Cancer Of Lung

>> so we're going to get started. it's a beautiful fall day outside. our first speaker is from the cancer of epidemiology and genetics. it's neal caporaso and the title is epidemiology.

neal, take it away. >> thank you. great to be here and welcome all those watching electronically. the title of this is epidemiology or why the governor of new jersey should not make it a public policy.

okay. so just by way of introduction, i work at nih specifically at the national cancer institute. i want to point out the other institutes are in epidemiology that is in population studies. and of course we are intramural. 85% of the cancer institute

funding extramural. so we are uniquely intramural and our division is focused on cancer epidemiology and that is the cause of cancer. and our branch is the genetic epidemiology branch. we focus on genetic components to the eat -- etiology ofcancer

but we have hormones, infectious disease, occupation, statistical methods and radiation. i'm going to give you a population perspective on cancer and i'll give you a few foundations. i'll highlight some tools that epidemiologists uses.

i'll tell you some of our challenges and i'll point out some future directions for this area of science. and i hope to run through this lecture a little bit faster than usual and i will certainly leave time for questions at the end. so some introduction concepts.

epidemiology is the science of the causes of health and disease in human populations. so, there's a derivation here but a key point is that epidemiology is an observational science. unlike experimental scientists. you don't get to pick in

epidemiology who are your expose people or unexposed unlike an experiment of an animal laboratory. people out there make choices about what they're exposed to. they choose to smoke or to consume alcohol. and so there's always this

factor which introduces some bias into epidemiology study designs. but we'll talk about that a little bit. even -- epidemiologists are prohibited from doing experiments on people. we observe large populations,

divide them into groups and assess their outcomes or statistical methods to relate those to interests. it's important because in the study of lung cancer and i'll go back to lung cancer often as a key theme. the idea that tobacco was the

cause for lung cancer, was vigorously fought by the tobacco companies. so what are the goals of epidemiology. we want to understand the causes of disease. once we understand the cause, we want to quantify those risks and

want to identify risk roots. we want to understand something about the mechanisms. it's not enough to say tobacco causes cancer. what is it in tobacco that causes cancers? we use epidemiologically search to gauge public health efforts

and costs and budgets. and we'd like to identify syndromes. epidemiologists typically emphasize prevention. it is much cheaper and more prevent a disease than to treat a disease. this presents a paradox in that

politicians who are the ones who provide the funds to deal with public health find that curing disease is a much more emotional appealing and immediate form of gratification. so you say we have a new cure, you tend to get funding much faster than if you say look, if

we have saturated in father of hamburgers by 02%, in 30 years we'll prevent 100,000 coronary or farctions. that may be true but it's just not as appealing been so politically it's a little built of a cause done number -- conundrum for epidemiologists.

as a side of effect not being able to do a an experimental design where i place the rat i exposed in one cage and rats i didn't expose in another. some are exposed some people are not. i'll give them a questionnaire. it may not be accurate.

for a lot of reasons epidemiologists are very concerned about different kinds of bias. and example is selection bias which means that the people in your study may not be representative of the entire population or the group that

you're most interested in. and this results in a problem because then your results may not be generalizable to the larger population. so remember years ago we did a large case control study in italy, a study of lung cancer and i mentioned before we're in

the intreu mueller program -- intramural program and they came and said we'll spend $2 million studies if you study lung cancer cases and controls, are your controls representative. and it's well you know, if you don't have at least a 50% participation rate, we're not

going to let you do the study which is a pretty big deal. so we did a phone service. we called up people on the phone and said hey if we give you $20, would you like us ask you questionnaire. it will take about an hour and give us 70ccs of blood, four

and-a-half tablespoons. what do you say. well less than 30% say they'd participate in that. no study. so then we isn't them an invitation letter. we followed up by phone. if they were in the hospital we

went there, we made advertisements. we increase the cash awards. we get a letter from their physician and study them either in the home or how much wherever it was more convenient, response rate 49%. still too low.

so then we got extremely charming and personable interviews. instead of money we gave them a gas coupon which was a big deal in italy. we had tv ads who was the johnny carson of italy at the time. we got a letter from the mayor,

we got an invitation letter that was written on a sixth grade level and we give them a toll-free phone line. at that point we reached the response rated we needed. this took us a year and i don't want to tell you how much it costs to do these kinds of

studies. but this is required to get a population study that's reasonable. controls. if you ever do a population study, you're going to hear about your controls and this is because many small clinical

studies looking at something or other will say let's just use the folks in our lab or those standing around here. those called convenience controls invariably introduce bias and are problematic. in you consult with your epidemiologist, he's going to

look at your controls and say they're no good, you need new ones. population controls are gold standard and this is because they're most representative of the population from which cases derive so they're most likely to give you valid results.

there are also some other bonuses. like you can calculate absolute risks instead of relative risks. and you reduce selection bias. so if you decide you want to do a population study for some reason, you go see an epidemiologist as your

consultant. they'll ask you about your design which is the first question. we'll focus on your control. they want you to collect covariate data. so this is super super imimportant.

covariate data just means factors that can influence either your exposure or your disease. and so these, if you don't take them into account in your design or your analysis, and there are two opportunities to do that. your results will be distorted.

power calculations are super important. your study must have the right size to be able to make statistical inference about what you want to study. and there are simple formulasto do that. if they include bio markers,

they'll ask how did you validate these. this is critical. one other point i can't help bring up the most common economy over many years, that is i have a grandmother who smokes and she drinks and she never exercises and she eats buttered toast with

sauce -- sausage and she has buried three of her doctors and how do you explain that. this is the fact that epidemiology is a probable science and not an externalness anything can happen to any one individual but we're worried about a population.

and so it's probabilistic first it's a deterministic issue and scientists are not as likely to ask this question. but when you speak with the general public that's here. so okay. so what kinds of tools do epidemiologists use.

one is a cancer map. this is something we used to do a lot more of but it's a very handy way to look at the distribution of disease and say something about risk factor. here's melanoma and i think it's pretty clear that you can see that the red area indicating

high levels of disease is across the southern tier of the country and that is concordant with the fact that some light is the key risk factor for melanoma. cancer maps are evolved into something else which is gis systems which is geographic information systems, which are

satellite-aided methods to establish databases and infer risks and learn something about exposure. another tool that epidemiologists use is seer which is surveillance, epidemiology and end results. this is covers a quarter of the

population and gathers data on incidence, survival, patient demographics and features of tumors. there's a too many commonly used and there's a website and you want information on a specific tumor type, you can get a lot of information from the

publications. i will not spend a lot of time on this but if you look at your data, you'll see a lot of interesting things. and here's an example of a bump in the cancer rate that was observed in men but not in women.

and so you wonder what exactly was going on here. and here it is. we break it down into specific cancer types. you can see that the other cancer types are relatively smooth but prostate shows a big bump and this is due to psa

screening which caused overdiagnosis of prostate cancer. so in a sense this was an artifactual increase in the rate of prostate cancer. so you always have to take into account diagnostic modalities when you look at changes in

cancer rates. and then epidemiologists say well if we're being to looks at a cancer rate like an incident, at the same time maybe we should look at mortality to see that those two are linked to make sure that it's valid. so here some rates in women,

that is the constant. here are some rates by race, and you can see that cancer in white men and african men. and i'd like to point out that the overall rates of men and women seem to be leveling off. and this is one of the last millennium.

and why is that? and this is one of the basic and fundamental features of this is because of the decline in smoking. and so you can see that what has happened over time is that lung cancer rates are rolling over. of the other cancers are

relatively stable. some cancer's been declining for decades. but what is driving down cancer death rates is the fact that we're getting less lung cancer bond we have blood snarking in the united states. i'll make one or two more points

here and that is if you have incidence much higher than mortality, one of the features that that could tell you is that you have a tumor that's responsive to treatment. and that is exactly what we have for a variety of pediatric cancers.

especially leukemia. so one of the issues in epidemiology is that we find associations. statistical associations. and a key question is so you found an association but we all know correlation does not indicate causation.

so how do you prove it. what types of evidence do you pull together as an epidemiologist to inculcate around association as a cause. there are a number of them. first if something is causal it helps it confers a higher risk. sack it should be consistent

across different studies. third, you should see a dose response. fourth, there should be a correct temporal order. the cause should come before the disease, obviously. but that is an issue in term of study design.

and finally, we'd like to have a mechanism. the biology should make sense. so here is some of the data that was marshalled to establish that tobacco was the cause of lung first of all in cohort studies you see remarkably consistent data in that the more cigarette

smoke the day the higher relative risk of cancer and it's extremely consistent across three very large cohorts. you also saw the temporal relationship. so as males began to increase their smoking, in the start of the 19th century, you saw no

smoking rise and after a lag of 10-15 years lung cancer increased as well. same thing with female smoking. you also saw animal studies, and these were studies by oscar auerbach who was inly and did studies? smoking beagles.

and he showed that the same changes of dysplasia and meta plegia in the epithelium and this was the kind of experimental evidence that was considered a very compelling complement to the population data. and finally, a very important

piece of evidence, you saw that in the cohorts, when you establish that somebody quit smoking, slowly over time, their relative risk of lung cancer steadily and consistently dropped so they osmototically. this is a nice piece of evidence that shows that tobacco causes

it. what are the accomplishments. identification of the general specific causes. role of advocates. eliminating the cause of cancer. i used tobacco as a causal factor in lung cancer. and i'm going to say something

about secondary tobacco smoke because that's something we all benefit from every day. so it was studies of epidemiologists that established that secondary tobacco smoke was a risk factor for cancer which resulted in a lot of the public health laws which keep people

from smoking in places like this room. and for which i'm sure we're all grateful for. and finally the rise. i'll not spend time on the general or specific, something you can certainly ask questions about or you can look this up.

it's really, but i do want to touch that age, environmental factors, genetic factors and combinations encompass the general causes of cancer. and roughly in a third of cancer due to tobacco, a third due to diet and a third due to pretty much all other factors combined.

most cancer is due to the environment, and the evidence for this comes from a broad international studies where we look at international variation in cancer rates, and what we see is that the high to the low ratio is strikingly high across different populations.

now, you may say well neal, that could be genetics. the different groups are widely disparate and maybe some of them have at risk genes and others don't. and genetics may contribute a certain a in some cases, but another class of epidemiological

studies show that it's mostly the environment. and what kinds of studies are those? those are migration studies. so when people my great from a high risk area to a look risk area. within the generation they

acquire the low risk, the rate the low risk rate. so this is an example of a high risk area in china xuan wei province. our group had done studies there. what's the risk factor. it turns out to be in door air

pollution from burning smokey coal and in door cooking. and oftens that are place -- ovens that are placed under the bed. this is a very high rate of lung cancer in this particular region. another example is a copper

smelter in montana and actually i had a cancer map which may have dropped out but on the cancer map there was a little red area in montana and that particular copper smelter spewed arsenic into the local environment and arsenic is a lung cancer risk factor.

so the smelter was eliminated and the hot spot over time disappeared. so back to tobacco. tobacco is the, still the major cause of preventive morbidity and mortality. rates in the united states are slowly declining.

worldwide they're increasing. this means these going to be more lung cancer, more cases of lung cancer over the next century. tobacco cancer a fifth of all outs -- u.s. deaths. all the cancers caused by

tobacco are difficult to treat. terrific costs, and inspite of the fact that of lung cancer, sorry, the tobacco is an extreme cause of our health risks. almost a fifth of american adults continue to smoke. and teenagers continue to take up the habit.

this shows the per capita decline in the use of different forms of tobacco. but this has now somewhat leveled off. and i told you i would say a word about environmental tobacco smoke. the first study was

never-smoking spouses of smokers. as they established that the relative risk to these powpsz -- spouses was in the range of 1.2 up to 1.6. little by little evidence accumulated.

there were meta analysis, and this, these studies as they accumulated, indicated a number of things. first was that there was no lower level of risk. in other words, even though you were exposed to not that much, you still acquired some risks.

so having a non-smoking section in an airplane with a little fan really didn't help you that much. you were still being exposed. and if you were a pilot or a stewardess on that plane, you were certainly going to encounter terrific risks over

time. so these kinds of studies began the movement to create legislation, clean air legislation that eliminated smoking from the workplace, from schools, from hospitals. and eventually even some restaurant bars and many other

public places. and so we really have these epidemiological studies to thank for the fresh air that we get to breathe most of the time in most places. so i'm not going to go into many of the other risk factors but alcohol is the number two

carcinogen after tobacco and it's associated with a number of ionizing radiation is a very well-known form of carcinogen associated with a number of but then a lot of studies a-bomb exposed individuals exposed to medical radiation. those exposed to radial oh

newcleoide. s and overall risks particularly conference. this is definitely a risk factor for skin cancer. one controversy is that some might think this enhances your vitamin d production. low vitamin d is a risk factor

for a number of cancers, somewhat controversial. on it out weighs the benefits. this is associated with a number of cancers. i won't go into it. there's been a lot of study recently in our group looking at hpv and oral pharyngeal cancers.

this is an area where molecular epidemiology placed a key role and essentially established that the cervical cancer, 100% of curve -- cervical cancer except for pathology is due to hpv we've done studies look at other infectious examples. this shows no association with

lung cancer but shows with gastric cancer. the design is critical there because if you do a case control study and look at tissue at the time of diagnosis all gastric cancer, the effect of the h pilori to the degree you can't detect them.

you get a case control study to study at the time of the disease it will be difficult to associate h.pylori and gastric lately micro biome study are looking for associations and the one is is -- colorectal carcinoma. s they are in the infancy now

but over the next few years after studying the skin micro bi-op we'll have more study. the major cause of human i guess one thing i would say there is that epidemiology has the issue that many of the things we study are strongly opposed by commercial groups.

so, as a diesel groups that produce diesel fuel are strongly, strongly fighting studies like this that look and so they have subpoenaed the records of our occupational studies group. they've called them in front of congress.

they've done all sorts of things to harass these studies. and so in some ways, epidemiologists, and particularly epidemiologists that have a little bit of protection working for the federal government are is against a hoe rendously

complicated environment where they can do anything they want and spew anything into the environment and degrade our health. so this is a continual battle many people are not aware of. but our occupational studies group is using lawyers and

fighting off the efforts of the commercial groups to stop their study. a little bit extra rest. and a candidate or agnostic study. in the old days we said well, i think that this specific gene is the one that's the cause so

let's look for it. but today we do genome wide studies so we try to cover the entire genome and look at all of them. this is really what has revealed the genetic basis from those and why look for an inherited basis to cancer in the first

place and the reason is that family studies clearly show that risks of virtually every specific cancer, and i showed you an example of lung cancer here, are increased in relatives of cases compared to relatives of controls. so having a family member

diagnosed a lung cancer increases or odds ratio 1.75, that's 57% increased risk. and that takes into account other factors such as age, gender, where you're living, your education level. a lot of detail on smoking, etcetera.

and genetic epidemiology looks at these genetic factors. i told you already if you want to look for rare genes by themselves confer high risks, you need families. i also told you here's an example of one of our privilegive disease families

with a lot of cll. and to date we have not found a gene. we're looking very hard. a lot of tumor suppressor genes were clones. so now i'm going to move ahead. this just shows you an example of all the different low

penetrate genes that we identified the genome-wide association studies. i'm going to move ahead here over to lung cancer and tell you a little bit about molecular deem all and then i'm going to move to some future directions. so the most basic kind of

epidemiology is looking at an so is association in a diseaselike tobacco and a disease like lung what we have done over time is evolved to a system where we don't collect the exposure data but we also try to collector bio specimens so that we can look at the steps that happen between

the exposure and the disease. so first we measure an internal dose. so we save the cigarette but when we measure it in your blood how much do he with a really find. and if we look at dna awe ducts which are the carcinogens to a

molecule like hemoglobin or dna, how much do we find there. how much actual dose do you get. and if we look at cells in your respiratory tract, how altered are these cells. and finally, can we find evidence of inside to cancer in those with exposure and disease.

so by interpreting all these components, we help to elucidate the mechanism. the idea behind therapy. so the study i told you about in oathly we actually got tissue as well. we'd go into the operating room and wait for the receptive

tumor, give it to a pathologist, the pathologist would cut us a piece and put it in liquid nitrogen and that's an incredibly valuable resource that we have today. and we of course would ask the questionnaire and going after diet we would of course gather

all the dietary information on what they eat but we would also exploash other features like doneness. we wouldn't ask them how much do you cook your meat because that was a hard thing for people to answer, we show the pictures. does your meat look like this or

like this. so we'd get a lot of information on that. molecular epidemiologists contributed a lot and it really would take almost another entire lecture to talk about that. i'll just focus on one thing, and that is that pretty much all

of the g was studies that have established the genetic component to virtually every tumor has been based on biospecimen collections into cohorts. the idea that we proposed in the 90's that you can't just give people a questionnaire, you've

got to get a tube of blood or vial urine and put in thefreeze i been and when you develop the assays we study them and get the dna from it look at the that's exactly what happens and it's a really important e vent. one other kind of epidemiology you'll hear about is integrative

what this says is it's not only the exposure but it's the behaviors that make you get the and in fact, tobacco smoking is so has a genetic component and we should study that. so an example of this, and i don't think i put in a slide to show you, is that we look at

nicotine dependency and we have a paper in jnci that came out a few months ago that says that oh, your time to first cigarette. how long you wait in the morning to smoke your first signify rit indicates how addicted you are in nicotine and if you are very

addicted and smokes ten cigarettes a day you absorb a lot more carcinogen than if you're not so addicted and smoke ten cigarettes per day. and that turns out by itself to be a powerful risk factor for lung cancer. at the other end of the spectrum

we look at outcome. so we know you have disease and you're alive when you have the disease, but how long do you survive and how do you respond to treatment. and this particular area disease to outcome is one where there are risk factors as well.

so for example, if you're a smoker here, your outcome is a lot worse. that's called integrative epidemiology and we put all sorts of questionnaire instruments to dissect the behavior part. and we incorporate treatment

survival and prognostic and clinical information at the other end. one of the ways we do these big studies is through consortia. that's why we think the cohort dr. collins is proposing might not be the answer. an alternative might be hey

let's gather more information on these folks because we already have the cohorts already. we just have to add components to make them work better. i'm going to skip ahead and i'm going to tell you about some future components to epidemiological studies.

because as i told you, we don't have a lot of the exposures that contribute to cancer. and we'd like to incorporate new technology that don't cost a lot or are non-invasive, are reasonably accessible to people, can be validated and used in large populations.

i'm going to cover eight of these in about two minutes because i want to have time. the first is sleep. sleep is clearly a risk factor for a lot of conditions. obesity is one, mental illness is another. and there are now, it's rather

difficult. there are some validated questionnaire instruments that it's difficult to get a good sleep history. you can give people a very many phone and new technology and put it under their bed and you can get the stages of sleep, the

time you spend in each kind of sleep the total hours of sleep. the movement during sleep. and all these factors are interesting ones to look at in relation to disease that haven't really been examined much in relation to cancer. physical activity.

this is a difficult area to ask how active are you. i'd have to ask you 50 questions, do you go to a health club, how about if you're vacuuming 9 rug or walk to school over day. their many ways to have physical activity.

so now, and there are a lot of technologies from fit to the new generation of apple and samsung software which can automatically collect different kinds of fitness information. use this for run keeper. there's a variety of them. little by little these are being

refined to the point where they can calculate not only activity but periods of inactivity. inactivity may be as important as your activity. calories, climbing steps, heart rate can be calculated with the number of electronic means.

and heart rate, heart rate visuality, arrhythmias, maximum heart rate, how you respond to diet or exercise are important. social factors. so it's become clear that who you interact with can have a big impact on different risk factors.

so smoking and obesity are two factors that have been shown to be related to who you're closest to in a social network in the framing ham study. this is an area that's increasingly of interest. location. so you saw what i showed you

earlier but your location can be related to a lot of other features. everything from local economic factors, local exposures, socio-economic status, zip code reviews by politicians to targets the kinds of ads in an increasingly sophisticated way

to get people to vote the way you want. and similarly it's a way to characterize cancer risk so smoking. so there's a lot of hunt for smokers using different applications, but there are also new kind of software which can

look at the movement of your arm and so not only determines a number of cigarettes that you smoke but how long you inhale and different features of how you handle the signify rent which can be related to weather and climate. a lot of cancer types are

related to features of vegetation, pollution, nutrition, water availability, insects, dominant air masses. and so incorporating climate would be something on the population level that will be and finally circadian variation. most of you are aware there's

been literature for decades that have looked at circadian variation called crohno therapy. they are thought to be at risk for a variety of conditions again. obesity is prominent but breast cancer's one that's been frequently mentioned.

in animal studies when you do various things to disrupt the circadian rhythm, you get implanted tumors that do better. so ir has determined that the variation is a risk factor for but these studies in human pop laces tend to use groups that are shift work or have their

circadian variation disrupted by error line stewardesses or individuals who are blind and like the cue from sunlight that resets the circadian. so recently we have proposed that a way to determine where you are in terms of body is to assay ma tabalomic.

we have markers from different times of day and we propose to use that for markers of circadian variations. that is probably enough and i will stop there, and i will be very happy if anyone has any questions. thank you.

i didn't say a word about clearly epi genetics is abridge between genetics and the environment. there are studies ongoing looking at a variety of indices of even -- epi genetic changes. anything that uses tissue epi genetics is incorporateinto it

and is going to be important. a recent example of a study like that e data and genetic data to identify any qtl's that are related to four of five -- lung they alter using methylation. that's just one example and there's a lot of information incorporated.

i'm sorry i didn't mention that. >> so there's a little bit of a paradox in something we don't understand about lung cancer in other countries. frankly it hasn't been studied so much. they often discuss the indian paradox where there's a lot of

vegetarian diet and yet rates of heart disease is very high. there are other cancers that are more common in india. it's a similar situation perhaps for china where there's pollution is a major source of but the relative risk associated with tobacco is lower.

so current smokers have a relative risk up to 0 -- 20 inthe united states in china it's lower like 5. in ind i'm not familiar with all the details in india but there's been a dearth of population studies in india. there really needs to be a lot

more research in that area especially the dietary front. because there is such variation. some vegetarians but there are also a lot of diets that more resemble the west. and so people have said well could be a lot of the spices that are used that convey some

protective effect. i thin it's just under studied. yes? so a number of years ago there was a large study which looked at cell phones -- and didn't find anything. a lot of these studies are difficult to do.

they've done a power line stud which is an area that's frequently brought up that low power electromagnetic raped radiation could have an effect. though studies have been negative. they've been criticized. anyway.

thank u. the. >> we have an announcement. after this next lecture, we're going to be going over to the pathology core to building 10. i will be walking people over if any of you are interested. and we have a schedule change.

we're going to have perez hussain today talking about pancreatic cancer. title is pancreatic cancer understanding d future challenges. >> thank you very much teaching your lecture with me. i appreciate that.

what i'm going to do today is to give you a brief overview of pancreatic cancer and some of the current challenges and current understanding and future challenges to improve disease outcomes in the patients with [indiscernible]. so let me start with this slide

which shows some of the grim realities that are associated with pancreatic cancer. it's the fourth leading cause of cancer death in the united states. median survival in advanced cases are to six months and it's an estimated for the 6,420 new

cancers [indiscernible] in 2014 alone. as you can see, everyone in 2014 who died pancreatic cancer the majority of them died within the same year. there's no effective treatment that is available for treating this is an estimate that was

reported in cancer research has indicated that pancreatic cancer is going to be second leading cause of cancer related deaths by 2030. as you can see here surpassing liver and colon cancer. but the previous length risk factor that includes exposure

and also inherited syndromes and family history. one of the major risk factors is smoking. it has been shown that smoking strongly associates with risk of getting pancreatic cancer. in addition [indiscernible] they are at high risk of pancreatic

and then in relation to genetic syndrome enhances the risk of getting pancreatic cancer by 50 fold -- alteration in urc01 -- two genes. and also the syndromes that has an alteration -- in addition this syndrome and alteration so what we have achieved so far

in terms of treating advanced again it's a disappointing progress in the treatment of in 1997, more than a decade ago -- was approved as a standard care of drug over -- the genes enhanced survival in advanced cases just by a month. then after a decade, the egfr

inhibitor was used in combination -- that increased the survival only by one and-a-half. the situation is so bad that it was approved by fda. then recently in 2011, a combination of -- which is called -- improved the

survival -- by about four -- that was highly toxic. but recently combination of albumin -- showed better survival as opposed to -- alone but only by two months. but this was less toxic. and they in fact used the information from the tumor

biology that -- high level of protein and albumin -- and so the drug is available to tumor. so as we know about the outcome in pancreatic cancer because of the diagnosis -- you don't have any protection marker that is specific or sensitive as but in about 18-20% of the

cases, they are identified at early stage and qualified for physical resection. but in the cases at earlier stain the medium survival is less than two years and the percent of these cases that show -- within two years. about 20% of these resectable

cases reduce the life of five years or evening longer. there are some clinical prognostic factors that are associated with survival. for example -- how are these not consistent because sometimes you will see that a stage one patient resected module zero but

they're -- one of the hypotheses is -- in tumor that determine patient's outcome. so how we can improve the outcome in pancreatic cancer patients. obviously we need effective therapy that will require identification of novel

effective targets. more nuclear subtypes so therapy can be used for that subtype and that includes and also including drug delivery. i'll show you an example that have taken advantage including the risking drug to the tumor, that has done better in at least

animal models -- that controversy. and then you need only detection marker. and only detection marker that can identify precancerous lesions and not the cancer because once the cancer is developed again the survival is

so to achieve this, we need to understand the pancreatic tumor biology which is the key to -- let me tell you a little bit about pancreatic tumor genetic alterations and that because of the leases. there are physical alterations. one of the most important one is

k-ras. you've seen more than 90% pancreatic tumors have actually -- and about the same 80 to the 0% have the p16 alterations. about 50% of the tumor you'll see -- and about 50% for so the most common lesions are

pancreatic neoplastic lesions or panning. and they have been categorized at panning one, panning two, panning three. and there are different timing of alterations in these pancreatic precursor lesions. one of the other important

characteristic, it is quite interesting in pancreatic cancer, is a huge inflammatory stroma. as you can see here in the infection. these are the pancreatic tumors carcinoma and all of here's imbedded in huge stroma.

this is a stroma cell first a fibroblast, pancreatic cells -- number of different group factors cytokines and chemokines. and they are responsible for really active tumor stromal interaction. the other important feature of

pancreatic cancer that it is highly heterogeneous. there are 68 on average genetic alterations in these tumors. and that constitutes alteration in the core signaling pathways. and also these signaling pathway was altered -- they have different gene alterations which

altered that particular pathway in two different tumors. so it was highly heterogenous. this they analyzed the gene expression in several tumors, they found that there was they divided the tumor into three subgroups. what they found was in fact they

had different survival. and when they did some of the envy trust studies where they have the same gene expression as these three tumor subgroups, they responded differently to the drugs. one other interesting observation in pancreatic cancer

as in other tumors, they alter the tumor metabolisms a requirement of the energy made for the tumor. there are several alterations that -- one is that increase glucose uptake and also enhanced glycolysis which is used to activate some of the enzyme of

the non-oxidative pathway of phosphates that enhances. they also have -- lose literally not only -- there is also alternate pathways with the use of maintaining -- the tumors. so i want to go more into details on that metabolism part. i want to go more into deal

about the complex of stroma because a number of attempts have been made to target the stroma to achieve effective treatment. so let me give you a brief overview what happens in the pancreatic cancer stroma. these are the tumor epidemiology

cells and the met kindal cells. there's immune suppression. and immune suppression happens because of the [indiscernible] tumor association macrophages -- of the tumor types. and this inhibits the cba they inhibit the anti-tumor. these tumor epithelial cells,

they secrete a number of different cytokine and chemokine but support the migration and attraction of these immune suppressive cells. then there's the pancreatic tumor cells which is stimulated-- these are very important in maintaining the stroma.

what we're interested is in these -- that are so much accumulation -- enhances the interstitial pressure on the blood vessels. and blood vessels are compressed. when they cannot be transferred of the drug.

i will show you one example so that what happens when they relieve the pressure, the drug into the tumor and then they better survival. so treatment of casualties to include disease outcome. the first is drug delivery copy fectiveness of systemic therapy.

these are been done on engineered mouse model mu indicated k-ras and mutated p a 3 which are specifically activated in the pancreas. very common genetic engineered mouse model which is known as kc model as you can see here. and this capitulates the stages

and progresses of ductal adenocarcinoma in humans. so what they did in one of studies by ken olive they inhibited hedge hog, they completed a stroma and improved survival in mice. so as you can see here that the mice that got hedge hog

inhibitor, ipi926 and also -- they had much better survival as compared to the -- so they are able to enhance drug delivery by depleting the stroma and the end result was [indiscernible]. but unfortunately this clinical trial has failed and it has not improved survival in patients.

so now coming back to decreasing the interstitial on as you can see here because of the accumulation of interstitial pressure which is much higher than the vascular pressure and it is compressed. when you use -- it relieves that pressure and you can see the

extension of these vessels and the drugs can go outside into the tumor. and here they have use both -- in addition to the -- and what they found again was enhanced survival in the mouse -- and this trial is ongoing and this is categorized as the drug -- so

hopefully we will see some better outcome with this approach. but it's not that simple because stromal target always have beneficial -- there are cancer cells in 2014 that showed quite -- lab here in cancer cell.

they also use -- that became more aggressive. and they completely had so much hedge hog as a tumor -- using the jetally engineered mouse model so you can see here survival -- is high and also -- is high and also the number of lesions that were higher as

compared to the pkc -- marker to trace -- another paper that is consistent with these findings came from -- and colleague. and they used genetically engineered mouse model that was expressing -- is the indication, indicator of activated -- and they also found that both in

early -- carcinoma they found high pathological scores here. it's high as you can see here. there are more differentiated so they are also found that the mouse has overall survival. it needed myofibroblast. so tumors in one interaction is complex -- stromal needs

caution -- pancreatic cancer. so it's complex and it's more investigation maybe there will be some tumor subtype that benefits from the lesion but some may not. so let me give you the last part of my talk so i will not go all the way one hour.

and that is the association with inflammation. this comes from several of the -- both epidemiological and molecular studies. one is that i mentioned chronic pancreatitis in pancreatic you see a huge inflammatory active stroma with a lot of

active inflammatory cells. always have a constitutively active signaling. you see a number of different inflammatory cytokines that are at high level. and it always shows an increase expression of cox 2 and nos2. we have an accumulate

ion of inflammation [indiscernible] you see the accumulation of inflammatory changes. it's very interesting streefdz of -- in journal of clinical investigation have shown that it's not only the mutated -- it's the -- active level of

and all of the k-ras does not have the oncogenic potential unless it is maintained at a level by inflammatory mediators. and what he suggested is that and he showed evidence that there is a feedback so that activation of mf kappa b and generation of inflammatory

mediators that maintain the highest level of k-ras. oncogenic k-ras. i can tell you a small piece that we are doing in our laboratory at the cancer institute. so there's an inflammatory mediation cytokine which is

expressed in -- inflammatory cells and inflammatory responses. increase the special -- in many tumors including pancreatic it enhances a number of different oncogenic signaling pathways. for example in kappa b --

enhances cox 2 and also -- pathway. and there are other functions it is considered as a connecting inflation of cancer. so the hypotheses is this contributes to pan accurate excancer progression and predicts these outcome.

what we found is tumor exposed the higher level of myth as compared to the normal -- non-tumor. and what we also found is that the tumor that express the higher level of mif, the patient arrived had a shorter survival as compared to those patients

with humor was in low level. and this we also analyze by both variate and multivariate -- independent of other providers. and then the availability of these findings in other cohorts, and again you can see the higher expression is associated with survival in the pancreatic

carcinoma cases. so what mif is doing. so one of the important even in pancreatic cancer progression and in cancer -- the paper in 2012 showed that inflammatory in pancreatic cancer -- mouse model can enhance -- so we get mif induces epithelial to

mesenchymal transition. and we did some interesting studies using pancreatic cancer cell by expressing mif in the and what we saw, this is a -- that we saw a decrease -- and increase in -- the marker of emt. and this is the broad analysis.

and then for the visualization as you can see in that expression a controlled cells. so then we are a step further. we use the -- mouse model to see increase in metastasis and tumor growth in these orthotopicmouse models. when we looked at the tumors --

was poorly differentiated. and the control, they are then moderately -- and found that there was achieved change in global gene expression profile and it showed the over expressing tumors i showed you. so what happened to the new removed mif from a genetically

engineered mouse model from that was the next question because the idea is can you introduce mif to treat so first we look at the model and as you can see it's expressive high level of mif in the tumor -- dozens of express mif.

what we found is this enhances survival significantly. enhancement in the survival. so now we are investigating the mechanism what is happening, what mif is doing and what mif deletion does that enhance survoifl. and the next step is how to let

the mouse [indiscernible] and then treat the mif, inhibit the so there are different inbitions available -- and then currently anti-mif anti-bodies is in clinical trial for several different solid tumors for the one trial. and -- collaborating in the

preclinical model to see if small molecule antagonists have any effect on the pancreatic progression. so in summary what i've shown you in our study that the higher mif expression is associated with poor outcome making the -- mif enhances growth and

metastasis of tumor xeno grafts and mif deficiency increases survival in kpc mice with little pdac and mif may be a candidate target for designing improvement. and i want to end with this slide where we need to understand tumor biology to

identify better detection marker and the effective therapy by identifying the target. i well stop here and i would like to thank you for your attention even as a second talk after one hour talk i appreciate that and i will be happy to discuss with you any questions

that you have. thank you very much. [applause] that's an excellent question. in fact we are doing at. what we did is we went back to humans and we subdivided tumors in two groups. one with high mif and one with

no mif. and then we have done the analysis for the gene expression and rna -- micro rna analysis and also metabolite. and we are currently analyzing those based on the findings and interest on the genotype. yes.

so that's again a very big it is found in tumor cells and also in many different cells, inflammatory. so one of the important questions that needs to be asked i show you we saw in auto kind but the tumor express mif -- tumors are believed -- effect

coming from the [indiscernible]. the that is one of the hypotheses if you are interested. that's a really good question. thank you so much. and i'm happy that we are going in the right direction with the discussions and the questions.