Breast Cancer Websites

lawrence brody:thank you, vince. i think i'm on. so, i was asked to talk about breast cancer today andalso to leave time for questions in case i don't cover areas that you're interested in,so my attempt will be to finish in about 40, 45 minutes and leave 10, 15 minutes at theend for questions. if there's a burning question, feel free to interrupt, this is informal presentation.i want to start the presentation where i'm going to end the presentation, so for thenon-radiologists in the room i think you can tell this is an abdominal ct scan. shown inthe circle, again for the non-radiologists, is a tumor in this woman's abdomen. shownhere, a month or so later, is the same region and for those of you in the front i thinkyou can see the tumor is gone, and this tumor

disappeared because this woman was treatedwith a drug that was designed using knowledge of breast cancer genes, not a drug that wasused as a general tumor or an anti-growth agent, but a drug specifically designed tothe genetic make-up of her tumor. most of the talk will be about where the geneticand the genomic advances can be applied to medicine and i'm going to bring in examplesfrom breast cancer, some of which actually -- and i'm sure some of you in the audienceactually are applying them, and others that are still in clinical trials, but how didwe get here, and what are we going to cover today? we're going to talk mostly about breastcancer genes but as vince mentioned especially for brca1 and brca2, these are also ovariancancer genes and i'm not going to cover that

today but they're very important componentsof ovarian cancer. in fact, some of the treatment data that i will show you comes from ovariancancer rather than breast cancer. so we'll talk a bit about what these genes are, howwe find them and how we use genomics and genetics to get to these genes, what do these genesdo in the body, at least some of them, the ones we know about, and then there will bea overarching kind of woven theme of how we can use this genetic knowledge to improvehealth. i just want to start with showing you thedata that i think everyone in this audience knows is the mortality by different causesbetween 1950 and around 2000, and the red bars are the mortality in 1950 for heart disease,stroke, pneumonia and cancer, and what you'll

notice, it's quite striking, is that we madegreat progress in these areas over the last 60-some years. cancer -- and it is unfairto lump all of cancer together, but for the purpose of this graphic we're putting allof cancer together -- progress has not been as impressive for cancer. one of the reasonsis because we don't know the mechanism for the cancers -- the others, obviously thereare lots of different types of cancer. i'm sure that everyone in this audience also appreciatesthat breast cancer is a very common disease, so unlike some genetic diseases which arerare, this disease has 200,000 cases or so per year, 40,000 deaths. the one in eight,one in nine number, if you're a woman, your lifetime risk of diagnosed. less talked aboutis the fact that if you're a female your risk

of death from breast cancer is a few percentcurrently. this is what the cancer rates have lookedlike for various cancers over the last 70, 80 years, and shown in this yellow line isovarian cancer, which has held relatively steady, although there's been improvementsprobably in diagnosis. uterine cancer, nice example of advances in medicine. in this case,mostly surgical treatment decreasing cancer mortality by a lot over this interval. breastcancer staying roughly the same over this interval in diagnosis and probably our worstexperience as far as not a victory here is this great increase of lung cancer in womenthat obviously correlates with environmental exposure from smoke. if you zoom in on thebreast cancer curves -- and shown for reference

on the bottom are the lung cancer curves -- thisis breast cancer over the last 30-some years, and many of you will remember in this intervalthere was great concern about these rising rates of breast cancer in the '80s and italso corresponded with a great increase in awareness of breast cancer. prior to this,breast cancer was not a high profile public disease even though it was still a commondisease. this rate seems to have leveled off and if you look over here, and i'll zoom inon this area, this is a real decline that started in the early -- early in 2000, 2003,and does anyone know what has been attributed to this decline in breast cancer? female speaker:[inaudible]

lawrence brody:so this is -- a decline is attributed to the decline in use of hormone replacement therapy,and that coincided with a -- practically right there with a landmark study that showed thathormone replacement therapy did not protect against heart disease. it was always knownthat it was slightly increased risk of breast cancer, but it was on balance thought to bea good thing if you protected you against heart disease which was more common. at thispoint in time, hrt therapy dropped through the floor as far as uptake, and probably canaccount for this decline in breast cancer risk. i bring this as an example not to saywe did the wrong thing, but an example to show you that in addition to genes, the environment'svery important, so here we have medical practice

influencing the rate of breast cancer andessentially decreasing it through advanced studies. so, why would we want to know about the geneticsof breast cancer? there's always reason to look at mechanism to try to better understandthe disease, but in breast cancer and ovarian cancer we hope that looking at genetics willhelp us with prevention, early detection, being able to better predict the course ofthe disease and tailored therapy, and so today i'm actually going to talk about advancesbrought on by genetics in these two regions. there are advances in both of these othercategories, but i don't have time to talk about them today, so i'm going to focus onthese two areas during the talk. before i

do that i want to do one advertisement andone definition for you. you may have heard that cancer is a genetic disease and i thinkit's well soaked into the culture now that cancer is a disease of genes, but you haveto really divide this into two separate areas and that's why we're having two lectures andtwo talks on it. the first area that cancer and genes are involvedis when cells essentially can acquire mutations that are associated with growth advantagesand they escape normal controls and essentially form a tumor. this is a disease where thegenome of the cells change. these are also known as somatic mutations. these are notinherited, they're not the topic of today, a month from now stan lipkowitz, who is here,if he raises his hand, will be talking about

what we're learning and what we're being ableto do in medical practice by understanding the genetic make-up of the tumor itself. it'sa critically important and probably one of the most active areas of cancer research,is understanding the genome of the tumor. today what we're going to focus on is whatcomes before that, and that's the genome of the individual. so this is the genome thatyou inherit from your parents that has a collection of genetic variants; some of those variantsincrease your risk of cancer. so at birth, based on what you inherited from your parents,your risk of cancer from one person can be different from another, and that's going tobe the focus of today's lecture. come next month to hear stan talking about the changesin tumors and tailoring drugs toward tumors.

this is the inherited variation. i want to just -- can you see the chromosomesthere? i -- look a little washed out, but these are an example of how risk mutationscan occur in a cell. here's two chromosomes, and i don't know if you can see this, whati've done is delete one portion of that gene and that cell, since it has another copy isprobably fine, but if it loses the other portion of this gene, it now has no copies of thisparticular gene and will go on to essentially form a tumor. this is what happens in sporadiccancer, where you lose both copies of a particular gene, these are called, it's up at the top,tumor suppressor genes. this happens rarely, but since you have several trillion cellsin your body, occasionally one cell suffers

these two hits. this is what we think is goingon in sporadic cancer. in inherited cancer, and again, i'm sorry you can't -- they'renot dark enough to see. here's the chromosome in a person with inherited cancer, where theyalready have inherited a mutation or a deletion or something that knocks out one copy of theirgene, so now in all of their cells all they have to do is lose the other copy and that'sa much more common event, and so people who are born with one mutated allele have a muchincreased risk of a particular cancer, and this just shows you that -- the deletion ofthe other allele. for the most part, the cancer risk genes,the ones that are inherited, tend to fall into this category. and in fact in -- we allknow that family history is an important risk

factor for all cancers, and especially breastand ovarian cancer. in a small percentage of families the cancer really does appearto be inherited as a mendelian trait, and so what does that look like? oh, and thataccounts for three to eight percent of breast cancer, so not a -- not the majority of breastcancer, a small amount of breast cancer, but a small amount of a very large number. andso what i'd like to do now is focus a little bit on our understanding of the topographyof breast cancer risk, and this kind of foggy, cloudy, slide shows you where we were about10 years ago. i'm going to very quickly get up to the identification of the breast cancergenes, work that took about 15 years and really, that could be done now, probably, in a fewmonths with the right families due to advances

in genomic technologies and genomics. so, what does the topography of breast cancerrisk look like? if you focus in on this graph for a second, this is the frequency of therisk. so up here, 30 percent frequency of risk would be quite high, a fraction of apercent down here, and this is the relative risk. so, essentially, if you have a particularvariant, if you inherit a particular variant from your parent, how much does that increaseyour risk compared to someone who didn't inherit the variant, and so what we can do is mapthe landscape. first of all, there's boundaries that can be put in. so, down here, where thereare things that are very uncommon and very low risk, they probably exist, but we don't-- we can't find them because it requires

studies of enormous size to find them. there'san upper boundary to this map up here and that we know that there aren't super highrisk variants out there in the population that are incredibly common, because if theywere then breast cancer would be even more common than it is, and so the genetic variantshave to fall into this landscape. what do they look like? so this is what a diagram of a sporadic cancer-- here's a woman with breast cancer, is this familial? cancer in the family, is it familial?not because this person is not a blood relative even though it's cancer in the family andso this is finding out that you have an aunt or an uncle with a family history is not thesame as mapping out the pedigree. this is

a relatively common manifestation. we knowthat, empirically, that if this woman had breast cancer, this woman who might be interestedin knowing what her risk. her risk has doubled just based on the pedigree alone. this isthe picture of inherited cancer. here are -- the filled circles again are women whohave had breast cancer and you can see multiple individuals in the blood line. i illustratethis particular pedigree because those of you that have looked at genetics, this is-- looks almost like a dominant pattern inheritance, with one exception. what about this guy? doeshe have the risk allele? i see nodding, he almost certainly does, but he doesn't havethe cancer because there's this -- there's a sex-limited trait for this particular cancer.so we have heard, from time to time, of people

being told, "don't worry about your cancerrisk because it's all on your father's side," and certainly there's no empirical reasonto worry about which side of the family comes -- if there's cancer on one side of the family,that counts as increased inherited risk. so, obviously, this woman would really wantto know what her cancer risk is, and your gut tells you that her cancer risk is probablya lot more than this woman's cancer risk, and that is true. using these families, we'reable to identify the genes that cause the high risk of cancer, and how was this done?shown here is the curve of rate of breast cancer and age of diagnosis. so, as you mightexpect, very little cancer is diagnosed early in life, breast cancer is clearly mostly adisease of later in life. if you want to find

the genetic basis for any particular condition,would it be best to look at the average presentation or is it better to look at the more severeor earlier onset presentation -- and so clearly the more severe and earlier onset is likelyto have more of a genetic component because as you get later in life, the combinationof your genes and environment tend to tip toward environmental influences. earlier in life, it's more genetic influences,and so to identify the major genes for breast cancer, what was done is women who were diagnosedat very early ages in these families were looked at and that led to the discovery, now-- 1994, sounds like a long time ago -- of the brca1 gene or breast cancer gene one.soon after, when these families that had lots

of cancer where typed, we realized that brca1did not account for all of the families, and so very quickly, within a year after, brca2was identified. these remain -- these two genes remain the major players for high riskfamilies and high risk breast cancer. when you study families that have breast canceronly and you ascertain breast cancer only, you don't notice that there are other cancersin the family until you start looking at family history and very soon after the genes wereidentified, or that there were breast cancer genes were identified, it was appreciatedthat there are other cancer associated with. so brca1 and brca2 are associated with roughlyan eightfold increase of risk. so if your risk is one in nine, one in eight, and goesup eightfold, it gets up to be pretty high.

it's also -- brca1 mutations are also affiliatedwith breast cancer in males, which is a rare, but not incredibly uncommon, disease. aboutone percent of breast cancer that's diagnosed is diagnosed in men. ovarian cancer, as i mentioned at the beginning,is a very major player for -- especially those with brca1 mutations in that carrying a brca1mutation increases a woman's risk of ovarian cancer 20 to 30 times, sometimes 40 timesdepending on how you measure it. had the studies been done differently, brca1 probably wouldhave been called an ovarian cancer gene, and in men who carry mutations in this gene, they'reat twofold increased risk of prostate cancer, that's been replicated a number of times.so, these breast cancer genes are also other

cancer genes as well. where do they fit inour topographic map of the landscape? they fall up in here. brca1 and 2, the number ofpeople who carry mutations are quite rare, probably, looking around, only one or twopeople in this room might carry variants, or deleterious variants in these genes, butthey are very strong risk factors so they are up here. the other gene that is up thereis the p53 gene, which is involved in li-fraumeni syndrome, which has other cancers involvedthan breast cancer. so, these genes, very potent risk factors for breast cancer risk.what do the lesions in these genes look like? i'm -- this is a slide that was originallymade in december of '94. i can tell you i made it as a slide, i think some of you probablyremember slides.

[laughter] and had to scan it in -- and shown on thisline is a picture of the brca1 gene and it was identified two months earlier, publishedin science two months earlier, and all these arrows show you were mutations have been foundin this gene in different families, and i think you can appreciate, within two months,we realized that these genes were going to have a lot of mutations. this is the untranslatedpart, so the gray area is the part that makes the protein, and so the protein is being hitall over the place in all the different families. so the families that i showed you, this extensivepedigree, in general they tend to have different mutations from one another, with a few exceptionsthat i'll talk about in a second. this has

led to, as many of you know, a genetic testing,made available in the u.s. almost exclusively by myriad genetics, to find out who carriesmutations for this gene, and if you do the calculation, these two genes are probablythe most sequenced genes in the entire genome. so worldwide, there has probably been 250,000,300,000 individuals who've had their genes sequenced. so we know a lot about the differentvariants in the gene. this is an updated picture of the gene, a little more colorful now onthe web from our database. this, just again, the gene at the bottom. shown here where themutations are. there are so many mutations in this gene that some of the nucleotideshave been changed to -- you know, there's four different possibilities at each nucleotide.if it's originally a t, we found that t has

changed to a c and also to a g. so we've almosthit a kind of saturation of this gene. what you also will appreciate from this is thatthere's no hotspots, there was no insight into function from looking at the mutationdistribution. for some genes, and in fact maybe some that stan will talk about, thereare hot spots in that that show how they're activating mutations. this is a gene that,brca1 and brca2, produces a risk by losing its function. as i said, there are lots of different mutationsbut there are some groups which have the same mutation due to a founder effect, and thisis just due to common ancestry. one that we've worked on is the ashkenazi jewish individuals,who have one brca1 mutation, another brca1

mutation and a brca2 mutation. so, in aggregate,one in 40 ashkenazi jews carries a deleterious mutation in the brca1 or brca2 genes. we usedthis effect more than a decade ago to do a study of -- and i don't know, did anyone participatein the washington area study? to try to figure out what the risk was associated with this,and it was carried out in the bethesda community. in iceland, there is one brca2 mutation thatis found in one of 170 people in iceland. there are dutch founder mutations, there arefounder mutations in these different areas of the world. this has the effect in thatif you are going to get a genetic test for brca1 and you're in a group that has foundermutations, it makes more sense to look for that first before moving on to the more expensivetest. i can tell you that the full sequencing

test now approaches $4,000 of a charge inthe u.s. from myriad genetics. the test for specific mutations is quite a bit cheaper,a few hundred dollars. if you look at the mutation types and don'tworry too much about the different kinds of types, just look at the total number of entries,so there, as vince mentioned, there are thousands and thousands of different mutations in thesegenes. distinct alterations means the different mutations, so there's nonsense mutations whichshut the protein off, there's -- at the time i made this slide -- there was about 176 differentnonsense mutations, but the ones that are found in one family only are 84, so in eachcase there's a lot of different mutations and about half of them are found in a singlefamily. this is why you have to test almost

every new family, because about half of themhave something we haven't seen before. the top three rows are mutations that we knowkill the protein, so if a woman has -- is found to have, or her family is found to havea nonsense, a frame shift, or a splicing mutation, we can reliably tell them that that mutationis probably associated with risk, you notice how i'm hedging a bit. the other case, wherethey're missense changes. these are changes that change one amino acid for another. sothey -- proteins, these proteins are very big, they're over 1,000 amino acids. you swapone amino acid for another, we have a hard time with those, and this has been calledthe unclassified variant problem or the variants of unknown significance problem, and so inbrca1 there have been almost 3,000 families

that have had these, there's 500 differentmutations. about 300 or so have been found only in one family, and when people get thesetest results they're told, "you have a variant, we don't know what it means, and sorry aboutthat," and that's a very active area of research because when you go to get a diagnostic test,if you have the pretest counseling you're prepared for getting either a good result,that you don't carry something bad, or a bad result, that you do carry something bad. it'sharder to prepare people for a, "we have something interesting, but we don't know what it means."and roughly about 10 percent of test results, five to 10 percent of test results turn upone of these variants that we just don't know the significance of, and so there's a veryactive group trying to nail these down and

figure out how many of them are deleteriousand how many of them aren't. i bring this up because, as you'll hear inother talks and you may have heard from dave valley [spelled phonetically] in the firsttalk, we're going to be sequencing lots and lots of genes going forward. the sequencingtechnology has become so spectacular that it's very easy to generate dna sequence information.we have to be prepared for generating information that we don't understand. so that's a quicksummary of the high penetrance, meaning they have a very strong effect but low frequency,genes and again, brca1 and brca2 are the major players. p53 produces li-fraumeni syndromewhich is usually not confused for just breast cancer syndrome. what about the low penetrance,meaning the low risk but high prevalence,

area, and so these may have a low relativerisk. if you carry the variant, it only increases your risk of cancer a little bit, but theymay have a high population attributable risk because lots of individuals carry them. can we find those genes? i told you that tofind these genes, individuals recruited very spectacular families and used the familiesto look via something called linkage analysis to find the genes. for these genes, we dowhat's called association studies. in association studies -- linkage studies are relativelycomplicated, association studies are very easy. you just do case control type studies,you look at people who have the disease, match people who don't have the disease, and youcount up how many variants are present or

how many genes are present in one versus theother. this is just an example of one that was done quite some time ago to give you anidea of the scale. you could identify the brca1 and 2 gene by having 20 nicely collectedfamilies. you could identify those genes. in order to identify these low penetrancegenes, you need large numbers of cases and controls, and this -- i show you the raw datato give you an idea. control is about a half a percent carry it, in the cases about twopercent carry it, and so you're looking for something of relatively modest effect. inthis case, about a doubling of risk. there are now genes that have been filled in tothis area so these are relatively more common than brca1, but still around one to two percent,and there's a collection of genes in that

area that we now know about in this landscape.so, the genes listed in here have relatively rare variants that are associated with roughlya doubling of risk. that still leaves this big area down here and that's an area that'sbeen filled in most recently by what we call whole genome association studies or genomewide association studies. this just shows you an example. as you compare the frequencyof, in some cases, a million markers in your cases compared to your controls, and certainmarkers -- this is the p value plotted in an inverse scale -- certain markers show outto be standing out that they're much more common in your cases compared to your controls. this has been done for breast cancer, thefirst one was published in 2007. you'll see

that the scale gets to be quite large becauseyou now need to pool lots and lots of cases, so there were 147 institutional affiliationson this paper, this was the first big association study of breast cancer risk, and that hasallowed us to fill in this region of the graph. it's likely that there will be more and morevariants that show up in here, and they have a very modest risk, a 30 percent increase,a 20 percent increase. remember that having a family history, first if your relative isa 100 percent increase, so they're very modest, they're in the scale of risk that's associatedwith certain diet choices and alcohol intake. so how can we use this, now, landscape andthis will be filled in a little bit more but i don't expect -- we're pretty sure thereare no more genes up here, we'll probably

have a few more here, this area will get populatedwith some more, but we're now starting to have the entire picture of what the geneticlandscape of breast cancer risk looks like. and so we now have a very clear picture; howcan we use it? the first i'd like to talk about is earlydetection. early detection is the product of screening, and that's generally probablyusing these and these, and so if we go back to this family, it's pretty obvious if thisfamily has an inherited cancer, whether or not we know they have a brca1 and 2 mutation.this woman's surveillance and screening methods cannot be what we suggest for the population,and so what is normally suggested is that mammography is done more frequently, startedearlier, depending a bit on the family history

in these individuals and prophylactic surgeryis considered as well. male speaker:how early? lawrence brody:how -- as early as 20-something in some families. i don't think there's a uniform recommendation,there are -- the standard that's been used is that the earliest -- 10 years earlier thanthe earliest diagnosis in the family has been used at some point, so if this woman was diagnosedat age 30, then you might push it to 20. so that's much earlier than we would do in thepopulation. the risk that these women face has been estimated originally at up to 90percent. by looking at different study types, the risk, if you carry a mutation of havingbreast cancer, it ranges depending upon how

you do the study and this just lets us knowthat there's some heterogeneity in the mutations, and so right now women who carry a brca1 mutationare generally told they have a 50 to 80 percent or 50 to 70 percent lifetime risk of breastcancer. so, i think we're getting a good handle on these high risk families, but they tendto be rare. what about the low penetrance, high prevalence?what about the -- are they clinically significant? and i think the answer is we don't know yet.are they significant to the individual health? we're not totally sure about that yet, butcould they be used as public health style application of medicine? and for that i wantto show you some modeling that's been done by paul farrow in the u.k. how do we do breastcancer screening now? you reach a certain

age, you start mammography. you do self-examand the assumption is that everyone of a certain age has an equal risk, so that's plotted herethat if you're a 50-year-old woman and i screen 20 percent of the population, i will catch20 percent of the cancer. if i screen 100 percent of the population, i should catch100 percent of the cancer. that's based on the idea that all 50-year-old women or all40-year-old women have the same cancer risk. we know that that's probably not the case.this is what risk probably looks like in the population. if i, again, go back to my entirepopulation of 1,000 40-year-old women, some of them will be at very low risk due to theirgenetic make-up. others will be -- sorry -- others will be at very high risk and risk is distributedin the population. genetic risk is distributed

in the population almost like a normal distribution. so if we go back to our screening algorithmand instead of making this assumption, this is what a bell curve would look like of risk,we assume that risk is not randomly distributed in the population, but we can identify whois at higher risk. if you could -- if you can implement screening programs that tookthis into account, and you screened the 20 percent of individuals who are most at risk,you would then pick up 60 to 70 percent of the cancers that way. this is not somethingwe're ready to do, we don't know exactly how to employ the genetics yet, but it is thepotential for where genetic risk assessment can be done and genetic risk assessment canbe done at any time in your life, presumably

if we could sequence your genome early onwe could give you a profile at a very early age of what that might be like. i'm not sayingthat we have to do this, but if you are going to employ resources that are relatively scarceand expensive, you might want to use the genetic risk profiles to guide them. the second area that these -- knowledge ofthese genes is having impact already is in the tailored therapy, the data i'll presentare really from clinical trials, they're not yet part of practice, but they rely on ourknowledge that brca1 and brca2 are actually dna repair proteins, so your -- every cellin your body has a genome in it. that genome is constantly having mutations and constantlyunder attack. your cells, and the cells of

actually all organisms, have a very elaborateset of proteins and gene products that are designed to repair that dna. took about adecade of work to recognize that that's what these two genes do in the cell. in fact, theyare -- actually participate in a very specific type of dna repair called double-strand breakrepair. so you could imagine -- think about dna as a double-stranded molecule. if youbreak one strand, it's still attached by its connection by the overlapping, but if youbreak both strands those molecules are free to drift apart and the cell cannot toleratevery many double-stranded breaks at all, and brca1 and 2 seem to be involved in double-strandbreak repair. keeping that in mind, a group in the u.k. thought very logically and said,"how can i use that fact to design a therapy?"

and so i mentioned that any kind of insultcan do dna damage, in fact oxidation does dna damage. so if you want to have your dnabe completely intact you should stop breathing, which is -- -- not really an option. every time a celldivides, it introduces mutations, so there are a lot of -- there's a lot of dna damagegoing on, almost all the time it's repaired quite efficiently. so, in normal cells, it'srepaired in cells viable. these things are to show you the different types of repair.if you don't have brca1 in a cell, or brca2, you have a defect in one type of repair, thisdouble-stranded break repair. if you take another product out of the mix called parpyou don't have a type of dna repair that repairs

single-stranded breaks. if you don't haveeither of these, then the cell dies, and so what the group in the u.k. did was designan agent that inhibits these enzymes, and so normal cell is okay, if i inhibit thisenzyme, i'm fine. if i inhibit brca1, viable, although it might be a tumor, but if i knockout both of those, the cell is dead, and in a person -- because you remember that theseare tumor suppressor genes, the only cells that are missing completely brca1 are thetumor cells, and so if you add this extra layer, you ended up killing the tumor cellsand the tumor cells only. that was the theory -- yeah? male speaker:is the brca at risk for all insults to dna

or is it more specific to a specific enzyme? lawrence brody:it's more specific to these double-strand breaks. in fact, there are -- you can insultdna 1,000 ways, mutations as well as backbone breaks, but it seems to be very specific fordouble-stranded breaks, and so, this may be a little bit of a difficult scheme to follow,but did it work -- and so shown here is experimental data, where these are tumors that are implantedinto mice that are either have brca, brca2 or don't have brca 2 and i'll not go throughall of the data, but these are tumors that kept getting bigger and bigger and biggerin all the control mice and all the mice that were treated, except for the ones that weremissing brca2, so the treatment plus brca2

made the tumors go away in mice, and thiswas work that was published several years ago, somewhat fast-tracked into clinical trialsand i'll just show you some of the data from the clinical trials. so this is olaparib, which is again this parpinhibitor, there are several parp inhibitors out on -- in various clinical trials. thisone seems to have had the most success, and the blue line here are those that weren'ttreated. here's the treatment regimen, red line was treated, so you can see that there'sprogression-free survival, is essentially better in those treated with this drug thathas relatively minimal side effects compared to normal chemotherapy. there was a secondtrial published last year -- from where i'm

sitting this doesn't look very good, but theseare ovarian cancer. people having another parp inhibitor on top of standard therapy,and again, showing an effect of the drug tailored toward the gene mutation. these trials arestill going forward; for breast cancer, there's been less success. and it maybe due to heterogeneityfor breast cancer. there's a lot of excitement in ovarian cancer world for using these totreat ovarian cancer because, again, many more ovarian cancers have brca mutations andmany more ovarian cancers are more brca-like than there are breast cancers. i think thecompany -- one of the compounds is an astrozeneca compound, i think they were quite disappointedin the breast cancer results because obviously for them it's a larger market than the ovariancancer. i think if you have ovarian cancer

you would be pretty excited about anythingthat might be an advance in treatment. from what i've heard, these are being used essentiallynow off label, even though the clinical trials aren't done, and i don't know if any of youin the room have experience with any of these parp inhibitors. but i just want to close with showing youthat these results again were driven by the genetics, understanding the genetic mechanism,understanding what the genes do, have had an influence on the way to target these tumors,and so by understanding mechanism and refining the characterization of the pathways, we'vebeen able to rationally design drugs to help treat the tumors. i think that -- i've toldyou that we can use -- in theory, use this

knowledge of genetics of breast cancer totailor early detection methods. prevention is still somewhat up in the air, i think manypeople reject prophylactic mastectomy as a preventative measure, but it does preventbreast cancer in brca1, brca2 families, as well as oophorectomy. there are some advancesthat have been had in prognosis in knowing who has a brca 1 and 2 mutation and i thinka lot of excitement, it's been in the last couple years, has been in the tailored therapyend. and with that i'd like to close and thank you for attention, i promised that i would-- [applause] -- leave time --

-- for questions. male speaker:if you look at environmental risk factors, alcohol, [inaudible], that sort of thing,are patients with brca1 more at risk if they have a large [inaudible] of risk factors ontop of it, or -- lawrence brody:it kind of overpowers the other -- so the question was, if you look at environmentalrisk factors, reproductive history, alcohol, dietary exposure, obesity, do the brca1 synergizewith those environmental risks and i think synergy is probably not there. it tends tooverpower it. there are -- there have been some studies done where oral contraceptiveuse seems to be a little bit more powerful

of a protection in individuals who have brca1data, but it's not in great -- very large studies, so they tend to be really overpoweringrisk factors. as much as we would like to say that we could remove some environmentalexposure and make breast cancer go away, i don't think we can; i think it's the internalenvironment. lawrence brody:right, so the question was what about hrt, hormone replacement therapy, and i think fromthe -- it's not suggested and recommended for people who carry the high prevalence mutationsand probably not recommended or suggested for almost anyone now, given the breast cancerrisk and the lack of cardiac protection. female speaker:you said that the [inaudible] protective,

but hrt is against diseases, is that right? lawrence brody:i'm sorry, i didn't hear the last part. female speaker:[inaudible] are protective, is that what you said [inaudible]. lawrence brody:the contraceptive -- oral contraceptives used early are somewhat protective from some epidemiologicdata. female speaker:and the hrt is harmful. lawrence brody:later in life, yeah. yes. male speaker:one [inaudible]

lawrence brody:yeah. male speaker:why is this gene so sensitive, so twitchy? lawrence brody:so, the -- male speaker:[inaudible] lawrence brody:the question is, why is this gene so sensitive? i don't think the gene is sensitive. i -- wehave a lot of mutations going on in our body. the gene shows us pattern of mutations thatis consistent with just internal environment causing mutations. some of these mutations,and i -- the founder mutations. some of those mutational events actually happen 2,000, 3,000years ago and are just carried through because

the breast cancer, as awful as it is, is notsomething that interferes, for the most part, with reproduction. so it's not necessarilyselective at -- disadvantaged to having cancer as a whole in that realm. male speaker:if a young woman with very strong family history is going to have repeated mammogram, is itbetter to do mri other than the mammography [inaudible]? lawrence brody:the current recommendation is to do mri, not necessarily because of the radiation risk,which has always been theoretical and controversial because radiation causes double-strand -- cancause double-stranded breaks in women who

potentially may have their double-strandedbreak repair compromised. the studies that have been done that i am aware of show thatthe mri's just more sensitive and detects earlier lesions, and so that's why they'reshifting toward that recommendation in brca1 or positive family history. male speaker:how early should it be done? at age of 25 or something? lawrence brody:yeah, i don't know the current -- and maybe stan, do you know, or -- i don't know thecurrent asco recommendations as to interval. male speaker:what i think you would do is annually and

then follow up as indicated, so a normal mammogram[inaudible] an annual or a normal mri, do an mri a year later, an abnormal one [inaudible]. lawrence brody:i do know there's a lot of diversity. some of these families, the women show up everysix months because they're very nervous about their risk. i think the hardcore data on efficacyis somewhat slim. there's a question over there. female speaker:just two points [inaudible]. mri is much more sensitive, but it's less specific, and sothe problem that we're having with mris is that [inaudible] lots of biopsy could turnout to be benign. of our patients who are

at the very high risk level, our brca patients,what i like to advise them is that they alternate at six months between mammography and mri,so in the winter they have their mammogram and in the summer they have their mri. so-- but these are patients who are at a very heightened level of surveillance and, at leastin our population, these patients are very much attuned to their level of risk and whatto do, but these are women who will have many, many biopsies and many, many benign biopsies,and the issue for them comes up with their insurance companies, and what is their insurancecompany willing to tolerate. but i do have a question, which is what correlation,if any, is there between the individual genetic analysis and the genetic oncotype analysisthat we're doing of the tumors themselves

and are there any implications for therapy? lawrence brody:so the question is, are there correlations between the individual or i can say the germline,the inherited profile and the oncotype type testing which actually looks at the expressionlevels in the tumors themselves. and there are some correlations in that the phenotypesof, especially in brca1, of the tumors tend to correlate with the mutation a little bit.i don't know in clinical care whether it makes any difference, so brca1 tumors tend to bethe -- more the triple negative, essentially estrogen receptor, a negative type tumors,they probably show up on the oncotype test as well. my guess was that if you had thegermline knowledge and the oncotype test,

the oncotype test would kind of overrule thegermline knowledge. you might predict the oncotype test result from knowing the germline,but you really want to know what's going on in the particular tumor. and that's a casewhere i think what's going on in the tumor at the time you actually run the test is probablymore important than the actual germline constitution. a lot of triple negative, a lot of basal tumorsoccur in women who are not brca1 positive, and so knowing brca1 status alone is not enough.question in the back. female speaker:how does increased risk in the brca1 or 2, you talked about heterozygous mutations orhomozygous mutations? lawrence brody:so the question is, when i'm talking about

increased risk in brca1 and 2, am i talkingabout heterozygous or homozygous. and it's an excellent question because the -- it'sreally a little bit of both. the inheritance is heterozygous, in that the person inheritsone mutated copy and has one good copy. so, in that case, the person is a heterozygote.what we think happens is that the tumor loses the other copy and becomes homozygous forit, but really when we're talking about risk we're talking about the heterozygotes. if you have two mutations in brca1, one inyour maternal and one in your paternal chromosome, you probably are not viable, and so mice are-- that is mutations you can make in mice and follow, they're lethal. in brca2 they'remostly lethal. there is one exception: there

are some mutations that you can have in brca2where you've got two alleles, so you're homozygous for mutation and that produces a differentphenotype called fanconi anemia, which is usually identified through other reasons inthe childhood time. so for the most part, one -- carrying one mutated chromosome increasesyour risk, and we're talking about it in the heterozygous state. female speaker:right, so i have a follow-up question. lawrence brody:[affirmative] female speaker:so does -- can you be heterozygous for both? lawrence brody:so the question was can you -- can you have

one -- be a heterozygote for brca1 and heterozygotefor a brca2 mutation, and the answer is yes. that's been documented most in the ashkenazijewish population, where you have one percent of one and one percent of the other in thegeneral population, so you'd expect one in 10,000 individuals to have both. and it'sfound at about that rate, so there's probably not synergy, in that you don't have worsedisease, but there's only been a very, very small number of people documented who havethat, so it's a little bit hard to say with any conclusion, but there have been peoplewho have a heterozygote mutation in brca1 and a heterozygote mutation in brca2, butthey're rare. [affirmative]? male speaker:a woman told me yesterday that her mother

and her sister are brca1 or 2 positive andshe has it, so does she have an increased risk of breast cancer over the population?and excuse my ignorance, but is this a test that you do just once in your lifetime orcan people mutate later in life and you have to test again? lawrence brody:so the question was, if a woman who comes from a family where her mother and her sistercarry the brca1 mutation, a specific one, and she is tested and she doesn't carry it,is her risk higher or is it the population risk? it is probably, almost certainly, infinitelycloser to the population risk. can we say it's not a little bit higher? no, but generallythe surveillance and the screening recommendations

drop down to the population risk. and thesecond part of the question was, do i have to get tested once, or can mutations happenlater? and so these are inherited mutations, we test them in the blood even though the-- your blood dna, even though it's not your blood that's the affected tissue, and youreally should only have to do it once. male speaker:could you comment on the use of ct scans of the thorax as a risk factor? lawrence brody:i'm going to defer to the radiologists. the question was could you comment on the useof ct scans of the chest -- female speaker:and the risk of radiation.

female speaker:this is a very hot topic right now, obviously, and i kind of have it in two camps, are theways that i think about it. one camp are the people who -- they have a symptom or an issueor they already have cancer or some other problem and they're getting a ct scan to answera specific question that will have an immediate diagnostic or therapeutic impact. the riskof radiation for those patients, i think, is -- and by the way, is based on extrapolateddata. we don't really know what that risk is. we think we know based on extrapolateddata, but nobody really knows what that risk is. those people, in my mind, their risk ofdeveloping cancer in 25 years is marginal compared to the benefit that they're goingto gain by having the exam. the second camp

are people who are having exams that we thinkare maybe of not such great use. people who have total body screening, for example, andi strongly, strongly discouraged this because there is no data to show that there are improvedoutcomes for these patients and obviously they're getting radiation that they wouldnot otherwise have. there's a certain amount of ct scanning thatis done, particularly in patients who are seen in the emergency department, and it happensbecause of the pressures in that department to diagnose and move patients quickly. andi think we could do a little bit more with better utilization in those patients and decreasetheir potential risk by having ct scans. but that's a much bigger problem that we can'tsolve on an individual basis and that is a

problem that will have to be addressed morelocally as to what is considered appropriate utilization. there are a lot of factors rightnow going into what is considered appropriate utilization. there's a whole liability stepagainst what we know as best practices, and that's going to get worked out over the next,i'd say, five to 10 years, as health care reform happens. but if you have patients whohave real problems, and you need diagnostic information, it's going to have an impacton diagnosis and treatment, and the patient is 30 years old, i certainly wouldn't nothave a ct scan because of the potential radiation to the breast in that example. male speaker:what of the smokers the ct [unintelligible]?

female speaker:well, there's all the new information shows that utilization of screening ct. by the way,all of our machines now and the way they set up our protocols, are set up to minimize radiationbut still be of diagnostic accuracy. there's a difference in imaging of what is diagnosticallyaccurate and what is a pleasing image to the eye. if you use more radiation you get a verypleasing image; it may not have any more diagnostic accuracy than the very grainy, noisy lookingimage, so there's a lot of sensitivity right now to using factors on -- in the scanningparameters to minimize the dose of radiation but still have diagnostic accuracy. lawrence brody:thank you.