Clin Lung Cancer

today we're having the pathology course, after six we're going to building 10. hang around. the first speaker today is john schiller, he graduated from the university of wisconsin in madison and got a ph.d. from the department of microbiology

at the university of washington, joined the laboratory of cellular oncology, national research fellow award, post doctoral fellow 1983, senior staff fellow in 1986, became a senior investigator in 1992, now he's chief of the neoplastic disease section.

his title, vaccines to prevent oncogenic hpv infection. >> hello. thanks for showing up in person. i know one person. i don't know the rest of you. it wouldn't hurt if i turned on the mic . what i'm going to talk to you

about is concentrating on vaccines to preven hpv infection. i'm not going to talk about any therapy, but rather talking about prevention measures. i'm going to start by telling you a little bit about the connection between hpv and

cancer. tell you about burden of disease, a little bit about mechanism but i'm not going to get into it too much and spend a good portion of the talk, talking about the current and future prophylactic hpv vaccine, tell you a little bit how they

work, how well they work and many why they work so well and if i have time get into the implementation issues where we start to deal with the problem that we have something that we think is a very good intervention to prevent cancer but the problem is getting

people to use it and getting it to the people who need it the most. so i think you're not getting any other lectures on the association between cancer and infectious agents so let me put up this slide to show you the infectious agents associated

with cancer to point out it's been estimated between 16 to 20% of human cancers are attributed to infectious agents, and what we're talking about today is human papillomaviruses which is responsible for 30% of all cancers that are associated with infectious diseases, mostly

cervical, penile and oral pharyngeal. the burden of disease is disproportionately in the developing world. there's the arrow for cervical cancer, also true for other diseases as well. we'll be thinking about

interventions, we have to not be satisfied with giving interventions just to the united states and highly developed countries but getting them down to the people who really need them. we look at the worldwide burden of hpv-associated cancers we

definitely focus on cervical cancers, it should be obvious here why, because it's both the largest number of total cases annually and also the largest attributable faction, basically 100%, almost no cervical cancers that don't have hpv in it, if you look carefully.

there's about 200 hpv types that infect people, but a relatively small number, about 12, are considered oncogenic. worldwide two types 16 and 18 account for 70% of cancers, in all regions of the world. after that there's a mix as to what comes 3, 4, 5 and 6, but

it's one of a group of about 5 types. if you look in the develop developing world the burden is different than worldwide, especially low resource settings o reasons. first of all, we have really good pap screening program that

reduced risk by 80%. without that we would have sky high rates of cervical cancer. the other thing we've had what's considered epidemic of hpv oral pharyngeal cancers, because of this it's estimated in the next few years, next coming years, the noncervical hpv-associated

cancers will be more than cervical, in gray a portion of these are in men, especially oral pharyngeal, for reasons not well understood about 65% are in men. the types are clumped in one the entire cycle takes place

only in the stratified squamous epithelial. so we believe that this lifestyle is mainly evolved to evade immune recognition. so in the lower levels of the infection is being maintained, there's very low levels of what are called the early proteins,

the nonstructural proteins, not virulent proteins involved in autonomous replication of dna genome, double-stranded, and transcriptional control to keep it tightly regulated. as these keratinocytes different differential rate, the virus response kicks into high gear

and starts making a lot of proteins, particularly the virions, shedding them to mucosal surfaces in the mouth and vaginal tract. histologically, in pink the protein induced, it's difficult to see e-7, expression is in these lower levels here, e-4 a

marker of making virus, you can see just made in the upper in the lesions, especially high risk types, often very hard to detect, a little white region here is hpv 16 infection. so now we can kind of understand cervical carcinogenesis from the point of view of the virus.

people new for quite a while cervical cancer was a sexually transmitted disease because prostitutes had relatively high rates of cervical cancer, nuns had low rates. we didn't know what agent was responsible. now we understand cervical cars

-- carcinogenesis. if you get a pap smear and have a low grade lesion by cytology or histologically cervical neoplasia grade 1, it's a manifestation of hpv infection. the entire layer of the epithelium is proliferating, associated with increased levels

of e-6 and e-7 expression, often integration of the viral genome into the host dna and progression to cancer. why does the virus have oncogenes? not because it's good for the life cycle, cancers are so dedifferentiated they don't have

the right signals to induce production of the virus. okay. it's a dead end for the virus as it is the host. the reason we think it has oncogenes is because it has a problem. trying to replicate its dna in

cells terminally differentiated, that means it doesn't replicate the dna. so it's got to induce the machinery for replication, the cellular dna by its oncogenes, so what it does is has one protein, e-7, that induces induces

aberrant prolifion. know protein e- 6 blocks apoptosis, introducing rapid genetic instability which leads to transformation and eventual carcinogenesis. now, both e-6 and 7, e-6 is 150 kilodalton protein, e-7 is a hundred, but even though they

are small proteins they interact with a lot of different elements, the main thing e-7 does is inactivates rb, if it sits on e2f it prevents e 2f from acting as a transcriptional activator to activate a lot of controllers of cell cycle. so it allows cell cycle to be

induced. e-7 again does many things, maybe even more, but again the primary thing that we think that's responsible for carcinogenesis is that it inactivates p53, it at a ubiquitin ligase to p53 not normally associate the with it,

ubiquit eighted and degraded, and activates tert. the interesting thing about hpv-associate the cancers, generally it happens at a very particular place in the epithelium. this is the place where stratified squamous epithelium

meets columnar epithelium, the transformation zone, the place in the cervix where cancers initiate. now, anal cancer also occurs at a transformation zone and sets the pattern. infection of the penis, vaginal tract, vulva, are just as common

as cervical infections, but they quite rarely go on to become cancers. we think the reason they don't is because there's not a transformation zone in those situations. that's why we think that cervical cancer is very common,

whereas hpv-associated cancers are less common in males because they lack a transformation zone. this carcinogenic progression, this is a picture of the transformation zone from simpler columnar which is red to this pinkish stratified squamous epithelium, it takes an average

of ten years, allowing a long time for detection of lesions by pap screening. the time line for cervical cancer progression is shown here. so hpv infections are very in younger people. lifetime incidence of genital

hpv infection is considered over 80% in the u.s having a genital hpv infection is almost synonymous with being sexually active. very common. most infections clear spontaneously, eliminating the risk of cancer.

if you have persistent infection, this is on the causal pathway for cancer. usually pre-cancers, high grade dysplasias take a decade to be acquired, cervical cancers come out a decade later and plateau. cervical cancer is a cancer that affects relatively young women

so one of the devastating things about cervical cancer oftentimes when women are most responsible for further families is when they get cervical cancer in developing countries. so pap smear screening which i've talked about is what we call secondary prevention where

you actually get an infection, it goes into, if it doesn't you get this abnormality and follow this complicated procedure, if you have pre-cancerous disease or frank invasive cancer you get treatment. we'll be talking about today what we call primary prevention

where you stop this whole process at the very beginning by preventing the initial infection. i hope you can appreciate that stopping it here is a much easier thing to do than to go through all these things, and in cervical cancer screening by the

pap smear although it works well in developed countries with organized screening, in most of the world it's too complicated and too expensive. so moving on to talk about the vaccine, we started working on this vaccine in the early '90s, and at this point what

we knew was that if you take real virus of an animal, let's say a cow or a rabbit, and you inject it intramuscularly where the virus doesn't replicate, generate antibodies, they prevent infection with the same virus type. what happens is with people, we

tried for the hpvs, we tried an immunogen by using denatured forms of the major virus protein which is called l1, for instance made in bacteria, the denatured form, or shorter peptides and they didn't generate neutralizing antibodies. we figured that the key to

generating an immunogen would be have correct l1, but the problem is there was no scalable source of authentic virus for making for instance inactivated vaccine like with hepatitis a. the only place the virus gets made is in termally differentiated epithelial cells,

they don't make virus n replicating cells. we thought let's try to make this l1 by itself an insect cell by a baculovirus. baculovirus expressed protein had been accepted for clinical use at that time. two, it makes a whole lot of

protein, this assembly reaction of 360 copies may be driven by concentration. lo and behold it spontaneously assembled into particles which for all the world looked like the outer shell of the virus, okay? but it didn't contain any viral

genes, the only thing being expressed was the l1 from a baculovirus, didn't contain any oncoproteins. more importantly it looked like the virus and when we injected into animals it made sky high titers the antibodies that prevented infection of tissue

culture cells. this is called virus neutralization. virus neutralizing antibodies is usually the key for most therapeutic vaccines. but since it didn't contai the virus genome, it was not infectious and nononcogenic and

provided a a applause plausible way of moving forward with the vaccine. std vaccines had been tried for herpes, hiv, and none worked, all you're going to do, if you have an infected partner you'll delay infection by two or three months at best. in terms of getting cancer, 20

years later, two or three months doesn't make any difference. so it was very hard to convince drug companies to go forward but eventually we were able to convince glaxosmithkline and merck, fortunately two of the best vaccine manufacturers in the world, to invest a couple

hundred million dollars to try to develop this vaccine. they did parallel independent development. the glaxo vaccine is bivalent, contains 16 and 18 which causes 70% of cervical warts and they use a proprietary adjuvant asl 4 which contains aluminum salts

and this is the first tlr4, the adjuvant is a milestone. merck developed a quadry valent and through in 6 and 11 accounting for 90% of genital warts, aluminum adjuvant made in yeast. they made it in yeast because they are already making

hepatitis vaccine in yeast. in the trials after licensure, the vaccine is delivered by intramuscular injection at zero, one or two months, and six months, three doses, okay? this slide depicts how long it took to develop the vaccine. if you want to do translational

research, the lesson is you better be in it for a long haul, if you really expect it to work out. because the virus was discovered by harold zerhouseen and colleagues, he got the nobel it took ten years to produce the vlps for various reasons.

we weren't smart enough earlier. but from this point even though they were two drug companies working as hard as possible in competition to be first in market, it took 15 years before the vaccine was licensed in 2006. importantly during this time

there was a series of molecular pathogenesis studies, also epidemiologic studies, hpv was the first necessary cause of cancer, this provided the impetus for both developing the vaccine and then using it in the general public. it was very -- it was critical

to understand the association between pre-cancer and development of carcinogenesis by the virus and cancer because we could never have used cancer as the endpoint of the trial. the only way we would have been able to sell the vaccine is if we had a label that said it

prevented cancer. the reason we couldn't have cancer as an endpoint of the trial is two-fold. first of all, it would have taken a 20 or 30-year trial, there's no way you're going to run a 20 or 30-year trial, we're trying to prevent the initial

infection, it takes a really long time for the initial infection on average to develop cancer, at least ten years. if there was active follow-up we would identify the pre-malignant lesion and they would have to be surgically excised ethically. you can't let people get

cervical cancer while you're watching them. the only way to move forward is have a biomarker for cervical cancer, in this case intermediate and high grade dysplasia because everybody believed if you prevented these lesions, you would prevent

and this got moved around, but so these vaccines have turned out to be remarkably effective. with these vaccines everything from pre-clinical days through the clinical trials to the latest ad i'm going to show you later, they have always exceeded expectations.

we really have been lucky. they have been really good. this showed some data. if you look at protection against the most cancer proximal outcome, cervical and intraepithelial grade 3, carcinoma in situ, and you look at efficacy of the vlp vaccine

against incidence disease by vaccine-targeted types, in pele who had no genital infection, women with no genital infection detected at entry, this actually is the closest that we can get to young, you know, young adolescent girls who are the main target of

vaccination. you see the average age here was 15 to 25, so most of these women were sexually active. but if they didn't have any hpv at entry, you can see the efficacy of gardacil, cin iii was 100%, gardacil was 96% preventive at preventing genital

warts. if you look at males, genital warts, protection is worse in men but we think the main reason is that it was more difficult to detect prevalent infection. we missed infections when we tested. it's easier to detect prevalent

infection infection at the cervix than male genitalia. there is some evidence against these other cancers as well. look at infection, we have some evidence against all of them. if you look at dysplasias, in addition to cervix, we don't for

penile lesions, because they are so rare. it doesn't happen often. trials have to be large to power it on that. we don't have it for oropharyngeal lesions, it's unlikely it will be licensed for oropharyngeal lesions but it's

likely it will prevent it. so what the vaccine doesn't do, they don't prevent infection or disease caused by most other types that cause 30% of cervical cancer not caused by 16 and 18. there's cross-protection for closely related types so we don't say they are type

specific. we say they are type restricted. i'm not going to show you that data, it's confusing. the other thing unfortunately they don't do is they don't induce regression of established hpv infections or prevent progression of hpv-induced

lesions. so if you already have your infection, you get vaccinateit'snot going to do anything. it's so important that for vaccination programs, you get the vaccine before you're exposed to the virus. that's when you get the most benefit out of it.

as i said, 16 and 18 account for about 70% of cancers, what about the other types? well, merck has been working on a vaccine which was just approved less than a year ago, and in addition to 16 and 18, 6 and 11, it contains the next five types associated with

cancer, from 4.4% to 2.5%. so if you look at the antibodies -- for this trial, they couldn't use a placebo control, because gardacil and cervix were licensed, you can't denying something of benefit to people in a study if it's f.d.a. approved.

this had to be a noninferiorty study for immunogenicity against 6, 11, 16 and 18 relative to gardacil, the comparison was gardacil 4 and gardacil 9. all they could do for types common is show it's as good as. so they showed immunogenicity was the same, okay?

that was at absolute criteria by f.d.a. then they looked at these other types here which gardacil doesn't protect very well, they showed 96% efficacy by the five additional types, you can add a bunch more types and still get good protection against the new types.

this now becomes the gold standard and it has the potential to protect against 90% of hpv types that cause cervical how does this vaccine work? it works better with coffee. but so we think that this vaccine probably mostly induces sterilizing immunity in people.

this is something that when we started, this was the big kicker. no one believed a vaccine could prevent infection, you never got invection despite multiple exposures. most vaccinees never test positive as measured by assays, when we get breakthrough

infections they tend to appear early in the trials rather than farther on. so why would a vaccine protect better in year three, four, five, six, seven than in year one? again, it's this idea of emergence of prevalent

so what we're doing in that first year is basically getting rid of the infections that were preexisting when we vaccinated. and then emerged during the vaccination period. it's not therapeutic, so we're not protecting against those. most of those breakthrough

infections happen in the first year, after that time protection has gotten better rather than worse. i'll show you why we think that is in a minute. we think antibodies are the only immune protection.

we see cross-protection in the clinical trials it mirrors the antibody mediated cross-protection, cross-neutralization we see in vitro. we can take transfer experiments, put the serum in a naive animal and challenge the

naive animal and get complete protection. that shows antibodies are sufficient because all you're transferring is antibodies. we can get complete protection. the other thing is this l1 protein in the vaccine isn't expressed in the basal

epithelia, so it's hard to get rid of cells that aren't expressing your antigen. that's why we think it doesn't act therapeutically. it's important to note that the various types are individual serotypes, we don't get much cross protection but they vary

in multiple loops, we think people are responding to multiple loops, not one immunodominant epitope. the reason this is important, it's hard for the virus to escape this pressure of antibody neutralization by the vaccine by just changing like one amino

acid on its surface. if you have three or four or five places that the virus antibodies can bind and prevent infection it's much harder to do one of the questions is how does the antibodies that we generate get to the cervix and induce protection?

this is an example of a completely mucosal virus, the only thing it infects is the epithelium, and yet we're not invoking any measures of mucosal immunity. so we think that this systemically inducessed antibodies get to the cervix and

protect by two mechanisms. one, the cervix is an unusual tissue, female reproductive, there's transudacian of serum igg, active transport into the cervical vaginal mucous, it doesn't happen in the gut or oral cavity. that's a break.

more importantly, we have good evidence that in order for this virus to infect it's got to bind, the basement membrane that separates the dermis from the epithelium, and then we've actually studied this. i won't go into any detail. these cervical -- the basal

keratinocytes migrate in and bind the virus and only then can we get infection. the fact that you need trauma or exposure of this region means that there's -- you'll get a direct exudation from interstitial antibodies, infects against a gradient of antibodies

come up a fair number of genital warts, this exudation is how it's protecting, we think that's sufficient. now, the vaccine is remarkle in its ability to consistently induce antibody responses. this is just a study, early study from gardacil where you

can see it's 100% to a rounding error. even more remarkable, this is data where it's going out over nine years, the antibody titers drop maybe 50-fold in the first year-and-a-half or two years, but after that it's basically level.

there's no significant drop in antibody titers after that. we reach a plateau level, generating long-lived plasma cells in the bone marrow pumping out antibodies and we think they may pump antibodies out for the rest of your life. after nine years, 100% of

vaccinees are sero positive at levels at least log higher than what you find after natural infection. this is really encouraging that this is basically everybody gets to this state. the reason we think the vaccine is so exceptional in inducing

high levels of antibody, b-cells recognize the dense repetitive protein array as dangerous microbial structures, the proteins in our bloodstream, even if they are foreign youget binding, week activation signals and low antibodies, most go away. when the immune system

sees this virus-like structure, it leads to cross-linking of the immunoglobulins which form the b-cell receptor on the surface, cross-linking in the strong proliferation signals, high levels of antibodies with durable response. we think the virus-like

structure is the key the immune system hones in on to generate strong immune responses that are going to last forever. so duration of protection, we don't know how long the vaccine is going to protect. we're at about ten years now, as i mentioned there's no sign of

waning protection. and no increase in the number of breakthrough infections. if the antibody levels are the same, the mediators of protection, you would expect that protection would continue to last. that's what we've seen so far.

let me turn to some more recent data that's come out of a trial that the nci has sponsored in costa rica. when these two companies decided to go and do these trials, these were two of the best drug companies in the world, as i mentioned.

we got to thinking is it worthwhile for the nci to run a separate trial? we debated back and forth, they are going to be expensive trials. we're glad we did because we generated data that would never have been generated by the drug

companies. i'll show you one example of this data. we were interested in more than just getting the vaccine licensed. we wanted to understand the vaccine and investigate other parameters.

so what this is, we decided, well, let's look at what happens in the women who get less than one dose. the drug companes would never -- if they get less than three doses prescribed they are out of the trial. we continued to follow them and

we looked at what happened. what this showed you is that -- it's important to note these women were not randomized to get they just happened to get one dose. it's not as strong a data as if you would do a randomized trial but it's very interesting

nonetheless. what happens is if you gave a woman one dose, turns out very unexpectedly they also got a plateau level of antibodies. if you looked at the peek titers there was more than a hundred-fold difference at peak. when you boost, you boost a lot

of antibody response, these are short-lived. surprisingly, if you look at the end of four years, there was less than a four-folddifference there's very little bang for the buck from the extra doses. for one dose it was only 100% at the end of four years, the

confidence intervals with not bad, 79 to 100, we're getting good protection from a single dose and the antibody titers are being maintained it looks like. and we have unpublish -- the data was published but we were looking at unpublished data, out to seven years, we have no

breakthrough infections, and the antibody titers are continuing to be stable. this shows some infection data. if you look in the new time points, between year four and year seven, when you may be thinking antibody protection is waning, you look at three, some

ladies only got two doses at zero and six months, some ladies got two doses at zero and one month, basically no infections between year four and seven, no matter what we used. there was some infections by these three types, we get some partial cross-protection, but

importantly they are still getting a good number of infections by genital types that aren't protected. so why this is important, they are still getting exposed to genital hpvs, not like they are stopped having sex or partners that are infected.

there's not any significant difference between those that got one, two, three doses. so looks like we're getting continued protection. and right now we're working in the planning stage of actually now comparing one and two doses. the reason we're not comparing

three is in some countries there are two doses that's the standard of care for younger girls and boys. if you're over 15. so this would be a four-arm randomized control, comparing one and two doses, and it would be a noninferiority trial,

showing that immunogenicity and protection against infections that last six months there's no inferiority between one versus two doses of vaccine, we think but very important because in terms of implementation around the world if we have to give one dose it's going to be amazing.

even in the united states, i did a calculation and this trial might cost $60 million, okay, or maybe more, but if you calculate what we spend on cervarix and gashed a sile gardacil, tree doses to one dose, you make it up in months. in terms of health care expenses

it could be a huge savings, in two months you would make up the money of the trial. we think it's worthwhile to do this. if we don't do a randomized trial then no regulator is going to license one dose based upon this post hoc analysis we did in

costa rica that i showed you some data from. so conclusion from hpv vaccine efficacy, it got screwed up a little bit on the conversion, the vaccines are highly effective. gardacil is also protective against genital warts in women

but the protection is type-stricted. the duration of protection is unknown but strong protection at ten years after antibody levels have plateaued i think is really encouraging. we think it's well worth the efforts to now see whether one

dose can give long-term protection and move the whole field into one dose. so now hpv, turn now to implementation stuff, the last ten minutes or so, so for a long e new kids on the block, wetime talking about hpv vaccines have to consider them now established products.

they have been licensed worldwide. there's been over 100 million doses of gardacil delivered over 25 million of cervarix. gardacil approved for men in 2009, for anal cancer in both in 2010, in women initially approved in 2006 and 2007.

the countries in gray here have introduced vaccination programs, some for instance in latin america just starting to get underway, but unfortunately the countries shown here have at best demonstration projects oftentimes sponsored by gavi, the problem is some areas that

have the highest levels of cervical cancer like india and sub-saharan don't have programs, you can buy them if you go to part of the normal immunization program, you have to spend your own $150 a dose to buy them, for that reason they are not widely used.

as i mentioned some countries are going to two doses, zero and six months, been approved in the e.u., several other countries. this was based on immunobridging studies where several studies have shown two doses in young adults, say 9 to 14 year olds, generates antibody titers that

are not interior to giving three doses to 15 to 26 year olds in which the efficacy was established, so the obvious reasoning is if you get as high antibody levels as these other people who are strongly protected, you should be protected as well.

this is actually seen with quite few vaccines that before you reach puberty you respond better than after puberty. so by the time you reach puberty your immune system is already going downhill. so in the last few minutes we'll talk about some outstanding

implementation issues about who to vaccinate, how to deliver to adolescents, what effect it will have on cervical cancer screening programs and delivery to economically disadvantaged persons. who to vaccinate? there's a lot of talk about

the bottom line, it's not one size fits all. mostly it depends on what your resources are and what type of screening program you have in place. ideally you give it to everybody, but even the united states it's way too expensive to

give it to everybody. you want to vaccinate girls first, because cervical cancer dominates the hpv-associated cancers worldwide. so i wouldn't start vaccinating boys at all unless you make sure all the girls have this strong but what type of age should you

continue vaccinating where you should stop is really difficult to decide because it's a law of diminishing returns. as a woman gets older the chance she isn't already infected or already exposed and is immune and will get infected later because new partners tend to go

down with age becomes greater, there's some chance she will derive benefit but it tends to and vaccinating males, it makes more sense in countries that, one, have poor screening programs, okay. i mean that have good screening programs because now men have a

greater proportion of the hpv-associated cancers. in terms of immunity, if you're not covering women well, at low frequency, vaccinating men may give you some benefit, if you have some men that will take it. in the united states, vaccination of boys was first

approved in 2011, so the cdc recommended routine vaccination for 11 and 12-year-old boys, up to 28, previously if you wanted to do it it was fine but the government's not going to pay or it and hmos won't pay for it. now it's being recommended so

your health care should pay for it, the government will pay for it if you can't afford it. the reason they made this change in recommendation, there was such poor uptake by girls, i'll show you in a minute, they figured if we can vaccinate boys we'll at least give them some

secondary protection even if not as good as primary. there was indication now that the vaccine prevented anal cancer, i said in 2011 it was licensed for prevention of anal cancer, it's hard to say a safe vaccine can't be given to somebody who has a chance of

getting a cancer. delivering three doses, it's important you vaccinate early. if you look at from the time of sexual debut, almost half of the women in both of these trials by two years had gotten an hpv you can't wait around too long to vaccinate.

uptake varies in countries can school-based vaccination like u.k. and australia, we're in the 70 to 80% range for one dose and two dose, which is good. this is so little that the numbers became small. the united states three-dose for girls.

and 54% are one-dose rates. most people consider one dose a failure but as i showed you i'm starting to consider it more of a success than failure but still the recommendation in the united states is three doses. if you ask me should i give my daughter or should i just get

one dose or three doses, i can't tell you to get one dose, okay? i can't tell you even to get two doses even though that's what's happening to young people in europe. in france, and germany, it's even more abyssmal that the united states.

of note we have disorganized clinic-based vaccination where you have to haul your child to a clinic three times in six months to get them vaccinated. in comparison to other vaccines given to adolescents that the vaccination rates of three, even one dose, are behind these in

boys just starting to come up so there's been a lot of we think provider hesitancy, so health care providers when they talk about these vaccines they say you need to get this, and for the hpv vaccine they say, well, it's available. if you want it, take it.

you know, whatever, because there's a big difference between the uptake of the other vaccines and adolescence, so a lot is the pediatricians and gps. which is really unfortunate because if you do distribute the vaccine and get 70 or 80% c overage you can get dramatic

decrease in hpv disease. this is a reduction in high grade dysplasia in australia, you can see after vaccinon in the younger girls it falls off the table, it doesn't change in the older women who weren't being vaccinated. the same is true for genital

warts, the youngest age group being vaccinated it fell like 92%, next age group by 73, but over 30 years old it didn't change at all. perhaps more remarkable, during the time in which boys were not being vaccinated, the rate of genital warts in the youngest

group under 21 males dropped by 82%, so this is an example where sexually transmitted disease vaccine can give you great herd immunity if you get good coverage rates. so what about cervical cancer screening programs? this is something, it won't help

women with established lesions, it's time restricted so dependen upon the vaccine it won't prevent 10 to 20%, pap screening is 80% effective, so basically changing one effective program for another. we need to convince vaccinated women to comply with screening

programs and we need to do something about screening because if you look at the total cost of preventing of hpv-associated cancers, or all the medical expenses associated with it, 80% is associated with cervical cancer screening, much more than any kind of treatment,

so if we can't to save money in health care we've got to do something about cervical cancer screening. what are we going to do? now we have this pap screen for the most part, which is relatively inefficient. if you get a pap smear, it's a

reasonable chance that you have an abnormality that's been missed and will go to a high grade lesion or cancer in ten years, that's why we recommend every three years you get screened. because it's not that efficient at picking up.

what we're moving towards is hpv dna which has a high negative predictive value. if you don't have an hpv dna, the chances of now getting cervical cancer in ten years is almost zero. we would be able to stretch out the length of time between

screenings. so what we hope to do is combine vaccination at this early age to decrease this whole big bulge of hpv infection and screen relatively few times in a woman's lifetime, maybe at 35, 45, and 55, and that this would then not only decrease the total

number of cancers detected but do it as a much more cost effective way. so delivery to economic disadvantaged individuals, this is where most cervical cancers occur, it's been estimated about 85%, the vast majority. how do we increase delivery to

economically disadvantaged individuals? this is something we'll have to atake in a multi-pronged fashion. both merck and gsk are committed to pricing. if you're in latin america you get the vaccine at a lower price

than the united states, and i think this is fair. in gavi, financing vaccine for the poorest countries, the companies are giving it to gavi for $5 a dose, slowly being introduced in the vaccination programs of the poorer countries.

one solution, partial solution, will be vaccine manufacture in emerging countries. look at hepatitis b, when the recombinant one was made in the west, by major pharmaceutical companies, it cost $100 a dose. cef for 30 cents atoday it's made in india and it's taken 20 years to get from

point a to point b and we're hoping to short circuit that and get from point a to point b at a shorter period of time. one of the companies that's farthest along in this is this company called xiame next. nnovax, they figured out a way to make vlp in e. coli, in phase

3 trials. it's likely before too long the women and perhaps the men in china will have access to a vaccine which they don't have currently. obviously if we can administer one dose, rather than three doses, this is going to be

transformative in order to really deliver this vaccine to the people that need it. so on that i'm going to quit and thank my collaborators, particularly doug lowy. thank you. i'm right on time. [applause]

yes? >> (inaudible). >> does it do anything for them? >> yeah, like -- (inaudible). >> yeah, in terms of can we use this as a pattern for other types of vaccines? i sometimes have a slide about that.

so most of the other viruses that will cause cancer there's a problem with developing vaccines. hepatitis b we have one. hepatitis c is very genetically unstable, hiv all over again, it's been difficult to develop a vaccine based upon antibodies.

now we have a triple drug therapy for hcv which looks like it's very effective. it's whoppingly expensive right now but we need emerging country manufacturers to make the drugs. that's going to be attacked by drugs. ebv has the best chance of

having a commercial vaccine, the reason is it can be licensed for preventing infectious m ononucleosis. jeffco hen in the building upstairs is developing a vaccine based on virus-like particles, 100 times more antibodes, neutralizing anti-bodies than

seen with the ebv vaccine. i think that's going to be a really good vaccine. some other ones associated with cancer, hclb causes adult t cell leukemia, it's unlikely that will be -- the other one that us quas causes cancer is helico pylori,

the regimens are long but effective, there's the drug versus vaccine. >> (inaudible) >> no, there's nothing expressed on the surface of the cell that would be susceptible to antibody. people have been working on this

quite a bit. we're some work as well. people have done trials against high grade dysplasia and they work but not too well. nothing's been licensed, in my feeling the biggest problem is that we've shown this in animal models, rena is working on this

some, if you vaccinate systemically it makes a lot of systemic antibodies, but they don't reach the epithelium where the lesions are being maintained. people have been fooled because they do an animal model, they take out the whole tissue, they

find t cells that are antigen specific, they are in the stroma, they don't get to the women have persistent infection, they mounted an immune response, t cell response, underneath the lesion in the dermis you'll see this nest of t cells but they

can't seem to get into the so the thing is too get a way of generating large amounts of interepithelia neoplage, there's plenty of lesions that will be around so both academia and in companies we're moving forward on that but getting the t cells where we need is going

to be the key. okay. [applause] >> thank you. >> again, at 6:00 we'll go over to building 10 to do the pathology tour. and our next speaker is dr. szabo, chief of the lung and

digestive cancer, m.d. from duke university, did an internal medicine residency at the new york university medical center, and then after completing a medical oncology fellowship at the nci she did a laboratory study in lung cancer biology, and now she's in the division of

cancer prevention, oversees the lung and head and neck cancer prevention trials. her title, nonsmall cell lung cancer >> thank you, terry. people can hear? this is a big topic, nonsmall

cell lung cancer. i'm going to give a brief overview of cancer treatment. it has come a long way in the past ten years more in the past years than the 20-plus years that i've been paying attention, and i'll talk about cancer prevention.

excuse me. the general concept, examples for two specific types of agents that we've been using, and the strategy and early detection and major progress. statistics, this year the estimate is 226,000+ new cases,

158,000+ deaths. this is the number one cause of cancer mortality as well as incidence in the united states. it's a huge global problem. more cases than breast and prostate and colon, the next three most common cases together in terms of deaths.

although the good news is that cancer deaths are decreasing, but ever so slowly. let's see. there we go. the peak in men happened, i don't know, ten years ago or so. but you can see that it's still many more cases than all the

other cancers, and women who had a lesser peak compared to other cancers, we're over the hump but just starting to come down now. you can't talk about lung cancer without talking about tobacco because that's responsible for 85% of cases of lung cancer. and this is one of my favorite

slides, because it has so much information on it. so there's the cancer. here is some emphysema. some fibrosis. you see that these little black areas is where the lung has been destroyed by smoking. here is the pack of cigarettes

on the ct scan. it is a horrible addiction, and that's been the cause of many, many deaths worldwide. as i said, risk factors, 85%, yes, a question? >> so it is tobacco is a huge component of it. i wouldn't say that the physical

hot products going in has no contribution, but it's tobacco. and tobacco in different delivery mechanisms still, for instance when you chew tobacco d so on, it's still important. and it's been many, many carcinogens, many upped when you light the cigarette.

tobacco, passive smoking, prior digestive malignancy, in conjunction with tobacco exposure. and chronic obstructive pulmonary disease, not as big as tobacco alone, also caused by smoking. asbestos, radon, that's why we

get our houses checked before they change hands during sales. chromium, nickel, inorganic arsenic, much lesser in terms of contribution, but tobacco. there are genetic predispositions. there's familial lung cancer, which has been associated with

635 to 25 abnormalities, the receptor subunits, which are associated increased risk of lung cancer and copd, so that's a common commonality. but bottom line is it's tobacco. just a very quick review, we tend to think about lung cancer in two big categories, nonsmall

cell and small cell. non-small cell is everything which is pathologyically not small cell. this is small cell, highly aggressive histology, about 20% maybe less of all lung cancers. and i really am not going to be talking a whole l about this.

this is the rest of lung cancer non-small cell, the most common is adenocarcinoma, this includes both tobacco-related lung cancer but also the lung cancer in never smokers tends to be adenocarcinoma. squamous cell carcinoma is 20%, used to be the most common form,

but as cigarettes have changed, you get deeper penetration into the smaller airways, there's been a shift in the histology and more adenocarcinoma from tobacco exposure than squamous cell. large cell relatively uncommon but not insignificant, other

nontobacco-related pathologies, carcinoid, sarcomatoid, relatively infrequent pathology. we still treat on anatomic basis, early stage, we take it out surgically. if it's more spread with involvement of lymph node and do surgery and adjuvant

chemotherapy, systemic if it's more spread to more distant but regional lymph nodes, stage 3a and b, we do combined modality treatment, we no longer just take it out but tend to do radiation and systemic chemotherapy together. and of course if it's

metastatic, spread outside the lung or multiple places in the lung, then you really need chemotherapy and then we save radiation for local control, for treatment of areas that are creating problems. on rare occasion we'll take out the isolated mets.

that's all non-small cell. small cereal is a different beast, metastatic at presentation even if you don't see where the metastases are, so we give chemotherapy to everybody. plus or minus radiation for limited stage so when it's all

in the chest and then we give prophylactic cranial radiation to prevent brain metastasis, terribly common in this disease. i just put this in so you have a framework. this is not what we're talking about. in terms of treatment there are

two main areas that have really revolutionized where we are today with lung cancer. one is targeted therapy. personalized precision medicine, it is lung cancer is actually the poster child in some respects for this. we now know that there are a

number of specific abnormalities, this piechart is for adenocarcinoma subtype. there are specific abnormalities that can be targeted for treatment and targeted therapies for a small number of these abnormalities, make a big difference.

what am i talking about? egfr, okay? it's mutations in this particular molecule occur in about half of nonsmokers with adenocarcinoma of the lung, and treatment with specific tyrosine kinase inhibitors that targeted mutant egfr protein is

highly effective although not curative. so erlotinib. it's more like 10 to 14 months, we have second generation inhibitors, third generation, about to go on the market which target specific ways of specific

resistance mutations, so we're making a huge progress in this disease. however, we do not cure anybody even with egfh mutant lung we simplyend their lives. eml4-alk, it's about 5% of non-small cell, mainly in never smokers, i don't think i said

that about egfr but it's true. there are specific tyrosine kinase inhibitors, crizotinib, 8 to 14 month duration of response. second line agents some are approved, so we're making progress in extending the life of people with this kind of

abnormality, the fusion protein that results. but we don't cure anybody. and multiple mechanisms of resistance, and so where you go for second, third line treatment is obviously a major area of research. now we're starting to get into

s, the ros1even less frequent rearrangement, less than 2%, abnormality that is once again responsive to crizotinib, mainly in never smokers, long duration of response, nobody gets cured. other abnormalities, her-2/neu, braf, ret, some you with target with specific drugs.

we're chipping away at small subgroups of the total pool of non-small cell with very significant therapies, but you can imagine that to do all the molecular analyses that you need to do to identify these not so common subgroups is making treatment of lung cancer very

complex, very expensive, and the path forward is a little difficult. so realistically, right now these kinds of -- this precision medicine in the setting of lung cancer really applies to less than 20% of all comers, which leaves everybody else without

those kinds of options. for squamous cell carcinoma, we're quite behind. there are certain abnormalities that are being looked at, fgf, fibroblast growth receptor, amplification, pi3 kinase mutations are somewhat -- are less common, ddr2 is another

abnormality, but none of these have gone forward to actual treatment, approved treatment modalities. this is in terms of research right now. what is new and is making a -- so that's one area of research. what's also new and making a

huge splash, you probably hear this every time there's a major meeting, are the immunotherapy approaches, which i'm, again, not going to do justice to by any stretch but the checkpoint inhibitors, the blockades, there are now antibodies that target this checkpoint receptor.

therefore upregulating t cell activation, and there are very important and approved drugs, and i'm just showing you this one curve here, which shows second line treatment, in this case squamous lung cancer, this was the previously approved therapy, this is overall

survival, with docetaxel, second line chemo treatment, versus nivolumab, an antipd1 antibody approved for second line treatment of. pembrolizumab is also approved. you see you have significant improvement of survival. it applies to a minority of

cases, but what you're seeing here is that for a long time out, these curves separate and you have these long-term survivors, again minority of all people with lung cancer, we've not seen that before. so for the first time you're seeing these extended responses,

sometimes complete, often what looks like partial, but probably with no viable tumor, and these people are living -- some living two, three, four years with metastatic nonsmall cell lung cancer, unheard of five years ago. and still doesn't occur except

in rare occasions with the targeted therapies. know how to use theseso this is really a new area we drugs very well yet, we give them to a lot of people but only 20% of people respond. so there's a lot of work being done and i suspect that when i give this lecture in five years

it will be given in quite a different manner. that's very exciting. that's not what i do in my day job, that's not what terry asked me to talk about. i'm going to now talk about how else can you improve morbidity and mortality from lung cancer?

i've talked about therapeutics, but it's really how can you prevent lung cancer or how can you detect it early so that it doesn't become metastatic, and i think this segues nicely from john shoulder and hpv. i wish we the causative agent for all of lung cancer, it's not

that simple, not that it's simple in the cervix but not s john showed the progressionsimple in lung cancer. in the cervix, they are similar in the lung, so that lengthy period of time between your genetic hit from tobacco and the time you get cancer affords an opportunity to prevent

progression, or for early detection. that's what i'm going to be talking about. so i told you that tobacco is the major offending agent. if nobody smoked there would be a heck of a lot less cancer. what if you quit?

should you quit? this is an interesting study, it was an emphysema study. but what you see here is that people who were followed for an average of 15 years, if they stopped smoking, okay, then the death from lung cancer was about half.

this was people who stopped smoking, people who continued to smoke, these are the people who intermittently stopped. this tells you that after 15 years of quitting, you can very substantially decrease your risk of lung cancer. what you don't see on this slide

is that after five years of quitting, there was no difference in lung cancer death. this is a long-term process. with tobacco exposure, you continue to ramp up, or you continue to get more and more abnormalities and you continue to increase your risk, and you

don't actually undo that damage easily, but if you don't continue to increase the damage, then you plateau off. that's what you're seeing here. that's why you're seeing it at 15 years of smoking, not at cessation, not at five years, and this is very different from

heart disease where if you stop smoking, your risk of heart disease goes down pretty quickly within a year or two. doesn't mean you shouldn't stop smoking, it's critical that you stop smoking whenever. however, the benefits in terms of lung cancer don't become

obvious until years down the line. so obviously a lot of people are working on ways to prevent or to help people stop smoking. and we need to make more progress there. it's not my area of expertise. i'm going to talk now about

cancer chemoprevention, the use of strategies to suppress or reverse the process, that lengthy process that starts with an abnormal cell, going through different levels of premia plasia, and ending up with invasive cancer. this is a long process, the

goals are to regress those early pre-neoplastic lesions, to suppress the recurrence once somebody has been treated successfully with surgery for early stage cancer and as you can imagine there are a lot of different biological processes

all wrapped up in this one concept. so the rationale for going this approach, metastatic cancer is still very rarely curable, and of right now has not changed so substantially over the past 30 to 40 years, okay? it's still about 16% five-year

survival. we know from other organ systems such as breast cancer that if you treat with effective therapies, breast cancer tamoxifen , you can prevent new cancers from occurring, we know from pre-clinical model, the last piece of rationale for

preventive approach, it's a long process it get from the initial damage to the actual invasive so if we knew how to treat curatively when cancer is metastatic, or at least very limited metastatic, that's the time to treat, right? because you know who is the

person who needs to be treated. however, right now we don't know, we don't have good therapies, and even with the excitement of immunotherapy, it's still applicable to only subset of all patients with lung so we do think that early cancer is much more curable than late.

if you take it out at stage 1, you have much better survival than stage 4, where you have zero per cent long-term survival as of today. we therefore think that the less molecularly complex precursor lesions should be more curable than even the early invasive

that's a hypothesis, not a known fact. could we prevent carcinogen induced damage, analagous to preventing hpv invention where hpv is the absolute cause for cervical cancer, but working against this is the fact there are multiple pathways for lung

carcinogenesis with different offending agents, with different molecular signaling pathways involved. that's one consideration. we think early needs to be proven. toxicity of the intervention is another major consideration.

we give chemotherapy, immunotherapy, it's acceptable short-term in the setting of advanced cancer, absolutely not acceptable in the setting of pre-cancer or high risk setting so you really need to go to strategies that are acceptable, that are tolerable, and people

who are otherwise healthy and who don't know that they are going to get cancer five or ten years from now. the target population is very important. all cancer, all smokers, 90 million in the united states, what did i say about how many

lung cancers? you know, it's 220,000 then out of 20 or 40,000 people living with metastatic cancer, the vast majority of smokers are not going to get lung cancer, they are not going to get it today. they might get it -- it's about 1 in 8 and it's of course a

dose-response. so many are at risk but relatively few get cancer any given year, we can't figure out who are the ones who are going to get it. and when it comes to the 15% of lung cancer that is in nonsmokers we have no idea who

are the nonsmokers. we're not even studying nonsmokers at risk for cancer. finally, cost and resources, you can give lots of preventive therapies but the more people you target, compared to the ones who get the disease, the more you spread out your resources,

the more you make people who are not patients into patients, a psychological impact, so there are lots of considerations when you should deliver interventions. >> the genetic link is hard to study because the vast majority is together with smoking

exposure, you know, there's a small number of families with certain kind of mutations, the t7 0, egfr and other abnormalities identifies, these families are high risk for lung cancer, and develop multiple cancers, but that's a teeny, teeny, tiny fraction of all.

so, yes, one would want to study these, but this is a different pathogenesis from the typical smoking associate. when we study smoking, we study lung cancer prevention, we focus on heavy smokers, typically 30 pack years, but even those people most don't develop cancer

in the years we study them. so it's a challenge. so but having said that, we don't give up. so how do we develop preventive strategies? we need to look at the risk and benefit. the benefit and preventing

cancer associated morbidity and mortality but the risk of any given intervention, lack of adverse side effects, that can increase morbidity and mortality from other diseases as well as tolerability. so some of you may be aware of the drugs vioxx, a cox2

inhibitor, similar tocelacixib, still on the market, very effective for pain in a colon cancer prevention trial, where it was given, and studied over several years, had been approved after fairly short-term studies, less than a year in duration, very effective, but when given

long term there was about up to a three-fold increase in cardiovascular disease, the drug came right off the market. trying to prevent one disease you have to make sure you don't cause another, okay? so how do we move forward? how do we identify agents

mechanism? if you know the mechanism you're golden, hpv and cervical cancer, that's not to say it was a huge amount of research and terrific conceptual insight, but once the insight was made the path was clear as to how to move forward. for lung cancer, many different

molecular pathways of molecular path genesis, not so easy how to predict who is going to get what kind of lung cancer and when. so we go to pre-clinical models, and we use those, for instance carcinogen-treated models, transgenic animals, i'll show you a couple examples.

we go to observational epidemiology, looking at people who have been treated with various drugs, whether you can decrease the incidence of a specific cancer, lung cancer of course is what we're talking about here. and in clinical trials, for

instance adjuvant clinical trials, to see whether we can prevent the incidence of second primary cancers, so this is sort of the main ways of identifying agents, and then you take into clinical trials. so i'm going to give you a couple of stories to show you

how we've moved forward in this area. one of them is the inflammation story, which has been around for quite a while. i'm really going back 40 or so years, in animal studies showing that corticosteroids can prevent cancer of the skin, in the lung,

especially delivery through the inhalational route, all in animal models. more mixed for lung cancer. mainly so you use steroids for asthma, those people are not such high risk for lung cancer, or emphysema, those people are at higher risk.

most studies have been somewhat short. there is at least one study in the literaure showing that people with emphysema who have -- who use inhaled steroids versus other classes of drugs have a substantially decreased incidf lung cancer.

not the strongest evidence, but it's what we've got. a number of animal studies were performed, actually at the division of cancer prevention under contract, this is one example using the drug budesonide, an inhaled steroid, oral and inhaled, in this case

given in the diet. what you see here is that with increasing doses in the diet, you really decrease the number of tumors after giving a carcinogen to animals by about 80%. and not only that, this is a model where animals develop

first adenomas, and then adenocarcinomas, and what you do is you shift back to an adenoa histology, so these 20% of humors human -- tumors are more likely to be adenoma rather than carcinoma, the way we chose to move forward is by focusing on

pre-malignancies in the lungs, in this case bronchial dysplasia which is leading to squamous cell carcinoma. and so with increasing grades of abnormality, this is the sequence that leads to invasive squamous cell carcinoma. the natural history of these

lesions is that actually a fair number of them develop cancer, this is a pretty recent study where people who had bronchial dysplasia were followed over time, about a third of them developed invasive cancer, between one and two years after the diagnosis of the bronchial

dysplasia. it's not just the precursor. also also a risk marker for the entire abnormal field damaged by tobacco expressure, 40% of cancer develop from that area of dysplasia, of 60% elsewhere from the bronchial tree somewhere else or as adenocarcinomas in

the periphery. so bronchial dysplasia is a precursor to some cancers, but it's very much a risk marker of a whole abnormal field. so the study that we performed with our collaborator stephen lamb, british columbia cancer agency, took 112 cancers with

dysplasia, underwent bronchoscopy, treated for six months with inhaled budesonide or placebo, underwent a second. to show you the complexity of the study to get the 112 on study, dr. lam screened over a thousand people with sputum, to identify those who have some

level of abnormality, he then did -- that was half of them. he did bronchoscopy on 350 people, out of the 540 who had abnormality 350 agreed to bronchoscopy, a third went on to study, very labor intensive. unfortunately this study was negative, we can talk about why,

the model, the animal model was looking at adenocarcinoma, we were looking at squamous carcinogenesis, bronchial dysplasia, but we found people also tended to have ct lung nodules, small nodules, in the early days of ct screening before it became an approved

modality. and there was a significant decrease in the number of persistent ct nodules in those who had inhaled budesonide versus placebo. this led rise to a study that looked at the peripheral lung, ct screened nodules such as this

little ditzel right here. our colleagues at the european institute of oncology in italy had a ct screening trial, we took 202 participants with persistent small nodules, were not felt to be cancer, or they went for treatment. questionable they went for more

diagnostic studies. so these were smaller nodules. they had to be persistent and were randomized for a year of inhaled budesomide versus placebo, had a next screening ct, looking for shrinkage of lung nodules. this is the first study that

we didn't know what we should be looking for. we took all long nodules that were persistent and what we found is that the majority of the lung nodules, 70% were solid, and over the course of a year if they were there prior to the onset of study they didn't change at all.

but the non-solid nodules which i'll describe in a second did appear to get smaller, and actually over time for the one year of intervention and then five years of follow-up, over time they continued to get smaller, whereas the placebo arm they did not get significantly

smaller. so this suggested to us that maybe we should be looking only at the nonsolid or partially solid nodules, the solid ones are fibrosis, they are nothing. why is that? so one of the precursors for adenocarcinoma, so this is a

nonsolid nodule on ct scan, is this lesion here. atypical adenoma hyperplasia, a well described pre-malignancy for the peripheral lung. and until ct screening the natural history was really not well understood at all, because you take it out, sometimes

because there was something in the chest, or it happened to be there when you resected for an adjacent lung cancer. but we now know that these non-solid nodules, between 25 and 50% of them are actually atypical adenoma hyperplasia, pre-malignancy, perhaps this is

the population we need to look at. one study showed that these capacities will grow over time, and if it's in smokers which is of course our chemoprevention study populations, they have a much higher incidence of progressing than if they occur

in nonsmokers. and so i'll show you just -- this is a little -- actually a shows you what happens with these nonsolid nodules and how variable they are. so 11 years ago she had some chest pain, went to the emergency room, they thought she

might have a pulmonry embolus, they saw this and you need to have it followed up, this could be a lung cancer, so three months later there it is again, it's there, different quality ct, maybe a little bit less obvious but not much changed. like all good people she went on

her merry way and didn't come back a year later, until 2010, five years ago, where she had chest pain again, wound up in the e.r. and then here is that nodule yet again, didn't have a pulmonary embolus, still ground glass, same consistency. she went back a year later for

follow-up. this is what you compare with this initial thing. and now there's this solid center, okay, still some adjacent hard to see ground glass opacity, ground glass carcinoma, pre-malignancy that has taken how many years?

really seven years before it was diagnosed as an invasive cancer. there's a lot of variability over how long this takes. follow-up is needed. we looked to see whether the noncalcified ground glass opacities are associated with cancer in the national lung

screening trial, ct screening study. oops. and what you see here is that people underwent three rounds of screening, so zero, 12 and 24 months, but their risk of lung cancer, if they had a ground opacity started to go up at five

years or so, suggesting that these ground glass opacities, some of them are cancer showed you in the previous slide. now, how are we moving forward? from our budesonide studies, the ground glass opacity is what we need to look at.

what drug? a lot of pre-clinical data, not so much clinical. aspirin has a new life in cancer this is a patient level meta analysis of multiple studies of aspirin, versus placebo, in this case all the studies were done for cardiovascular or stroke

but if you look at risk of cancer deaths, okay, it starts to take off five years, aspirin is substantially lower than the control placebo, and at five years you start to have a significant decrease in lung cancer death. about a 20 to 30% effect over

time. so this suggests to me that within five years you go from having a pre-malignant lesion to death in lung cancer. if you have an invasive cancer most people die in two to five years. this suggests we should look at

aspirin in the setting of pre-malignancy and we're doing which that our colleagues in europe who have ct screening and we're looking at asymptomatic current and former smokers, and giving them with persistent nonsolid or part solid nodules, giving them a year of aspirin or

placebo, and looking by ct scan specifically at those nonsolid nodules. and we're about a third of the way done right now. we're also asking whether aspirin can modulate the abnormal field, and whether you really need to give it every

day, compared to intermittently. so there's data that you don't have to give chemo preventive intervention every single day, you can do a week on, a week off, three weeks on, three weeks off, different regimens. why would we want to do this? because the less drug you take,

the less side effects you have. so if you can be off for a week, or off for three weeks, you can regenerate your stomach mucosa, you don't get as much bleeding, which is a problem you get with aspirin. so we're doing this study which is looking at readout of

abnormal gene expression. it's actually quite a long story, but we're looking at nasal swabs, looking at gene expression, smoking-related signature, in the nodes with aspirin given for a week on versus week off, or comparing that with aspirin and zileuton

and play placebo this . this shows the way forward. myo-inositol is a dietary compound, studied minimally in diabetes, but it's very appealing because it is generally regarded as safe, graphs using f.d.a. terminology, and there are a number of animal

studies showing that it can inhibit carcinogen-induced tumors in mice, both with carcinogen, with mainstream and side stream smoke. so this is an agent that we got interested in, because of the work of lee wattenberg, one of the giants of cancer

chemoprevention who passed away last year, in his 90s. and so we performed a phase 1 clinical trial again with stephen lam, based on these animal data, showing that we can tolerably deliver 18 grams per day, a large amount, and in those completely uncontrolled

dose finding safety study, we saw that there was very considerable regression of bronchial dysplasia, compared to historical controls. so this is the worst kind of study, but preliminary evidence that maybe we should look further.

what was interesting about the study is that dr. lam did bronchial brushings from the normal bronchial tree. and at boston university gene expression analysis from people with bronchial dysplasia treated with myo-inositol or placebo, adjacent lung cancers, smokers

and nonsmokers, with and without dysplasia, there was increased activation of the pi3 kinase pathway in the normal epithelium of smokers who had dysplasia but not healthy smokers who did not have documented dysplasia and myo-inositol inhibited this in the normal bronchial airways,

people with adjacent dysplasia. this gives us a way forward, molecular analysis, to answer many questions. first question, are those the highest risk people, those who have pi3 kinase activation, are those who we ought to be association with the

pre-malignancy bronchial dysplasia, allows the brushing without having to do multiple biopsies, a little bit easier than having to do multiple biopsies, but more importantly it allows us to potentially identify the right cohort. and it gives us a new clinical

trial model, one where we would be looking at shorter interventions and smaller numbers of participants, looking at the high throughput kind of analysis, such as done with the gene expression pi3 kinase signature. so this is a potential way to

move forward which we're really exploring right now. and so the follow-up to this small phase 1 study is this phase 2b placebo-controlled blinded study, giving myo-inositol versus placebo for six months, 85 people were randomized, and we're doing the

analysis as we speak. so this is how we move forward, i think. i'm going to do five minutes on early detection, because unlike chemoprevention where we have no approved therapies, for early detection we have made huge strides.

lest anything be too easy, okay, issues in screening, i don't know whether you get electro or not, but the three big bias which make life difficokay, lead-time bias, if you screen you can identify earlier disease but it doesn't mean the people will live any longer.

survival will appear longer but people still die at their pre-appointed time. lead-time bias because you identify the disease earlier but you're not actually postponing death. you just think you're making progress.

length bias, there is lots of disease out there, screening tends to diagnose the more indolent disease with a longer clinical face, phase, thehighly virulent cancers that people die from quickly you don't find by small cell lung cancer very aggressive, moves quickly, you

don't find that with ct finally, overdiagnosis, lots of things look abnormal, but that aren't necessarily clinically this was well documented with autopsy studies looking at prostate cancer. most men by the time that they are in their 60s, 70s,

probably by the time they are 80, almost everybody, everybody has prostate cancer. but in fact very few people, not very few but few people relatively few people die from prostate cancer because most of this is clinically unimportant so overdiagnosis is a major

we used to think there was no such thing as overdiagnosis in lung cancer, that's not quite true. those are the potential bias. then you have to remember that if you're going to screen, there's going to be morbidity from the procedures, follow-up,

right? abnormal things, some mortality, it's going to be taken from the very carefully controlled centers that do the trials to the general community. and there's always cost, a monetary cost. so just two studies that i want

to talk about, plco trial, this is the chest x-ray versus nothing, usual care, before screens, looking for lung cancer, decreased deaths, and of course nothing. the chest x-ray is not useful for screening for lung cancer. this is the definitive study

that had been many others previously. but the national lung screening trial is quite different. this was a ct screening trial, 53,000 smokers, current and former, 30 pack years, had to quit within a certain amount of they had three ct screens versus

chest x-ray, the results were very positive. 20% reduction of lung cancer deaths. so this isn't lung cancer incidence, it's actually deaths. overall there was 6.7% reduction in all cause mortality, lung cancer is a significant cause

for part of mortality. but 24% of the cts were, quote, positive, there were a lot of positives that didn't turn out to be lung cancer. this shows the decrease in -- a pair is the cumulative number of lung cancers, they continue to

go up. this is lower dose ct, screening was here, deaths come later. a lot of work being done to figure out which of these nodules are the important nodules, the ones that are associated with cancer as opposed to all the other things

that happen in your lungs, and trying to move this to identify those people who are going to be at highest risk when we could target for prevention, this is just one example of the risk model, the brock model, brock university in canada where they looked at the risk of lung

cancer in nodules from one single ct, and if you put into this model the age, sex, family history, emphysema, and various parameters related to the nodule, ground glass versus other, et cetera, you can actually identify the highest risk smokers who develop -- who

are most likely to develop lung cancer within the next three to four years and that's the population that we hope to target for cancer prevention this is, again, a work in progress, but it brings together the early detection and prevention components.

so i'm done. lots of progress. this lecture is, you know, for many years i was reusing the same lecture year after year, and it's delightful to be able to have to change it. so tremendous progress has been made.

precision medicine is on the map for lung cancer. it's applicable to significant but still small subsets of advanced stage cancers. cancer patients. increased survival but no cures. these are the early days of immunotherapy.

i'm sure you'll have good lectures about that. prolonged survival, but in a small subset of patients. early detection is also making a mark, decreased lung cancer mortality, we're still figuring out in the best cost effective manner, but it is an approved --

i say that in quotes, but it's approved by the u.s. preventive services task force and medica pays for it. and as we're making this progress, we are identifying new targets and tools which we can then move to chemoprevention to prevent those cancers in the

first place. and i think that's it. i'm happy to answer any questions >> right. there's going to be a registry associated with these ct rollouts, and so in terms of identifying populations, and so

on, yes, there will be a lot of work done. in terms of interventions, not really. there's not an easy way to interface with that. however, as more ct screening is occurring we'll have all these people followed.

we're working with various big centers to try and capture that population. it's a little early. that's why we went to italy, because they have the ct screening, but it's a little early for any given center to have enough years of follow-up

for us to be able to put trials in there. but that's our hope. we're going to be seeing lots of these people with persistent nodules and a lot of work going on there. >> there are. i don't think we've seen a

signal particularly for lung, but yes, people have -- colon, prostate, but, you know, some studies show, some studies don't show, so it's the episignal has been not the most consistent. there's some animal data for colon, levy's work, that suggest there may be some benefit.

it's been a little bit difficult to study because so many people in the united states are on statins, so there was an attempt to do a colon polyp prevention trial, phase 3 study, i think that study has closed. you know, you can make the argument, well,

antiinflammatory, statins, go for something that's anti-inflammatory first, if you want to look at statins, downstream signals study statins, but not for -- i like to study the most effective agent for the putative mechanism i'm tried to study.

that's my approach. thank you for staying to the end. >> i have a comment. i opened my "time" magazine friday, on the third page there was an advertisement for merck for the immune checkpoint inhibitor for nonsmall cell lung

cancer.