Breast Cancer I

vence bonham:good afternoon. my name is vence bonham. i’m the senior adviser to the director of nhgrifor genomics and health disparities, and i’m pleased to welcome you here this afternoon.this is a unique lecture today because it’s a collaboration between the intramural researchprogram of nhgri and the genomics and health disparities lecture series, so it’s an opportunityfor the two groups to come together to have an opportunity for you to hear from our speakertoday. the genomics and health disparities lecture series is sponsored by nhldi, niddk,nhgri, and the national institute of minority health and health disparities, and the officeof minority health for the fda. it’s an opportunity to bring different parts of thefederal agencies who are interested in genomics

and health disparities together to bring speakersto nih to explore issues from the perspective of how genomics can play an important rolein understanding health disparities and health equity issues. i’m pleased that this afternoonmy colleague, dr. nigel crawford, is going to introduce our speaker, and we look forwardto our talk this afternoon from dr. olopade. [applause] nigel crawford:good afternoon, everyone. as vence said, my name’s nigel crawford. i’m a tenure-trackinvestigator in the dir side of this equation. it’s my really great honor to be able tointroduce dr. funmi olopade to be giving this joint seminar today. dr. olopade is the walterl. palmer distinguished service professor

in medicine/human genetics, the associatedean of global health, and the director of the center for clinical cancer genetics atthe university of chicago, school of medicine. dr. olopade went to medical school at theuniversity of ibadan in nigeria; subsequently, she came over to the u.s. to do a fellowship-- sorry, a residency in the cook county hospital in chicago, and following this, she went onto do a fellowship in the university of chicago. dr. olopade is really very much an internationalleader in cancer genetics. she’s an expert in cancer risk assessment and developing individualizedtreatments for both breast and ovarian cancer. her work very much focuses on identifyingpatients that are at risk of the most aggressive forms of these diseases and then implementingnovel management strategies based on individual

genetic assessments in these populations toperform early interventions and essentially reduce mortality in these high-risk groups.she approaches the study of aggressive cancers using a number of different approaches. she’svery interested in the mechanisms of familial cancer, the mechanisms of tumor progressionin high-risk individuals, and, very importantly, genetic and epigenetic factors contributingto disparities in cancer outcomes in diverse populations. i really think that the powerof dr. olopade’s work and her contribution to the field is well-highlighted by her workwith the brca1 and brca2 genes where she’s shown distinctively that these two genes increaseincidence of breast cancer and cancer outcomes in women of african descent.

overall, her lab uses highly innovative anddevelops innovative approaches to studying aggressive forms of cancer. for example, sheuses whole genome technologies to reduce global disparities in cancer outcomes. she’s beenthe recipient of numerous awards throughout her career. i can only list a few of themhere, but some of the highlights would be the macarthur foundation genius fellowship,the doris duke distinguished clinical scientist, and exceptional mentor award, and she’salso been elected a member of the american academy of arts and sciences. so, dr. olopade,it’s my great pleasure to be able to introduce you, and we’re looking forward to hearingyou talk. olufunmilayo olopade:thanks, nigel.

-- on, and knowing that this is a really mixedaudience, i tried really hard to just give you snippets of how we’re thinking aboutbreast cancer genomics and health disparities. and, you know, the bottom line of what i reallywant to emphasize at the end of -- conclusion of my work, my lecture today is that we reallyhave a unique opportunity in precision medicine to get rid of black/white differences andbegin to really talk about getting care to individuals at the personal level, and toget to that, we all have to work together collaboratively. i know that this lecturehas been sponsored by the office of minority health, but really, when you talk about individualand personalized medicine, we’re all in the minority because it’s one genome ata time, it’s one drug at a time, one person

at a time that we have to take care of asphysicians. so i just want to share my journey to trying to get precision medicine for all. it’s really because i came to this countryto study cardiovascular medicine and i quickly gave up on that dream to become an oncologist.i think that i was really inspired by what i was seeing in cancer at that time when wewere really doing subset analysis of different chromosome changes in leukemia and lymphoma,and we were cloning one gene at a time. and janet rowley and francis collins and all thegiants in human genetics at the time said, “why don’t we just do the human genomeproject and just map every gene, clone every gene, and then once we have that, then wecan really begin to think about risk assessment

and best therapies.” and in the ‘80s thatwas really a broad vision and it materialized, and now we’re past that. so the questionis how do we actually get to -- get in the genetic information that will allow us todo the best risk assessment and provide the best therapies so that we can, in fact, eliminatehealth disparities and improve health equity? and so, you know, i started in lymphoma/leukemia,and i still think that, you know, there’s really just arbitrary division of these differenttypes of cancers, that if we really begin to think about pathways and think about howwe treat individual diseases we may find that there’s actually more commonality amongall these different types of cancers. having said that, i think the reason why theoffice of minority health and center to reduce

disparities really want me to give this lectureis because when you talk about incidence rate, and ever since, you know, people have beengiving -- have been talking about data, it’s always been that you have more breast canceramong african-americans and, you know, even though the incidence now has approached theincidence in blacks, the mortality is really what is so disturbingly much higher amongafrican-americans. and as we began to think about how people are diagnosed and what we’vedone with population screening, we now actually are having a debate in the breast cancer communityabout over-diagnosis of breast cancer. how many of you have heard about the debate aboutover-diagnosis and over-treatment? raise your hand. okay, so that’s been talked abouta lot, but what’s actually not been in the

public press is the fact that there are manypeople who still get diagnosed with regional and distant metastases at the time of diagnosisdespite the fact that we’ve done large-scale population screenings over the last 30 years.so the question is who are these individuals? well, we began to sort of break it down theminute we started talking about, you know, different types of breast cancer, and thefact that hormone receptor-positive, her-2-negative breast cancer is the most common type of breastcancer, and it’s the one that’s been studied the most. it’s still also what kills mostwomen who develop breast cancer. the reason why we’ve all been very excited about triplenegative breast cancer is because we don’t have a target for it, and as oncologists andcancer researchers a lot of our effort has

been focused on finding new treatments fordiseases and less of prevention. when i came to this country, the whole mantra in our medicalschool was that prevention was better than cure because you have low resources, and ifyou have low resources you have to really prevent diseases, keep people healthy insteadof spending a lot of money to treat diseases. so it was really surprising to me that atcounty hospital, when i started, all we did was treat people who came into the hospitaland there was not really a lot of public health work being done in chicago. but now we knowbetter. we know that when you actually collect datayou actually see that african-americans are more likely to get triple negative breastcancer. it wasn’t always the case that we

knew that. we only knew that because my goodfriend jeff trent, when he was at nhgri, really thought we should do gene expression profilingthat was a new technology, and he applied it to a very few numbers of patients. he didn’thave a lot of tumors to look at, but with -- i think it was maybe seven or eight, youknow, brca1-associated breast cancers, brca2-associated breast cancer, and sporadic breast cancer,he actually was able to tell us that your germline genetics actually could determinethe type of tumor you had. this was profound at the time, and of course the stanford groupand everyone sort of really began to see that your germline genetics determine the kindof tumor you had. when this paper was published in the new england journal our lab read itand, you know, as a good oncologist, what

i said was, “well, it really wasn’t thebrca1 signature that they were talking about.” we wrote a letter to the editor that whatthey probably did was -- “these were all er-negative tumors; these are all er-negative;these are all er-positive; majority of sporadic breast tumors are er-positive, so what theyare really describing is a signature of er-negative tumors and not necessarily a signature ofbrca1,” because that was the level of understanding we had at that time. however, with subsequentwork, with our work, a lot of work of the breast cancer linkage consortium, it becameclear that, in fact, brca1 patients actually develop a disease that was a well-definedphenotype. they tended to have a medullary -- atypical medullary, very high mitotic rate,and one of the things that sort of struck

me was the fact that they had high proliferationfractions. if you are coming in as a lymphoma doctoror a leukemia doctor, you know that when patients come in with a high tumor burden that -- and,you know, with my internship at university of ibadan, if you have a patient with burkitt’slymphoma who came in with large jaw masses, big abdominal tumors, our job at that timewas to make we treated them immediately and with that little red drug called adriamycin,these tumors will just melt away over the weekend. we didn’t even care about tumorlysis at the time because these were usually young children, but the tumors went away.and i was really fascinated why these big tumors melted away, but when we talk aboutbreast cancer, these tumors that were growing

fast -- we were still really thinking aboutcutting them off, doing surgery, and a lot of things. anyway, to cut a long story short we thenrealized that these tumors, of course, had the worst outcomes, and african-americanstended to have tumors that were sort of described like this and also had the worst outcomes,and that’s what really got us thinking about “can we do a cheaper version of gene expressionprofiling?” now, this is a very good pathology and morphology of breast cancer. if you thinkabout global health, most people would not be able to do pathology of breast cancer likethis in most parts of the world, and so most patients that are treated with breast cancerdon’t have appropriate diagnosis of their

breast cancer, and as a result of that, theywill not have a proper therapy of their breast cancer. it doesn’t matter where they live;if you cannot get a good diagnosis, you cannot get the best treatment. so we know that estrogen receptor is the firstdruggable target. it was identified at the university of chicago. charlie huggins dida very simple experiment, took the ovaries of women with metastatic disease, and if theywere premenopausal and they had an oophorectomy, their tumor melted away because you had massiveapoptosis and you starved them of estrogen; and these women did well. okay? that’s atreatment that actually works. it’s very effective for any woman with er -- estrogenreceptor-positive breast cancer. along the

way came another target for therapy in breastcancer, which is her-2-positive breast cancer. dennis layman really was also a student andresident at university of chicago and thought if we could just target her-2, we’re goingto be able to treat a lot more aggressive breast cancers. these two targets are reallywhy every woman with breast cancer wants to know is it er-positive or her-2-positive,because we know that if you treat women appropriately they’re going to -- if they don’t survivetheir breast cancer, they’re going to live very long lives. when we were transplantingpatients with her-2-positive breast cancer they didn’t respond any better than whenwe didn’t do bone marrow transplant. however, the minute we got her-2-targeted therapy,a lot of women who would have died within

two years of breast cancer became long-termsurvivors because of the target. so, fast-forward. we then said, “let’sgo and look at the patterns of breast cancer in nigeria,” and this is the last time iwill show this slide, because when we first did the experiment we didn’t believe that,in fact, the pattern in nigeria, senegal, and among african ancestry could be so differentfrom what has been reported in japanese and white women in europe. so anytime i talk aboutdisparities, the thing that i always like to say is that the real gap is really theknowledge disparities. right? if you don’t do the research, you’re going to come tothe wrong conclusion. so the office of minority health, whatever the nih’s address is, we’vejust got to just do the best science, it doesn’t

matter where, and the best science will giveus revelations of what we should be doing because we can’t assume that we know everythingabout breast cancer. but that was what we found, and then, once we did publish the paper,there was all sorts of disagreement about the fact that the way we, you know, did theanalysis; the tumors were, you know, in the fridge too long, and it couldn’t be possible.and then subsequently -- all that’s in africa, north and east africa. everybody started lookingto say, “okay, what’s the estrogen receptor status of our tumors? because now we can dothis research.” you know, a number of studies -- i mean, when we did the summary prior tothe aortic meeting in morocco i was really impressed that -- 30 studies in egypt, sudan,libya, morocco, tunisia, ghana, mali, nigeria

-- everyone trying to figure out what is thetype of breast cancer that we find. and what we find is that there’s really -- you know,the data are all over the place, from in northeastern regions having 63 percent that were er-positiveto studies in west africa showing 35 percent that were er-positive and south africa showing60 percent er-positive. i show this slide just to give you the impression that, in fact,if you look at african-americans, 64 percent in the u.s., and you look at u.s. seer caucasians,80 percent are er-positive. there’s definitely heterogeneity in who’s showing up and whattype of breast cancer they have, and that’s why we can’t say one size fits all thatwe’re going to have a population approach that treats breast cancer the same way nomatter where they are.

and this is really important because wheni do my practice -- and i have the benefit of being a doctor -- i see different typesof patients, and they come in different shapes and sizes and colors. this is a 64-year-oldwhite woman with interval triple negative breast cancer. based on population data, thiswoman should get a mammogram every other year. okay? and she went, predictably, march 2003and march 2005. the only -- she was only two days late in terms of two years to the time,and yet her breast cancer was so aggressive, so highly proliferated, she only lived oneyear and didn’t respond to any treatment we gave her. okay? when you read the literature,you will think this only happens to black women who are poor and have no access to care.that’s not true. so, now the u.s. supreme

court and a lot of things that really haveopened up how we approach and what we do can be interpreted however we want it. it says,you know -- it takes 1,900 mammograms of women in their forties to save one life, but only1,300 of women in their fifties. it’s more cost-effective, so just wait until you are50 and stop calling us the death panel. right? that’s been some of the things that’sbeen argued as we try to address the affordable care act and what’s covered, what’s notcovered, who should have access, when should we rationalize health. i actually -- you know,anytime you’re thinking about any public health intervention, we always want to docost effectiveness analysis, and sometimes that cost effectiveness analysis doesn’tget us to the personal level where people,

in fact, really want to have information aboutwhen should i get a mammogram, how should i get a mammogram, is it going to help mesurvive breast cancer or not? so to move forward, i think that we’re actuallynow post the mammogram’s age, and i -- there’s a nice walk that we’re now doing in hydepark with one of my african-american patients who decided that we should -- you know, hermantra is “beyond mammography” and she’s trying to do a walk to let african-americanwomen be aware of the facts that because you have had a mammogram doesn’t mean that you’rehome-free. so this is a very important 45-year-old womanwho had a perfectly normal mammogram, even at the time that she felt a lump in her breast.because she was savvy and because she knew

what to do, in fact, because she was awareof her cancer risk because of her family history, she didn’t stop there. she went, she gotan ultrasound, and she got an mri, and here’s this very tiny triple negative breast cancer.some of the stories i hear when i go on this walk with women on the south side of chicagois the fact that they will show up to a doctor with a lump and they will say, “well, you’retoo young to have breast cancer,” and so it doesn’t get evaluated, or “well, youknow, maybe it’s a boil.” when we were in nigeria, they would show up at the primaryhealth center, and these women have just started -- you know, maybe they’re lactating, maybethey have just had a child, and somebody will tell them, “well, you know, it’s not breastcancer because you’re too young to have

breast cancer.” well, the data asserts that women of africanancestry tend to get breast cancer at a much younger age, both here and throughout thediaspora, and many of them will get breast cancer at the time that they’re also lactatingand having babies. that’s some of the epidemiology data that our work in nigeria has shown, sovery often these young women are misdiagnosed or not sent for appropriate diagnosis. however,if you are flat-chested, you’re not really endowed with a size d-cup, you might be -- itmight be a lot easier for you to actually feel that lump. a lot of women do not havethe luxury being small-breasted and being able to feel a lump at the time when it’sthis tiny and it’s picked up by both an

mri and an ultrasound. so the question iswhat is the value of teaching anybody to do breast exams? what is the value of tellinganyone to get a mammogram when it doesn’t work 30 percent of the time? so all our publichealth interventions actually will fail women who develop aggressive triple negative breastcancer. so then what are we going to do about it? my view of the world is why not just sequenceeveryone’s genome? right? if we knew who was going to get that type of breast cancerand we knew it by age 30, then maybe we can actually begin to personalize care. how did i come to that conclusion? i cameto the conclusion because we actually know that there are some high penetrant mutationsin genes like brca1 and brca2 that predictively

-- they’re very rare, but they will causebreast cancer at an early stage. and because brca1 is part of that gene pool, we’ve beenreally doing a lot of work trying to understand how brca1 and brca2 and some of these genescontribute to the early onset. they’re rare alleles; they’re pathogenic. if you inheritthem, either one -- through your father’s side or your mother’s side, you’re predictablygoing to get early onset breast cancer. and then, of course, you have your atm and othergenes that now we have the ability to test for that are -- have -- they’re not thatrare, but they contribute to increased breast cancer risk. then, a lot of people have donegenome-wide association studies trying to look at common alleles that are going to bedistributed in such a way that they may contribute

to cancer risk. so, this is a good model for us to think aboutin terms of population risk stratification. we started doing this work by looking forfamilies. families are rare. if you’re working in a place where people don’t even have-- know what a diagnosis of cancer is, people are going to show up not even knowing whattheir mother died from or what other diseases are in their family. if you only, you know,focused on family-based ascertainment -- people are adopted. there are so many reasons whypeople may not know their family history, and yet when these genes have been testedfor, the majority of -- 60 percent of women who at the time of cancer diagnosis are identifiedto have brca1 or brca2 mutations do not record

a family history of breast or ovarian cancer.so how do we pick those women up? so, the question then is, you know -- we askthat question in my clinic, because we have, you know -- we’ve had a tumor bank, african-americanscoming in and -- or -- you know, we’ve had data from european women, everyone who’sreally done linkage studies and who has case-controlled studies, they’ve all been able to use next-generationsequencing to look at the estimates of patients who are in their cohorts who develop breastcancer and have mutations in one of these genes. we collaborated with mary-claire kingand we said, “okay, let’s just look at 289 african-americans from the south sideof chicago that either i or my colleagues at the university of chicago saw,” and wellover 80 percent of -- 22 percent of them had

mutations in one of these genes that actuallyconfer a high risk of breast cancer. so clearly this is not the end of this story, becausewe only picked the ones that are definitely deleterious, pathogenic, and so when we takeit from having breast cancer as an ascertainment, one in five have a deleterious mutation. then we did a study in nigeria where -- thatstudy’s been in the field since 2002. we also wanted to know what is the burden ofmutations in these genes in this relatively young cohort of women in nigeria coming into have their breast cancer diagnosed. and you can see lots more variations, lots moregenes that are in the dna repair pathways, but the predominant genes causing breast cancerare still brca1 and brca2. and these are consecutive

cases, and when we look at what’s -- youknow, at this -- you know, they -- are they founder mutations? of course not. each onehas their own personal mutation, right? and the mutations are in this large gene; 1,863amino acids all across the gene. so there was no -- you know, initially we thought,“let’s do it like we can do in women of ashkenazi jewish ancestry where three foundermutations explained all of their cancers.” for african genomes, you can’t do that becausethey’re not founder mutations, and you have a whole diversity and a spectrum of mutationsthat are explaining breast cancer in these women. then we asked, “do these women have a strongfamily history? were they diagnosed” -- you

know, we’re always looking for a shortcutway to actually find the ones who have mutations. in the literature it was described that earlyonset breast cancer -- so we said, “okay, let’s only sequence women who had breastcancer under the age of 40,” and then we got an answer. but now that we can do next-generationsequencing, we said, “why limit it to women under 40? just sequence everybody.” andwhen we sequenced everybody what we found was that, yes, whether they were young orold, we found people with mutations, right? so under 45 -- about 12 percent of the patientswith mutations, and over 45, you still had mutations identified in these individuals.most people didn’t report that they had a family history; however, if you had a familyhistory, nearly 22 percent had a deleterious

mutation [unintelligible] they didn’t reportany mutations. these are, you know, cases and controls, and we’re continuing to actuallydo this work so that we can really do population estimates of these rare deleterious allelesin -- as a way to look at who should be screened and who should have access to genetic testing,because if we wanted to use the rules that we used in america to tell people to go andhave genetic testing, the majority of the individuals who didn’t have a family historywould not show up and would not qualify for genetic testing. this then tells me that, in fact, there areindividuals among women in the african diaspora, and i can go on and on about studies in thebahamas looking at women who walk through

the door and clearly 25 percent of them havetheir breast cancer because they have a founder effect mutation in the bahamas, which is anisland. so there’s a lot more work that we have to do across the african diasporato actually estimate the burden of disease that these young women who are getting themost aggressive breast cancers have and where we have no strategy for cancer control inthis population, whether they’re living in the u.s. or they’re living in their homecountries. now, let’s look at tumors and the cancergenome atlas. the cancer genome atlas, for those of you who are not cancer genomicists,this is one of the best things that’s happened where we used sequestration money -- moneyduring the financial crisis to actually develop

large-scale repositories and to do big science.so, initially, when the call was to go and sequence all the tumors, well, it turns outthat the tumors that were available to be sequenced were mostly tumors from white women,and they had to do an extended call to get tumors from black women to be part of tcga.so while there’s been a lot of publications about tcga and, you know, the genomic landscapeof tumors in tcga, we’re now just beginning to analyze data in women of african ancestry.and so we happened to lead that analysis, and what’s really very important -- thisis actually for my good friends, because anytime we discuss culture and race and identity,the question is should you use genomic race to do your analysis, or should you do self-reportedrace to determine, you know, what it is that

you are? in tcga there were people who reported thatthey were black, they were white, and they were asian, caucasian, and of course the africanancestry yoruba. and you can see, when you do principal component analysis, that theafrican ancestry group is really nicely clustered here, and the asian ancestry, nicely clustered,and then you have your caucasian. but then there are the individuals around here who,you know, depending on how they self-report, they can either say they’re black, white,or whatever. but one of the things we wanted to do with tcga was to actually look at -- youknow, if you do genomic race and the patient has self-reported as black, so what are they?if you look at this, most people who say they

are black, whether you do it by genomic orby self-report, there was actually a very good concordance. there were very few misclassificationswhere somebody thought they were white but they actually are black, based on their ancestry,but there was no one who really thought they were black that turned out to be white whenyou do both genomic and self-report. so the question then is what is the best approximationwhen you are trying to figure out whether ancestry matters in terms of the proportionof african ancestry, or whether just ask the patient, “are you black or white?” maybethat’s the best way to do the answer. and i think -- there’s still a debate in theliterature about that, but the thing we really wanted to do was we wanted to know withintcga people get different types of breast

cancers. the most favorable breast cancer is your luminala breast cancer which is estrogen receptor-positive and which is treated with hormonal therapy.then you have luminal b, which is estrogen receptor-positive, but it tends to be a littleresistant to hormonal therapy. then you have her-2-enriched, and then you have basal-like.you can see that the reverent -- the most favorably treated breast cancer -- the moreafrican ancestry you have, the higher the proportion of aggressive breast cancer thatyou are going to have, based on tcga data. okay? so when we unadjusted african-americanwomen more likely to have basal-like cancer, more likely to have her-2-enriched breastcancer, adjusting for age, more likely to

be basal, more likely to be her-2 comparedto your slow-growing luminal a breast cancer. no wonder population screening for mammographyactually doesn’t work, especially if what you do is to have mobile mammographies goto churches and then these women have to then wait to see who will take the breast tumorout or who will do the biopsy. so this is really hard data. then we asked -- there are some, you know,synchronous polymorphisms that have actually come out from gwas of breast cancer, and wewanted to know if those snips were genotyped in tcga, is it possible that there would belittle differences between the individuals who are genomically characterized as african-americanand those who are caucasian or other races,

and clearly 45 of the 56 snips had differentallele frequencies by race. so, at the population level, we know that the allele frequenciesof every gene that is important is going to differ based on your ancestry, and the diversityacross the african diaspora makes it even more challenging. then we wanted to know,are there some alleles that actually are going to determine subtype? some of these have nowbeen verified that they’re actually some -- you know, like the babam1 snip that predominantly,depending on the allele that you inherit, you might have er-negative cancer versus er-positivebreast cancer. okay? so looking at allele frequencies, then weasked the question, “is it possible that subtypes” -- the different types of breastcancer that you get -- “that it might actually

be heritable?” now, the challenge for thisanalysis is that the number of samples that we have in tcga to actually do this complexanalysis is only about 110. you need a larger data set to be able to really nail down theresults. but whether we -- when we looked at it, what we found was that, at least whenyou compare basal versus luminal a, or luminal b versus luminal a, her-2 versus luminal a,there’s a very high likelihood that the differences in the different subtypes of breastcancer actually is determined by differences in the kinds of alleles that you have, againshowing that your personal germline genome determines the type of breast cancer thatyou have, and i bet that it’s also going to be the same for all types of cancers. theseare really -- and then we asked, you know,

“how about for estrogen receptor? is it-- are there -- is it a heritable factor?” and you can see that, in fact, based on tcgadata, you’re more likely to have less expression of estrogen receptors if you have africanancestry genome versus not. it’s not so compelling for her-2 breast cancer, but theseare really sort of big data science that is allowing us to get some insight about thebiology of breast cancer and how it’s distributed across different race/ethnicity. so let me transition to what we have donetrying to actually replicate data that we get from, you know, european ancestry womeninto an african ancestry patient. so this is -- we’ve now formed a consortium tryingto look at, you know, everyone who has samples.

just give us your samples. so, samples frombarbados; african-americans; baltimore; philadelphia, pennsylvania. i’m showing you this justto tell you how challenging it is to actually do research that will be meaningful amongwomen of african ancestry or populations of african ancestry, just because the data setsare not there. and because it’s easier to get european ancestry data sets, we’re goingto have a lot more knowledge about the diseases that impact those populations before we getany knowledge of minority populations. that’s just the way science is. you have an idea;you have a resource to do it; you do the experiment; you publish your data; end of story. but thestory is actually just beginning because if we’re going to get to precision medicine,we need to do a lot of things.

so this is a very busy slide, but i show itbecause every time we try to replicate what others have done in european ancestry becauseof the linkage, you know, the short ld blocks in african ancestry, we can’t do it untilwe repeat the experiment, we do our own analysis, and then we can trust the data that we have,just because -- these are the top index markers in european populations, and when we lookedfor it in a polygenic risk [unintelligible] for african ancestry, it just didn’t doanything. however, when we mapped our own best markers in an african ancestry population,then we’re beginning to see alterations that are actually meaningful. and so thatreally has informed the kind of work we do, because what we want to do is to now poolall the samples.

this is the african-american breast cancerconsortium. we called our own african consortium the root because africa is really the rootof all of genetics. and so, in our discovery phase we had more than 1,500 cases and controlsfor nigeria, and then we used the amber consortium -- which is the breast cancer in black womenstudy -- to replicate some of our findings. so clearly, we’re looking for snips thatpredict er-positive versus er-negative, and by doing that, we’re beginning to actuallyidentify snips that have genome-wide significance that are able to distinguish women in thepopulation with er-positive breast cancer and er-negative breast cancer, and we arereally hoping that by doing this work we’re going to be able to develop a polygenic riskmodel for individuals of african ancestry.

so let me end by then thinking about the bigscience that we’re all doing which is really the signatures of mutational processes inhuman cancer. the most important thing that’s happening in oncology now is immunotherapy.the fact that there are all these new antigens that are created and that you can use thesenew antigens to develop therapies, immunotherapy, vaccines, it’s just unbelievable how wecan, in fact, advance the field. i’m excited about it because the genetic -- the epidemiologistshave always said that some of the risk factors or some of the genes that actually make youget cancer should also be targetable, right? and so when we’re looking at the mutationsignature, so everybody’s looking at the significantly mutated genes in breast cancer,and some of them, just like er/pr, they’re

druggable, and there’s a lot of new drugdiscovery based on this mutational signature. but what i think is actually really excitingis that we’re able to do that experiment in our nigerian cohort as well, right? becausethe genome technology is there and we can ask -- if you have er, pr, and her-2-positivebreast cancer, and you walk through the door in nigeria, do you have p53 mutation? of courseyou do. do you have her-2 amplification? of course you do. gata-1, all of the genes thatwe can test for right, is it important in your breast cancer? of course it is! so thequestion is, if it is, how do we get drugs? how do we get treatment? how do we actuallymake sure that these patients have their tumor personalized? so in the u.s., the same thing;the women who have been diagnosed in alabama,

in louisville, wherever they are, what isthe mutational signature of their tumor? if we really are serious about reducing healthdisparities and improving health equity right, what tools would they have to be able to dothis? i love the way that the aids activists actuallytaught us what we needed to do in cancer, you know, with funding. everybody went outand we had point-of-case diagnoses, and now we’re talking about the end of hiv/aids.in oncology we’re just not serious about it yet because we’re still only now escalatinghow we treat cancer patients by finding more and more druggable targets that extend lifefor a few months, where women who actually have diseases that can be cured with targetedtherapy are not being treated. so we’re

looking at mutational signature when you havetriple negative breast cancer versus not, and you have p53 mutation, what actually causesthat mutation to occur. it’s -- we knew in studies looking at aflatoxin, in hepatocellularcarcinoma, that there’s actually a particular mutation that occurs in p53 because peopleare exposed to aflatoxin. for our nigerian breast cancer study, we’re looking at kataegisand looking at the hypermutation that occurs in particular regions of the genome to giveus insight into what exposure might actually cause these cancers in this place where theexposure of a woman with breast cancer in nigeria is going to be very different fromthe exposure of a woman with breast cancer in chicago. what is common among those twoand what is different, and will that allow

us to begin to understand population geneticsand then gene environment interactions. so we’re really quite excited because whenwe then put everybody together and we look at if you have triple negative breast cancer,and here’s the west african breast cancer study -- and we’ve only done 40 exomes bythe time i put this together -- we have tcga black or african-american, 22. we had twiceas many nigerians that we’ve been able to do. and then tcga, 78, because there’s justmore triple negative breast cancer in nigeria, and that’s where we went to try and getit. it doesn’t matter; you can see their pc53 -- they mutate p53 at about the samerate. there’s no difference. why is that the case? however, if you look at all theirgenes -- take this in tcga; you’re going

to find that this gene is more highly mutated,and you can see that the spectrum of mutations among white women is different from the spectrumof mutations among black and african-american women, whether they’re in tcga or not. theseare small numbers, but they’re telling us something about how genes interact with theenvironment. so let me end by saying that there’s a lotof things that we need to do population stratification. who has breast cancer that just needs to haveroutine mammography? who has hereditary breast cancer? who needs genetic risk assessmentand needs to have more modern modalities or new modalities to screen them? this is wherewe are now, where we’re just telling every woman, “when you’re 50 or 40, go and geta mammogram.” it doesn’t work. we have

to do better. now we’re doing genetic riskassessment and we need to spend money where the money is to get people the care that theyneed. and i think that, at the end of the day, when we don’t have just one versionof what women should do, we’re going to actually be able to get to precision healthcareand get better outcomes. there’s very exciting work now being done in radiogenomics whereyou can actually classify from the first image that you get whether, you know, the prognosisof a tumor. so all of these things really give me hope that we will get to the pointwhere we can have precision management of individuals at the highest risk, and thenpeople who have low risk, maybe they don’t have to do much.

so a lot of people have heard about angelinajolie. i’ve been doing brca1 mutation analysis now since 1997 and she’s actually not thetypical woman who has the brca1 mutation. most women all over the world are not dyingto have their breasts removed. in fact, it’s a barrier to having testing done if you tellthem the only option is to go and have a bilateral mastectomy. yet in the u.s., the only waypeople know about things is when celebrities talk about them okay? and then they can’treally relate to that celebrity because they’re not going to have their breasts removed nomatter what. so what we know now is that not all women need risk-reducing mastectomies;we know that because we don’t have a way to screen ovarian cancer that we can actuallyreduce ovarian cancer death rate by having

people tested. this is why ob-gyns are now,all through their practices, asking women to get genetic testing, and we have a debateabout that. now we know that women with brca mutations actually need to be treated withparp inhibition, and now a lot of oncologists are doing -- are testing positive trials inovarian cancer, prostate cancer, and even head and neck cancer. the challenge we’regoing to have in oncology now are all these people who get tested because they were dyingof cancer, and then their at-risk relatives have inherited the same genes. what are wegoing to do? how are we going to treat them? so precision medicine -- i’m all for it,but i think we need to really think about how these individual characteristics can actuallybe learned.

so this is the ad from my hospital; they wantyou to come in and have, you know, specialized care, genomics, and personalized medicine,but you’re only going to come in if you have insurance, right? and then you’re onlygoing to come in if you can get into cutting-edge research. so how do we do cutting-edge researchin the places where people actually need it, right? how do we develop innovative clinicaltrials? how do we use genomics for preventative care and treatment? because, certainly, whatwe’re doing now isn’t working, okay? every oncology clinical trial, if you do subsetanalysis, will tell you that there’s black/white differences that the drugs don’t work. isit the drugs don’t work, or is it the wrong medicine for the wrong treatment?

so let me end by saying that i’ve had theprivilege of being a physician and a scientist, and it’s really allowed me to really havea different perspective of what it is we’re going to. i think we need to really beginto engage in community-based research and networks so that we can get to where patientsare, we can deliver care where people can actually access them, and we need all sortsof research in terms of implementation science, and we need to really think about how we dothis global cancer research so that we all benefit, because america is an immigrant country,and people’s ancestry does matter. so my lab is very large, but we have a lot of collaborators.the gwas data -- we had to put a consortium together. a lot of people contributed to it.and then, of course, the tcga data is the

work of a lot of people. so for team science,we need diverse individuals who are engaged, and then we need access to funding and thepopulations. nhi -- nhgri can lead the way. and thank you all. nigel crawford:that was a lovely talk. any questions? olufunmilayo olopade:any questions [laughs]. male speaker:thank you for the talk. i’m [unintelligible]. so i have a question. when you mentioned africanancestry, how much of that applies to the entirety of africa, given that there’s somuch diversity there? and even in the united states, you know, as we grow to become a morediverse nation -- people’s genetics, you

know, people from different ethnic groupsintermarry -- how do you think that will change the landscape? olufunmilayo olopade:yeah. so, the question is the analysis we did, we had to sort of see how much africanancestry are we going to really say is sort of driving the disease, and we came to “well,if you have 50 percent ancestry, then we’re going to categorize you as african ancestry,”whereas with european ancestry it was clear. i think those are the kinds of challengesthat we all have to solve as scientists. you know, when you do your analysis, are you goingto do analysis based on what? right? is it one percent ancestry that’s driving it,or is it just a rare allele? so i think the

challenge is really that there’s a lot ofwork to do. you know, i see that there are, you know,good statistical geneticists here, and population geneticists -- they need to help us, becausethe allele frequencies are different -- just talking about the u.s. then, if you go toafrica, every tribe has a different allele frequency, right. there’s so much different.so, to do black and white, it doesn’t quite get it. i hear you, and america is a meltingpot; there is going to be more and more intermarriages so that the distinction by color will getsort of morphed. and then the question is does it matter? you know, brca1 is brca1.if you have that disease -- if you have that mutation, if you get the drug that targetsbrca1, it doesn’t matter whether you’re

black or white or asian; it works, right?that’s why i’m thinking that precision medicine will actually get us to know whatwe’re talking about and then match diseases with the right treatment. female speaker:thank you for funmi for such a interesting talk. you mentioned in your somatic profilinganalysis, and so you mentioned that you found the patterns of the driver mutations varyby race, and so that might be related to the different exposures. and so i just wondered,like, you know, whether you could give us some idea about, you know, what kind of thoseexposures, and in your analysis, how you can link these somatic alterations with exposures?

olufunmilayo olopade:okay, that’s your job. [laughs] i’m a doctor! but that’s the thing, is that whenyou look at this and you see the pattern that actually varies, then the next question iswhat’s the exposure? is it common? and that, i think, is the beauty of global health. theapobec signature was first discovered and a lot of the shanghai breast cancer study,and the alleles and the different types of breast cancer in asia may be different fromwhat we’re talking about, but i think now we know that we should go and find the exposure,because now we have this big data showing us that something is smashing the chromosomeand causing this mutagenesis. what it is, we haven’t the opportunity to do it beforebecause we’ve always asked people about

their reproductive health history, their 'this,'and their 'that.' maybe that’s the wrong question that we’ve been asking. we haveonly 110 patients in tcga. we’re not going to be able to use tcga to answer that question,but, boy, i hope we’ll get money to go and ask the question in other cohorts. female speaker:right, and in the epidemiology study, i guess, you know, now this is the direction, and sowe would like to actually combine, integrate the exposures in the somatic profiles, butthen, like, you know, to begin with, of course, you can only do a certain number of the patients,like, you know, for the dystronomic [spelled phonetically] analysis because they’re verycostly to do. but then, like, you know, what

kind of exposures, like, in your mind -- becausewe have -- we will collect the epi data, the risk factor data, but what kind of the exposures-- olufunmilayo olopade:yeah, so, i mean, the important thing is that -- so one of the things we learned from ournigerian breast cancer study was every time we asked the questions about what we knowabout reproductive health risk, it was always in the opposite direction, right? the womenhad lots of children and they breastfeed for, you know, 20 months, versus african-americans,where the average number of months of breastfeeding was two months. so they’re not obese, they’rereally very lean, and they haven’t eaten a lot of hamburgers, so many of the risk factorsthat we associated with african-americans

here doesn’t apply in nigeria, and that’swhy every day we’re looking at our data and putting different hypotheses. and so whatwe want to do is an ecological study. what are these women eating, what are they doing,and why the breast cancer in a low-resource setting? so there’s a lot of questions,really, that are, i think, just now ready to be answered because we have the tools todo it. female speaker:thank you. female speaker:thank you for a great talk. i really liked your points about healthcare disparities andthe fact -- or the challenge of delivering healthcare to those who may need it most.and i noticed in your data a lot of your samples

come from nigeria, but there’s a lot ofcountries in sub-saharan africa with very unsophisticated healthcare systems and lotsof barriers to care, so i was wondering, like moving forward, what do you see as kind ofa vision for addressing cancer diagnostics in those areas? is it collaborations withgenetic information from those countries, or is it developing point-of-care testingthat may be applicable in those places? olufunmilayo olopade:yeah. actually, my point is about local/global. there are parts of this country that don’thave any sophisticated way to diagnose breast cancer, and we still have to think about howwe equitably distribute, you know, cancer care, even in this country. so i’m throwingit -- you know, i met with the fellows at

lunch and i’m throwing it at everybody.we all have to think innovatively, even for those of us who have big hospitals and knowthat it’s too expensive for us and we can’t afford it. our patients are going into financialtoxicity because they can’t afford the drugs that we have worked so hard to deliver tothem. so i think that we all have to work on this together. i think about the low-resourcesetting and how we can accelerate progress by getting the studies done quickly and gettingpoint-of-care diagnoses so that no matter where you are, you have a chance to survive.that’s just how i see it. when we started this, you know, delivering her-2-targetedtherapy, even in chicago, northwest indiana, there would still be people who managed tocome to our clinic and had been seen by their

local oncologists. there’s a 20 percentdisparities gap between where you’re treated and your outcomes, just because there arenot enough resources. so equity, health equity, is a question that we all have to answer.we have to think outside the box about how we all do it together, whether it’s in nigeriaor it’s here. but we’re in this global community and we have to work together acrossafrica, across latin america, asia. i mean, there’s just a lot of places that need help,including my neighborhood. female speaker:thank you for a great talk, and it’s actually got my mind exploding about -- in trying tomake a message to the public based on the data you showed today. if 30 percent of thetime a mammogram isn’t going to show a cancer,

yet in a such a high percentage of women you’retesting who already have breast cancer they’re showing that they have mutations -- in yourcontrol group it wasn’t very high. i mean, women who don’t have breast cancer don’ttend to have these mutations, especially the age variation, so what’s the message? if-- because i think in our attempts, the public health service, to send the message out thatfor most women -- most, whatever that means -- not stratified, under a certain age youdon’t worry; over a certain age you go more often; and what you’re saying is the mostaggressive cancers are in the young women that may have mutations. so what’s the soundbite that we give to the public, if there is one?

olufunmilayo olopade:well, you know, the prevention task force actually already gave you the sound bite,but we rejected that as a community. there’s no data that mammography helps anybody right?and when we looked at the data and we looked at it and we massaged it, politically we couldn’tsay, “don’t get mammograms.” we said, “talk to your doctor and have your doctorrecommend when and how you get mammograms.” so now the affordable care act and a lot ofcancer prevention guidelines say for every woman, do risk assessment. the ob/gyns havebeen very aggressive about it. they pushed out information to all ob/gyns and said, “testeverybody,” but the geneticists were not ready. we said, “there are too many variantsof unknown significance; it’s too expensive;

this is a -- we can’t do it, we have todo the experiments first.” so i don’t think we have a public health message otherthan what the task force has said, which is, “talk to your doctor [laughs].” and mostdoctors don’t even understand the nuances, and most people are so afraid of genetics,and yet, you know, you know your cholesterol level. you know, you know, if you’re obese,your risk factors. in cancer, it’s the genetics that we knownow, and until we know all the other things that we can measure, it’s a hard sell, buti tell -- i mean, my clinic is a cancer risk assessment clinic. many people are embracingthat, and i’m trying to help families to know when to go for screening. mary-claireking and i have actually come up openly to

say every woman at 30 should just have a genetictest done, and everybody said, “oh, it’s too expensive.” well, if a test is $200,i’m sure everybody wants to do it, but if it’s $3,600, then we can’t do it. so thecost is coming down and i think at some point people are going to have the data to knowthe genotype/phenotype correlation, and people will be able to have access to it. the questionis will it come down enough for everybody, and will -- i mean, point-of-care diagnosis,just like cell phones -- we all use cell phones. in the most remote places in africa they havecell phones, so hopefully the technology will really drive, and the innovation will drivehow we move this forward. but the time is right to actually do it. i don’t know ifi answered your question [laughs].

female speaker:you just opened so many more, but [inaudible]. olufunmilayo olopade:good! that’s the whole point. that’s why i’m here [laughs]. male speaker:fumi, that was a lovely talk. there’s such diversity from each tribe within africa, youwonder if you’re better off to go to a place which has much less diversity. i go back tothink about the japanese migrants. when japanese women had low rates of breast cancer, butwhen they come to the united states their risk goes up over the time of a generation,probably not because their genes are changing. is that an epigenetic phenomenon? are therespecial genes that change or mutate? what’s

happening there? olufunmilayo olopade:yeah, thank you. that’s actually part of my talk that i didn’t get to, because i’veonly really laid the framework for genetics. i’m coming back for another talk in octoberand then we will talk about the epigenetics, because, in fact, it’s only -- even if itell you about, you know, one in five or 22 percent for brca1, we’re finding that, infact, the regulatory region and the epigenetic changes in these tumors might actually justbe related to exposure and to other things that are not based on your -- that you wereborn with a mutation. so the field is exploding, but you don’t get epigenetics without yourgenetics, and that’s why we really need

to do a better job with measuring exposure,measuring population differences, and then figuring out what exactly should we be measuring.so i think, you know, the genomic revolution has given us new biomarkers. it may be thatthere would be blood bond biomarkers that we can look at that would be better predictorsthan doing dna analysis. but i think that this is what we have now, this is where weare, and the field will advance the more we are able to do these kinds of studies. female speaker:hi, fantastic talk. so i like the direction we’re going, because we’re opening a discussionabout risk and knowing the differences between risk in people from african descent in america,but we also need to look at more druggable

targets. how can we have that conversationwhere we are looking at other targets and getting targets that potentially are alreadyfda-approved, and seeing if we can use some of those already to look at the triple negative,like efgr, for example, and maybe -- i know some of the triple negative have -- claudinsare involved. so are you looking at some other targets, and how can we get that conversation?because, at the end of the day, you want to treat the cancer. olufunmilayo olopade:yeah, thank you. what we actually found -- my first experiment at the university of chicagowas to look at our database and to look at women that our surgeons had treated a longtime ago, because chemotherapy was sort of

-- you know, after the manhattan project sortof folded, leon goldberg was an oncologist that was really treating leukemia patientsand cancer patients with chemotherapy, and if you didn’t even get any new drug to themarket, but you treated everyone with triple negative breast cancer with chemotherapy,and you treated them -- that woman i showed you with a small triple negative breast cancer-- she had [unintelligible] chemotherapy; her tumor disappeared. and when you have pathologic[unintelligible] remission with triple negative breast cancer, it never recurs. it’s justif -- you know, 50 percent of them don’t recur. so the issue is there are drugs thatwork, even for triple negative breast cancer. the problem is most of it occurs when it’salready too advanced for us to actually use

effective therapies. if that woman had been on the south side ofchicago and she felt the lump, the time it would take her to get to me or to get to anybodyto treat her would probably be another six months to a year. and the thing is these tumorshave a high proliferation rate and they double, right? they’re not slow-growing. so whathappens is the mutation burden then increases; the tumors are beginning to mutate and havedifferent changes, and then they become resistant to therapy. so, i think that implementationscience means what we know works now; let’s do it, and let’s not wait for a perfectsituation to start implementing what we know works.

so if we can use chemotherapy, use it well,support women in the communities, a lot of women with triple negative breast cancer willbe cured of their breast cancer, and those that are not cured, they should be part ofa research to find new therapies, and there are certainly new therapies. immunotherapy;vaccines. there’s just a lot we can do, but we need to organize ourselves where women,wherever they live, can have access, because not every woman can come to the universityof chicago or come to the nih. so how do we equitably distribute resources in communitiesso people can get care? on that note -- nigel crawford:thank you.

[music playing] [end of transcript]