Breast Cancer, Moving Ever Closer to Cure for All

Transcript

• 00:00I'm doctor Mary. I'm lustberg.
• 00:02Thank you for joining in person and
• 00:05for those of you joining online.
• 00:10I'm pleased to introduce Doctor Louis
• 00:14Pushti as today's ground round speaker.
• 00:19Doctor Pushki is professor of medicine.
• 00:22And Co director of the genomics,
• 00:25genetics and Epigenetics
• 00:26Research program here at Yale.
• 00:29He received his medical degree
• 00:32from Semmelweis University of
• 00:34Medicine in Budapest and his Doctor
• 00:37of Philosophy degree from the
• 00:39University of Oxford in England.
• 00:42His research group has made
• 00:44important contributions to establish
• 00:46that estrogen receptor positive
• 00:48and negative breast cancers have
• 00:51fundamentally different molecular,
• 00:53clinical and epidemiological characteristics.
• 00:58He's been a pioneer in evaluating
• 01:01gene expression profiling as
• 01:03a diagnostic technology.
• 01:05To predict chemotherapy and
• 01:09endocrine therapy sensitivity.
• 01:11And as shown that different biological
• 01:14processes are involved in determining
• 01:16the prognosis and treatment response
• 01:19in different breast cancer subtype.
• 01:23His group has also developed new
• 01:26bioinformatics tools to integrate
• 01:28information from across different
• 01:30data platforms in order to define
• 01:33the molecular pathways that are
• 01:35disturbed in individual cancers
• 01:37and could provide the basis.
• 01:40For individualized treatment strategies.
• 01:45Doctor Pushki is a trusted colleague
• 01:48here at Yale and is a principal
• 01:50investigator of several clinical
• 01:52trials investigating new drugs,
• 01:55including immunotherapies for breast cancer.
• 01:58He's published over 250 scientific
• 02:02manuscripts in high impact medical journals
• 02:05and is among the top 1% most highly
• 02:09cited clinical investigators in medicine
• 02:12over the past 10 years.
• 02:15Today he will speak on breast cancer,
• 02:18moving ever closer to cure for all.
• 02:21Thank you so much Doctor Pushkar.
• 02:29You can go ahead and start using this.
• 02:31Thank you, Mary.
• 02:32I'm so if you're OK with you,
• 02:34I will take this mask off because
• 02:36having a mask, my accent and my
• 02:37voice would be really serious.
• 02:39Triple hit against me from the get go.
• 02:42So I hope it's OK with you.
• 02:44It's delighted to see that some
• 02:46people are in the auditorium because
• 02:47I actually forgot how to get here.
• 02:50So I really sympathize with those of
• 02:52you who are actually online with this.
• 02:54So I think I need to start
• 02:58with my disclosure slides.
• 03:01And then before I start my slides,
• 03:03I would actually like to make a
• 03:05confession to you and admit a weakness.
• 03:07It's not chocolate,
• 03:08but I do feel like a child in a
• 03:10in a candy store surrounded by a
• 03:12lot of really delicious and very
• 03:15interesting scientific questions.
• 03:16So my weakness is that I have a really
• 03:18eclectic and very broad range of interests.
• 03:21And don't be scared,
• 03:21I'm not going to talk about all
• 03:23of these questions,
• 03:24but these are the type of questions that.
• 03:26My group has been studying in the
• 03:28past few years and I showed this here
• 03:30for you to forgive me and understand
• 03:32why I don't show up to most of the.
• 03:37Administrative meetings,
• 03:38so these studying things like
• 03:40cost effectiveness,
• 03:41what's the best cost effective
• 03:42strategy in the new adjuvant
• 03:44setting for for breast cancer,
• 03:46why some preoperative chemotherapy
• 03:47regimens produce high response rates
• 03:49but very little improvement in survival
• 03:52and other regiments to the opposite
• 03:54small improvements in response,
• 03:56large improvements in survival.
• 03:57Why there is some women develop breast
• 04:00cancer 20-30 years before the median age?
• 04:03Could we develop some sort of a tool to
• 04:05sum up all the genomic abnormalities?
• 04:07From germline and somatic regions that
• 04:09would actually describe the capture
• 04:11the totality of abnormalities in atom.
• 04:14How comes that summer stragen receptor
• 04:16positive cancers recur as they are negative?
• 04:18You know some ER positive cancers
• 04:20are not fully ER positive,
• 04:223040% positive.
• 04:23So what are the rest of those cells
• 04:25which are ER negative?
• 04:26What's their relationship to the
• 04:28ER positive cells?
• 04:30What novel therapeutic strategies one could
• 04:32dig out from high dimensional genomic data.
• 04:35So what is the molecular phylogenetic
• 04:38relationship between different
• 04:39metastatic lesions and the primary tumor?
• 04:41Is these different for synchronous
• 04:43mats against asynchronous?
• 04:44That's you know why some Kansas are
• 04:47immune reaction immune poor was the
• 04:48difference between the immune rich ER
• 04:50positive and PR negative terms is there
• 04:52a difference in the microenvironment
• 04:54that's race influence this so really
• 04:57study all of these things and.
• 05:01You can look at the publications on them.
• 05:02So I'm only going to focus on a
• 05:04few which I think have a longer
• 05:06trajectory and contributed to the to
• 05:08this remarkable events that happened
• 05:09in the past 20 years that breast
• 05:12cancer survival and mortality decline,
• 05:14mortality decline by about 50%.
• 05:17I think this is primarily driven
• 05:19by new treatment strategies based
• 05:20on better understanding of the
• 05:22disease and the new
• 05:23classes of drugs that we developed.
• 05:25And I think the journey is
• 05:27just just about to begin.
• 05:29So how new treatment strategies
• 05:32could influence outcome?
• 05:34So in the early 2000s,
• 05:36I was in the right place at
• 05:38the right time at MD Anderson,
• 05:40we were interested to explore
• 05:42period preoperative chemotherapy
• 05:43for women who actually had operable
• 05:45disease and we assumed that they
• 05:46would end up with a better cosmetic
• 05:48outcome as smaller disease.
• 05:49And at that time,
• 05:51it was a pretty controversial idea
• 05:52and there was really no good way
• 05:54to either define the response.
• 05:56How do you measure the efficacy
• 05:57of these preoperative regiments?
• 05:58Do you measure it by response?
• 06:00On imaging or we measured by
• 06:02the extent of residual disease.
• 06:04So we proposed the the definition
• 06:06which eventually become the standard
• 06:07of care definition that you have
• 06:09no residual invasive cancer in the
• 06:11breast or lymph nodes and that's kind
• 06:13of the best outcome that you could get.
• 06:15So with this definition it pretty
• 06:17quickly become available become
• 06:19obvious that individuals accomplish
• 06:21this complete pathological response.
• 06:23It really well regardless of what
• 06:25type of breast cancer they had,
• 06:26they are positive or negative
• 06:28or too positive.
• 06:29Those who had residual disease didn't do so.
• 06:31And this immediately defines you what you
• 06:33actually want to accomplish in the clinic,
• 06:35right?
• 06:35You want to put more patients
• 06:37into these pathologic CR category
• 06:38and you want to hurt harm.
• 06:40Do you wanna help those who are
• 06:42in the residual disease group?
• 06:43So we did that in the past 20 years.
• 06:45So you see the evolution of the chemotherapy.
• 06:50Regiments,
• 06:50in 2008 when we published this
• 06:52paper on the survival curves,
• 06:54the best chemotherapy was
• 06:55Taxol anthracyclines.
• 06:56It produced about a 3035%
• 06:58response complete response rate,
• 06:59in particular negative disease
• 07:01and now we have doubled that.
• 07:03So now we actually accomplish
• 07:04about a 63% complete response rate
• 07:07by adding an immunotherapy drug.
• 07:09And you also learn that adding other
• 07:11chemotherapy agents like carboplatin
• 07:13improves the pathologic CR rates.
• 07:15We have regiments that don't
• 07:16include the anthracyclines that
• 07:18some of my colleagues think that.
• 07:19Is the chemical incarnation of the devil.
• 07:22Also there are even single agent therapies,
• 07:25targeted therapies like PARP inhibitors
• 07:27that produce pretty respectable
• 07:29pathology company eradication of
• 07:31the cancer before surgery in in
• 07:33germline Brockhampton patients.
• 07:34But we also made him really important
• 07:37improvements for in the life of
• 07:39those who have residual disease.
• 07:41So those are three randomized clinical
• 07:43trials that established the value
• 07:44of giving capsidae in chemotherapy
• 07:46for those and the residual disease
• 07:48with triple negative cancer.
• 07:49And the Olympia study showed that
• 07:51that whole party improves the
• 07:53response within a similar population
• 07:55if the average germline Broca's.
• 07:56And the Catherine study did the
• 07:58same for the record TDM one or
• 08:00Godzilla for her to post the disease.
• 08:02But I want to spend a few minutes on
• 08:04how do we get there, in particular,
• 08:06how we actually came about to establish
• 08:11the value of immunotherapy in.
• 08:14In breast cancer. So the roots of
• 08:16this idea that immunotherapy might
• 08:18work in breast cancer has been
• 08:20long rooted in preclinical studies.
• 08:23But also in the early 2000s a number
• 08:25of of groups reported that even in
• 08:28patients who only receive surgery,
• 08:30the amount of immune cells in the tumor
• 08:33microenvironment is hugely prognostic.
• 08:34So this is what the the first half of
• 08:36this slide shows you survival curves
• 08:38for patients who did not receive
• 08:40any other treatment than surgery,
• 08:42they were stratified into three groups.
• 08:44Little high immune presence,
• 08:46intermediate in presence or low
• 08:47immune presence and you see that
• 08:49that the the immune cells have a
• 08:52massive prognostic value in all three
• 08:53categories of of breast cancer subtypes
• 08:56including the the ER positive patients.
• 08:58And what we used in this particular
• 09:00study was gene signature to define
• 09:01the immune richness.
• 09:02They're in the same time German
• 09:05investigators showed that that
• 09:07the presence of immune cells also
• 09:09predicts the probability of complete
• 09:11pathological response.
• 09:11But this slide shows you 32 important things.
• 09:14One is that in the red circles you
• 09:17see the pathologic computer response
• 09:19rates by tumor infiltrating into side.
• 09:23Presence.
• 09:23So they grouped the cases into
• 09:25no lymphocytes, some lymphocytes,
• 09:26lymphocyte predominant and you
• 09:28see that the pathologic CR rates
• 09:30these numbers in the in the little
• 09:32blood red circles increase as you
• 09:33have more and more lymphocytes.
• 09:35So for example in the blue,
• 09:37so the square or highlighted
• 09:40area and ER positive disease,
• 09:43we know lymphocytes,
• 09:44it's a very small 6% PCR.
• 09:45If you have a lot of lymphocytes,
• 09:47it goes up to a respectable 23% and you see
• 09:50this same trend across all the subtypes.
• 09:52So of course these observations lead
• 09:54to a lot of other questions then.
• 09:55So why some breast cancers are immune,
• 09:57originalists don't is the immune
• 09:59microenvironment differ between
• 10:00the primary system and the maths,
• 10:02it's a different by ER subtype or by race?
• 10:05And ultimately the the most important
• 10:07question is this a causal relationship
• 10:09or immune cell presence is actually
• 10:12responsible for the good outcome or
• 10:14it's just an association that reflects
• 10:16some other underlying biology.
• 10:17So when these papers were published,
• 10:20you couldn't really test this in people,
• 10:21there were no chemotherapy drugs.
• 10:23But now we have and we actually have
• 10:24the answer to most of these and I
• 10:26put there some of the publications
• 10:28that that address these these issues.
• 10:32So I want to share with you some results
• 10:34which I think really informed a lot of
• 10:37my thinking about the the value of the
• 10:39role of immune system in breast cancer.
• 10:41So a few years ago Anton Sofronoff
• 10:44was a medical student here at. Yeah.
• 10:47At that time took on this project,
• 10:49but downloaded all the CG data or an
• 10:51AC DNA copy number, mutation data,
• 10:54germline snips and ask this question.
• 10:56So what drives the immune infiltration
• 10:58and breast cancers?
• 10:59So we looked at Chrono Heterogeneity,
• 11:01mutation load, new antigen load,
• 11:03copy number variations,
• 11:04germline snips,
• 11:05single gene somatic mutations,
• 11:07pathway level abnormalities,
• 11:09which of these is associated with
• 11:11high immune presence,
• 11:12whether you think the results showed?
• 11:15So. Gosh.
• 11:22So the results are actually
• 11:25quite counterintuitive.
• 11:26So what this shows you is a correlation
• 11:28matrix of about 12 immune gene
• 11:29signatures that we use to define the
• 11:31immune presence or absence or in your
• 11:33richness and about 6 genomic features.
• 11:36So the darker brown shows a higher
• 11:39correlation value and the darker
• 11:41blue shows a negative correlation.
• 11:44And you see right away that
• 11:45the immune gene signatures are
• 11:47highly correlated one another,
• 11:48whereas they are not correlated
• 11:49very closely at all. In fact,
• 11:51they are anti correlated with many of the.
• 11:53Economic features.
• 11:53So and you see this across the
• 11:56board in all the three subtypes.
• 11:58So in in primary breast cancer greater
• 12:01chromo heterogeneity and higher mutation
• 12:03and neoantigen loads are associated
• 12:05with lower immune infiltration.
• 12:07So there was such a weird finding
• 12:08that we actually teamed up with
• 12:10with the A colleague from Germany,
• 12:12Thomas Cohn to really confirm this
• 12:14in an independent data set data sets
• 12:17and we find the same same result.
• 12:20So why is this interesting?
• 12:23Because even though we found no share
• 12:25genomic alterations that drive the
• 12:27immune infiltration in breast cancer,
• 12:28we really find a strong supportive
• 12:30evidence that there is an active
• 12:33immune editing in early stage disease,
• 12:35right.
• 12:35So a lot of immune cells in actually
• 12:37called remove chromo heterogeneity
• 12:39and that's why you have a chromoly
• 12:42simple tumor and actually a lower
• 12:44your antigen load because the cancer
• 12:46cells with the high neoantigen load
• 12:48are removed by the immune system.
• 12:49So that's really attractive.
• 12:51Hypothesis and it makes testable predictions.
• 12:54So one prediction is that even tumor cells
• 12:57sort of undergo medical transformation.
• 13:00Some of it could be eliminated
• 13:01by the immune system.
• 13:02So if that's really true,
• 13:03then then actually immunotherapy
• 13:05should work as chemoprevention.
• 13:07Of course, it's too toxic to do that,
• 13:08but the concept is important.
• 13:10So we're going to test this in
• 13:12an ongoing large event trial that
• 13:14uses symbolism for a year to see
• 13:16whether it alters contralateral
• 13:17breast cancer events and also
• 13:19whether it alters breast density.
• 13:21Which is sort of a somewhat
• 13:24validated risk predictor.
• 13:25But the most important consequence is this
• 13:27that when we actually diagnose these cancers,
• 13:29there may be a quasi equilibrium fight
• 13:32between the immune system and the cancer.
• 13:34So when there are a lot of immune cells,
• 13:35it's kind of indicate that the
• 13:37immune system is having almost upper
• 13:39hand and that's why it actually is
• 13:42associated with better prognosis.
• 13:43But at that stage you might actually
• 13:45help tip the balance towards the
• 13:47immune system by chemotherapy or by
• 13:49immune checkpoint inhibitors and then.
• 13:51Do not have the drugs to test this.
• 13:53And we actually launched 4 studies
• 13:54to to address these questions
• 13:56and three of them have results,
• 13:57and I'll show that to you.
• 13:59But the third prediction is also interesting,
• 14:02right?
• 14:02So if you really follow this logic,
• 14:04then the metastatic disease should
• 14:06really arrive through an immune escape.
• 14:08So we did a series of studies
• 14:10to compare primary
• 14:11exams and maths, and it's among the
• 14:13first groups to show that actually
• 14:14metastatic lesions in breast cancer
• 14:16are profoundly immunocompromised.
• 14:18And we also looked at whether there
• 14:21is subtle variations by sight.
• 14:23So now these are all sort of
• 14:25relatively valid accepted principles.
• 14:26I I thought I showed this to you,
• 14:29especially for those of you
• 14:31who are younger investigators.
• 14:32So there are risks of being coming up
• 14:34with an idea too early or too late.
• 14:36So this particular idea came
• 14:37on a little bit too early.
• 14:38In 2012, about a month of Tiki came here.
• 14:41I approached Merck to do 2 large
• 14:43studies in the curative setting.
• 14:46What was the neoadjuvant trial to see
• 14:48whether we could actually push the PCR?
• 14:49It's up based on the associations that
• 14:51I showed you to test the causality.
• 14:53The other one was an adjuvant study.
• 14:54We could actually improve the outcome by
• 14:56giving people liberalism out and eradicate.
• 14:58Micromedex and this is what they said,
• 15:00sorry you're unable to avoid the drug
• 15:02and the monetary support at this time
• 15:04due to unclear regularly path forward.
• 15:06But it was three years later they
• 15:08actually realized that there is a
• 15:10path forward and they actually run
• 15:11both of these studies or or agree to
• 15:13do it and they to their credit they
• 15:15actually invited me back to their
• 15:17steering committee of the new adjuvant
• 15:19trial and I lead the adjuvant trial.
• 15:22So what do these studies show it?
• 15:24This is just the selection that is
• 15:26representative of the findings from
• 15:28the neoadjuvant immunotherapy trials.
• 15:30And they were lounged in triple
• 15:32negative disease because of the
• 15:34really strong association of immune
• 15:36cells with pathologic CR or strong
• 15:38strong association with prognosis.
• 15:40And by and large triple negative
• 15:41cancers have a higher in your presence.
• 15:44So all these studies took place
• 15:45in in that space except one,
• 15:47the ice spy all talk to you a
• 15:49little bit more about it.
• 15:50So what this study shows is that the
• 15:51the computer response rates improved.
• 15:53Didn't have as much as we thought.
• 15:55So the largest study keynote 5 to 2,
• 15:57the Merck study showed improvement
• 15:59about 7 percent, 56 to 63.
• 16:01Really underwhelming because chemotherapy
• 16:02trials could do double digit improvements.
• 16:05Yet the chemo studies actually
• 16:06didn't really improve the event
• 16:08free survival that dramatically.
• 16:09Oftentimes it didn't deal with
• 16:10it all to a significant extent.
• 16:11But keynote 522 did.
• 16:13You see the same in an even smaller study,
• 16:15paranormal.
• 16:16They're also showed a 9% even PCR rate.
• 16:18Not even significant,
• 16:20but the event free survival was significant.
• 16:22The other?
• 16:23Important finding in this sort of
• 16:25or observation from these studies
• 16:27is that in metastatic disease,
• 16:29again parallelism have improved the
• 16:31outcome when combined with chemotherapy.
• 16:33But this was only seen in the pediatric
• 16:35and positive patients whereas in
• 16:36the early stage setting you don't
• 16:38need to have Pedialyte and one.
• 16:39So that confuses a lot of people.
• 16:41But I think there is a really
• 16:42simple and elegant
• 16:43explanation and it comes from the
• 16:44slide that I showed you previously
• 16:46from the fact that the metastatic
• 16:48lesions are immunocompromised or really
• 16:50immunosuppressed immune attenuated so.
• 16:53And the only stage setting I think a
• 16:54small amount of immune presence that
• 16:56you could miss with the biopsy and they
• 16:57actually miss it oftentimes with biopsy.
• 16:59So this is a work that Adriana Khan,
• 17:01one of our fellows showed and we presented
• 17:03the San Antonio Breast Cancer meeting.
• 17:05So even a few period like in one positive
• 17:07cells that are intermixed with the
• 17:09micro environment and missed the initial
• 17:11biopsy could be enough to actually
• 17:12ignite an immune response and the same
• 17:15way chemotherapy ignites sort of like
• 17:17one expression in the more massive scale,
• 17:20but you don't see the same thing
• 17:21in in in the metastatic setting.
• 17:24So the other question was this really.
• 17:27This thing observation that why small
• 17:29improvements in Pathologic CR really lead
• 17:32to large improvements in survival whereas
• 17:34in other setting it doesn't happen.
• 17:36So that brings me to another sort
• 17:38of debate that used to rage and
• 17:40the the breast cancer community and
• 17:42we spent a lot of time on it.
• 17:44It's really prompted by the 1st
• 17:46initial new adjuvant trials and shovel
• 17:47power to show improvement in PCR,
• 17:49but was woefully underpowered and
• 17:51included all subtypes to to really
• 17:53show improvement in survival.
• 17:55So this matter analysis by the FDA
• 17:57and showed very little in fact
• 17:59no relationships at all between
• 18:01improvement in PCR and survival.
• 18:02They confused a lot of people,
• 18:04but it would have to fly against
• 18:06the totally common sense.
• 18:07Observations, Taxol improved pathologic,
• 18:09sciarid improved survival receptive
• 18:11improved Pathologic CR,
• 18:13it improves survival.
• 18:14Platinum improved Pathologic CR
• 18:15it's and now we know that it
• 18:17improves survival as well.
• 18:18And of course the immune checkpoint
• 18:21inhibitors improved pathologic
• 18:22security improve survival.
• 18:23But nevertheless it's really true
• 18:25that at the individual trial level
• 18:27the relationship between the PCR
• 18:29change improvement and the improvement
• 18:31in PFS is hugely variable.
• 18:33So that's the next question to
• 18:34study why and I actually have a
• 18:36good explanation for you.
• 18:37And I think it's very elegant and simple.
• 18:39But to understand that you need
• 18:42to familiarize yourself with this
• 18:44concept of of a continuous metric of
• 18:46of outcome or pathological response.
• 18:49So again in 2007 we developed this
• 18:52metric called residual cancer burden
• 18:54to capture the pathological residual
• 18:57disease as a continuous variable.
• 19:00We did that because continuous
• 19:01variables are more powerful to
• 19:03identify genes that would be associated
• 19:05with outcome or not but.
• 19:07So eventually it took sort of
• 19:09traction in the form of categories,
• 19:11so you can use this continuous score to
• 19:14create bins of 0 being complete response.
• 19:17Another bin.
• 19:18That's the minimal residual disease
• 19:19or RCB 1 moderate amount or CB2
• 19:22and a large amount of RCB 3.
• 19:23But the truth is that this is really
• 19:26a continuous scroll and that's
• 19:27why we did it so.
• 19:28Be teamed up the deal I spoke
• 19:31to investigators because
• 19:32this continuous sort of score,
• 19:35I thought actually could reveal
• 19:36some really interesting things
• 19:37about how different drugs work.
• 19:39So what you see here is actually a pretty
• 19:42cool picture of the continuous RCB scores in
• 19:45seven different arms of the eye spy study.
• 19:48So the eye spy is randomized trials,
• 19:50the control arm is always staxel ACC,
• 19:52and but you see here is the RCB values
• 19:55from zero to 50 is complete response.
• 19:58Five is expensive.
• 19:59Single disease.
• 19:59This kind of shows you the the the
• 20:01prevalence of the density or the
• 20:03frequency with which you encounter a
• 20:05particular RCB value in the trial arm.
• 20:07So the black is the control and the dotted
• 20:11lines are various experimental drugs.
• 20:12I just want to look at you the two
• 20:15panels which are labeled so I don't
• 20:18think I can use a A.
• 20:21Sort of a pointer,
• 20:22but you probably see there
• 20:23that the bottom panel,
• 20:24which is regimen 7,
• 20:25you have a large improvement in PCR rates,
• 20:28right, because the the initial
• 20:29zero values are much higher.
• 20:31That's where the curves start.
• 20:33But you also see a massive shift towards
• 20:35the smaller values across the board.
• 20:37If you look at the Regiment 3 on the
• 20:39top instead of right hand corner,
• 20:42then you see that that regimen
• 20:43also improves PCR rates.
• 20:44But it does it by moving the RCB 1,
• 20:47the little residual disease group,
• 20:49into the PCR company response.
• 20:52And that is very unlikely to
• 20:53affect survival like it doesn't.
• 20:55But this particular regimen didn't
• 20:56affect at all the higher residual cancer.
• 20:59So we thought that actually measuring
• 21:01the the distribution of the differences
• 21:04in residual cancer burden scores could
• 21:06capture the efficacy of a regimen.
• 21:09And we developed a new statistical tool
• 21:10that you can find in this paper and
• 21:12you can even play with it if you have
• 21:14a breast cancer on this open website,
• 21:16we call it treatment efficacy
• 21:17score and it basically compares
• 21:19the distribution of RCB scores.
• 21:21Cross through trial arms in that
• 21:23particular metric actually really
• 21:25correlates quite well with event
• 21:27free survival which is what you see.
• 21:29There's a significant difference.
• 21:30There is an event free survival improvement.
• 21:32Is that all significant improvement
• 21:33in this test score then you don't
• 21:36have significant improvement
• 21:37in event free survival.
• 21:39So we're going to validate this
• 21:40within with the other groups.
• 21:42So we're not move to this other question
• 21:44that these studies show up, right.
• 21:46So pembrolizumab is expensive and 15%
• 21:49of the patients have severe toxicity,
• 21:51so.
• 21:51He entered into this race to find
• 21:54predictive markers that define the
• 21:57patients who need pembrolizumab and
• 21:59this is a slide from from us from
• 22:02a group in Germany civil libel.
• 22:03And one of my former lab members Thomas Kuhn,
• 22:06who leads their translational research arm.
• 22:08And what they show in this randomized
• 22:11immunotherapy versus chemotherapy alone
• 22:13ARM study that there are a number
• 22:15of molecular variables that predict
• 22:17response to any if you have them like
• 22:19high commutation burden or a high.
• 22:22Energy and expression or high P like in
• 22:24one expression or high till comes you
• 22:26have higher PCR rate with chemotherapy,
• 22:28chemotherapy but also with
• 22:30chemotherapy plus immunotherapy.
• 22:32But the improvement by immunotherapy
• 22:34happens in both groups,
• 22:36the remediation low and high,
• 22:38the PD low and high or the field
• 22:40count low and high groups.
• 22:42So these are these one of these
• 22:44metrics are selective to identify
• 22:46who actually needed the panel,
• 22:48but we have an idea who actually
• 22:50might benefit from Pedro.
• 22:51So we teamed up with the investigators.
• 22:53On the build who previously suggested
• 22:55that MH subclass 2 expression in tumor
• 22:58cells might actually identify a group,
• 23:00the group of patients who
• 23:01really need it Pembroke.
• 23:03So I need to see class to is is
• 23:05mostly expressed in immune cells and
• 23:08participates in antigen presentation,
• 23:10but it can be induced to be expressed
• 23:12in cancer cells and epithelial cells
• 23:14by interferon gamma, for example, so.
• 23:17Have you run this immunity chemistry,
• 23:20a simple immunity chemistry for
• 23:22emission classical expression on
• 23:23cancer as opposed to the immune cells.
• 23:26And we actually confirmed that what
• 23:29Justin Balko originally reported
• 23:32that the cancers which were positive
• 23:34for MHC Class 2 expression actually
• 23:36had a higher pathologic CR rate when
• 23:38Pembroke was added in the ice spy study.
• 23:41But the pathologic CR,
• 23:42it was the same whether they were
• 23:44MHC Class 2 high or low if they
• 23:46only got chemotherapy and so.
• 23:48They really strong interaction,
• 23:50marker treatment interaction
• 23:51in that study and parallel with
• 23:54this completely independent.
• 23:55Another set of former lab member of mine,
• 23:57Jean-paul Bianchini showed the
• 23:59same thing in their new adjuvant
• 24:01trial without the salesman.
• 24:02You know,
• 24:03I highlighted for you the
• 24:05interaction between Italy,
• 24:06the expression on epithelial cells that
• 24:09actually predicted higher odds ratio for PCR.
• 24:12Vidot is always the map but didn't
• 24:14have any sort of significant other
• 24:16ratio with chemotherapy alone,
• 24:17but the same.
• 24:18Study our immune cells didn't carry this.
• 24:21So it's a really cool project there
• 24:23and we just got funding from the NCI
• 24:25to kind of test this and validate this
• 24:27in a larger trial them S 1418 that I,
• 24:31I mentioned to you earlier.
• 24:33But again,
• 24:34so this study is the fascinating thing
• 24:35about science is that every advance
• 24:37actually throws up new questions,
• 24:38even more interesting questions.
• 24:41So one question is why some cancers
• 24:44are important in reach, right?
• 24:46A lot of people are struggling
• 24:47to find answers,
• 24:48how you make a cold against the heart.
• 24:51But we thought we ask something a
• 24:52little bit more original and maybe
• 24:54something that that could be easier to crack.
• 24:56And that's the question,
• 24:57why doesn't all immune high
• 24:59cancers actually accomplished PCR?
• 25:01Why is the PCR only 63%?
• 25:03And 100 or 90 that's a project
• 25:05that Kim actually came women led
• 25:07and we compared the immune reach
• 25:09triple negative disease that had
• 25:11the PCR versus those that did not.
• 25:13And we find really pretty interesting
• 25:15stuff that I think could lead
• 25:17us to some leads about what
• 25:19combination therapies,
• 25:21immunotherapies could really be
• 25:23make embolism and more effective.
• 25:25So just to summarize this let's we
• 25:27found that the teacher have better if
• 25:29one teacher beat is high in the immune
• 25:31microenvironment even if you are in reach.
• 25:33You don't accomplish PCI and a
• 25:35lot of innate immunity markers
• 25:36are also associated with it.
• 25:38The innate immunity markers actually
• 25:41are macrophage and K markers and when
• 25:43you look at the cytokine milieu then
• 25:46you really see this very strikingly
• 25:48so cancers it raises your disease.
• 25:50The dominant cytokines are actually
• 25:52cytokines which are involved
• 25:54in chemotaxis and activation of
• 25:56neutrophils and macrophages.
• 25:58So we hypothesized they're blocking.
• 25:59Some of those would actually improve
• 26:01the outcome or the efficacy. Of.
• 26:05You actually went pembrolizumab.
• 26:07So interestingly I just put that
• 26:09asterisk for you to to that.
• 26:11It's so beautiful because it congruent.
• 26:13So we find that a lot of these very
• 26:15same cytokines that we see highly
• 26:17present in immune rich non responding
• 26:20TNBC at the very same chemokines
• 26:22and silicones that we find in the
• 26:25microenvironment metastatic disease
• 26:26right in that paper that showed that
• 26:28the metastatic microenvironment
• 26:29is more immuno attenuated.
• 26:34Just instead of finish these sort of
• 26:36series of questions and immunotherapy off.
• 26:38So if immunotherapy works
• 26:40beautifully entrepreneur disease,
• 26:41could it actually work in a
• 26:43subset of ER positive cancers.
• 26:44And we think that it will work because
• 26:47we noticed in the eye spy trial data
• 26:50that in three arms that included
• 26:52immunotherapy including the door volume up,
• 26:55Olaparib arm, the Iliad,
• 26:57the Penrose Metaxa arm and the pembrolizumab
• 26:59and it's all like receptor antagonist.
• 27:02Arm in all of these three arms
• 27:05independently we saw that among the
• 27:07ER positive here we call them HR
• 27:09hormone receptor positive cancers.
• 27:11There is a group that is characterized by
• 27:15routinely reported sort of molecular feature,
• 27:18the ultra high mammaprint status.
• 27:21So all of these patients had
• 27:22to have high mammaprint result.
• 27:23High MAMMAPRINT defines patient
• 27:25superficially benefit from chemotherapy
• 27:27but within that high mountain
• 27:28group you can devise an agent,
• 27:30they actually introduce their system.
• 27:32The device to group smaller print
• 27:34high high and some Withrow high.
• 27:36So the small print we throw higher
• 27:38MP two group is the subset among the
• 27:40ER positive patients who benefited
• 27:41and it's really, really elegant.
• 27:42You can't see that right.
• 27:44So the HR positive MP1,
• 27:46there's no difference whether
• 27:47you get chemo plus durva,
• 27:49but if you are MP two then
• 27:50Nirvana improves your PCR.
• 27:51It's same for pembrolizumab
• 27:53with the other two arms.
• 27:55And what's even nicer when you
• 27:56look at the molecular features
• 27:58of these empty two patients,
• 27:59the area are positive but
• 28:01their ER signaling and.
• 28:02Yeah,
• 28:03sort of the gene signatures that typically
• 28:05associated with endocrine sensitivity,
• 28:07this is low.
• 28:07So that's the group let's see
• 28:09are positive but least likely to
• 28:10benefit from endocrine treatment.
• 28:12They have sort of a higher proliferation
• 28:14signature which also makes sense.
• 28:15So they are more sensitive to chemotherapy
• 28:17and we also saw this in the the,
• 28:19the chemotherapy arms and but
• 28:21we didn't really see a major
• 28:23difference in the immune micro
• 28:25in in immune signature genes.
• 28:27So again we hope to launch the prospective
• 28:31study that would validate this concept.
• 28:32With the routinely available essay we
• 28:35could actually identify a group that
• 28:37will benefit from the same way as
• 28:39triple negative disease benefited from
• 28:42including immune checkpoint therapy.
• 28:44So just to summarize these clinical
• 28:46partially the paradigm shift that
• 28:47happened in the past sort of 20
• 28:49years is that the best way to treat
• 28:51most stage two and stage three
• 28:52triple negative patients is new
• 28:54adjuvant chemotherapy and the best
• 28:56PCR rates are accomplished about
• 28:57two third of the patients having a
• 28:59competent navigation of the cancer,
• 29:01the same happened in her two
• 29:02positive disease.
• 29:02Don't talk about this because it's
• 29:04really predated at least by 1015 years,
• 29:06the immunotherapy revolution
• 29:07and there are a lot of really
• 29:10interesting studies that will push
• 29:12the survival even further among
• 29:14those who have residual disease.
• 29:16So there are new studies that
• 29:17are launched in that space that
• 29:19I kind of highlighted for you.
• 29:21So what's next,
• 29:22right.
• 29:22So what's going to be the
• 29:23next paradigm shift in the next 10 years?
• 29:25And I think the this is really.
• 29:28I I see two really potentially very
• 29:30high impact fields which we could
• 29:32improve again survival within the
• 29:34next 5 to 10 years and which is.
• 29:36So wait a second.
• 29:41Yeah. So what is coming up with this
• 29:44concept that could we detect molecular
• 29:46relapse in solid tumors the same way as
• 29:49we detect molecular relapse in leukemia.
• 29:51So if you see that with PCR that
• 29:53your genomic abnormalities returned,
• 29:55then a second round of treatment
• 29:56at that point would actually
• 29:58cure some people from leukemia.
• 29:59So could the same paradigm apply to
• 30:01to sometimes it didn't really have
• 30:03good ways to catch this and we didn't
• 30:05really have good effective drugs
• 30:07either 5610 years ago to test this,
• 30:09but now we have we have most molecular.
• 30:11Essays that can pretty reliably
• 30:13identify and the SEC DNA is
• 30:16particularly tumor informed C DNA.
• 30:18So if you have a high C DNA level
• 30:20that's starting to rise while you
• 30:22are in the surveillance of follow
• 30:24up stage of the initial curative
• 30:26therapy as the city then rises,
• 30:28unfortunately it's almost sure bad that you
• 30:30will have a recurrence clinical recurrence
• 30:32within the next seven or eight months.
• 30:35So could we intervene at that point
• 30:37when people are still sort of
• 30:39micrometastatic but the micrometastasis
• 30:40is raising its ugly head?
• 30:42So that's an idea of a second line.
• 30:44I look in therapy and we
• 30:45actually lead a study.
• 30:47We have a study in that space that that's
• 30:49exactly this idea in your positive
• 30:51patients who are receiving endocrine
• 30:53therapy but start to have a rising CDN,
• 30:55they randomized the full
• 30:56Western public cycling and.
• 30:58And we'll just continue with their standard
• 31:00of care treatment and get treatment
• 31:03when they become clinically symptomatic.
• 31:05So the other potentially paradigm
• 31:07shifting idea is really that they
• 31:08could cure some metastatic disease.
• 31:10So you have metastatic disease kind of the
• 31:12current dogma is that you will die from it.
• 31:14It may take many, many years,
• 31:15but ultimately people die.
• 31:17I'm not sure that this actually
• 31:18has to happen like this.
• 31:20So what happened in the past five,
• 31:22six years is that you really
• 31:24understood much more clearly that
• 31:26only that there are multiple.
• 31:28Different types of meds,
• 31:29not just some medicine.
• 31:30Disease doesn't exist.
• 31:31There's a homogeneous entity,
• 31:33just like the breast cancer doesn't
• 31:34exist to looking. It doesn't exist.
• 31:36It's a useful concept.
• 31:37But practically really these
• 31:38are all very there are many,
• 31:39many different types of leukemias that
• 31:42require different approaches and treatments,
• 31:44different types of breast cancers.
• 31:45And the same way like metastatic
• 31:47disease is also heterogeneous.
• 31:49So the novel stage for disease is unique
• 31:51because it never received any prior therapy.
• 31:54That's obviously very different
• 31:55from somebody relapsing and
• 31:56having a metastatic disease.
• 31:58After they went through all the
• 31:59treatments that I showed you
• 32:00in the new adjuvant setting,
• 32:01the chemotherapies was embolism and whatnot.
• 32:04So curing those folks with
• 32:06existing therapies is a long shot,
• 32:08but curing those folks who never had
• 32:10any therapy with the combination of
• 32:11drugs is probably not such a long shot.
• 32:13And there are many case reports and
• 32:15oncologists who practice for a long time.
• 32:17All have anecdotal cases of
• 32:19metastatic patients,
• 32:20particularly with her two positive
• 32:21disease because her two positive
• 32:23disease had the best drugs initially
• 32:24the her two targeted drugs,
• 32:26but now we have good drugs for
• 32:27for triplet disease as well.
• 32:29And also for your poster disease,
• 32:31so this paradigm that really
• 32:32kind of put into the mind of many
• 32:35practicing physicians that some her
• 32:36two positive cancer can be cured.
• 32:38I think it's kind of increasingly
• 32:40applicable to the other subtypes as well.
• 32:43So we hope to do a study that would
• 32:45actually focus on covad especial
• 32:47group of her of metastatic patients,
• 32:50they de Novo newly diagnosed
• 32:52metastatic patients particularly
• 32:53with oligo metastatic disease,
• 32:54so that we could really get rid of all
• 32:57the known homicides and what's left.
• 32:59Is micromass,
• 32:59but we can deal with micro Mets.
• 33:01That's the success story that I showed you.
• 33:03That's how adjuvant therapy improves
• 33:05survival after removing the the primary
• 33:07breast cancer in the lymph nodes,
• 33:09the systemic therapy.
• 33:10Washes and and and kills them
• 33:12at the Micromax.
• 33:14So I think this better than probably
• 33:15will hold up in stage four disease
• 33:17and the vision is very simple.
• 33:18So in in five or ten years you don't
• 33:20call these patients the oligo metastatic
• 33:23stage four patients stage four,
• 33:24but you call them stage 3C.
• 33:26Because they are deep, sorry.
• 33:28Because then they will be curable.
• 33:31So I'm going to move on to some other
• 33:33projects that I also find amazing and I
• 33:35just wanna share you some of the results.
• 33:37So why do some women develop breast
• 33:39cancer 20-30 years earlier than the
• 33:41average or median age even in the
• 33:43absence of any germline mutation?
• 33:45Actually that's the majority of
• 33:47young women with breast cancer.
• 33:48It's only a minority who has broken
• 33:51mutations rather identified mutations.
• 33:52So we had two ideas.
• 33:53One was that each is the strongest non
• 33:56genetic risk factor for breast cancer.
• 33:58So could you actually sort of
• 34:00hypothesize that young women?
• 34:02Could be breast cancer actually
• 34:04experience an accelerated epigenetic
• 34:06age of their breast.
• 34:07So this was an idea that Erin Hofstatter,
• 34:09our former colleague picked up and
• 34:11we did a series of publications
• 34:13that actually suggests that this
• 34:14is indeed happening.
• 34:15So it shows you this insert from
• 34:17the the clinical epigenetics paper
• 34:19in 2018 shows this the most sort
• 34:22of simply and clearly.
• 34:23So what you should what you see
• 34:25there is each acceleration in the
• 34:28normal breast tissue of women who
• 34:30had breast cancer later and the.
• 34:32Epigenetic age acceleration of people
• 34:34who never develop breast cancer.
• 34:35So we did this with the Susan Comment
• 34:37Tissue Bank and with some tissues from here.
• 34:40So you see that there is a
• 34:42significant acceleration.
• 34:43So epigenetically speaking based
• 34:45on the methylation signature,
• 34:47the breast normal breast tissues of
• 34:49woman who subsequently developed breast
• 34:51cancer is older than their chronological age.
• 34:54And we don't see this to such
• 34:56extent in the control patients.
• 34:58And then and then we had some follow
• 34:59up patients which really kind of
• 35:01papers that explained that it's mostly.
• 35:03Polycom related genes whose
• 35:04methylation pattern is associated
• 35:06with this age acceleration,
• 35:08and this last paper on the review in
• 35:11science advances shows that actually every
• 35:13cell proliferation adds a little bit of
• 35:16epigenetic aging to to to the tissues.
• 35:18And there is a share of epigenetic
• 35:21signature between cancers and and normal
• 35:23cells and it relates to aging and it
• 35:26relates to ultimately cell divisions.
• 35:28But it's probably not the full story though.
• 35:31So what's the rest of the story?
• 35:32So family history is a predictive risk
• 35:34factor even in the absence of any
• 35:37detectable hyper reference gene mutations,
• 35:38right? So something you inherited
• 35:40increases your risk,
• 35:41even if it's you can't see it so.
• 35:44Polygenic risk scores that use individual
• 35:47snips that are individually associated
• 35:49with risk to a very small extent,
• 35:52sum them up and you've made them
• 35:53by the risk that they confer.
• 35:55That's a polygenic risk score.
• 35:56However,
• 35:56even the best ones today using several
• 35:59100 risks polygenic risk and have
• 36:01a lot of missing heredity in them.
• 36:03So they don't explain this complete story.
• 36:05So we have this other idea that could
• 36:08the combination of non recurrent rare
• 36:10germline variants and cancer relevant
• 36:11genes determined individual risk.
• 36:13So because they are not recurrent.
• 36:14Missed them in in indigenous studies,
• 36:16right,
• 36:17because they start out finding individual
• 36:19snips that are associated because
• 36:20they are recurrent in the mental
• 36:22state of India's cancer population.
• 36:24But if it's not recurrent,
• 36:25you won't see it.
• 36:28So this is an idea that really kind of
• 36:30wanted me for quite a while since this
• 36:32paper came out from the 1000 Genome Project,
• 36:34which showed that all of us
• 36:35here have different faces.
• 36:36And the reason we have different faces
• 36:39is this amazing set of variation in
• 36:41Snips and Jermaine Snips and other
• 36:44genomic variations that we are born with.
• 36:48So an average person carries about
• 36:5020 and 50 to 350 genes that have
• 36:54a loss of function.
• 36:55That's probably the reason why I have
• 36:56this poor voice and small stature.
• 36:58But anyway,
• 36:59so the point is that this low
• 37:01frequency events that occur in unique
• 37:04combination individuals might set the
• 37:06stage that what additional events
• 37:08matter or cause the transformation.
• 37:11So it's a combinatorial effect, right?
• 37:14So.
• 37:16We put these hypothesis forward
• 37:18that really that functional germline
• 37:20variants as potential Co oncogenes.
• 37:22And this actually I think there's
• 37:25something that covers on the screen.
• 37:27Yeah, so you can't see this well,
• 37:29but this model,
• 37:30the the nice thing about models is
• 37:31they predict testable hypothesis, right.
• 37:34So this particular idea that the
• 37:36Germans polymorphisms all of them
• 37:37together said this theme stage for
• 37:39what counts as an oncogenic event and
• 37:41eventually this is the totality of
• 37:43abnormalities that lead to cancer.
• 37:45So it's this sort of testable leads
• 37:48to this testable hypothesis,
• 37:50right that cancers in younger patients.
• 37:52This is correct.
• 37:53They should have more germline variants
• 37:55because they need fewer somatic
• 37:57events to reach a threshold, right?
• 37:59The sexual disturbance that
• 38:01pushed them over to to
• 38:03become malignant.
• 38:04And theoretically you could also
• 38:06use this idea to develop a cancer
• 38:09gene systems integrity score that
• 38:10captures how far a cell or tissue is
• 38:13from this malignant transformation.
• 38:15So we started to study that.
• 38:17And this is a paper that
• 38:19touching postdoc in my lab did.
• 38:21So we asked this really fundamental
• 38:24simple thing that amazingly not a
• 38:26lot of people actually studied before
• 38:27that what's the relationship between
• 38:29the person's age of that each of your
• 38:32diagnosis of cancer and the germline
• 38:35variant load in cancer relevant genes.
• 38:38So what are cancer relevant genes?
• 38:39So we just put from the literature
• 38:41and from from review articles about
• 38:431500 genes which are experimentally
• 38:46validated that they alter.
• 38:48They've played an important
• 38:49biological role in cancer.
• 38:50And when you see here,
• 38:51it's actually pretty obvious and
• 38:53it's really beautiful, right.
• 38:54So people who develop cancer at
• 38:57an older age have fewer germline
• 39:00alterations in these cancer relevant genes.
• 39:02People who develop cancer at
• 39:03younger age have a much higher,
• 39:05these are age bins by years of 10 and the
• 39:08opposite is seen in the somatic space.
• 39:10So people will develop cancer at their ages.
• 39:12Prostate cancer folks
• 39:13have a lot of mutations,
• 39:14whereas people who develop cancer
• 39:15at an early age have fewer somatic.
• 39:18Positions,
• 39:18and we knew this from the
• 39:20pediatric literature actually.
• 39:20Pediatric cancers don't have
• 39:22a heck of a lot of mutations.
• 39:24So that's actually a really nice story
• 39:26that that supports this idea that
• 39:27somehow that's the combined effect.
• 39:29And if you have a lot of germline hits,
• 39:30you need need a fewer random
• 39:33somatic hits to push you over.
• 39:36In this paper view,
• 39:37it kind of did you think a
• 39:39little bit deeper and you know,
• 39:40so cancers which actually are highly
• 39:43linked to environmental factors for
• 39:45lung cancer for example that they
• 39:47actually tend to have a lot more
• 39:49somatic events and some somatic
• 39:50mutations from somatic origin,
• 39:51from germline in other cancers
• 39:53kind of coffee.
• 39:54So in between and some of them are
• 39:56actually like testicular germs,
• 39:57atoms are dominated by germline
• 40:00hits rather than somatic hits.
• 40:05But then this this location OK so
• 40:07why 1500 genes so probably are
• 40:09there more genes related to cancer.
• 40:12So we we asked this question whether
• 40:14what's the what's the totality
• 40:16of cancer relevant human genes
• 40:18and the name we came up with the
• 40:21really simple concept that if.
• 40:23Core cancer genes are important
• 40:24and we define core cancer genes
• 40:26actually from a clinical panel,
• 40:28the MSKCC impact panel that's clinically
• 40:30used to define actual permutations.
• 40:33So these hypothesized the
• 40:35genes that interact in a.
• 40:38Putting putting interaction network or
• 40:40the string network that there's a lot of
• 40:42different ways to measure interactions.
• 40:44So genes that interact with the core
• 40:45genes will be somewhat important and
• 40:47genes that interact with this one step
• 40:49remove genes will also be important
• 40:50to some extent but probably less.
• 40:52And then those which are three four
• 40:54steps removed are even less important.
• 40:56So we wanted to test this hypothesis,
• 40:58but as you get closer to the close genes
• 41:00then you would have increasing connectivity.
• 41:02That's one mathematical way to measure
• 41:05the importance of gene as you get closer.
• 41:08So one step.
• 41:09Both from from core cancer genes
• 41:10then it's going to be more important
• 41:12than survivability.
• 41:13We can check this in genome
• 41:15wide CRISPR and ASARONE screens.
• 41:17Also predicted genes which are
• 41:18one step removed,
• 41:192 steps removed are more important than
• 41:21those which are three steps removed
• 41:23in terms of having large number of
• 41:25somatic mutations in in Kansas and
• 41:27that they will be under a stronger
• 41:29negative selection in the germline,
• 41:31right,
• 41:32because they are important.
• 41:33And in many of these genes that
• 41:34are important,
• 41:35cancer are important in many other things
• 41:37and that's exactly defined in this paper.
• 41:39And this just shows you the numbers though.
• 41:40So one or two step remove genes in
• 41:43our genome is about 10,000 genes.
• 41:45So actually probably the cancer 11
• 41:47genes space is much much bigger,
• 41:49just don't know about a lot of these
• 41:51and of course they're importance is
• 41:53not as important as a P53 mutation but
• 41:56nevertheless they contributes very
• 41:58likely contribute to the biological disease.
• 42:01So where are you going with this?
• 42:03So what you actually want to do
• 42:05really is so address cancer as a
• 42:08cellular transformation as as a a
• 42:10defect in a in a in a complex system.
• 42:13So complex systems fail through unique
• 42:16combinations of individual non lethal events.
• 42:18I mean just think about this if
• 42:20you would run the statistics on
• 42:21what's causing plane crashes,
• 42:22even find anything.
• 42:23Because even though flying through
• 42:25a storm is a risk,
• 42:26but many many planes fly through
• 42:28storms have any problem, you know,
• 42:30pilot sleeping or not.
• 42:31Been trained,
• 42:32it's a lot of happens that that
• 42:34despite of this sort of human errors,
• 42:36the plane survives,
• 42:37you don't even know about it.
• 42:39So it's really a unique combination
• 42:41that brings down points.
• 42:42And so that's the thing that we
• 42:43actually try to see whether we could.
• 42:45So some of these unique combination
• 42:47of Germany and some
• 42:48of the events into a score and
• 42:50they ultimately visualize it.
• 42:51They did a little bit of a sort of
• 42:54preliminary kind of effort in this
• 42:56few years ago with wavey she trying
• 42:58to kind of map all the molecular
• 43:00abnormalities that particular cancer.
• 43:02As and visualize it in a standardized way
• 43:04in in these papers we try to resurrect
• 43:07this really delighted that Susan Coleman
• 43:09actually accepted this challenge for
• 43:11their hecaton in March next year.
• 43:13So we're going to lead A-Team to to
• 43:16try to develop this Kansas score. Umm.
• 43:22So the new classes of drugs, right.
• 43:24So that's the last piece that I'm
• 43:26actually going to talk to you a little
• 43:28bit because I'm so excited about it.
• 43:30So metabolically, right,
• 43:31rewiring is a major hallmark of cancers,
• 43:34yes. Yeah, we don't have any
• 43:36drugs that exploit it.
• 43:37So remember, a lot of chemotherapy
• 43:38drugs interfere with DNA synthesis
• 43:39because you need to double your DNA,
• 43:41but you need to also double your lipids.
• 43:43You also need to double your proteins.
• 43:45So why don't we have drugs in that space?
• 43:48So we started off with the computational
• 43:50biology project to look for.
• 43:52Most of isoenzyme diversity in cancer
• 43:55compared to corresponding normal tissue.
• 43:57So isoenzymes kind of more or less sort of
• 44:00could catalyze the same chemical reaction.
• 44:03But they are different genes and sometimes
• 44:05they are located in different compartments.
• 44:07So what you want to look at is is a
• 44:09particular isoenzyme becomes cancer dominant.
• 44:12So this isoenzyme diversity gets lost because
• 44:14out of the three or four isoforms that
• 44:17produce the same sort of chemical reaction,
• 44:19one becomes dominant.
• 44:20That may be actually important.
• 44:22Analogy.
• 44:22So if you're looking for is this sort of
• 44:25change that the normal cell has kind of fun,
• 44:27actually both sides of enzyme one
• 44:28and two and the cancer actually
• 44:30one of these becomes dominant.
• 44:32So we asked how many are these
• 44:34in the human genome?
• 44:35So we again went to the TTC share
• 44:38data and called all the human enzymes
• 44:40which have less than 5 isoforms
• 44:42to find to look for a pattern that
• 44:45showed this cancer dominance.
• 44:47Once we find this,
• 44:48then we looked whether we can see the same
• 44:50in the CLA the cancer cell line encyclopedia.
• 44:52Just to make sure that this is really
• 44:54happening at a cellular level,
• 44:55not at the tissue level because the
• 44:57TCG's tissue level and it also then
• 45:00once we confirm those that they are
• 45:02also dominant in a cancer cell line
• 45:04that enabled us to really check
• 45:06whether this particular isoform is,
• 45:08is survival critical in the depth
• 45:10map data which is CRISPR.
• 45:13You have no card database.
• 45:15And then the final hit you wanted to confirm,
• 45:17so this is what we found.
• 45:18So there are about 136 cancer breast cancer
• 45:22dominant isoenzymes that we find in the CG.
• 45:25About 81 of these are also cancer
• 45:27dominant in breast cancer cell lines,
• 45:29but 53 are important for survival.
• 45:33When you knock it out,
• 45:34you can sell lines, survival improves.
• 45:37And about 44 of these,
• 45:38the locking out the the particular
• 45:40isoform is more important than knocking
• 45:42out the other one and then you actually
• 45:44meet all these three criteria then
• 45:46you end up with about 17 potential
• 45:49targetable isoenzymes in breast cancer.
• 45:52But we did this for a whole bunch
• 45:54of cancer types and the the most
• 45:56shared sort of cancel them in a
• 45:58nicer form turned out to be a C1
• 46:00or acetyl coenzyme carboxylase.
• 46:03And this little uncertainty,
• 46:05the things the right side for
• 46:09you shows the
• 46:10the actual pattern expression pattern, right.
• 46:12So the red one is a potential target and the
• 46:15first column or the first sort of set by
• 46:17the line start is the normal tissue and the
• 46:19second column is the corresponding cancer.
• 46:22So you see that the blue goes
• 46:23down because it's lost in cancer,
• 46:24but then the red stays up.
• 46:26So we actually looked at why this is
• 46:27happening. It's maturation driven.
• 46:30And but what is a C?
• 46:31So C1 and C2 are actually the first literally
• 46:34the enzymes in fatty acid synthesis.
• 46:37They pre they are immediately
• 46:39before fast or fatty acid synthase.
• 46:42They convert acetyl coenzyme to
• 46:43Malaya coenzyme and this C1 is
• 46:46actually in the cytoplasm.
• 46:47C2 is the mitochondrial membrane
• 46:49also regulates fatty acid breakdown.
• 46:51So if you block ACC,
• 46:53you block fatty acid synthesis and
• 46:56accelerate fatty acid burning.
• 46:58So it turns out that actually this
• 47:00wasn't real skin of pharmaceutical
• 47:02companies for a long time because
• 47:04because as a target for Nash,
• 47:06which is non Alcoholics started
• 47:09hepatitis or fatty liver and it's also
• 47:12actually one of the major targets for
• 47:14herbicides that we use in agriculture.
• 47:16Turns out that Pfizer actually had a drug
• 47:18that worked amazingly well in people.
• 47:20They put it through several clinical trials
• 47:22and they established that it actually works,
• 47:25it blocks the novel fatty acid synthesis
• 47:27as you see on that curve that ports.
• 47:29The percent of the noble lipogenesis
• 47:31in people, it was also safe,
• 47:33except for one thing.
• 47:34It caused a little bit of
• 47:36hypertriglyceridemia and made and
• 47:37caused a drop in platelet counts.
• 47:39You know, we play the games on 400,000,
• 47:41so the politicians,
• 47:42not the 200,200 thousand is actually,
• 47:44it's a 50% drop.
• 47:45But we don't even count this as a
• 47:48toxicity in chemotherapy because
• 47:49it's a very safe level.
• 47:50Nevertheless,
• 47:51Pfizer felt that that this
• 47:53warrants discontinuing the drug.
• 47:55So we reached out to them and we actually
• 47:57got the right to test this drug.
• 47:59In.
• 47:59In preclinical models and hope
• 48:01to bring it back to the clinic if
• 48:04these little promising,
• 48:05but did the preclinical model look promising?
• 48:08So I don't really invitro data
• 48:11because the invitro,
• 48:12you know metabolism is highly sort of.
• 48:14Dependent on how much fatty acid
• 48:16and one that you have in the media.
• 48:17So this is the in vivo data in mice.
• 48:20So this is PBX to macros that we
• 48:22contracted out for Jackson lab and
• 48:24you see that this ACC inhibitor
• 48:26actually inhibits the growth although
• 48:28doesn't strike completely.
• 48:29The MDA MB 468 Genographic did here
• 48:32at Yale shows the same thing but the
• 48:34most striking thing was synergy,
• 48:36the doxorubicin and Vina Robin and
• 48:38also with the collaborator is
• 48:40interested endocrine sensitive CVD
• 48:42and resistance to develop the food.
• 48:45Strand resistant MCF 7 cell line,
• 48:48she also showing you know xenograft
• 48:50model that there are actually
• 48:51inhibited the growth.
• 48:53So this looks pretty promising to
• 48:54us and we do some additional studies
• 48:57to really figure out more about the
• 49:00synergy between chemotherapy agents and
• 49:01we hope to get this back from Pfizer.
• 49:05But how does this work?
• 49:07So the most interesting thing was that
• 49:09when we looked at what transcriptional
• 49:11changes occur after exposure to this drug,
• 49:14what really was.
• 49:16Striking is the that there was a
• 49:20dramatic increase in genes that are.
• 49:22Mediating and involved in unfolded
• 49:25protein response and upregulate
• 49:27endoplasmic reticulum stress.
• 49:29So our working hypothesis thereby
• 49:32inhibiting the Novo fattiest synthesis,
• 49:34you actually alter the membrane
• 49:36composition of the endoplasmic reticulum.
• 49:38You know proteins have to find a threat
• 49:41through the membrane to get into the
• 49:44endoplasmic reticulum for secondary
• 49:46modifications and we think that by
• 49:48changing the endoplasmic reticulum lipid
• 49:50composition we change this process.
• 49:52Of of protein synthesis and in user
• 49:55unfolded protein response which
• 49:56eventually overwhelms the cell.
• 49:58So that's the project that we do in the lab.
• 50:00Look at the lipid membrane composition
• 50:02of of the endoplasmic reticulum as as
• 50:04far as we can and the lipid alterations
• 50:06in the cells exposed to this and also
• 50:09some some reporter systems to nail
• 50:11this as the mechanism of action.
• 50:15So I'm going to summarize this really.
• 50:18So for those of you who are clinical fellows,
• 50:20you know every clinical dilemma
• 50:21that we discussed in a tumor boards,
• 50:23it's a research question asking for a study,
• 50:25some movies disheartened then
• 50:26people come about saying that OK,
• 50:28what should I research?
• 50:29I mean what you should
• 50:30research is all around us.
• 50:32You just need to open your eye.
• 50:33And so recognizing the prognostic
• 50:35importance of Pathologic CR residual
• 50:37disease has left new treatment
• 50:39strategies and improved survival in
• 50:40triple negative disease and her two
• 50:42positive disease and I showed you how so.
• 50:45Molecular offices of these issues
• 50:46also gives some idea that how
• 50:48we could make it even better by
• 50:50studying the difference between
• 50:52the nonresponders and responders.
• 50:54So immunotherapy established its
• 50:55value in breast cancer and Robinson
• 50:57is now approved as as neoadjuvant
• 50:59therapy together with chemotherapy
• 51:01for all three primary disease.
• 51:03It's also approved as first line
• 51:05therapy for PD like 1 positive
• 51:07metastatic breast cancer.
• 51:08And I think we have a reasonably
• 51:10decent explanation why you need the PD
• 51:13ligand one in the metastatic disease.
• 51:14So we are about to launch studies
• 51:16to demonstrate that similar benefit
• 51:18could be seen in a subset of
• 51:20molecular defined subset,
• 51:21small subset of ER positive breast cancers.
• 51:24And we also have some promising
• 51:26markers that could actually make this
• 51:28whole strategy safer and more cost
• 51:29effective by tailoring the treatment
• 51:31to those who really needed it.
• 51:33But these you need validations and
• 51:35I think the most exciting sort of
• 51:37things on the horizon clinically is
• 51:39CDN surveillance and interventional
• 51:41homophone macular relapse that might
• 51:44ultimately reduce further metastatic
• 51:46recurrences and this understanding
• 51:48the molecular phylogeny of metastatic
• 51:51disease really prompted this idea
• 51:53that because the.
• 51:54Synchronous mats are very similar to
• 51:57the primary tumors might be they are
• 52:00responding to the same way and the
• 52:02micro mats that remain after eradicating
• 52:04those are also similar to the to them.
• 52:06So that the microbes that remain after
• 52:08the primary tumor is being resected
• 52:10that may be approaching the same these
• 52:12disease with the same strategy that
• 52:14we very successfully used in stage
• 52:16three disease might actually cure a
• 52:19small subset maybe 10% maybe 30% of of
• 52:22the Novo metastatic stage four disease.
• 52:25And.
• 52:25There's a really deep portfolio
• 52:28of new classes of drugs.
• 52:30And that's my last slide.
• 52:32I apologize ahead of time for people who
• 52:34actually didn't make it to the slide,
• 52:36but I ran out of space.
• 52:37But these are the various people
• 52:39who worked in my lab and contributed
• 52:40the work that I showed you and
• 52:42students and other collaborators
• 52:43and collaborators within Yale.
• 52:51So.
• 53:02Yeah, so. If you have any
• 53:05questions then feel free to.
• 53:07Ask yes, silly. I have.
• 53:14Saying that, we were going to.
• 53:18And you mentioned, right and when you
• 53:21talked about the model especially.
• 53:24Negative. I want to know if you will
• 53:28consider rate in that model and it's so.
• 53:33So actually Kim and and some other
• 53:35previous lab members did they really
• 53:38nice analysis trying to see whether
• 53:40there is a immune difference between
• 53:42triple negative breast cancer by race.
• 53:45The hypothesis was that that.
• 53:48Stress and this sort of this weathering
• 53:51that that unfortunately many people
• 53:53with African American or Hispanic race
• 53:55have to suffer would have an impact
• 53:57on your immune immune system, right.
• 53:59So the truth is that if there is
• 54:01such a thing, it's really subtle.
• 54:03We find some some really intriguing
• 54:05things around macrophages things,
• 54:07but whether this really holds up,
• 54:09I'm not quite sure yet.
• 54:10So I can send you the slides
• 54:12and we have some things,
• 54:13some references there and we we see
• 54:15some things but I'm not sure that it's.
• 54:18It's really detectable.
• 54:18There are other things that we haven't
• 54:20looked at but we plan to do which is
• 54:22like inflammatory markers in the blood.
• 54:24But that's also kind of
• 54:26biased by comorbidities.
• 54:27So if you have a lot of other diseases,
• 54:29then it's just going to be high anyway.
• 54:30And in terms of the models,
• 54:32you know,
• 54:33so Pathologic CI is equally good
• 54:35in terms of metastatic recurrence
• 54:38regardless of race.
• 54:39In fact,
• 54:40I personally have a really serious
• 54:42doubt that there is any major
• 54:43genetic sort of explanation
• 54:45behind disparities and outcome.
• 54:50So models that include in survival
• 54:53rates are problematic, right,
• 54:54because it perpetuated a risk
• 54:57factor that that maybe not true.
• 54:59So if your social,
• 55:00social circumstances change.
• 55:03Is there a question from online?
• 55:06I should call you back.
• 55:11So there's this question online that.
• 55:14Umm. Somebody's relevant regretting
• 55:17their choice that they're not breast
• 55:18oncologist and they agree with that.
• 55:20That's the do patients with inflammatory
• 55:22breast cancer have higher response rates
• 55:24to checkpoint inhibition and the agent
• 55:26setting regardless to applying results.
• 55:28Yeah, that's a good one.
• 55:29So you know inflammatory
• 55:30breast cancer is a misnomer.
• 55:32It's really, it's a clinical description
• 55:34that people came up and whatever
• 55:36maybe the 19th century and because
• 55:37the breast looks like inflamed,
• 55:39it's red and hot and and and swollen,
• 55:42it looks like a skin infection and
• 55:44very often primary care physicians.
• 55:46Give it antibiotics and it just gets worse.
• 55:48So inflammatory breast cancer
• 55:49actually is not particularly rich.
• 55:51In fact it's pretty poor in immune cells.
• 55:54But we did.
• 55:55Actually the first whole genome
• 55:57sequencing of inflammatory breast cancer,
• 55:59hoping to find something and
• 56:01disappointed we didn't find
• 56:03anything that actually defined this
• 56:06autonomically at the DNA sequence space,
• 56:08but we find some interesting things.
• 56:10Again, TGF beta macrophage
• 56:13related markers show up there.
• 56:16As potentially contributing
• 56:17to the poor outcome.
• 56:19But yeah,
• 56:20so inflammatory breast cancer
• 56:21is all the four subtypes and
• 56:23as far as we can tell today,
• 56:25there is really no proton
• 56:26nomical genomic alteration.
• 56:31So what type of preventive
• 56:33interventions do you foresee for
• 56:35patients with high cancer score.
• 56:36So if you already have validated and
• 56:40really effective prevention drugs,
• 56:42right, the moxen aromatase inhibitors and
• 56:45food and other drugs, the I type drugs,
• 56:48but they have side effects and and I
• 56:51think one way to use these cancer score
• 56:53would be to if you're high risk that
• 56:56you are close to this tipping point,
• 56:58I should say you that we
• 56:59don't have that score.
• 57:00It's working on it.
• 57:01But it's the idea that if you
• 57:02can tell that these biopsy,
• 57:03tissue biopsy shows that you are
• 57:05close to this tipping point and
• 57:07maybe you are willing to put
• 57:08up with some additional.
• 57:10Umm.
• 57:10Discomfort from a prevention drug.
• 57:18All right. Let's go ahead, Andrew.
• 57:20A lot of times with the
• 57:23people who have even PCR,
• 57:25they can relapse in the brain.
• 57:27And people sort of say that's
• 57:29due the blood brain barrier,
• 57:31but are there molecular alterations
• 57:34that predict frame labs or can you?
• 57:38No, I can't. But you know,
• 57:39I mean, that's the reason why I
• 57:40don't go to many of the meetings,
• 57:42because there are so many
• 57:43interesting things to study.
• 57:45I just enjoy them more but yeah so,
• 57:47so people tried that but they didn't find it.
• 57:49But what you bring up is illegal one right.
• 57:51So the pathologic CR is really good
• 57:53but it's not a perfect predictor and
• 57:55for for there are many reasons why
• 57:57there should be a disconnect with
• 58:00Pathologic CR improvement in survival.
• 58:01So you can't cure people twice.
• 58:03So if you enroll a lot of people
• 58:05that are on stage one breast
• 58:06cancer and the surgeon cure them,
• 58:07it doesn't really matter whether
• 58:09they are chemosensitive or not.
• 58:11But in terms of recurrences look to Silver
• 58:13Point out something that many oncologists.
• 58:15Even breast oncologists may
• 58:16not be totally familiar with.
• 58:18So there are a number of studies
• 58:20that show now that the first
• 58:22sight of recurrence of the PCR,
• 58:24half of the time it's the brain.
• 58:26When you have no PCR residual disease,
• 58:29then the brain is the first site
• 58:31in about 10% and it goes along
• 58:33with this idea that the brain
• 58:35is somehow a protected site.
• 58:36And the question is then how they
• 58:38actually can break this protection and
• 58:41really help avoid brain recurrences.
• 58:43There are some some really good
• 58:45initiatives in the in the her two
• 58:47positive space and some of the ADC
• 58:49may get in there triple 90 disease,
• 58:51but what actually would define
• 58:53high risk for brain recurrence
• 58:55in terms of molecular markers?
• 58:58But they could find that in a reproducible
• 59:00and accepted sort of widely accepted way.
• 59:06Thank you. Thank you for all
• 59:09of you who have joined both
• 59:12in person and virtually.
• 59:13This concludes our breast cancer
• 59:15awareness month grand rounds.
• 59:17Thank you so much.
• 59:39Yeah.