Breast Cancer Clinical Trials

>> john welch: so this may be a little bitless biology-oriented than previous talks and a little bit more process-focused. so please, i hope you've had enoughcaffeine at this point in the day. i know you're from all differentbackgrounds and all over the world, different degrees of education,so i want you to pretend that you're all physicians really interestedin investigating cancer in a clinical setting. so i realize that you may be in the lab or youmay be studying something complete different, but for this lecture we're all, you know,trying to run a cancer trial with patients. so i'm going to start offthis with some definitions

so we're all talking about the same thing. a little bit about how we do -- or how cancerdrug development has evolved in the u.s.. a little bit about our own internalstructure, because that structure is going to be similar whether you're doingtrials in the u.s. or anywhere. there's a core infrastructure that you need. about our current internationalcollaborations and how we make them happen. about running trials. i think the intention of this, when we said lowand middle income countries was to really talk about how to do a treatment trialin a resource-constrained setting.

now, nci itself is more resource-constrainedthan it has been in the past, so we feel your pain. and finally a case study, [inaudible]consortium study that's performed in sub-saharan africa and which is active now. and then some concluding remarks. so, i don't very much think aboutlow and middle income countries as part of my day-to-day job at nci. so i had to go and look up what it meant. and the definition actuallycomes from the world bank.

and they divide economies intofour categories, high, medium -- with an upper and lower medium -- and low, according to the grossnational income per capita. this really doesn't exactly reflecthow much people make in a country. it's sort of a general gestalt of thecountry and how much money is coming into it and is being saved by it. and it really sort of penalizescountries that have a large population, because that's the denominator is per capita. anyhow, a lot of the worldis in the high category.

a lot is in the medium, and a lot is in the low. and of course we'd like to be ableto do trials in all of these areas. â  so, when we're talking about clinicaltrials, just to define that term, clinical trials are trials that involve people. so versus an animal model or something donein plastic, a clinical trial involves people. and that raises issues about human subjectprotection, irb approval and that sort of thing. we try to find an answer, whether it'sto prevent, or diagnose, or treat cancer. from the nci's perspective,it's all about cancer.

when i'm talking about clinicaltrials today, though, we mainly talk about treatmentclinical trials where we're trying to develop a drug or someother form of treatment. it may be radiation, it may besurgery, but treatment rather than concentrating on those other areas. and that is just my focus. i mean, the other areas are also veryfruitful and may even be more appropriate in some research-constrained settings. all of these trials build on previous trial.

even the negative trials are very helpful. they give us an answer and let us move ahead. so as i mentioned, other types of trials. in the u.s., we spend quitea bit on treatment trials. and sometimes these treatmenttrials are pursuing cancers that we already have some treatment forand we're looking for better treatment. we're very focused on treatment becausewe have a relatively elderly population. the life span in the u.s. is relatively high. people tend to die not ofearly infectious disease

and other things, but thingsrelated to lifestyle. it may be heart disease, butalso in large proportion cancer. so from a public health perspective, it'sworth our time and effort to try to refine, even if we have a good treatment,and try to make it better. that may not be the same publichealth motivation in all countries. in countries were communicable disease ismore of an issue, it may make more sense to put more public health dollars into that kind of a clinical trial thana cancer clinical trial. likewise, these other areas -- prevention,early detection, screening and so on,

make a lot of sense in a -- if you havea limited amount of resources to expend, you might be better off trying to prevent thecancer or find it early than trying to treat it when it's very difficult andresource-intensive to treat. so i don't want to neglect those otherareas, but my talk is on treatment. so we'll be focusing there. so, i'm going to kind ofstart at the basic level because i'm not sure how much you've heard. if this is old hat for you all, letme know and i'll move on more quickly. but just to discuss treatment trials.

we don't go with a drug thesedays right into people. we need to know that the drug has someactivity and that there's a rationale for it. but beyond there just being agood idea that the drug may work in people, we need to see some evidence. so whenever a concept comes up for --being deployed in a cancer center or in one of our national networks, we want to see what isthe pre-clinical information behind that drug? we want to know what have you done in plastic? what's been done in animals,especially [inaudible] models? and we want to have some idea that you havea good idea about the starting dose to use,

how you'll detect that thereis an effect and so on. so typically that's done beforethe true clinical trial phase. within clinical trials, we referto four phases of clinical trial. the first one, phase one, is just dose finding. trying to figure out how muchof the drug to give to people. at one time, the drugs were all [inaudible]. their main job was to kill cells. and cancer cells proliferated more quickly, sothings that were good at killing cells were good at killing cancer and everything else, butmaybe a little bit more towards cancer.

and we looked at that -- that degree ofselectivity and tried to improve it in trials. nowadays, we're not always trying tofind how much can we maximally put in? we want to put in as muchas we need, but not more. so for a lot of targeted agents, thismaximally tolerated dose really isn't the dose that we'd recommend into a phase two trial. so phase one; dose finding. the other things that are sometimes addedinto phase one are pharmacokinetics, looking at the levels of drug in the bloodand understanding the metabolism of the drug to get an idea about howoften you should dose it.

is giving it once a week okay? do you need to give it three times a day? does it get in orally, and so on? and pharmacodynamics, which is looking atsome kind of biological readout of the drug. is it inhibiting the enzyme that you want? so all of those can be done in phase one. we generally don't go into phase one at all unless we think the drugis going to make it to phase two. so you have to have a good amount of evidencefor us to spend the time doing phase one to know

that it will -- if we can find a gooddose, then it will go into phase two. so in phase two, we're looking for activity. and this can be sometimes acrossmultiple different tumors. it may -- we might test it in patientsthat have relatively advanced disease, say end stomach cancer and lung cancer. or we may pick one histology,if we have reason to do that. and we're looking just to see,does it stop cancer from growing? does it slow it down? is it likely that this will pay off ifwe really invested the big resources

in a phase three trial? so phase two is a very critical step. it's the make or break, go or no go decision. and we make it at nci forthe government-run trials. companies do the same thing when theytry to figure out what they're going to take forward into large-scale development. so finally, phase three. these are the big studies. hundreds of patients and they'resupposed to be game changing.

if you have a positive phase threestudy, the drug or the treatment in that study is likely tobecome a standard of care. so really, that's kind of the end of the line. phase four, i've greyed out a bit. phase four is post-marketing. once something is already on themarketing, it's received its approval from whatever regulatoryagency, in our case fda, sometimes we want to collect more information. and we know that if we watchit over an even wider audience,

we'll get more informationflowing back about the drug. anyhow, that's the general flow. as we go through it, the sizes ofthe trials get bigger and bigger. the resources that are committedto the trial get bigger and bigger. risk goes up, in terms of what'sbeen laid out in terms of resources. and this is true whether we're performinga trial domestically or internationally. so just a little bit of backgroundon nci cancer development. nci really got into cancerdevelopment because -- cancer drug development becauseno one else wanted to.

companies would not go nearcancer drugs in the 1950's. they felt it's not going to work and it'snot going to produce any kind of profit. so at that point, nci was synthesizingits own drugs, it was looking at a lot of natural products, and it was doing massivescreenings trying to find active compounds. so you needed a large central infrastructureto do that, to run all of these compounds through cell lines and so on trying tofind the one in a thousand that would turn into an effective cancer drug and then tryingto find the right kind of cancer for that drug. or sometimes working in reverse, starting withthe cancer and looking for the appropriate drug. with growing, i guess through the 1950's, '60s,'70s, there's more and more concern about cancer

as a public health concern and nci began tocoordinate national networks to perform trials. very often, a given institution is not goingto see enough say of the sort of lung cancer in a given year to perform a phase three trial. probably not even close. so you need the large network. and in those networks, everyone hasto work off the same instructions. so we have what are called protocols. it literally is a recipefor how to do the trial. you know, here's the list of peoplethat can come into the trial.

they have to meet these conditions,have this cancer, have this kind of performance status,have blood counts that are good. it could be anything. they're going to get this kind of atreatment at every single institution, worldwide potentially, participatingin the trial. usually it's one kind of therapy andthen another that has a modification. i mean, it's essentially applyingscientific method that you'd use in the lab in a clinical setting. you change one variable.

everybody does the same thing. and the more regular that isacross the entire network, the more clear the resultwill be at the end of it. and what we use on a clinical trialfor a readout is how the patient does. do we see on x-rays the tumor shrinking? do we see patients living longer? i mean, that really is theultimate, ultimate question. do the patients live longer? do they have better quality of life aswell, on either the standard of treatment --

the standard of care treatment or on anew treatment that we're testing out? and that's where we try to go with these trials. so, being government-sponsored, weare interested in certain trials, and industry right now isinterested in other trials. we're always motivated bypublic health concerns. so our trials may do things that areconsidered too high-risk by industry. we may combine drugs fromdifferent pharmaceutical partners that industry wouldn't do on its own. we may look at things which don'thave a clear profit motivation,

looking at optimizing surgery, or incorporatingradiotherapy, or optimizing radiotherapy. or bringing in special populations. there's a subset of cancer trials that typicallywouldn't be done by industry that there's a role for the public health system to do. and there are parallel organizations throughoutthe world that have essentially the same mandate and government support to do the same thing. so one thing i wanted topresent was a restructure, and that's a somewhat boring topic excepti'm presenting it to show you the sort of infrastructure that we've evolved andthat other organizations like us have evolved

and that you would need to really rollout clinical trials on a large scale. and that infrastructure couldcome from a government. it could come from ngos and charitable sources. it could be outsourced tocontract research organizations. it doesn't have to be donein exactly the same model. maybe just a discussion about what ctep is. so ctep is the office that i'm from,the cancer therapy evaluation program. we're sort of the central office thatreviews studies that are proposed and we help prioritize what getsdone on the national network level.

we do not do the trials ourselves. we are not in the hospitals. we're an office at nci. but we interact with thenetwork that does the studies. and we review the protocolsand we try to optimize them. so our network in the u.s. consistsof about 3,000 institutions. some of these are academic, you know, very high-end academic institutionslike universities and so on. but many of them are community sites.

private doctor's offices or communityoncology practices where they see patients and occasionally put a patient onto a clinicaltrial and follow the protocol for that trial. so we've got a network of about 14,000 peoplein the u.s. and canada participating in this. and we put on, you can see on the bottomline the number of patients per years vary, between 27,000 and more recentlyon the order of 20,000 per year. so that's quite a -- quite a large through-put. and about 80% of the studiesthat we do are phase three, by number of patients because there are studies. so within ctep, i wanted to talkabout some of the different sections.

one is the clinical investigational branch. nci doesn't so much synthesizeits drugs anymore. now we get them from commercial sources. so there has to be some agency withinour office that has relationships to the pharmaceutical industry andis able to provision the drugs. one thing that the investigational drugbranch does for investigators in the u.s. and i believe worldwide, is make drugsavailable for preclinical studies. now sometimes it's difficultto find drugs to put into your [inaudible] model or your mouse model.

so relatively -- not enough to do aclinical trial, but enough to do those kinds of models can be provided bythe investigational drug branch. if you google ctep agents, in fact,you'll get an excel spreadsheet which lists the different agentsthat are within that branch and which are available potentiallyfor collaborative use. i'm sure there's some sort ofagreement that has to be signed, but i know that we've shippedthem worldwide in the past. they take care of a lot of agreements withthe pharmaceutical companies to get the drugs and then they run trialsthat are very drug-centered.

these are the early-phase trialsand pd and pk trials, which are -- from the perspective of the drug, if youhave a drug or if you have a class of drugs, what is the best drug in that class? and we'll take that one andwe'll develop it and go -- try to figure out which tumors itwould ultimately be matched up against. that's what they do. i'll contrast that with the clinicalinvestigation branch, which is our medical corps where we start from the disease perspectiveand try to find a drug in that direction. this branch deals with all of the -- thekind of operational acts of clinical trials

and the groups that perform them. i can't emphasize enough how importantit is to have a good regulatory core. when you run a clinical trial, thereis a lot of government paperwork to do. even though we are the government, wehave to fill out a lot of paperwork. drugs are regulated not by nci, but by thefda in the u.s. so we have to comply with fda and there would be parallelorganizations in other countries. there are a lot of contracts with the sites thatperform the studies, with industry about drugs, about diagnostics, about intellectual property. things that will curdle your blood, theregulatory affairs people take care of.

and it really is a important niche thatwe couldn't do clinical trials without. the clinical trial monitoringbranch is another important section. as i mentioned, these are trials that involvehumans, so we have concerns about complying with u.s. and regulations about involvinghuman subjects, but also worldwide standards. if you're familiar with thedeclaration of helsinki, it defines what good practiceis in clinical trials. you know, early on in the development of cancerclinical trials, or clinical trials in general, people played pretty loose with the rules. they would give people drugs.

they wouldn't tell them what they were. they might switch them from onedrug to another and not tell them. there were quite a few abuses. this is somewhat remote history now, but itresulted in the drafting of that document. and the signatories to that documenthave essentially said that, worldwide, will comply with being upfront with peopleand telling them what's really going on in the trials, telling them what the risksare, and helping them watch out for themselves. the principles that come from the declarationof helsinki are somewhat formalized in ich. you may hear this acronym if you areworking in international clinical trials.

that's the international conference onharmonization, which was a conference that involved u.s., japanese, and europeaninvestigators and regulatory agencies that drafted what industry has formalizedas a standard for how to do clinical trials. so if you're involved in a commercialtrial, they are conducted under ich. gcp are guidelines that correspond to ich. some countries have taken gcp,which is good clinical practice, and have taken it almost directly into law. other countries have regarded it asguidelines about how to perform trials. but -- not only is human subject protectionin the informed consent sort of way,

but in the process that youfollow in performing a trial in the sense that you will keep good records. you monitor compliance to the protocol. you will flag any kind of incident that happenson the trial that may put somebody at risk, and you'll take corrective action. so, it's not just the u.s. that has this aspart of its law and is required to follow it. this is a concern around the world. in china, in fact, the international studiesare not allowed in early phase clinical trials because of the concern that they don't wanttheir population being exposed to drugs

that may be coming from outsidethe country, from foreign companies that may be exploiting theirpopulation to test out these drugs. we don't know if they're safe ornot safe, so it's their perception that it would be better to limit that. they may also feel that it protects theforeign company because if something did happen to their citizens, that wouldreflect poorly on the company. in any event, this is a worldwide concern. and the clinical monitoring branchis our way of inspecting sites and reviewing the protocol itself andthe informed consent and making sure

that it all complies withthese international standards. in conducting trials, we have apharmaceutical branch responsible for getting the drug out to the site. and this is really crucial. you can't do a clinical trial -- adrug clinical trial without the drug. and when we're working in an internationalsetting, they do have the capability to ship the drugs worldwideinto clinical trials. but very often, we get into a lot ofissues with import/export at the border, trying to clear customs with drugs.

it can be quite difficult. so they have that expertise. biobanks for biological specimens, aclinical trial support unit which is sort of a central computerized facility that allowseveryone involved in our trials around the world to put their patients -- register themon a trial using a web-based interface and allows the remote entry ofdata and so on so that [inaudible]. so those elements are wherever clinicaltrials are done, there has to be something that supports those different functions,whether it's ctep or some other organization. some more.

yes, more. grants and contracts, the protocol informationoffice that -- that manages that document, the protocol at the heart of theclinical trials and [inaudible]. all right, so i know that'sprobably more than you wanted to hear about it, but how we do things here. let me talk about international. so most of our trials we do-- well for several reasons. it may be science that we have a trialwhere the sample size is just so large that we couldn't accrue it in the u.s. alone,or we couldn't accrue it in a reasonable amount

of time, or there's a population that we don'thave and another group does have overseas. so those are common reasons for doing this. sometimes there's specific expertise overseas. they can do something better than we can. there may be genomic variations thatwe want a worldwide sampling of. pharmacogenomics could, you know, be differentdue to different [inaudible] variations. or we might want to validate somethingfound in one place in another place. so those are all scientific reasons. of course, if you have morescience doing the study,

you'll probably get it donefaster, so the speed. and cost. this isn't necessarily a driver because doing international studieshas its own overhead in terms of cost [inaudible] regulatory and so on. but in principal, if you dividethe project amongst many people, you may spread the cost of the project. and just getting it done faster really helpscut down on the overall cost of the project. and everyone is sharing resources,so it may be more efficient. so there are lots of challengesfor international clinical trials.

i'm not going to go into all these. maybe i'll just highlight a few of them. and then this is sort of a checklist that werun through when we're considering, you know, should we enable this collaborationor is it going to be too difficult, or what do we have to do? but the take-home message is we've managed toovercome all these issues and it is possible. i mean, it's really a matter of how muchdo you want to do the trial and you know, what resources we're able to put into it. these are kind of -- these weredevised more for collaborations

with europe and asia and developed settings. so maybe we can add someto this as you go through. so first just the administrative issues. typically our protocols havebeen written in english. in some cases, we've done localization. in a setting where there is --english is not the native language, there would have to be translations made andthose translations would have to be validated. and also informed consent documents haveto be provided in the local language. as i mentioned, getting the agent intocountries, getting it distributed,

having it labeled in the local language,conforming to the local regulatory requirements. you know, we have fda. every country has something else thatregulates how drugs are distributed. we have to show that the drug isexactly the drug that we say it is. it has to be tested. there are all kinds of quality controls on that. i'd be happy to talk more about it, butgetting drug [inaudible] is quite an issue. in a constrained-resource setting,we also have the issue not just of the study drug but of all the support drugs.

so sometimes, in generic drugs, weneed to get into low income countries so that they can perform the trial in auniform manner across all other sites. things like, for supportivecare, [inaudible], things that -- drugs that support blood counts and so on. that is something that we typically wouldhave within our normal cost of medical care that would probably have to be provided in asetting depending on the drugs that are used. patient protections. these are going to be prettymuch similar worldwide. each country may have itsown requirements in addition

to the kind of core concepts that we have. data collection we have to work out howthe data flows back and is analyzed. it's probably also an issue forinternational trials how that data is protected if the data is coming tothe u.s. auditing of trials. i mentioned the clinical trial auditing branch. insurance is a big issue. in the u.s., we have insurance -- hospitalshave liability insurance and so on. but if a patient is injured on a clinicaltrial, there may not be a safety net in countries that have lower income.

so we want to be sure that patientsthat are participating in a trial that they're not harmed andthat there's some recourse. unfortunately, the way we're structuredas a federal agency, we are not allowed to provide indemnity insurancefor studies that are run overseas. and this has become part of the issue in someinstances, and it effectively excludes us from trials in india due to some recentclinical trial legislation there. and finally, we have to workout all the science, but the science we always have to work out. so that's one model, that we have these u.s.groups that have some overseas affiliate sites.

another is to have one of our groups in the u.s.partner entirely with an external organization. so there are large networks,[inaudible] in europe, for example, the [inaudible] in australia. those groups have structures that are similarto us and we're very parallel and it's easy to have a trial where wehave one protocol between us. and that's great because we'reall working off the same sheet. it's very difficult in a regulatorysense to get that to work, though. so the next best thing would be having parallelstudies, where we have a study designed and running in the u.s. and a study designedoverseas and we both accrue to our own studies,

the protocols are essentially thesame, but they're one length removed. the next best would be a pre-plannedmeta-analysis, where we agree that we're going to do separate studies butin some way pool the result. so those all are -- if we're trying todo a study here and in an external site. but that's not always the case. there are cases where we have studiesthat run entirely outside north america. and i think that's probably morerelevant to what we're talking about here, where we want to have a study that is notjust something that we would ordinarily run in the u.s. but something that really isadapted to a low-resource setting and maybe more

to the needs that are there and not here. one thing to mention, cancer centers,individually, have the ability to pair up with other cancer centers in other countries. and while we're trying to come at thisfrom a federal government perspective with nci behind it, sometimes it's faster forindividual cancer centers to move on their own. very often, somebody from a cancercenter spends some time in the u.s., or someone here spends some time overseasand they establish a relationship. and over time, that relationshipbuilds and the two centers form sort of a sister hospital networkjust between the two of them.

and in that case, there can besome interesting collaborations that can happen relatively rapidly. so that's another possibility alittle bit outside our typical system. so for the most part, nci is constrained byits mandate to work on u.s. public health. but there is a higher mandatefrom the nih director. when francis collins came in, heidentified five areas that he really wanted to work on and put nih's resources. so nci is part of nih, so nih is bigger. we have to listen.

so the fourth one down onthat list is global health. so you may be aware that at nci, anoffice was created about a year ago. the office of global health. and part of their job is to see what we can doin terms of involvement in middle and lesser -- middle and lower income countriesacross the spectrum. not just treatment trials, butincluding treatment trials. so there's an interest in tryingto get some of these trials going. i have to tell you that there's ahandful of them that we've done, that are actually up andrunning that i can show you.

so i wanted to pick one of thoseand walk through it as an example. and you know, you can thinkabout it but until you do one, you don't really realize what's going on. so we'll talk about amc 068. the aids malignancy consortium is a -- anothernetwork that also is coordinated through nci, but through the office of hiv malignancy,oham, which is separate but related to ctep. and we work together in developingprotocols and reviewing them. so the networks based out in losangeles, and has core sites in uganda, kenya, south africa, and zimbabwe.

so according to that who -- sorry, not who,world bank list that i showed you earlier, south africa is a middle incomecountry and uganda, kenya, and zimbabwe are in the lower income bracket. so we've finally gotten to thepoint that we've penetrated to the lower income bracket,at least this trial. this is a randomized study. so yeah, it's a straight comparison,this oral regimen versus chop. so patients that come into the study are adults. that's part of our general clinicalpractice guidelines that we won't do a study

on a vulnerable population, including children, until we've got some experienceunder our belts with adults. different dosing and so on,clearer result with adults. hiv-positive. so this is an issue, because the definitionof hiv-positive for a clinical trial, again, trying to stick to protocol, is that they havea licensed test, usually an antibody test. it could be an rna test, that'sconfirmed by another approved test. and the approval could be a national approval. it could be something that's recognized in theu.s. but getting those two tests, you know,

there's some barrier there just toget the test and pay for the test. but to make sense of the study, we had toknow that they definitely had, you know, hiv. it's an hiv -- an aids-associated malignancy. we needed a biopsy of the tumor. and that biopsy had to be reviewed locally,which meant that the people reviewing it had to be certified that they were allreading the slides in the same way. and we have the same thingin the u.s. that makes sure that our pathologists are all on the same page. samples were also sent overseas forlater review, for central review.

and they were a match. these are patients with advanceddisease, never treated before. they're either on an effective anti-viralregimen or have been switched over to one, and not one that is immunosuppressive. they've had a lumbar puncture,which is a procedure that can i think be done pretty muchanywhere, but it has to be done steriley. so that may also, in termsof consumables be an issue. and finally, they have to have act scan before they go on trial. so -- and this is a fairly expensive testand maybe not something available everywhere.

but the trial required two ct scans, one toknow how big the tumor was at the beginning, and one to know how big it was at the end. and we didn't really see any way ofgetting a clear answer without the ct scan. so patients were assisted in travelingto wherever the ct machine was. in some cases, it was a fairly large trip,but we really tried to hold down the number of times that they had to travel. and pregnancy tests for women to make sure that we weren't giving themchemotherapy while they were pregnant. so this is the treatment regimen.

they both work out to 18 weeks. there were six three-weekcycles of a chop combination or three six-week cycles of the oral therapy. at the end of that they would get their ctscan to know if the tumor had shrunk or not. and that is one of the readouts from the study. patients would then be followed, primarilyclinically because we realize it's not going to be practical or economically liable tohave them come every few months for a ct scan. so we had to know how good is the clinicalfollow-up and is it possible for these patients to be followed effectivelyjust by clinical examination?

in some cases, they havethe capability to come back at two years for a ct scan or it can be done. so patients would be followed out two yearspast their treatment in terms of hiv control and up to five years to knowabout their overall survival. so that really is the bottom line. are these patients surviving longer or not? so having started a year agoand taking two years to accrue, it's going to take us maybe anotherfive to 10 years to get an answer from the study, but at least it's going.

i won't into the nitty gritty of the statisticaldesign, but just to say that they're looking at a number of endpoints beyond justour standard clinical endpoints related to the question of how effective thisis, but also how sustainable it is as a general approach in the setting. so a number of items had to bemodified from the standard protocol in order to work in this setting. every protocol contains guidelinesfor dose modification. we know that some patientsare going to get sicker than others when you give them chemotherapy.

so there was a cut made early in theprotocol for patients that are poorer risk than other patients, patientsthat we think are going to be more fragile in the face of chemotherapy. so those poor risk patients received a lowerdose and we're intentionally testing it that way, because we realize that in clinicalpractice, when the doctor sees someone who doesn't look like they'lltolerate the therapy, they would cut the dose backprobably themselves. so we're doing that in the protocol sothat if we show something on the protocol, that can be part of standard practice.

there are different guidelines depending on if stem cell factor isavailable to support blood counts. there are guidelines because there areanti-retrovirals in the picture of how to -- sorry, adjust the antiretroviraltherapy depending on renal function that may be impacted by the chemotherapy. as i said, extensive supportive careguidelines related both to the cancer and the chemotherapy, as well as hiv. and finally, a lot of work went into assure compliance that the staff at these centers were trained to make sure thatthings did happen the way it's in the protocol

and that the patients would take thepills, because these are oral pills. unlike iv therapy, we don't know that we putit in when the patient has to take these. in some cases, it's observed, butin most cases, the patient goes away and comes back the next month for the treatment. and what we ask them -- what wenormally do in the u.s. is we say, "you show us your pill bottle,"and count the number of pills that are left, and does that make sense? but we also in the u.s. typically ask people tokeep a diary of when they've taken a pill or not and we had to adjust that for literacy reasonsthat not all patients can settle the document.

so sometimes we have a -- afriend or a family member do it. sometimes we had to justgo by their recollection of what they've taken andhow many pills were gone. so, that's one of maybe five or sostudies going in lower income countries. most of them are hiv related. some, however, are related to [inaudible] andthe simplex virus which may exist in africa with or without actually aids association. and we're contemplating some otherstudies, one [inaudible] carcinoma. another in gall bladder cancer in latin americawhere there's a strangely high incidence of it.

but we're still finding our way,and probably some of you who are from those settings may know better thani about, you know, which direction we need to go in, and how to adapt our systemto be more interactive with middle and lower income countries for clinical trials. so final comments. nci does have a directive toimprove both u.s. and global health. we have an infrastructure that could bereplicated and would have to be modified and probably be replicated ina resource-constrained setting. we're happy to partner.

ideally in these partnerships, each ofthe partners get something out of it. i think we learn a lot from the trials thatwe're doing in resource-constrained settings. there are many models thati shared for collaboration. and final advice, since i assume at theend of the fellowship you'll be dispersing to all corners of the world,is try to make a lot of friends because very often these collaborationsand long-term things get started because of a personal contact betweensomeone in an overseas cancer center, and someone in a u.s.-based cancer center,and those often blossom into training programs and exchanges and finally,clinical collaborations.

so, that's my advice. so i will be glad to take questions [applause].