director at national cancer institute and

over the next 45 minutes or so i'm going to talk about the first half of the talk is some general principles about the biology and treatment of lymphomas and then i'm going to talk about some of the research that we have been doing in our branch just to give you some

sense of the translational work that we can do in lymphoid malignancies. so in terms of how common lymphomas are, in males they're the fifth most cancer type and in females they're the sixth most cancer type. the incidence of lymphoma is

increasing. there are a lot of reasons for that. the first reason is probably because there is an increase in the elderly population. so as people get older, they have a higher chance of developing cancer, and that's

true for lymphomas, particularly for certain sub of lymphomas. also certain diseases are associated with lymphomas. hiv is associated with it. if you're hiv positive, you've got about a 100 times increased chance of getting the most common type of lymphoma and a

thousand times chance of getting a rare lymphoma. we talk about hodge kin lymphoma and non-hodgkin's lymphomas. they make up 83% and hodge kin makes up 17%. so in the u.s., i'll focus on non-hodgkin's lymphoma. the incidence rises with age and

that's the biggest reason why we're seeing an increased incidence of lymphomas. shown here are data -- showing you the period from 1992 to 2010 and the incidence of nhl in both males and females. and as you can see, there has been an increase in the

incidents of nhl in both males and females during that period of time. but the good news is that there has been a reduction in death rate from lymphoma during that same period of time in both as you can see from these curves.

so for most people who develop lymphoma, there is no factor, but there are certain factors that put you at increased risk of developing lymphoma. people who have altered immunity have an increased risk. some of the examples of that are diseases where you have

inherited altered immunity. as i said, acquired altered immunity, patients with hiv have got a much increased risk of developing lymphoma. people who have eye tragenic altered immunity, such as who are on i immuno suppressive drugs.

they have pplt or post-transplant disorder and people who have had certain types of chemotherapy are also at risk. at the nih we study show gren's syndrome. when you can do blind biopsies you find that almost all of

these patients have got evidence of mucosal associated lymphomas. autoimmune stimulation is an increased risk for lymphoma. acquired immune stimulation is, h.pylori, and in some cases if you eradicate the organism, you can see regression of the lymphoma.

that's not in most cases, but in some cases. and hepatitis c and hiv are also examples of this. our understanding of the link between chemical exposure and lymphoma is not very well established. but it would appear that certain

chemicals like organic solvents, agriculture chemicals and maybe others are associated with there's a lot of interest in the moment at micro bome. while we've certainly under covered a certain organisms, there are a lot of other infections that possibly cause

lymphoma but this has not been definitively established at this point in time. so of the infectious etiologies that have been associated with various lymphomas, burkt lymphoma, you see three different types of this. you see endemic burkett lymphoma

in regions of the world like africa, it occurs in children in ages 3 to 5 and it also presents with masses and it is linked to ebv in 100% of cases. the other two types of burkett lymphoma sporadic. both sporadic and burkett are ebv positive in about 40% of

cases. hodgkin's lymphoma can be linked to ebv, particularly after kidney transpants and heart transplants can be linked to ebv. age-related nhl is linked to hiv. there's an unusual leukemia

lymphoma called adult t cell which is called by a virus and this again is seen in distinct regions of the world, particularly in the caribbean. and these patients develop atll which is a very aggressive lymphoma and is largely incurable and most of the

therapies that we use for other lymphomas don't work for this type of lymphoma. primary effusion lymphoma is seen almost exclusively in patients who have hiv and it is linked to the hhv8 virus as is multiple myeloma, i talked about linked to held coback ter pie

lorry. so the most important thing, when a patient has a newly diagnosed lymphoma, the most important thing is the pathology and getting the pathology right. and really the tissue needs to be looked at by a pathologist who is really expert in

hematopathology and in looking at these lymphomas and has seen these before. because they are very complicated group of diseases. so i think it's best if the -- general pathologist who looks at all different types of cancers, but really focuses on lymphomas

and hematological diseases. and for all cancers there's a w.h.o. manual which is very helpful in making the diagnosis for pathologists and it lists all different types of lymphomas and their different characteristics under the microscope and also clinly how

they present. and we're in an era now where a lot of lymphomas should not only be diagnosed by how they look under the microscope, but also how the patient presents. so this is called a clinical pathologic diagnosis. so when we have a new patient

with lymphoma, we look at their biopsy with our pathologist, they like to know how old is the patient, what did they present with, what does the ct scan look like. so it's important for them to know that, because that helps in establishing the diagnosis and

in doing that they can request special tests on the tissue that will be guided by the clinical presentation. and then lymphomas are classified according to whether they're t-cell or b-cell. about 85% of lymphomas are b-cell, 15% are t-cell.

their location in the lymph node, specific features and clinical features, what the history was coming in and what disease sites the patient has. and if you look at the breakdown of non-hodgkin's lymphoma, the most dmon type makes up about 40% or 45%.

and after that there are a number of different types. but they're really pretty rare as you can see here. so for the most part when we see b-cell or foe lick lar lymphoma. depending where it arises from, it can be very different. for example, you can have

marginal zone lymphomas, which at rise in the marginal zones and folicular which obviously arise in follicles. so this is important for diagnosis. and most lymphomas are b-cell and you can correspond b-cell lymphoma to different stages of

b-cell differentiation. you'll see a little bit later and then the main lymphomas that we see a little bit later, and i'll talk a little about this in relation to the b-cell because that's the most common histology. and when the pathologist is

looking at a new lymphoma under the microscope, they use immuno histochemistry to decide what type of lymphoma it is. the lymphoma has a particular appearance in terms of its morphology with hne. but immunohistochemistry is helpful.

this is something that wasn't around 20 years ago. so for example if it looks like b-cell lymphoma -- all of these stains will be very helpful in telling you that that is the definitive diagnosis. and likewise there are t-cell markers which can tell you it's

a t-cell lymphoma. so immunohistochemistry is very important. cytogenetics are done in certain for a lot of b-cell lymphomas, frequently they have a translocation leading to overexpression of bcl-2. the lymphomas that are

associated with helico bacteria, they can have translocation, if it's coming from a stomach, it's something they can do to confirm. and burkett lymphoma has a mikry arrangement. that's present at 100% of burkett lymphomases.

burkett lymphomas are an easy lymphoma to diagnosis. it's got a starry sky appearance because of the high proliferation of the tumor cells. but if you check for a mic arrangement, it's positive. anaplastic large cell lymphoma

and it's got a 2, 5 rearrangement. so the pathologist decides to do these tests based on suspicion or where they're going with the they're not done on every single it's guided by how it looks under the microscope. so then when you have a patient

with newly diagnosed lymphoma, how do you evaluate them? the first thing is you ask them what has been going on with them. what have this presented with. the most common presentation in lymphoma is having lymph adenopathy.

people usually notice this in their neck, under their arms, in their groins. if you have lymph nodes in your abdomen, it really takes a long time before you notice them. so most of our patients when we ask them what have you been complaining of, they usually say

i noticed a lump in my neck or under my arm or in my groin. and these are usually caused by infection, but when these are caused by lymphoma, they don't go away with antibiotics usually and they're also usually nontender. infectious lumps tend to be

tender. most patients who present with lymphoma will tell you it doesn't hurt at all. in the history we ask specific questions about how people have been obviously, but there are symptoms called b symptoms, and those are fevers, night sweats,

weight loss of more than 10%, and those are important because they put patients into different prognose stick categories. it might be that you have a more aggressive lymphoma and it's associated with the words prognose than if you don't have b symptoms.

anybody that comes in you should examine them, obviously. that's very important and do blood tests. all of the basic tests should be done. there's a test called lactate dehydrogenase which is a liver enzyme but it's a nonspecific

test of tumor bulk and it's often elevated. once and everybody who presents with lymphoma, they should have a hiv test done. because of the association with and they should also have hepatitis screening done. they should have hepatitis

screening done for two reasons. one being that hepatitis can be associated with lymphomas and two, certain types of hepatitis can reactivate with chemotherapy and rituximab. everybody should have imaging of their chest, abdomen and pelvis, that's called staging and there

are four different stages, stage 1, 2, 3 and 4. stage 1 or 2 disease meenls that the disease the is one side of the diagram and if it's 3 or 4 it's on both sides of the diaphragm. and that's really becoming less 20 years ago it was important

because people who had very early stages of the disease, they either got chemotherapy but then they got radiation treatment if it was in one particular location. nowadays radiation is used less and less frequently. staging may be becoming less

almost all patients should have a bone marrow biopsy done. in the b-cell lymphoma is the most common one. in the follicular, 15% to 20% of cases diffuse b-cell lymphoma and other studies should be done as indicated. there's a lot of work looking at

the role of fgg pet scans, a lot of doctors will do these and depending where the lymphoma presents, for example, if you have a lymphoma in your brain, it's best to do an mri scan. and certain types of lymphomas or certain presentations should have their spinal fluid checked.

so if you have aggressive b-cell lymphoma and if you have involvement of what are called extra nodal sites, sites that are outside the lymph nodes or if you have involvement of the bones or if you have a very high ldh level, there are certain criteria that suggest that it's

really important to check the csf. because the treatment that we give to patients, it doesn't really get into the central nervous system. it doesn't for the most part been trait the blood-brain barrier.

you want to know if the patient has involvement of the spinal fluid. it's rare to have it, but it's associated with certain clinical factors. so if a patient has those, then we would recommend that you check the csf.

feel free to interrupt me at any point if you have questions or i'm happy to answer questions at the end. this is the staging system that i told you about. i won't go through it too much, but there are four different stages and one being one lymph

node area and 4 being diffuse or disseminated involvement and involvement of extra nodal sites. and as i said, the morphology of lymphoma is important. this is diffuse large b-cell lymphoma morphology. and typically you see large

cells that are about twice the normal size of a lymphocyte. they're large cleefd or -- they may be rich in t-lymphocytes. so over recent years there's been a huge amount of interest in the tumor micro environment. it might be something some of you are focusing on, but it's

really important in certain types of lymphoma. not just the tumor cells, but the infiltrating immune cells around the tumor cells. and in follicular lymphoma and diffuse large b cell lymphoma if you close what these info cells are and what their genetic

expression is, you can predict how certain patients will do. the tumor micro environment is becoming very important. we're starting to develop therapies that actually target the micro environment. not just the tumor cells but also the micro environment.

okay. so what about the outcome of patients with lymphoma? how do they do? are most of them cured? and for this talk i'll focus on diffuse large b-cell lymphoma quite a bit, because that's the most common type and the type

that you see the most when you look after your patients with so these are markers and they show on top overall survival and progression survival on the bottom. and you can see here that the interval is in months. so that's 100 months which is

quite a long time. and these are patients of all ages who had diffuse large b-cell lymphoma and they received the standard treatment, which is a chemotherapy regimen called chop. and r is rituximab. over the past 20, 25 years the

most significant advancement has been the addition of rituximab to chop chemotherapy. that has increased the cure rate by about 10% to 15%. so it's very important to give rituximab. and it's interesting because the adult world and the pediatric

world of lymphoma are very separate. and the pediatric world has been hesitant to give rituximab but they really haven't been giving it to everybody up until a few months ago where they were doing a big randomized study in children and they were

randomized to get rituximab or not to get rituximab but they have recently shown a little bit earlier than they expected that there was a significant cervical eiffel advantage to giving so for now everybody, adult and child who has diffuse large b cell lymphoma gets rituximab.

i think what these curves tell you, if you're a new patient with dlbcl, and these are patients can be divided into low-risk ipi, intermediate risk or intermediate high risk ipi. it's called the international prognostic index. it looks at your age, the

disease you have, whether you have an elevated ldh, your egoc performance status, and the presence of extra nodal sites and that's what the low risk, intermediate risk and high risk is. but these curves will tell you that with standard treatment, so

all around the world if you go into an oncologist with diffuse large b cell lymphoma, you get r shop. you can see for a significant portion of patients, therapy is not adequate. while most people have a long overall survival, a lot of

people don't and a lot of people progress as you can see from the curve on the bottom. so our field has really been trying to improve the cure rate for glbcl and a lot of different strategies have been tried. what have people been doing? they have said well, we usually

give six cycles of r chop and let's give eight cycles. it's associated with more toxicity but no increased benefits. they've tried therapy intensification. so r chop is typically given every 21 days.

people have tried giving it every 14 days. in germany about 10 years ago there were some studies that suggest it that if you gave it more intensely it was associated with a better outcome. about a year ago a group in france and in the uk did large

studies of patients where they gave half the patients r chop every 21 days and half the patients every 14 days and there was no difference in both studies and the patients who got it every 14 days actually had more toxicity so that wasn't people have tried doing bone

marrow transplants when patients go into remission. it's not been of any benefit. try giving maintenance rituximab that's where people finish their rituximab and chop treatment and they continue to receive rituximab for 18 months or two years after that.

and studies have again half the patients had maintenance rituximab, half had not. and there was no difference between the two groups. one of the promising strategies? other better chemotherapy platforms than r chop. the french developed a regimen

and in young patients who have a low ipi score, that was better than r-chop. but that was a very small proportion of patients who get the lbl and our group has been doing a lot of work with those and we're waiting the results of a ran omized study.

we have a lot of novel agents that are based on the identification of new pathways. and different types of lymphoma and i'll talk about thoes in a little bit. i think that offers the most promise for improving cure rate is to incorporate some of these

effective novel agents into up-front standard regiments like r-chop. and i feel it's really critical to develop novel agents and approaches in the context of biologic subtypes and factors. so in that respect, historically we always thought that dobcl was

one disease, it was just one entity. and this is work that was done by the leukemia molecular pro filing project. it's a big project and it was led by a lab based at nci. and about 20 years ago they applied gene expression pro

filing to cases of lymphoma. and so they looked at the differential expression of tens of thousands of genes in a large sample of dlbcl cases. and where they did this, they were able to break down most of dlbcl cases into molecular subtypes.

most of them fell into one of two molecular subtypes. a small percentage of them fell into another type which i'll talk about at the end of the lecture. so this is a subtype of dlbcl, but if you look at its genetic makeup, it actually has a lot

more in common with hodgkin's it shares about a third of its genes with us. so it's distinct from the abc or the gcb subtype. so you know, you may ask well, this is very nice, it's very nice to be able to genetically divide this disease into at

least three diseases, two of them that are shown here. but is that helpful for patients? does that make any difference clinically? and it does make a difference clinically. so this is a study that was

published last year from a group in british lumia. they looked at the outcome of over 300 cases of dlbcl. shown here is progression-free survival and time and years. so you can see that this group of patients have been followed for a very long time.

and they did something called an assay. profiling is expensive and this assay has been validated with gene expression profiling. it looks at the expression of about 20 genes and there's close to 100% concordance with conventional gene expression

profile. this is easier to do and much less expensive. patients who had the abc subtype had a much worse outcome than and just to go back to this, the gcb, abc and pmbl, they're really distinct diseases and they have distinct mechanisms of

oncogenetic activations. pmbl is a totally different disease. it arises in the mediastinum. if you look at the oncogenetic characteristics, they're distinct. gcb which makes up the majority, you often see a translocation of

bcl2 in the abc subtype. there's activation of a pathway and that has led to the identification of many interesting targets that are currently being pursued in clinical trials. and pmbl has got activation of pathways.

there's a lot of interest with immune checkpoint inhibitors and immune oncology is a tremendous earia. there's an early study of about 20 patients but it responds wrl to immune checkpoint inhibitors. so it's very important in pmbo. just talk a little about the

biology of different subtypes of dlbcl. so the most common subtype is so if you look at children and young adults who get dlbcl, they almost always have the gcb subtype and the abc subtype starts creeping in when people get older.

so if you look at people over the age of 40, about 5% or 10% of them have the abc subtype. if you look over 60s, it jumps up to 20% or 30%. if you look over 80, it's over 50%. so the gcb is the type that you see in both young and older

people, but pretty much all young people. and about 10% of those have got a mic rearrangement. and i think it's interesting to think about that in the context of whether the lymphoma is gcb or abc. this is i fuse b cell lymphoma

but these are almost always of gcb origin. whereas cases that have high mic expression but do not have a rearrangement are almost of abc orange in. about a third of cases have got a bcl2 rearrangement. again these are almost always

where as cases that have got high expressions but without a rearrangement are almost always of a abc origin. so in our field one of the most controversial areas are the double his and double express sors lymphoma. patients are asking about what

the best therapies are and they talk about them a lot. double hit lymphomas are those that have a mic rearrangement and a bcl-2. and high bch2 expression and high mic expression. but this distinction between double hit and double expression

is really important in our field in relation to the cell of origin. gcb or abc. because it really is starting to help us decide what therapy patients should get. and it's really important then to distinguish the double hit

lymphomas from the double expressed lymphomas that do not have any hits. because pretty much all the double expression lymphomas that do not have any hits are the abc origin and the double hits are of gcb origins. the context of these

rearrangements because they make up a high proportion of new so the abc subtype is associated with a worse outcome. so there's a lot of interest in that, because we really need to augment the cure rate of abc lymphoma and the key characteristic of this subtype

is that there's activation of the nfcappab pathways. there's a lot more red on the left. the nf kappa b target genes are much higher. so nf kappa b is highly overexpressed and therefore is a potential target in abc, dlbcl.

and over the past 15 or 15 years, a lot of work has been dodge in understanding the mechanisms in understanding the nf kappa b activation and the abc subtype of dlbcl. so for about 15 years we've known that about 10% of patients have got a card 11 mutation.

but recently over the last four or five years, it's been identified that about 20% of patients who have abc dlbcl have got a mutation in 79a and b in the b cell receptor. and about 30% of patients have got a mutation -- and both overactivation of the b cell

receptor and over activation can activate nf kappa b. so these are very important pathways in thinking about what drugs should one use to target abc lymphoma. and we now have a lot of inhibitors of various components of this pathway or these

i could probably put up about 20 to 30 drugs that are in clinical trials or in development that specifically inhibit, small molecule inhibitors that inhibit various components of this pathway. i'm not going to talk about all of those, but i'm going to talk

about inhibiting btk. so it's a key ends mattic step in tumors that have got chronic active b cell receptor signalling. it really makes sense to think about inhibiting that. and you've probably heard about a btk inhibitor.

there are a lot of others. this is really the first one. and it covalently binds to the btk-active site. and this is a study that we were involved in, but the hypothesis was that so in abc dlbcl, btk is a critical enzymatic step, so therefore inhibiting it might be

an interesting therapeutic strategy. so to preclinically confirm this was something that this might be worth doing, both abc and gcb cell lines were treated with ibrutanib. and different types of abc cell lines were used.

those that were card 11 mutated and those that were hire up than card 11. you would expect that with ibrutanib that the card heaven mutated ones would not respond to it and that gcb would not respond to it and that's exactly what happened.

so the abc cell lines that did not have the card 11 mutation, they were upstream of that, they responded well to ibrutanib. and that led to a clinical study of ibrutanib in relapse in refractory dlbcl. they had to be relapsed disease, and they received single agent

ibrutanib until they had progressive disease. and as you can see from this waterfall, orange is gcb and blue is abc, patients with abc had a much better response. this is -- those on the right had a shrinkage, those on the left had growth over zero or

below zero. but you can see the drug is much more active in the activated b cell subtype. but in terms of responses, close to 40% of patients had a c or a p or and just one patient had a t or with gcb. so it wasn't really effective

with gcb. so this proved that ibrutanib does work in abc much more than it works in gcb. this was the survival again. you can see that the blue is doing significantly better than the orange and that's the abc versus gcb.

this was one patient from the study and with single agent ibrutanib you can see they have had a very good response. one thing unique about this study was that when the results were being analyzed, a proportion of tumors had mutational allowances done to

see which tumors exactly had the best response to ibrutanib. and as you might have expected, those that had a card 11 mutation didn't respond. but the tumors with the best response were those that had both a cd-79b and a my d8d8 mutation.

and there's now a study in newly diagnosed patients, this study has actually been completed and patients with the abc subtype went on to this study and they were randomized to get r chop or so we'll know in a few years' time whether it improves cure rate.

studies, one adding an immuno mod lar agent, that appears to be much more effective in the abc than the gbc subtype. and these studies will help us in the coming years to see if the abc subtype is susceptible to adding one of these novel agents or small molecule

inhibitors to standard regimens like r-chop. just going to talk a little about primary cnf lymphoma when i'm talking about abc. so primary cns lymphoma that starts in the brain for the most part, it can start in other areas of the cns as well.

and over 90% are diffuse large b cell lymphomas. these tumors they resemble the abc subtype, but they're not exactly similar. they're not exactly the same. and broadly you see activation of nf kappa b, card 11 is mutated in a significant

proportion. it's really interesting that shown here are my-dad8 mutations. and a subtype called primary cutaneous lymphoma, primary -- and while we have thought of these as abc subtype and similar biologically, when we look at

mutational status, they are very different. and you can see that over 90% of primary cns lymphomas have got a and if you see cd79b it's also mutated in over 50% of primary cns lymphoma. so most primary cns lymphomas got the mutation.

and why is that interesting? it's interesting because what about using ibrutanib in primary cns lymphomas? so for patients with primary cns, although their diffuse large b cell lymphoma, they're treated in a different way and the reason for that is because

the treatments used in systemic lymphoma don't cross the blood-brain barrier. for that reason this group of patients have received very different types of treatments. and they've received drugs or approaches that overcome not being able to cross the

so the two drugs that have been are still standard in this disease are -- they're not very effective drugs for dlbcl. we would never use them in systemic dlbcl but they have been used a lot in this disease because they cross the blood-brain barrier and cranial

radiation has been used a lot and transplant. but these mow dalties have been used because of the failure of other agents to be effective in this particular disease. and this disease has been very challenging to institute trials in.

it's rare, it's difficult to biopsy patients and a lot of these patients have got a very poor performance that is much more so than systemic abc so we designed a new treatment for this disease. we took it effective regimen that is good for systemic large

cell lymphoma. and we thought about it and the drugs that did cross the blood-brain barrier we included those and if drugs didn't cross the blood-brain barrier, we substituted those for drugs that did. so atopazoid does cross it, so

we did include that. prednisone -- we don't know how effective it is in dlbcl so we left that out. temazolamide does cross and has been used in lots of brain tumors, doyar -- and we've included rituximab. ibrutanib, you know, as i said,

because of the high rate of mutations, we thought that would be a good idea, but we didn't have any information that it crossed the blood-brain barrier. nobody had done studies or done pk levels. so we included that as well. so we called this regimen -- and

combined an agent that targets b-cell receptor signalling with agents that penetrated the cns effectively. this is a regimen that we used. so we gave patients six cycles of this. and in the very first vehicle we gave ibrutanib as a single agent

for 14 days to see if it was effective in this disease and also to do pk studies in the csf and in the peripheral blood. and assess getting into the csf and at what concentrations it was getting in and if it was effective or not. and what did we see?

these were -- we've treated 18 patients, but these are our first two patients and we saw that it did get into the csf for a significant amount of time and achieved meaningful csf concentrations. the two top curves are plasma, ibrutanib, and me tablits.

and i'm not going to show you the pks, but this shows you that the drug gets in there at pretty good concentrations. this is over a 24-hour period. so we've treated patients that have really had good responses with the ibrutanib alone and with the suggested teddy or.

this was a patient who was 67 back in 2010 she presented with slurred speech, noemic afascia. gait instability. mris showed a leaguesian in the left cerebellum. had a complete response but then in 2014 she relapsed with left temporal lobe disease.

and when she came to see us, her doctors were considering hospice at that point. so this is our mri scan before and after she got 14 days of so at the very beginning of the study she didn't get any chemotherapy, just ibrutanib for 14 days and then before and

after. you can see this lesion has regressed significantly. this was another patient with relapsed disease who had had a lot of treatment in the past. she had a mass in the brain and after 14 days of ibrutanib she had a good response.

so this is illustrating to you that this drug works very well in primary cns lymphoma. we've had a number of patients who were very refractory to other treatments and are doing well and remaining in remission. she's in remission now for over two years after this.

so just to finish up. i'm going to talk about another type of mediastinum b cell lymphoma that affects younger people. especially younger females. so in younger people, the gland is much more prominent than it is in older people.

and classical hodgkin's lymphoma comes from -- and this just shows you there's a spectrum of these b cell lymphomas -- and if you look at the morphology and immunohistochemistry, you can for the most part say one is hodgkin's and one is pmbl but there are diseases in between

that are intermediate in torms of morphology and immunohistochemistry. and there's a lot of interest in these. i'll just quickly go through this. the therapy of this disease has been controversial.

most patients who get this disease, they're females in their late teens, 20s or 30s. and historically the way this disease has been treated is by giving chemotherapy followed by radiation. but of course as these patients are young females for the most

part and media stein al radiation is a effective treatment, but down the road it's not a good thing, we now know from studies in children that if you get mediastinum radiation, if you get high doses of it and if you look at patients 10, 20, 30 years later,

a very high proportion of them develop cancer and ischemic heart disease. so for females, the rate of breast cancer is about 20% starting 10 years after the end of treatment. so this is a disease, because it africas younger people,

predominantly females, you really want to think about that when you are selecting up-front therapy because you might cure the patient but you don't want them to get breast cancer in 20 years' time. so i'll quickly show you this in the interest of time.

so there are ai lot of different ways of treating this but most of them include mediastinum radiation with my colleague, we did a study and hypothesized if we gave that we would not need to give radiation. because it's a rej man we believe is much more dose

intense, it's dose adjusted and i won't go into it again. this disease just like classical hodgkin's disease. they both affect young females predominantly and they historically benefit from those intensities. we started our study and said

that this treatment is going to be effective and we don't think that patients are going to need radiation afterwards. and that would be a significant shift in how we treat this. because if you could take radiation out of the equation, that's a big deal.

because then you eliminate the chance of these secondary complications like cancers and particularly breast cancer happening. so we treated 53 patients and as you can see from these curves, the evidence of survival was high.

just three patients did not have complete remission with this approach and two of them got radiation and they're both doing fine and one of them, she had a small residual mass left behind and we resected it and her disease never came back. and for our ash meeting this

year we've updated these results. the results are lalt less, the efs is about 90%. we've shown that you can effectively treat this disease without radiation with this regimen. i'm not going to talk too much

about that. i think for the last few minutes i'll take some questions before the next talk. thank you. [ applause ] >>. [ away from microphone ] >> just like in several other

cancers, the treatment is changing at a very fast rate. i would say in aggressive lymphomas, particularly hodgkin's lymphoma, the checkpoint inhibitors, are there are two, these really have made a huge difference, especially in hodgkin's lymphoma but in also

in primary mediastinum lymphoma which is less common. that's one class of agents which i think are going to change the field significantly. i've shown you some results with btk inhibitors. i think that's, throughout a lot of lymphomas and leukemias,

certainly ibrutanib has been effective and now there are a lot of other btk inhibitors so that particular pathway and inhibitors of it are very, very effective. and then there are new transplant treatments. some of our colleagues at nci

are doing a lot of work with something called card t cell therapy where they take t-cells out of patients and genetically modify them and reinfuse them and they have all different types of card t cell treatments that target specific antigens of a tumor cell so there's been a

lot of work. and now they're working on other card t cells. so i think their examples of some approaches are very promising and effective. the thing about lymphoma is the cure rate is very high. we have effective strategies at

the moment. these new strategies are very effective, but bringing them in to the setting is challenging because it's not easy when in some of these diseases when you have cure rates of 80% and 90% to bring these new promising agents into up-front studies.

so that's the challenge. i think the field has changed a lot over the last few years and i would expect will continue to change at this pace with the drug development that is happening at the moment. >> the car t-cell treatment? that's still pretty new.

and in lymphoma it depends on the type of lymphoma. i think the field is still trying to figure that out. because i mean, it's at the point of being commercialized at up until 2 or 3 years ago, a lot of groups around the u.s. were really developing their own type

of ct-cell therapy. there are a lot of different vectors that you can use and none of these approaches are exactly the same and for the most part people in their studies included all different types of lymphomas. at the nci, there was a paper

published and it was smaller, 12 or 13 patients but there were two patients who had a very long remission. they actually two patients had had multiple lines of therapy before and after they got car t cell treatment had a very good response and at 3 or 4 years

when he published his paper, they were still in c or. over the next few years we'll appreciate a lot more where the sort of niche is for ct-cell treatment and which types of patients we should be thinking about. >> thank you.

>> okay. so an announcement is that i've lined up now the tumor board. and the core facilities. so i'll be sending you all an e-mail later this week. and if you want to visit these extracurricular sites, you can sign up for them, return the

e-mail to me by september 30th. we'll start doing the rotations in october. so our next speaker is tina annunziata, she's a graduate of georgetown medical school. she came to the nci oncology training branch. she's now moved on to ovarian

cancer. she is with the women's ma nancys branch. ovarian cancer. >> thank you for inviting me to talk. i'm just going to give an overview of ovarian cancer and the genomics era because a lot

of work has been done. what is genomics? i would stay a study of the genome. study of chromosomes and study of gene expression but on the global level instead of individual events. when did the genomics era start?

probably in 1959 when there was a study of chromosome to identify an abnormality. they called it the philadelphia chromosome because that's where they were. and it was in chronic leukemia. so what they did was looked at a number of cases of chronic my

lodge news leukemia and noted that most of them had this abnormal chromosome as you can see by this arrow here. this abnormal chromosome, and they saw that that was the recurrent thing in patients with this cancer. it wasn't until 1973 that janet

rowley published this was a recurrent fusion between chromosome 9 and chromosome 22. and then in 1984, it was sequenced and found that there was a fusion between the bcr green and the able kinase and in 1996 finally we came up with a drug that blocked that able

kinase and could be used for treating cml and then eventually this drug became approved as imatinib. that was a long period of time in the initial findings in 1959 and the approval of the drug in 1996. that was originally, i would say

it was based on more of an observation rather than a functional experiment. so what is the functional genomics? how can you find out what part of the genome is functional. so another type of approach, in 1981, she discovered the gene

her-2-nu because it was carcinomas and neuroblastomas that they took the dna, chopped it up and introduced it into fiberglass that were benign and saw which of these genes would transform the fiberglass and could grow in vitro and could grow in mice.

and it wasn't until later on they found that nu was related to the epidermal growth factor receptor. they found this by a southern blot looking at these her b related sequences. and then in 1985, it was found that this gene, now that they

were calling her-2 was found on chromosome 17. so you can see in these cases, these individual chromosome 17 from various cancers, there is an amplification of chromosome 17 here which is recurrent. and that is associated with the her-2 application.

and then in 1987, it was found that actually the amplification was recurrent in certain types of breast cancer. so now we have affiliated this with a disease. now, if you remember at the beginning they actually used neuroblastoma, which is not

really specific for breast cancer, however it is most common that we use today in breast cancer, the her-2. so, and then i didn't put the slide in here, but eventually the her-2 targeted therapy herceptin was approved for breast cancer with the her-2.

this is actually a targeted therapy based on a functional approach that discovered it. not just random and observational studies but a functional approach and i'm going to come back to this later on. so what about ovarian cancer?

do we have any drivers? have we done any functional experiments and do we have any drivers in ovarian cancer? so just to back up and give you a little background on ovarian cancer, it's the most lethal gynecological disease this the united states.

it's the fifth most common cause of cancer deaths in women. this is because 70% of ovarian cancers are diagnosed at advance stage. and less than 35% of those with advanced stage will be alive at five years. this is sort of emphasizing that

point is that the incidents of stage 1 and 2 ovarian cancers is 25%, where the survival is 65% to 90%. whereas stage 3 and 4 is the majority, 75% is diagnosed in stage 3 and 4, but the cure rate is less than half, or so probably around 40 or less.

so what is the treatment for newly diagnosed ovarian cancer. first complete surgical staging followed by optimal cytoreductive surgery, which means less than 1 sent meter of the disease left followed by chemotherapy and clinical trials.

so chemotherapy, just to highlight, is platinum and toxane and based on many years of studies that have compared various regiments, the standard of care is cisplatin and -- and i have highlighted here interperitoneal is stage 3 with an optimal reduction.

and that is based on this timeline, which, you can see here in the 1960s when we only had chemotherapy there was nobody who lived for five years. 1970s we have the introduction of platinum single agent, which led to a small benefit in 5% of patients that were alive at five

years. combination therapy was brought into the clinic in the '80s, with a 15% response -- i'm sorry, 15% five-year overall survival. and then in the '90s we had introduction of taxane. and they were giving us more

increment of survival. 35% of patients survived five years or more. and then the 2000s we had the introduction of not a new chemotherapy but a new administration of the chemotherapy, bringing the five-year survival up to 40% in

the 2000s. so what do we know about ovarian cancer? so we treat ovarian cancer as one disease. as you saw from that sort of timeline of chemotherapies. ovarian cancer occurs in the peritoneum.

however, ovarian cancer is not as you can see here, even histologically, you can see there is five different subtypes of ovarian cancer. so based on this, you would say it looks really different, it must be a different -- it's a different cell type for sure.

but we still treat it all the same based on its location in the body. we don't even really know for sure the tissue of origin. and i'm going to go into this in a little detail, but we're not even sure of the tissue of origin in the giynecologic tract

for the different types of subcancer. not a clear differentiation. overall it does come from some part of the epithelium that occurs in an extra uterine site. but again, very unclear really where the tissue of origin. many studies have been

undertaken on a genomics level. underlyi underlying progression will create new avenues for treatment. what do we know? i'm going to give a brief overview. i have references in the bottom

if you want to look into them. this is a brief history of genomics in ovarian cancer to date. clear cell cancers, 5% to 10% of all cases. this is much less common than serous pathology which is 70%. clear cell cancer ps tend to

have a worse response to standard chemotherapy. which was based on the serous cancer which is the predominant cancers. clear cell cancers up to 40% of situations can be found associated with end me tree yoes is in the peritoneal cavity.

in 2010, a paper published in the new england journal of medicine, sequenced rna and they sequenced dna from samples. and they went on to immunostain their predominant findings in the 455 additional samples. so really a large effort with a number of cases involved here.

and what they found was that there were recurrent aberrations, whether they were mutations of the 1-a gene. so what is arid 1-a? it's a chromatin remodelling gene. it deleted in 6% of all cancers. so when you see things that are

deleted, have nonsense mutations, these are potentially tumor suppress sor genes. basically you're losing its function in the cancer. in fact, they did find that in cases where arid 1 mutation occurred, there was loss of arid protein suppression.

this is showing you the protein product of the arid 1-a gene and in cases -- this is a little bit confusing, 70% had lost. so going up is actually more loss. 70% had loss of the gene in clear cell and 50% -- whereas if they did not have a mutation,

there were much fewer that had but here you note that it is highly specific to clear cell and endometrioid histologies and there was none found in any of the serous cancer that they looked at. so very nice isolation of this gene specific to endometrioid

and clear cell cancers. over 40% of clear cells, 30% of endometrioid but less than 1% of serous. however, it was not found by a functional test, right? so what did they look at? they looked at just shotgun sequencing of a lot of cases,

but there was no sort of drive to find out if it is actually functional. there's unclear therapeutic utility in this case. of course there's diagnostic utility, but unclear therapeutic utility as of yet. it's not to say that tumor

suppress ors have not been targeted and we have done successful targeting, but in this case, not a functional test and not a functional outcome. mutinous cancers. we know since 2004, this is showing you mutinous cancers have worse overall survival than

regular non-mutinous cancers. significantly worse. so in 2006, actually this came out of the nci, they looked at gene expression profiling, so global gene expression ovarian surface epithelium this is nontransformed normal patient samples.

serous low grade malignant potential. so serous cancers that are sort of grade 1. serous grade 3, so high grade serous cancers, mutinous higher grade and mutinous lower grade. so mutinous cancers, you can see right here, very much grouped

together, showing that they had a distinct gene expression profile than the serous cancers. and the course the serous cancers grouped together. so showing that actually yes on a genomic level have a different pattern. very small study here show that

mutinous tend to have k wrap mutations similar to low-grade mutinous, which also have been shown to have mutations. very small studies, 15 of 22 low-grade serous cancers, 31 through 51 precursor leasions. so is this a potential target? this was not really a genomic

level, but this was a single gene level. but here is summarizing that the serous borderline, low-grade serous, low-grade serous. none of high-grade serous and maybe a few endometrial. so at least k rab is a potential driver because it does

something. can we target it in ovarian low grade serous. here you see the pathway. so rath is here signalling downstream of the receptor and signalling down to erk through mec. so there was a study of 52

patients with recurrent low-grade serous ovarian cancer that did a study that received a mec inhibitor. so they're trying to block this specific downstream pathway of the raft mutation. they had 52 patients, 8 patients responded, 34 had stable disease

greater than four months which is not that surprising for low-grade cancer. so here we see, okay, well we had eight patients, let's see if those are the patients that had the mutation, can we correlate the mutation with in response to the mec inhibitor and the answer

is actually no. because here you see that five of the 18 patients that did not have a mutation actually responded, whereas only two of the patients responded. so it didn't quite correlate there. so what happened?

well, we all know that k rath signals through many pathways. we were focusing on this pathway, but probably there's a lot of other pathways that feedback. so we never had evidence beforehand that mec inhibitors would actually work, and they

didn't. so there you go. what about high-grade serous cancers? this is where we focus mostly on, in our clinic, and i'm going to show you some results of our charts. but let's go back to what about

the genomic analysis of high-grade serous cancers? they're the most common, 70% of ovarian cancers. cancer genome atlas. the goal was to identify molecular abnormalities and constitute therapeutic targets. what they did was micro array

analysis. 489 cases where they looked at mrna expression, micro rna expression -- and 316 samples. so? functional experiment? no, it is not a functional experiment. again we are broadly looking at

shotgun sequencing expression and not at functional mutation. so what they included newly diagnosed patients, no prior treatment, and they had to have a companion normal tissue specimen which could be adjacent tissue -- or -- what they found, probably you guys are all aware

of this already, but i'll summarize quickly. this is just from the original population where they really don't have that much other to look at besides glrks loiblastoma. you can see here the copy number these are the glioblastoma

blue is deletion, red is amplification. you see glioblastoma, quiet here. a lot of amplification, recurrent, and a lot of deletions here, recurrent across all of the cases. what do you see in ovarian

cancer is a complete mess across the whole genome. so you see losses, gains, everywhere. every single chromosome is involved. you have some abnormality across so what does that mean? actually, that is quite

interesting, because that has led to thoughts about is it a recurrent genetic abnormality or is it just the fact that there is this global disarray and can we use that as a target. significantly mutated genes, not that many. you can see here, kind of a

defining feature is the p-53 mutation which has been known, and that was sort of verified whereas pretty much nearly all of the patients had either a mutation or a loss of p-53. the other recurrent abnormality, again we knew this, the rca1 and 2.

they are breast cancer related genes but they also lead to the breast cancer and ovarian cancer so they can occur in women with this mutation, occurs quite frequently in mediterranean populations, especially ashkinazi jewish. interestingly, there were

homologous, vrca -- if you remember the combination is the most reliable way to fix double-strand dna breaks. other dna repair pathways generally focus on single strand breaks, which is how we can target. so interestingly, up to 33% up

cases, either with or methylation of the gene and this is also downregulation of gene expression. so what they looked at here was the brac mutated cases had a better overall survival, better response to standard chemotherapy with plat yum and

taxane. interesting because platinum is a very strong dna platinum agent. they respond well to platinum where as had a worse prognosis. one of the summaries of this paper was that there were many altered pathways in ovarian

some of the predominant ones focused on dna repair. so what next? what do we do with all this information? are there new therapeutic approaches? we have 50% with depekts and we can treat those with park

inhibitors which i will talk there were also some other commonly deregulated pathways such as the rb pathway, and not signalling possibly and then the quote from the paper is that inhibitors exist for 22 genes in regionons of recurrent amplification and those trials

are all on going. but again as i mentioned, perhaps it's not individual's genes that we should focus on, but that their network and pathways that need to be targeted in order to get a selective therapy. and we sort of learn that from

the story in that we can't just target one part of it, because possibly there's going to be feedback that the cells can recover their signalling from. so we need to target a network rather than an individual gene in ovarian cancer. so targeting homologous

the brac mutation was not discovered for ovarian cancer, it was discovered in 1987 published in 1990 by king. and that was found by sort of a functional approach looking at what was similar, what was happening in these women who had families of multiple generations

of cancer. she was mainly focused on breast cancer, but again ovarian cancer is also in this syndrome and you can see here this family with ovarian and then breast. so as i mentioned, in the cancer genome atlas, they found brac1 and 2 schematic mutations, also

found methylation, but then also found other proteins involved in home ol gus recombination that could potentially lead to a defect. so almost 50% of cases in cancer genome atlas in the high-grade serous cancers have potentially a depekt in homologous

recombination. it's possible that this defect is what's leading to this vast chromosomal instability and aberration. so this all sort of makes sense from a molecular pathway standpoint. so many groups have been

interested since then in homologous recombination. this is just one example showing you that people can up with a signature for homologous recombination signatures. one deficiency and came p with this little signature of 230 genes and when they look at this

signature in patient samples, this is breast cancer, that patients with a defect in homologous recombination actually do better. again, similar to the ovarian cancer story, than patients with intact homologous recombinations.

this is where the park inhibitors come in when you treat the cell loans with the parp inherpttors, the ones with deficient homologous recombination actually respond with a lower ic50 than the ones with intact homologous so what is a parp inhibitor and

why do we care about it? in your typical cell you have your dna and it is damaged by multiple causes. it can be damaged by metabolism, environmental exposures, sunlight. what happens is that parp, in one of the basic rescission

repair pathways, it comments in and marks that area for polyadp ribose polymerase. it marks that area and you get assembly of the dna repair complex. but if you have a parp inhibitor, that single strand break does not occur.

in replicating cell, as you know the replication process causes a nick in the dna to unwind it so a single strand break gets converted to a double strand break. now you have a double strand break of course you can repair this by homologous recombination

and then the cell can go on to have a normal replication. if you have a cancer with a brca deficiency, it's defective and the cells have a slower pathway of repairing double strand breaks. most of the time that does not suffice when you have a lot of

damage so you get cell death. so this is the idea the development of parp inhibitors. this is approved for the treatment of women with brca mutant ovarian cancers. it is orally -- it is not novel anymore. it is orally active parp

inhibitor and it had been shown to have synthetic -- with vrca1 or vrca2 deletions. so we ran a phase 1b study. i mentioned that carboplatin is something that causes a lot of dna strand breaks, mostly single strand breaks, so i thought we would augment the single strand

break with the olaparib and be able to have good responses in current ovarian cancer. we had three cohorts on the study. the first cohort was anyone with a brca mutation with breast or ovarian cancer. or they could have a high

likelihood of having the brca mutation without having tested. so if they have a strong family history. we had another cohort with triple negative breast cancer and we had a third cohort with high grade serous ovarian cancer that were known to have normal

brca. and these are all both -- this one is published already and the other two are clinical cancer research. so this cohort actually we were able to dose escalate olaparib twice daily and carboplatin, we did have to give them a

two-break. this cohort only escalated to 4. probably because they had more prior therapies and the same with them one because they probably had more prior therapies, a lot of regimens beforehand. so in the brca mutation cohort,

we had 45 patients, 37 had ovarian cancer and eight had breast cancer. we had 30 patients in the dose escalation and 15 in the 1b expansion. as i mentioned the carboplatin was 5 on day one and olaparib was 400 milligrams twice daily

on days one through 7, every 21 days. these are the dose levels that the patients were given. dose level 6 is the one that most patients were in because that was the expansion cohort. but even on the lower dose levels, this is the response

curve. so we had patients with 30% chance would be considered a response. we had stable disease or responses in pretty much all of the patients with the brca this is remarkable. these are heavily treated

patients that had multiple therapies were for the most part resistant or refractory to platinum agents and we gave them the platinum plus the parp interprettor and we got at least stable disease if not complete responses from some of the patients.

and interestingly we have the three patients with the response still on study after five years. so this is actually very remarkable. and in breast and ovarian this is just a summary. we had complete responses at the time of this publications but

sense then we've had three patients with a complete partial response and stable disease in nearly all of the patients, six of them did have progression. with an overall response rate of 44%, clinical benefit rate of 82%.

so that was great. and that was based on genetic so a conclusion was that olaparib is well tolerated in combination with carboplatin. that was one worry that people are going to be more susceptible to having side effects with low blood counts because their bone

marrow cells, their normal cells really just have one good copy of braca. but that was not the case. it was well tolerated and we could give it up to a standard dose of carboplatin in the non-braca mutated community. highly activated -- greater

activity of course at the higher dose. this is positive group of concept that the activity and the tolerability of genetically defined targeted therapy with olaparib and brac cancer. so this is a great result. but again, that was not a

functional experiment and so we were lucky enough to have the drug that was targeted towards the homologous recombination pathway that was figured out before we actually started the but what can we do about a functional approach? how can we more efficiently

identify targets for ovarian so people, the common thing nowadays is to do these consequencing for actionable but what does that actually meenl? there are a number of commercially available tests and they give you back these long

reports, we've sequenced all these genes, those are possible benefit and unlikely benefit. so should we use this drug, maybe we shouldn't use that drug. it's a little bit unclear. basket clinical trials such as the nci impact trial are looking

at sequencing individual genes and putting patients in baskets for various drugs based on their so, for example if they have a k rath mutation they might go to a or others. so they're assigning treatment based on mutations. but typically there's no

functional link. this is again sequencing. we're looking for things that have been found in other cancers, but we don't know if it's going to be functional in your particular cancer. and there was a very well-written i think article in

the journal of clinical oncology in 2015 and they said that the word actionable depends in large part on the strength of the data linking the target to the targeted therapy. and that's where i think -- we don't have a lot of functional data.

this trial designed to work, two key conditions -- the tumor must depend on the target pathway and the targeted therapy must reliably inhibit the target. and they said achieving both goals can be a matter of some complexity. somewhat of an understatement.

so i think, i still think we need a functional experiment. what can we do with functional what we have is a retro viral library, and in this paper, actually this paper was looking at this library in lymphoma, but we went ahead and used it in so this library has shrna

targeting 125 genes with at least 12. there are 10,000 constructs in the library. they're all sequence verified. they contain a bar code sequence so you don't need to do an individual aside plate. uh-uh actually just do a pool

because they're all individually bar coded. the library targets all approaching kinases. and then differentially assessed genes in lymphoma because that's where the library was developed. but it does look at apop ptosis regulators and oncogenes.

so the experiment is this. the experiment is what you would library into your clels of interest. so you have your little fhrna into your bar code, in some cases we have the shrna in an inducible construct and we can induce it for 21 days or not.

and then we look for genes that dropped out. some cases we have library or no or library or control in the cases where it is not inducible. but in this case we have an inducible situation. we're looking for bar codes that drop out of the induced that are

present in the non-induced. and then do -- it used to be a bar code but now we sequence the bar codes because that's more envogue. so what we did, we put this library into four different ovarian cancer cell lines. we use two serous cell lines and

two non-serous cell lines. and what we found was we looked for the genes that were overlapping in three out of the four cell lines. we found that 63hrnas representing 55 genes were common to at least three of the so what are these genes?

these are the whole mix of them. there are many. and we did follow up on these in a sort of validation screen in the individual si-rnas with two individual si-rnas per gene. we're looking at things when you knock them out does it block cell growth.

we did follow up in six additional cell lines like i said with the individual firna plates we could follow up with those genes and we selected -- we focused more on druggable targets for drugs that actually already existed. we also looked at specific

subgroups, whether it was serous versus non-serous, et cetera. we prioritized this particular gene because we found that it was -- it was one of our very, very top hit in all four of the cell lines and it was actually over expressed in nearly all of the ovarian cancers in the

so check one, the check is involved, as you would imagine, actually in the dna damage so dna damage occurs and check one actually signals downstream of atm it actually signals to arrest the cell cycle in g2m until it can be repaired. so that kind of makes sense

because if ovarian cancer is having a ton of dna damage and susceptible to that, it would need to upregulate this gene to stop the cell cycle so they can repair their dna. just briefly, one figure from our paper, we found that actually the high-grade serous

line were much more susceptible, had a lower ic50, multiple other experiments and eventually ended up with a clinical trial. we are showing remarkable response rate and i just had another patient in clinic today that had a confirmed response almost complete eradication of

her cancer. highly pretreated, many prior therapies, refractory to platinum but is responding to the second inhibitor. highlighted by a functional genomic approach. so, to summarize, i went a little bit fast.

ovarian cancer genomics has revealed that there are multiple areas of potential targeting that are specific to various cell types in ovarian cancer. and the homologous recombination is the big one. not a single gene but a pathway in a high-grade serous cancer.

so as i mentioned in 19d 81 a functional approach discovered her-2. in 2000 and whatever, 2010 i guess, whenever we started the experiment, 2016, a functional shrna library screen found check 1 in ovarian cancer. so if we can do a functional

genomic screen, identify a driver, aberration or pathway, ultimately improve therapy for and with that, these are the women's cancer team. these are the translational scientists. these are the collaborators for the shrna library and of course

the funding is from the national cancer institute. i will take any questions. [ applau yes? so the serum 125 is a marker it's knot approved for it's approved for monitoring recurrent disease.

the cm125 marker is interesting. we just had a very large study for using that for screening. it's not very effective and the reason why i think it's not effective is because it's looking for serous cancers. it's not looking for all so when looking for -- we're

looking for a marker to identify all ovarian cancers and we're using a marker for serous in my opinion that's why it does not work as a marker. if we were looking for a marker, i think it would have to be -- you would have to be looking for specific cancers.

and i think those studies should be ongoing. right now they are not, as far as i know. another opportunity i think would be looking for circulating tumor dna if we knew a recurrent mutation like the arid1a. those are not ongoing as far as

i know. it is a dilemma. because right now the symptoms are really just kind of gi symptoms, they're very nonspecific and that's why people don't report -- nobody is coughing up blood or that kind of thing.

it's very difficult. but i do think we need novel approaches, maybe circulating dna, maybe more specific looking at different cancers. >> how many subtypes of ovarian cancer do you think there are? >> oh, my goodness. i think all cancers now are

getting to a very personalized level. so even all triple negative breast cancer is not the same. so everybody is subgrouping, subgrouping, subgrouping. so in some sense, you have to lump in some sense, because you have to say who should get this

therapy. so maybe people with a check amplification should get the check inhibitor. but there are certainly hike subtypes within high-grade certainly there are the five histologic subtypes. everybody doesn't respond to the

same therapy. so we're not getting 100% response rate. and until we get a 100% response rate, then we can't really say we're treating them all the same. >> oh, yes, yes. i think there's a lot of

promising research going on. so in the clear cell cancers, what i didn't mention is they have a very vascular -- they're very vascular. so all the androgenesis inhibitors are looking promising for the clear cell cancers. and then i mean people are

definitely looking at the pi-3 kinase pathway for the endometrioid cancers. there are many avenues that i didn't talk about. yes. the brca1 and 2 the hot topic is the harp inhibitor. and looking at the homologous

a lot is going on with respect to looking at seeing if we can find other people that have a braca-like homologous recombination defect that might respond because there were some patients in the early studies that didn't have a braca mutation that did respond.

so that's kind of the hot topic and then -- what else was i going to say? i lost my train of thought. so within the braca field, i think that is certainly the area to focus on. all right, thank you.