NIH CLINICAL CENTER GRAND ROUNDS
Episode 2009-013
Time: 54:38
Recorded April 8, 2009

CONTEMPORARY CLINICAL MEDICINE: GREAT TEACHERS
Hereditary Cancer Predisposition: New Challenges
Judy E. Garber, MD, MPH, Director, Cancer Risk and Prevention Program
Dana-Farber Cancer Institute, Associate Professor of Medicine, Harvard Medical School

ANNOUNCER: Discussing Outstanding Science of the Past, Present and Future – this is NIH Clinical Center Grand Rounds.

(Music establishes, goes under VO)

ANNOUNCER: Greetings and welcome to NIH Clinical Center Grand Rounds. We have a special "Contemporary Clinical Medicine, Great Teachers" lecture for today's podcast. The topic, "Hereditary Cancer Predisposition: New Challenges" Our speaker is Dr. Judy E. Garber, director of the Cancer Risk and Prevention Program at the Dana-Farber Cancer Institute, and Associate Professor of Medicine, Harvard Medical School

We take you to the Lippsett Ampitheater at the National Institutes of Health Clinical Center in Bethesda, Maryland, where So Dr. Mark Green, chief of the Clinical Genetics Branch in the Division of Cancer Epidemiology and Genetics will introduce our speaker.

GREEN: It really is a privilege and a pleasure to -- for me to welcome her back to the NIH. She's back and forth a lot actually so there are big gaps in time when she's here for one reason or another. She's a long time friend, colleague and member of the NCI research community who, as you will soon see for yourself, really does embody all that is implied by the appellation “great teacher”.

Her medical training began at Yale School of Medicine where one of many present decisions Judy combined in a M.D. MPH program. This was back in an era when the term population science was regarded as an oxymoron but she was very much ahead of her time and that public health training has played a major role in her career as it has unfolded. During her time in medical school actually she spent some time in The Public Health Service Co-Step Program where she and I first met. During which she was formally exposed to the world of cancer epidemiology and genetics working as a student in what was then NCI's division of cancer etiology where she met Joe Fremeni and Fred Lee who became her career role models and life long mentors. In fact, her first peer review publication came out of that experience. It was entitled, autoimmune disease in a family there by foreshadowing the path in the career would take.

After medical school she abandoned New Haven for Boston, a city in which she has ever since spent her career at various Harvard affiliated institutions including the Harvard Medical School, Brigham and Women's Hospital and the Dana-Farber Cancer Institute where she's currently associate professor of medicine and director of the cancer risk and reprevention program.

During her time at Dana-Farber Judy published more than 120 peer reviewed articles and some 30 book chapters and journal editorials. She's mentored 12 post-doctoral students, four who are now professors themselves. She served as a key faculty member on 24 major CME board review and cancer genetic educational courses. As you can tell we epidemiologists love to count things. She's lectured widely and frequently at home and abroad, collaborated extensively with other investigators both CME and international and has -- domestic and international and has emerge as one of the most articulate thoughtful advocates for women's health in general and her red tear breast cancer in particular.

She's one of a handful of investigators in the U.S. who have become what I like to think of as triple threat investigators. These are people who can combine clinical oncology, genetics and epidemiology, a set of expertise that is really hard to come by, into a remarkably successful and productive, truly multi-disciplinary career that encompasses the desk where those of us in this arcane business spend too much of our time, the research lab, the clinic, the classroom and the community. Dr. Garber will speak to us this morning about evolution of clinical cancer genetics in breast cancer and I would ask you to join me in welcoming her to the NIH today.

[Applause]

GARBER: Unfortunately I can't be that articulate. Thank you very much for this wonderful opportunity and this remarkable title to be called a great teacher even before the lecture. So I'll try to live up to that and I'm delighted to be back here.

I do come because I'm on the board of scientific counselors now so that's not a reason to be nice to me today. I will not translate that into your review. I do have one conflict of interest to disclose, one relationship I'm doing a clinical trial with Astra Zeneca and I will mention that at the end a paradigm trial but not offering uses as anything else. Though in my clinical life at my institution has many collaborations, I couldn't tell you who makes which drug so I don't list all of those.

I was asked to begin with a vignette. This is from a patient that I saw in last couple of weeks. And there are many possibilities, this was just one to think about, maybe a different way to think about cancer genetics for today. This is a 38-year-old internist married, two children, has a remarkable family history of which I'll show you a small portion in pedigree form, breast ovarian and pancreatic cancers whose cousin was found to have the BRCA-2 mutation. She she decided she didn't want to be tested because she wanted to have another baby. This is the family. Her age is incorrect. There is her cousin. With the mutation on the paternal side, you can see all the features that would help you think about this as a family where someone should be tested. She did agree to initiate breast surveillance and on exam actually had a breast mass. And this was confirmed on MRI, turned out to be a very small ERPR positive her-2 negative breast cancer and she will be treated with ma second to my and have her ovaries removed so afraid there will not be a third child. She will have testing since it's likely this is related to the mutation.

What we're going to tuck about today is not why she should have started her screening earlier or how all of this fits into her management but the fact that now when we think about breast cancer or think about genetics we really are trying to think from the beginning, from keeping healthy people healthy to the fact that we sometimes do find early tumors and we hope that that also will ensure their long term survival and that their treatment will be appropriate. Even though it includes right now a lot of difficult decisions. So this is a more classic family even e though hers certainly was classic. This family has many features one wants to recognize today. The early onset clustering of breast and ovarian cancer, the presence of cancer in multiple generations, and fact that this slide still is the ASCO slide, it show this is in the paternal lineage, because every time I take it out somebody asks if you can get breast cancer from your father's side so I leave it in now. We started to study her cancer syndrome in the NIH, Mark and his career earlier before he came back to the NIH. Fred we certainly partly in recognition the magnitude of the risk inferred by her hereditary was likely to war dwarf the contribution to risk that you see from other exposures we knew about from epidemiologic studies. This is no longer the only focus of cancer genetics but in 2000 when this slide was made that was true.

What else did you need and the other risk factors and that has been challenging. We started this all as small clinical activities. We checked families. We did genetic testing under research protocols to keep everything out of the medical records at that time when there was no real protection against discrimination. Here we are 15 years later and if you lived in New England you would have been part of a television ad campaign convincing people that they should be getting testing and making sure testing was offered by internists and gynecologists preferentially over genetic program. So genetics is out there for all of us to see and we'll talk about other ways that's true. At this point we've learned several things over the course of time and for those of you who might actually see patients still here in the clinical center which looks very new since I saw it last, it's easy when someone comes in there with a mutation in the family, that reminds you whatever you knew about their history there might be reason if it's the right side of the family to think about testing.

Early onset breast cancer still a clue early onset disease for any hereditary syndrome. In these families ovarian cancer, fallopian tube cancer, that's connectible. We'll talk about that more. People do not come with specific family histories always. They tell you about stomach cancer that was at least in this case you have to think it could have been ovarian or pancreatic. Now we know multiple pancreatic cancers are clearly part of the syndrome and we also begun to think about incorporating histologic features of the tumors in our thinking. Some pedigrees come looking like this where here is this woman at 32 who I'm sure didn't realize there was increased risk because what I showed you is a family where everyone was male. In the recent generations she might not have realized this was being passed from father to son first. And for this young woman who had ovarian cancer but mother might have had a THPO in the past which removed ovarian and breast cancer risk and she might not have realized her branch of the family was still at risk and why her daughter's cancer occurred younger would be a question for many collaborations to address. That colon cancer of course could have been colon cancer epidemiologic by frequency abdominal cancer should be colon cancer in this country but in a family like this maybe that was ovarian, hard to know.

So the prevalence figures haven't not changed or been refined very much over the years. Population estimates about 1 in 800 overall in founder populations of which there are numerous but the most frequent in the U.S. the Jewish population, the prevalence is 1 in 40. That affect it is prevalence of mutations in any other group by breast cancer overall this is a small component of the syndrome. Breast cancer diagnoses, maybe only 5%, maybe 10% in Jews and the place to look is young people. So when women are diagnosed at a very young age where breast cancer is a much rarer disease, then it's reasonable to ask is it genetic. But still only 10% will be BRCA-1 or 2 which is surprising. By the time you get to 50 which is not that ole, 3% genetic and ovarian cancer prevalence always much higher so raises questions about whether we should be testing all ovarian cancer patients or whether we should be refining that by other family history as well. These are some of the histology issues just because I mentioned them.

There has been an observation that BRCA-1 associated tumors are frequently ERPR and her2 negative and BRCA-2 tumors are more frequently ERPR positive, they're also her2 negative. This is always interesting, these tumors class -- cluster with the basal like cluster here on microarrays, they are in our hands indistinguishable from other basal like tumors. Others have debates about the basal keratin. We did a study at Harvard looking at about 180 women who had been seen in the breast practice given blood for the spore tissue bank and had tumor and we looked at their features and showed that though it's hard to tell from this slide, that the chance of having a BRCA-1 mutation was increased 3 to 4 fold in women based on what you would estimate by family history if their tumor was triple negative. BRCA-2 carriers can have triple 2 care tumors so it doesn't change your practice completely but should raise the probability when you see the tumor. Sometimes people with the histories are negative. This remains a challenge.

We don't have BRCA-3 though the original prediction by Francis Collins would be six months from the identification of BRCA-2, that was almost 15 years ago so not so easy. We have other syndromes. One of the most interesting turns in the past was the identification of the connection between FRANCONI anemia and the BRCA-1 pathway. A child I will not see since I only do adults but the facial feature, thumb abnormalities and radiated chromosomes forking bizarre chromosome formation. Observation on the left by Alan's group, that it was possible to have two BRCA-2 mutations that would give you the FANCONI anemia subtype D-1 so these are compound heterozygotes. There have not been a true homozygote and not in the ASHKANOZI population either so there's debate whether these BRCA 2 mutations are low penetrants which is not totally clear. But in this syndrome, in the family where this child has FANCONI anemia, the parents had a different BrCA-2 combination. And found other tumors in these children with -- who are compound heterozygote medulla blastoma and wimm's tumor reminding us there's a lot we don't know about the rare condition but raising the question from the reproductive standpoint, one more aspect to consider, patients with a BRCA-2 mutation considering reproduction may need to think about typing their partner. It's rare but this is not a place you want to go.

This is a diagram from Mary Claire King looking at the relationship among these pathways. Here is the fanCONI cluster but these are part of the same pathway. Besides BRCA-1 and 2, PALB-2 has been found, David Livingston's lab, it binds to BRCA- in the pathway, not so commonly associated with her hereditary breast ovarian cancer but in a science publication in pancreatic families. Check 2 for those who paid attention, all these things together not clear why some are more breast ovarian and some are other cancers. Those things still have to be worked out but now if you find a gene by studying DNA repair you have to look and see if it's a fanCONI gene and you study a fanCONI gene you have to see if it's a breast cancer gene. I don't think we would have anticipated that interconnection in the past.

These are examples, for published to date. Here is LI-FRAUMENI syndrome which got me into the, this is an example family with all the features of LI-FRAUMENI, multiple affected individuals in an autosomal dominant pattern, very early breast cancer as par of the syndrome with the rest of the malignancies. This is rare, less than 1% of breast cancer, not something we find often though Mark and I are now about to collaborate we hope on a project looking how to screen these families more effectively. Syndrome which exists, it is associated with mutations in P-10 and has a number of features, a challenge to identify but the risk of breast cancer in cowdens is in the 25 to 50% range so we're identifying. CDH-1 carriers are individuals with diffuse gastric cancer, that is primarily the manifestation of CDH-1 mutation. But they get lobular breast cancer, I'm not a pathologist but they look alike. It's an adherence molecule, this would be true.

We have looked at a series of individuals with lobular breast cancer and no gastric cancer in their history. In the small seriess we did find one CDH mutation but in the larger series friend presented at ASCO the prevalence is low. This is not a public health hazard because individuals can't be screen, they have to have prophylactic gastrectomies to avoid the cancer. Craig YAGERS, you think of as a GI syndrome has a breast cancer component. 45% in the large UK series from 2006. The problem is we have only explain ad small amount of breast cancer so there are other genes raised as issues. None of these have been used clinically in the US.

This is CHEK 2, the gene Dan identify when studying a large -- and he found the 1100 mutation in a family that looks like it could make it LI-FRAUMENI criteria, this is analyzed in thousands of the cases. Associated with small risk of breast cancer, male breast cancer, colon cancer but a small risk, about 10% increase overall and we have argued that this does not ready for clinical testing because we're not sure what to do with the information. Basically I think this is a gene that's associated with failure to immigrate to the U.S. Because it's very common in Scandinavia, less than 1% in the U.S. population so we done have to worry about it too much but our colleagues do. So we have not explained this. As you know, there are large collaborations here at the NCI, Steve Chanic and others working to find the other genes for breast cancers and there have been results of those publications from studies looking at SNPs. One is an identifiable gene, the others are still being clarified.

In work not done by us to show one of these at least is a modifier of BRCA-2 penetrant so there is at least a link. These do look to be genes related to ER positive breast cancer by and large. But much more work to do to sort this out. I think the estimate is still there probably are 50 more genes like this, low penetrants but accounting for a larger attributable risk of cancer in the population. You don't have to wait. You can go to 23 and have yourself typed. You can go to NAVAGENICS and be typed for multiple diseases, go to Harvard site and join George church and putting your whole genome on the web for whatever reason you want to do that for research he says. Our genomes are coming to be part of our lives.

Fortunately I have not yet had anybody bring me their print out in clinic so I can figure out what to do with a relative risk of 1.2 but I love this cartoon from the New Yorker because it speaks so well and because it has a mutation in it. A little typo let's go back to the BRCA-1 and 2 carriers.

What do we do when we find them? That was the question at the beginning, why would anyone want to know, and how much risk they have and what to do about it? How much risk is expressed in large ranges of risk. And the motivation for some of the large collaborative hunts for modifier genes has been to try to refine these risks because as you can imagine, it's -- everybody understands hi but it's not reassuring to get these 50 to 80% numbers. The ovarian risk may affect management too, much higher for BRCA-1 than 2. You'll see this come up in the talk. Modifiers we have learned I think it is fairly crude but interesting. Genetypic variation, there are the mutations in the ovarian cancer cluster region. The large central axon is both true for BCCA-1 and 2 but not validated well. Modifier genes. The cohort affect is one one of the most interesting observations repeatedly seen. Women born in the first part of the last century having the same mutations as their daughters and granddaughters have had their breast cancers at later ages.

This is not anticipation where the gene itself changes when being passed from parent to child. These must be other factors, either other mutations or lifestyle factors and pretty much everything from diet and exercise have changed between our grandmothers and ourselves so this is not easy to sort out but encouraging at least that it's not just your genes that dictate your destiny. Exercise is now shown in two studies to affect breast cancer penetrants and then the huge affect of hormonal modification. So just not to forget to mention oral contraceptives because for clinical audiences there's always a question whether kids can take birth control pills if they have breast cancer in the family. Generally the answer is yes, it has little affect with modifications but in BRCA-2 families there maybe a substantial increase in risk for women who use birth control pills for a longer period. They have been shown to reduce ovarian cancer risk in these families so this is a management issue and personally as the mother of a reasonably young daughter I think preventing teenage pregnancy may trump all of this but it is certainly there as a biological clue to the evolution in these settings.

We have learned to do some surveillance for women. Everybody has a slide where they have a mammogram that doesn't show a cancer and MRI that does so that's no trick but it is true from the data that you can affect sensitivity at least. These are the six large studies, none in the U.S., the closest one is Warner in Toronto. The rest are European. But they have done studies where individuals are screened annually with all this technology together to compare sensitivity. I think generally you can see fairly disturbing lack of sensitivity for mammogram, and much better in general sensitivity from MRI.

So Barry Kramer who is sitting here, -- would not be happy if I didn't remind everybody that you done judge screening technology by sensitivity. We have to have change in outcome and no study has shown that though it's hard to believe that you don't have a stage shift and based on hope and optimism this has become part of at least the NCCN guidelines and the European guidelines for managing individuals with mutations. We don't know whether these data come because women are generally younger and we know mammograms are less sensitive in younger women with denser breasts or whether their tumors are different and they don't show on mammogram. At the moment we don't know the difference. It may not be so important clinically but it is a valuable question to consider. The guidelines published for MRI and breast -- cancer in general are based on the data you have just seen which you realize is quite limited and really very limited only to mutation carriers but it has been extended to syndromes that are arer and harder to study like LI FRAUMENI syndrome, other inherited disorders, radiation where there's studies ongoing but all small. I hope MRI better, I think these studies have continued and the other guidelines say basically we don't know whether we should be using MRI for women with denser breasts because we can't see well with mammograms. That remains clinically a challenging area at this time.

The screening issues do raise the question of whether MRI alone is okay. In Germany for example where they -- Kristina cool's data shows almost 100% sensitivity with MRI and little with mammogram. They want to get rid of mammogram overall. We worry about radiation risk. Mark Robertson had a paper quantifying the risk where he still recommends mammogram and we don't know whether DCIS is better found with mammogram or MRIs. So we have work to do though we are at least partially comfortable recommending MRI.

Mark Green 1099 study that the large screening study with the GOG will hopefully provide the material to help us do early detection for ovarian cancer at the present time we really have no decent ovarian cancer tool for early detection, neither ultrasound nor circulating markers. We don't recommend anything but the patients gynecologists recommend regular screening so they spend a lot of time recommending normal image. Nothing you can do about that. But part of this situation has led to recommendations, strong recommendations for prophylactic hysterectomy. Here are data from our large collaboration, looking how effective it is in reducing the risk of ovarian cancer. From the top number from the combined data which is what we have always had, that the reduction in risk is about 90%. The worry is always peritoneal carcinoma from cells. In BRCA-2 where the baseline risk is 15 to 20%, the surgery is almost completely protective in this data set but there's individual risk in BRCA-1 carriers who began with earlier risk. The surgery works but the issue is timing.

This is an observation from our group at Brigham headed by Chris crumb looking at the fallopian tubes of women with BRCA 1 and 2 associated prophylactic oopherectomies and these lesions with P-53 loss and other features, they are raising the question whether in these women the tumors arise in the fallopian tubes or always in the fallopian tube. It required a change how the tubes were sectioned. They had to section them longitudinally to look at the FIMBRIA and you can't go backward and see the old tube and figure out whether it was the women with subtle abnormalities that developed the peritoneal carcinoma but hopefully going forward this type of thinking will be informative in trying to fill this desperate need for ovarian markers. There is debate whether we should be doing -- taking the uterus or not in these women. There's certainly varying feeling around the country and around the world. There's the usual gynecologic reason, no specific endometrial risk associated with these tumor, even PAPILARY carcinoma is controversial in this setting. It makes it easier for women to take estrogen alone, hormone replacement in the uterus is gone.

From the women's health the data is that it does not increase risk and large studies suggestion that estrogen alone may even reduce risk. Is that enough reason to take the uterus, it's controversial where I live. But surgery has gotten easier. The issue of inducing menopause remains a tremendous problem. This is still the largest series trying to assess the affect of MRT after in women with mutations looking to see if anything reduced the advantage, modulated the protective effect of the oopherectomy and there was none but the numbers are tiny when you look at how many women who otherwise participated in research.

We're busy collecting amounts of epidemiologic -- reproductive data hormone replacement and supplementation, all the things patients do to get to a large enough data set to have a meaningful comment and what you find instead at the moment is wide practice variation and whether women are given hormones are not, which a big deal when you're 35 and suddenly Menopausal surgically. I'm sorry not to have reproduced them in larger form. These are concerning but again, a small epidemiologic study, again from Steve Narod’s group in which women with BRCA-2 mutations who took Tamoxifen had increased risk of endometrial cancer. Only BRCA-2 carrier, not 1 and the use of Tamoxifen was for whatever whether treatment or prevention. That's disturbing because the 2 carriers with the ER positive disease more likely to take tamoxifen or may be taking it for prevention but this needs to be reproduced. Tamoxifen reduces the risk of contra lateral breast cancer.

In this treatment trial, so women who had in this case this a study in Canada who either had oopherectomy for whatever reason or Tamoxifen versus those who had neither were less likely to have a second breast cancer. This is a confusing study. For some reason estrogen receptor was not typed so we don't know whether these were ER positive tumors given Tamoxifen which you expect since it was a treatment trial or not. And this speaks to this question which I'll try to deal with more of whether hormonal interventions also protect against triple negative tumors of BRCA-1 carriers. These are the original first prospective data from New York versus our larger episeries looking a women with mutations in the upper line those who had prophylactic oopherectomys pre-menopausal and other cohorts an regulartives who did not, showing 50% reduction in breast cancer risk for women who had pre-menopausalooperectomy.

These from the same analysis looking at this overall showing in the combination of BRCA-1 and 2 patients which is all of our studies up until now because we didn't have enough to separate them, the 50% reduction is reserved. In BRCA-2 the reduction is almost 70%. In BRCA-1, the reduction is not statistically significant T. question is, in the old studies when everything was combined where two-thirds of the patients are always BRCA-1 in the U.S., does everybody have some affect? We just done see it well here, or was it really all along a huge effect in BRCA-2 and not an effect in BRCA-1?

This I think is an important issue because many of the patients we see come for testing because they're ready to have their oopherectomy and hoping to get breast cancer protection out of that as well and we would like to be clear to reassure them that's happening that it really is happening. If this is true, and we are not by these data but by meta analysis published in the other individual series that are beginning to separate BRCA-1 and 2, where several still see a very clear reduction in risk in the BRCA-1 carriers then the biology of these tumors is different and this is a place where the role of estrogen has to be clarified in the management and biology of BRCA-1 associated breast cancer.

Here is one depiction of the number I keep referring to in which in this series from will folks, women diagnosed with breast cancer and BRCA-1 mutation 445, 80% had ER negative tumors, whereas 80% of BRCA-2 tumors were ER positive. Genes almost sequential in the pathways of DNA repair, leading to very different tumor biology, not well understood. And this question still that the proportion of ER positive tumors does seem to rise with age. So are these BRCA-1 associated breast cancers or sporadic breast cancers on background of BRCA-1 mutation? We maybe limited by technology but we'll see. Important data from the large collaboration, this time analyzed by Susan Doncheck asking do we change mortality for the surgery we're doing to people? The answer is yes, both breast specific, ovarian specific mortality and overall mortality in this group.

So we can at least say to women that if the bottom line is survival, surgery will get you there. That's not the only thing they are thinking about at the time. And while that maybe a good surrogate for ovarian cancer diagnosis in many cases for breast cancer that's not necessarily true. Many women survive their breast cancer so this is an important thing, this is what we're aiming for but how we're getting there is still a challenge.

So here is the prophylactic mastectomy data from our large collaboration with Tim showing just if you like curves instead of figure, numbers, the marked reduction in risk from prophylactic mastectomy. The figures are about 90% reductions in all these series. And the issues now are how do we deal with women who need to have this surgery? So these are pictures from the Internet, not from patients of mine. We don't do that. But the issues are in patient whose are found to have a mutation at diagnosis, should they -- if they choose to have mastectomies can we do part of mastectomy, do we have to give radiation, it affects COSMESIS. This needs to be studied.

There's techniques to improve the cosmetic outcome. These are young women, very often. They are thin, they don't have a lot of available autologous tissue for recreation of the mound. You need to have better implants and that seems to be happening. There is debate about whether you should take lymph nodes in women with prophylactic mastectomies what's the okay cult rate of breast cancer after negative images? It's quite low. Not zero. You see various surgeons with various opinions about this, usual when there's not enough data and the question is can you leave the nipple, this is a huge issue from women from the cosmetic point of view as you can imagine. No one has nerve function to give them sensation but at least the cosmetic outcome will be different.

This is just from a project we have been doing with Barbara Smith at Mass General where we're trying to really clarify the anatomy of the nipple to see whether it's possible to carve out the nipple from underneath and still leave enough tissue for the tissue to be viable but be able to reduce the risk by removing the duct. Women have very different feelings about that little tiny bit of residual risk, if you're going to have the surgery why not do it all. If you're going to have the surgery does it matter if you leave a little bit? Hard to know.

There are data on satisfaction with prophylactic mastectomy. These are the Mayo Clinic data looking at women with surgery more than 20 years ago. I shouldn't say it that way but most women are satisfied in these series. The major factor determining the satisfaction was women deciding to have the surgery themselves not forced by well meaning surgeons or anyone else, but of course they're invested in being satisfied. And the disturbing part of these data was the number of women who needed reoperation to be able to be satisfied with their outcome.

We did a small study looking at women who were tested at diagnosis to see if their genetic outcome -- test result would affect their surgical decision, how many went on to have bilateral mastectomy versus those who didn't. The biggest predictor of that decision was genetic test result but the next bigger is whether they had an MRI and age and referral from the surgeon. This speaks to an unintended consequence of this work in genetics. These are data from Tuttle looking at the increase in contralateral mastectomy at diagnosis, women with breast cancer who do not necessarily have genetic risk but this shows a 5% increase in bilateral mastectomies over the years ending 2003 and he has continue third degree work. There's no sign this trend will abate. It may be that in Boston when I started out there were no surgeons who had ever done bilateral mastectomies for prophylaxis. They had bilateral mastectomies some but mostly breast conservation.

If a woman asks about the other breast because she didn't want to go through it again, they said risk is small and you can have a good result, now they say not that hard to do, okay. So we have a huge increase in contra lateral mastectomy which is not necessarily driven by true cancer risk. So we are not doing a good job communicating about this. So this speaks to the fact that although oncologists don't tend to think about it so much despite all of Mark's work in educating them, there are implications of BRCA-1 and 2 mutations for cancer patients as well as their healthy relatives. There is not so much concern about radiation. This is certainly a concern, whether the radiation, the affects of therapeutic radiation would appear in survivors who had a BRCA-1 or 2 mutation and on the left, in data with Lori pierce you see not for the IPSA lateral occurrence. But the contra lateral cancers of course are increased and that's what you would expect in a condition where every cell is at risk in a sense.

So there was reassurance that radiation was not a problem and for those patients who chose breast conservation they had no reason to expect an inferior outcome or an increased risk particularly of sarcomas or other cancer bus these numbers are not huge.

These are the data from Israel, one of the only places that could really do this at this time. They looked at women treated 20 years ago and divided them by typing their tumors because they could look at the 3 founder mutations which you then can do in fixed tissue without having to have full genes for germ line analysis. And they separated BRCA-1, BRCA-2, and non-carriers and show in all groups outcomes are the same as long as BRCA-1 patients got chemotherapy. That fits to the triple negative phenotype we have been looking at. And that's very reassuring those women got CMF 20 years ago so there was not high test chemotherapy not because these tumors are necessarily more aggressive. Because of the understanding of the function of the genes though, it's become possible to think about targeting therapies to these specific defects and asking whether they extend further.

So this is just a schematic diagram I stole from Alan Ashworth looking at double strand break repair. Here, so wherever we looked at the fanconi anemia chromosome, you can see the same chromosome types when you radiate BRCA-1 deficient cells. These are cells that have no functioning BRCA-1 in this case compared to the normal cells of a carrier who at least had one function BRCA-1. It was to some extent the observation that these defects were similar that led to the linking.

This is from a basal like tumor with an intact BRCA-1. I'm asking the question whether there were -- there's a BRCA-1 phenotype in a tumor that might be related to other genes not just BRCA-1. This is a model this our group has been thinking about, this is with David Livingston and Dan Silver and Andrea Richardson asking can you have tumors triple negative that have DNA repair defects with a BRCA-1 mutation? Of course but there are others with intact B RCA-1 that show the same defects. So are the tumors similar? These tumors are the triple negative phenotype and there are many possible ways to target treatment in this.

This is a very hot area because these tumors have a worse prognosis than others. We looked at the question whether there's something to learn from the fact that BRCA-1 associated tumors are triple negative and DNA defects we can exploit in treatment. These are data showing radiation can cause this issue to cells missing their BRCA-1, causes double strand breaks in the DNA, similar data.

Here is data from cell line, not our group showing in a BRCA-1 one minus cell there's more sensitivity to -- the drugs we usually use for breast cancer treatment any way. And that you can show that sisplatin sensitivity in the lab, this is Dan's data.

So we started thinking about whether we should consider a trial with platinum, a drug we don't use in breast cancer much. These are data in o ovarian cancer showing improved outcome, BRCA-1 or 2 underlying ovarian cancer versus sporadic colleagues and only oncologists think this is a good outcome but it's better. Was that the platinum? So we went back at platinum in breast cancer, it was developed about the same time as taxane which had a better therapeutic index.

The drugs were protective but not if given after other chemotherapeutic agents so they were dropped from breast cancer treatment. We decided to go ahead and designed a neoadjuvant study we did based on data showing that in patients with neoadjuvant treatment, chemotherapy before surgery, those who had a pathologic complete response disappearance of all tumor did very well. Whereas those whose tumors did not disappear with treatment had a much inferior outcome particularly when they were triple negative. So we designed a small trial as proof of principle.

We were all very worried about this, giving platinum to women with breast cancer since it wasn't done much but we had women with triple negative tumors. We gave them four cycles of chemotherapy after a biopsy we took for research, we did another one part way through and they when to surgery and we took their tissue at the end. What we showed was that a quarter of them almost had a pathologic complete response which is the same rate as has been seen in other large series using multiple drug and the clinical responses were also quite acceptable as a rate. And that in this study thee were triple negative not chosen for mutations but the two mutation carriers in the cohort achieved pathologic complete response consistent with the hypothesis and then at least some other people also did. So that was also consistent with the hypothesis.

Now, there are many arbitrary things in a trial like this, the number of cycles, duration of treatment, the fact it's one cell, all of these things could affect outcome. But it was enough to go forward. This is just the example of the fact the tumor did disappear even by MRI in some of these patients. Steve Nayrod took this observation and did a small trial in women with mutations in colon. I believe an update of this trial will be presented at ASCO this year. But they just had ten women with BRCA-1 or two mutations and gave them the same regimen and 90% of their patients achieved pathologic complete response assuming that all of that assessment has been up to regular standards. I don't mean to dispair Raj this at all, it was a trial in Poland and the paper is not full of details but one likes to believe this might actually be the case.

We have done a second trial looking at the combination of platinum with AVASTIN, one of the drugs targeting the vasculature and we did that based on data that ERPR negative tumors are sensitive to such agents and maybe that would potentiate the affect. This is the trial, at this time Paul Ryan was the head. We collaborate with Mass General and Beth Israel we try to rotate the chairmanship of these studies.

In this study of 51 women, we did not improve the response rate unfortunately. But it has led us down to look at the PARP inhibitors you may have heard about before. These are drugs that are being developed by multiple companies, the company with whom I have the conflict, Astra Zeneca has had as their strategy trying to develop these in individuals with BRAC-1 or two associated tumors. The idea is since those tumors depend on double strand break repair, that is defective in those cells the back up system might use break excision repair and the PARP enzymes which there are several are part of that mechanism. If you block those in the tumor then the tumor has difficulty repairing DNA errors. You can show that PARP inhibitors do affect -- effectively kill cells by themselves when there's either no functioning BRCA-1 or 2, whereas even in the hemozygoat cells that cells are less effective.

These data from AFCO last year showing the drugs given alone to women with ovarian cancer were effective which is encouraging. It led to the first trials done internationally since we can only study mutation carriers in a large -- in large collaborative studies. These were from Australia to Europe and the U.S. These were two trials, one in ovarian, one in breast, where eligibility required that women have a documented mutation. So I think that's the first time these data will also be presented at AFCO this year. Based on that, we have a -- the successor neoadjuvant trial which we'll look at the combination of platinum alone versus platinum plus PARP inhibitor in the mice that is developed with the BRCA-2 mutation, the combination is more effective than platinum alone so hopefully we'll be able to see that. That will be a multi-center trial and for reasons I can hardly believe I'm back doing therapeutics. The gene law will come into affect in several months which has made it possible to do genetic testing with less worry about the implications of knowing one's genetic status. We put this in the medical record. We didn't before. This law only protects against health insurance discrimination, not life insurance or disability insurance discrimination so it's hardly perfect. But it's at least passed so we'll have to give the last administration credit for something.

We do have to remember whether we test our cancer patients or their relatives that these are families and that many different circumstances will arise. If I neglected to put in a plug for genetic counselors I can it will you it is certainly possible to get your genetic testing through your internist and gynecologist but if you really want to do it all the way you should see the genetic counselors. There are many here at the NIH and the genome center and you should use them. They're amazing. For physicians in general we now have to think about genetics as part of life. We have to figure out whether to discuss this with certain patients, whether they bring it up or we do. When do we do it at diagnosis, after they're through with all of the many decisions a cancer diagnosis makes or later. We have to advise them about sharing information with their families and the research side we figured this out before, we can't find their families, they're scattered around. I don't know what happened. And thank you.

Finally, it is possible now to be involved in malpractice problems for failing to document that you have discussed genetic testing. I don't do this but I get asked frequently if I look at these awful cases where people say they were not told that their family -- their cancer diagnosis could have been genetic. So duty to warn is a real thing, maybe Ken office talked about that here so now you have to at least say you talked about it even if patients are not interested. So now, you too can be geneticist, after a small talk like this. I would like to acknowledge the very large group of people at Dana-Farber and collaborators in Boston who made the most recent part of the work upon, a huge team, all of my international and national collaborators on the other breast cancer work and just to remind you that we still are largely -- that guy behind the counter with our patient sitting in front of us figuring out after all this what is it that I want to do. Thanks very much.

[Applause]

Five minutes of questions, maybe seven. Or no minutes of questions. Whatever you would like.

QUESTION: Very nice talk. Thank you. I want to ask what you council your BRCA-2 ASHKANAZI Jewish patients these are the Tamoxifen given results of NAROD, et cetera.

GARBER: Do we advise people to use Tamoxifen to reduce breast cancer risk? You mean -- all the results? I would say you know, we do discuss Tamoxifen and RELOXAFIN with our patients. Most begin risk reduction with oopherectomies. That's not all but many women who are worried about ovarian cancer risk. If they're young enough they get breast cancer risk as well. The other ones done get that 50% reduction so we do talk about Tamoxifen and we do use it but I would say Tamoxifen has been a controversial drug and these patients many of whom are struggling with the Menopause affects of oopherectomy need a few years before they take on something that could bring that all back. So it isn't a good combination but we do use it in some patients. Okay then. Have a good lunch. Thanks for coming.

[Applause]

(Music fades in, under VO)

ANNOUNCER: We were pleased to bring you a special "Contemporary Clinical Medicine, Great Teachers" lecture recorded April 8, 2009. The topic, "Hereditary Cancer Predisposition: New Challenges" Our speaker was Dr. Judy E. Garber, director of the Cancer Risk and Prevention Program at the Dana-Farber Cancer Institute, and Associate Professor of Medicine, Harvard Medical School. You can see a closed-captioned videocast of this lecture by logging onto http://videocast.nih.gov -- click the "Past Events" link, or by clicking the "View Videocast" link for today's podcast at the Grand Rounds podcast page at www.cc.nih.gov/podcast/grandroundpodcasts.html. The NIH CLINICAL CENTER GRAND ROUNDS podcast is a presentation of the NIH Clinical Center, Office of Communications, Patient Recruitment and Public Liaison. For more information about clinical research going on every day at the NIH Clinical Center, log on to http://clinicalcenter.nih.gov. From America’s Clinical Research Hospital, this has been NIH CLINICAL CENTER GRAND ROUNDS.  In Bethesda, Maryland, I’m Bill Schmalfeldt at the National Institutes of Health, an agency of the United States Department of Health and Human Services.