Immune-Based Cancer Treatment Biomarkers – The Pathologist’s Critical Role

April 01, 2022

Information

March 31, 2022

Yale Pathology Grand Rounds

Robert A. Anders, MD, PhD

ID7647

To CiteDCA Citation Guide

00:06OK, well it's 1232. Let's get started.
00:10Welcome everyone to pathology grand rounds.
00:12It is my distinct pleasure to introduce
00:16my good friend Bob Anders to the group.
00:20Today. Very excited about his lecture.
00:24Doctor Anders earned his medical and
00:28molecular biology graduate degrees at the
00:31Mayo Clinic Medical and Graduate School.
00:34He then completed his anatomic and clinical
00:37pathology residency and a fellowship in
00:40gastrointestinal and liver pathology.
00:42And postdoctoral training in immunology
00:45at the University of Chicago.
00:47With our good friend John Hart.
00:50He joined the faculty at Johns Hopkins
00:52School of Medicine in 2005 as an
00:55assistant professor in the Division of
00:58Gastrointestinal and Liver Pathology and
01:00is soon to be full professor in that lab.
01:04He also serves as the Co director of
01:07the tumor microenvironment laboratory
01:09in the Bloomberg Kimmel Institute
01:11for Immune Cancer Therapy.
01:13But in addition to his duties as a
01:15practicing GI liver surgical pathologist,
01:18a Bob is one of those rare breeds of
01:21diagnostic pathologists who runs an
01:23independently funded research lab
01:25that focuses on tumor immunology.
01:27His interest in immunology developed while
01:30he was studying liver immunology and
01:32regeneration at the University of Chicago.
01:35And in his early years at Johns Hopkins,
01:37his current interests are
01:39in tumor immunology,
01:40and specifically interrogating the
01:42immune microenvironment in tumor tissue.
01:44Obviously a very hot topic.
01:47He first began examining human tissue for
01:50the expression of our good friend PD1,
01:53and PDL one in 2006.
01:55At the encouragement of his K
01:578 mentor doctor Liping Chen,
01:59who's currently at Yale.
02:01And he continues this research on
02:03human in inhuman and marine cancer.
02:05I mean analogy using single and
02:08multi color immunostaining couple
02:10to digital image analysis.
02:12In addition to all these accompanying
02:14this area of interest,
02:16his CV boast hundreds of publications
02:18and he is invited to something
02:21like 50 lectures a year,
02:23a giving courses and and interacting is very,
02:27very active and sought after collaborator.
02:30So really, the model of physician,
02:33scientist in pathology and I'm
02:35delighted to welcome you to his
02:38lecture entitled Immune based
02:41cancer treatment biomarkers.
02:42The pathologist's critical role.
02:45Bob,
02:45thank you.
02:46Yeah, thanks Marie,
02:48that's very generous.
02:50Introduction that would
02:51make my parents proud.
02:53Thanks. So this is a great.
02:57Time to be a pathologist,
02:59and that's sort of my my message here.
03:03So I receive funding and and
03:06consult for these in these.
03:13Institutions, but it won't affect what I say.
03:18So the plan here is to talk about
03:20biomarkers in cancer or talk
03:23about immune based biomarkers,
03:25and then you know kind of the.
03:26So the story of how DNA
03:29based immune biomarkers.
03:33Developed and then something
03:35that shouldn't be, you know,
03:38foreign to to most of you with the efforts
03:41of Doctor Shelper and and and Doctor Rim.
03:44And that's you know how do
03:46we find new biomarkers?
03:48So biomarkers it.
03:49Turns out from the beginning.
03:52There's confusion.
03:54So prognostic biomarkers that's
03:57about how patients will do,
04:01essentially untreated,
04:02or with definitive treatment like surgery.
04:06But most of what we talk about these days
04:09is is really about predictive biomarkers,
04:12and that's and that's personalized medicine.
04:15It's what it's what we've
04:17wanted for the longest time.
04:18It's how to select the the patients
04:21that will benefit from a given therapy.
04:25And this is great.
04:28This is great for pathologists
04:30because you know human human tissue
04:33is a prognostic biomarker, right?
04:38This was so clear even it was even
04:42clearer to a surgeon back in 1932,
04:47right with the Dukes classification.
04:49You know, if a if a colon cancer
04:53penetrated deeper into a colon.
04:56The patients didn't do as well.
04:59Fast forward Banta 2006,
05:02where Jerome Galand showed that.
05:06The tight density and location
05:09of immune cells were oops.
05:12He didn't mean predictive.
05:13He meant prognostic of of clinical outcome,
05:17and I will tell you that I presented
05:20this paper at our Journal Club.
05:22And and there were people who walked out
05:27of that room saying this is nonsense.
05:30And why were they so perplexed?
05:34Essentially,
05:34you had patients with stage one or
05:38two that's shown in the in the in
05:41the red and green regions of this
05:44Kaplan Meier curve who didn't do well?
05:51After the diagnosis,
05:52you know that that sort of didn't make sense.
05:55Well, it turns out that our immune
05:58system really is quite important,
06:00which will be the message for the
06:03rest of the of the lecture here.
06:06So you can go back as far as 2012
06:09and you know 120 articles and
06:11in in 20 different cancer types,
06:15showing that you know the type of
06:18immune cells that are infiltrating
06:20at patients tumor or the till.
06:22If you will, is really important
06:25and you know it's not surprising,
06:27it's great to have CD 8 cells that
06:30are secreting interferon gamma.
06:32We would call them TH one.
06:35In your in your cancer compared to
06:38you know things that are suppressive
06:42of of the immune based response.
06:45OK, so that's prognostic.
06:47What about human tissue as
06:49a predictive biomarker?
06:51You know,
06:52this is really talking about
06:54whether patients will respond
06:56or benefit from from therapy.
06:58And you know,
06:59we've been promising patients
07:00this for a long time.
07:02And you know,
07:03I breakdown the predictive biomarkers
07:05into in the three generations.
07:07First,
07:07it was important for oncologists
07:09to know whether something was a
07:12lymphoma or a colorectal cancer.
07:13I mean, this really changed.
07:15How they treated a patient.
07:18The second generation really
07:20came on and this is,
07:22you know work that Doctor Rim
07:24is is doing in really trying to
07:27get down to how much a cancer
07:32expresses a particular biomarker,
07:35not just any biomarker,
07:36but actually a biomarker that
07:38we can target and her two new
07:40would be a perfect example.
07:42And now we're into the third generation.
07:45You know, by my definitions of.
07:47You know how many immune cells are present?
07:50What are what's the immune contexture?
07:53What immune inhibitors are present?
07:56Which ones aren't present so,
07:59as pathologists are our goal actually,
08:02or are, you know,
08:04has gotten much more complicated?
08:06You know,
08:07telling some telling us something was a,
08:09you know an adenocarcinoma versus
08:11lymphoma is a lot different than saying,
08:14Oh well, there's a lot of
08:17tumor infiltrating lymphocytes.
08:19So immune therapy.
08:20What is it? Well,
08:21it turns out this guy knew a lot about it.
08:24This is Doctor Coley,
08:27and if you don't know much about him,
08:28I highly encourage you to Google him.
08:31You know he.
08:33He essentially was rubbing.
08:36Bacteria into the skin of patients
08:39with advanced cancers and showing
08:42that their cancers were regressing.
08:45Now he was a surgeon,
08:47so he was much maligned for his approach,
08:50'cause essentially he was
08:52giving chemotherapy,
08:53which went when you know conventional
08:56thought was that this should be
09:00more of a surgical approach and
09:02you know it was so hard for him.
09:06To get his message out,
09:08he had to travel all the way
09:10down from New York,
09:11which is where he was from down
09:13to Johns Hopkins in Baltimore,
09:15and this is a picture of him
09:18lecturing in Heard Hall.
09:19And this is exactly where pathology
09:22grand rounds is still conducted to the day,
09:26and I can identify this as as as.
09:30Heard Hall because we have these,
09:33you know, flags displayed in at the
09:36margins of of the lecture system.
09:39So what's immune therapy?
09:41You know, it's it's.
09:43It's basically the idea that a patient's
09:46immune system can eliminate a cancer.
09:49This is very patient empowering.
09:51I had the misfortune of spending a
09:54lot of time in a in a cancer immuno
09:57or in a cancer therapy setting and
10:01it was amazing to to to witness how
10:06patients responded to the fact that.
10:10They were mounting in.
10:13Immune response to their cancer.
10:15They just needed a little help removing.
10:20The break.
10:22So cancer immune therapy is is best,
10:26you know understood with you know
10:28anyone who drives a car, right?
10:32So we thought the problem with cancer
10:36therapy or tumor immuno tumor immunology
10:39was just that we we needed to press
10:43the accelerator harder but anyone who knows.
10:47Anything about you know driving a car
10:49if you have your foot on the brake,
10:51it doesn't matter really how hard
10:54you press the accelerator.
10:55It's all about releasing the brake.
11:00And that was recognized formally
11:03with the Nobel laureates.
11:06Here, Doctor,
11:08Allison and Honjo.
11:10Day.
11:10And you know,
11:12essentially both of these
11:15individuals were were studying
11:17immune based therapy in terms of uh,
11:22immune inhibitors,
11:24not immune stimulators,
11:26but immune inhibitor inhibitors.
11:28So what I've done here is,
11:31I've lined up the immune synapse between
11:34antigen presenting cells and and and T cells.
11:38And I think all of us recognize
11:41that we need a signal,
11:43one that's MHC engaging the T
11:46cell receptor and then signal 2
11:50here is what's responsible for,
11:53you know, driving the immune system.
11:56But then it's these breaks.
11:59It's the brace.
12:00It's the inhibitors of the
12:02immune system that went on to win
12:05the Nobel prizes here,
12:07and I would point out
12:08very important fact here.
12:12And I have the utmost respect
12:14for Doctor Handjo, who personally
12:15invited me to to give a talk.
12:20It it was, it was actually Doctor
12:23Liping Chen, one of your tumor,
12:26immunologists, who really figured
12:27out what PDL one did it in in.
12:30In fact it was misnamed, it was.
12:32It was thought to be a program,
12:34death ligands and it was Doctor
12:37Chen who figured out that that PDL
12:40one was actually a very critical
12:43immune inhibitor and ultimately the
12:46key to unlocking tumor immunology.
12:51So how do we reconcile CTL A4 versus PDL one?
12:58You know, for me,
13:00CTL A4 works in the periphery,
13:02it's an early part of the immune
13:05response where PDL one is a little
13:08bit later and occurs in the tumor.
13:10That's the key then to the
13:15pathologist because.
13:16Our ability to look in the
13:18tumor is very helpful. Why?
13:21Why are we even talking about this?
13:24I mean the the rate of.
13:27Of FDA approval for cancer
13:31therapies is unprecedented.
13:33In the last 10 to 15 years,
13:37the rate of cancer therapies approval
13:40with the FDA keeping in mind this
13:44is a very conservative group.
13:47Has just accelerated, you know,
13:50you know I kind of stopped updating my
13:53slide here about 2019 and it's 26 more.
13:56It's probably 50 more by now.
13:59You know this is really.
14:02Frighteningly fast approval in
14:07in terms of of tumor immunity.
14:11OK,
14:11so let's let's take a step back
14:14and you know the first person
14:17to look at A at a cancer.
14:21That had undergone PD1 blockade,
14:24and biopsy is me.
14:28You know, so how did we get here?
14:30So let's talk about PDL one
14:33as a as a biomarker.
14:36So we have this paper here which
14:39you know back in 2012. Explored the.
14:43This was really a safety paper.
14:48This was a phase one trial.
14:53About can patients even
14:56tolerate PD1 blockade?
14:59And.
15:01You can see that very important you
15:04know individuals to to Johns Hopkins as
15:07well as as to Yale or are listed here,
15:11because this is really a landmark paper.
15:15And I was fortunate enough to
15:17be in the right place at the
15:19right time and and what what?
15:21My laboratory at the time was really
15:23good at measuring was the expression
15:25of PD L1 protein in the tumor and
15:30the results of what we found are
15:33summarized in this graph here.
15:35And you can see that if patients
15:38were deemed to be PD L1 positive,
15:41we can argue about what that means.
15:44They had about a 5050 chance of response,
15:48but. And very important here.
15:51If patients were deemed to
15:53be PDL one negative.
15:58They had no chance of responding.
16:01And that's very helpful as an oncologist.
16:05If something has a 100%.
16:08Predictive power that means
16:11that it's as an oncologist.
16:14You move that patient on
16:16to something different,
16:17and this publication here is really the
16:22key to springing everything forward
16:25to PDL one as a predictive biomarker.
16:29Now it turns out that.
16:32That 100% predictive negative predictive
16:34value was was a little bit overblown,
16:38and it we can talk about the details of that,
16:42but it wasn't completely wrong either.
16:46So what I'm showing you here
16:47is one of our brilliant,
16:49now few faculty member members,
16:52Jill Sunshine,
16:53and what they did is they looked at
16:57all the different Dwan treatments
16:59there are listed on the top.
17:01All the different cancer types.
17:03Listed on the bottom and the
17:06response rate listed on the Y
17:08axis here and what you can see is.
17:11That patients who are PDL 1 positive
17:14again we can debate about what that
17:17means had about a 50% response rate
17:20and if they were PDL one negative.
17:25They had about a Tanner of 15 response rate.
17:28OK, so this is a little different
17:31than 100% negative predictive rate,
17:33but it it helps.
17:36So the most common question that I'm asked.
17:39Is is PDL 1A good biomarker?
17:43Well.
17:44It's not, it's it's not an end,
17:47it's it's not a perfect biomarker.
17:49It's imperfect.
17:50And what it does is it
17:53enriches for responders.
17:55So if you're an individual
17:57patient and you're deemed to be
18:00PD L1 positive or negative,
18:01you know Doctor Robert or myself.
18:04Look at your specimen in our
18:06expert weighs and and and call
18:08something positive or negative.
18:10You you probably have about a 50% fifty 50%.
18:14Or or or?
18:16I'm sorry,
18:16a 5050 chance of responding.
18:21But if you take 100 patients and you
18:26pick only the patients that are PDL,
18:291 positive again by an expert reviewer,
18:32you pick out a lot of winners.
18:36So is PDL 1A great biomarker?
18:40It depends if you're a physician trying to,
18:45you know, treat an individual patient.
18:47No, it's not. It's a 5050, you know.
18:50Sort of coin toss.
18:51But if you're a far if you're
18:53a large pharmaceutical company.
18:54And you wish to find the patients
18:57that are most likely to benefit.
18:59Then it's a fabulous biomarker,
19:02because you're a you're a
19:05affectively picking 50% winners.
19:10OK, so maybe the maybe the issue
19:13is we just need to do better.
19:16We need to do PDL 1 determination,
19:20immunohistochemistry, test performance.
19:23Better you know,
19:25maybe when it was first done in my lab.
19:27You know back almost you know
19:3110 plus years ago.
19:33You know, I hadn't considered everything,
19:35so the thing about PDL one expression is
19:38we're asking pathologists to do things
19:41that we've never asked them to do before.
19:44First off PDL one can be expressed
19:47on different types of cells.
19:49It could be in a histiocyte,
19:50a malignant cell or even a lymphocyte.
19:53It can even change its location.
19:55Sometimes it can be in the cytoplasm,
19:57other times it can be in the membrane.
19:59And then we're even further asking
20:02the pathologist to consider all these
20:05different cell types and say how many?
20:09How many cells are positive and this is
20:13really outside of what pathologists can do.
20:17And then on top of that we have
20:21different scoring systems so again
20:24I'm going to go over this quickly.
20:28Sometimes it's the number of
20:30tumor cells that are positive.
20:32Sometimes it's the number of tumor
20:34cells and inflammatory cells that
20:37are positive over a denominator.
20:39That's just the cancer cells.
20:41I mean, this is starting to get crazy.
20:43I mean,
20:45it's not practical for us to
20:49expect pathologists to be able to,
20:52you know.
20:54Find down this granularity and
20:56this was beautifully displayed
20:58in a study by Doctor Rim who,
21:00which I was fortunate enough to be part
21:04of and and basically what we showed
21:06or what doctor Rimm showed is that.
21:11If you ask us, the proportion,
21:14that's the percentage of tumor
21:17cells that are positive.
21:19Houses pathologists we can agree on this,
21:21but if you start making the equation
21:25more complicated and say well you should
21:28include immune cells but not plasma cells.
21:32The Inter rater of their agreement falls
21:36apart precipitously, so a .86% or a
21:41.86 here for inter rater variability.
21:45You know that's reasonably good,
21:48but something close to .2.
21:50That's absolutely terrible.
21:55Alright, so that's created
21:57an opportunity for us.
21:59In terms of pathologists
22:01because which PDL went antibody,
22:03there's actually many out there.
22:05What's staining pattern?
22:06Is it tumor cells or is it inflammatory
22:09cells and and then a cut off?
22:13Even changes between the different.
22:18The different stains and then how many
22:22pathologists do we need to agree?
22:23We've never even even considered this.
22:28So Doctor Robert and I have been
22:31fortunate enough to be to to spearhead
22:35an effort to look at ways to improve PDL.
22:391, staining, evaluation and what
22:41we've developed is an international
22:44collaboration looking at PDL 1 staining.
22:49And the the goal here is to
22:51measure how well we can agree.
22:54I've got a hint for you.
22:55We don't agree very well.
22:57And then more importantly,
22:59how can we do better and what digital
23:02systems may be able to help us?
23:04So that's something I wish I could
23:07share with you a little bit more,
23:09but we're in the process of of
23:11of putting together that data,
23:13but I promise you that it will be very
23:17revealing in that we really need to
23:20do better in order to help patients.
23:23OK, So what about the development
23:25of of of biomarkers?
23:27So it turns out that mismatch repair
23:31deficiency MMR proteins or MSI?
23:34You know,
23:35by PCR is a really important marker.
23:39A biomarker in terms of immune therapy,
23:44so I would like to invoke this
23:47individual Margaret Fuller.
23:48She was a transcendentalist female
23:50author who wrote this very important.
23:53Book.
23:54And what I wish to highlight
23:57out of this book is.
24:00Nature provides exceptions to every rule.
24:05And that's not how we generally
24:09think about cancer therapy.
24:11So what I want you is young,
24:14possibly physician scientists is.
24:17Study the exceptions and I'm
24:20gonna show you exactly why.
24:23So pharmaceutical companies
24:24have a very top down approach.
24:26They would love it if 100% of patients got
24:31a treatment and 10% of them responded.
24:35Because 100% of them need to be treated,
24:38but what I'm encouraging
24:39you to think about is not.
24:42That 90% that didn't respond,
24:46but actually thinking about
24:48the 1% or 10% that did respond.
24:53So the thinking is is very upside down here.
24:58As pathologists, what we want to think
25:01about is we apply a therapy and if there's
25:04one responder we need to drill down.
25:09And focus on that responder
25:11and figure out why.
25:13And I'm going to tell you exactly why here.
25:16So if I had a dollar for every time
25:18I heard this, well since immune
25:21based cancer therapy doesn't work
25:23for colon cancer so very early on.
25:26In my career.
25:30Immunotherapy didn't work for colon cancer.
25:35And and that goes all the way back.
25:37Look at look at the years on this here.
25:39This was 2006, 2009.
25:43In which we did a phase one again,
25:46this is safety. Study.
25:50And that was conducted here at Johns Hopkins.
25:55And you can see that the the
25:57cancers that are listed here.
25:58So Melanoma, lung cancer, renal cell cancer.
26:03You know what? These are?
26:04All immune based cancer therapies.
26:07They kind of tend to respond.
26:09In fact, one in four patients
26:11did respond to this therapy.
26:13Again, this was a safety trial.
26:17We are focusing on, you know,
26:18can patients tolerate this and suddenly in
26:22this 2006 study we're talking about response.
26:26This was crazy. And then.
26:29The interesting thing is,
26:31as I was hearing the chorus of well colon
26:34cancer doesn't respond to immune therapy.
26:37There were 14 patients who were
26:39who were enrolled on this study,
26:41and you know what?
26:44It wasn't hard to find them right.
26:46'cause colon cancer is among the most
26:48common cancers and the take home was yeah,
26:51you treated 14 patients with immune therapy.
26:54Nobody is responding.
26:57It clearly doesn't work.
27:00Well, not so fast.
27:02So it turns out that one patient did
27:07respond and that patients Histology
27:10ended up under my microscope and
27:13it looked exactly like this.
27:16And in fact, if any of you are are
27:19familiar with Johns Hopkins,
27:20it actually looked exactly like this.
27:23So we were at one of these multi
27:26headed scopes and Doctor Pardo,
27:28a very well known tumor,
27:30rheumatologist.
27:30Here I was.
27:31I was at this scope here or the head of
27:36the scope and doctor Janice Tower was at
27:39this scope and we looked underneath the
27:42microscope at the one of 14 patients.
27:47Who did respond to therapy
27:49and this is what we saw.
27:51This is the Histology,
27:54typical Histology of mismatch
27:56repair deficiency right all it's
27:59almost medullary in quality.
28:01There's a bunch of tumor infiltrating
28:04lymphocytes. And you know what?
28:07It's not surprising that
28:09this patient responded.
28:10To anti PD one therapy. Why?
28:13Because everything is in place
28:16and I'm I'll go through here and
28:19and and and and map this out.
28:23Mismatch repair deficiency.
28:24They have a lot of antigens because
28:26they have a lot of mutations.
28:28How many mutations?
28:29A lot right?
28:31So here's a graph showing
28:34the the frequency of of.
28:37Of mutations and even in
28:40something like a mutation,
28:41associated cancer even MSI
28:44here is orders of magnitude
28:48greater in terms of antigens.
28:51You know the way I think about
28:52it is every one of these antigens
28:54is is is is kind of like a.
28:59Is kind of like a ticket to Lynn to
29:01win the lottery and these people.
29:03These people have thousands of tickets.
29:06They also have lots of lymphocytes.
29:08Lymphocytes.
29:09We've known that every GI pathologist
29:11known knows that we actually went in,
29:14and our laboratory and measured it.
29:16We measured it in the in,
29:19in the tumor cells at the interface
29:22between the tumor cells and the and the.
29:27Are immune cells and this was work
29:29of a fellow and and junior faculty
29:32member at the time, and we were
29:34able to count them and guess what?
29:37It's not a surprise people with
29:40MSI or mismatch repair deficiency.
29:42They have more tumor infiltrating
29:44lymphocytes. Well, what about PDL one?
29:47Well, we were in the process of measuring
29:50PD L1 in solid tumors like this.
29:54Gastric cancer here. And what? Doctor.
29:58Thompson and Basharat T here found was
30:03that there's a lot of expression of
30:06PD L1 at the tumor infiltrate edge.
30:11The interesting thing is it's actually
30:13not on the tumor cells which are up here.
30:16It's actually at these immune cells
30:18and then we went and quantified this
30:20with Doctor Cruz lossa and by G here.
30:26Everybody that I've shown by the way,
30:28just as a as a as an aside here everybody
30:30that I've shown who's who's rotated
30:32through the lab or worked in the lab.
30:34They're all faculty members.
30:36So if you know if you're not even interested
30:39in tumor immunology, but you want to,
30:41you want to become a faculty member.
30:42Just study tumor immunology.
30:44The odds are in your favor,
30:46so in the end, you know it's obvious.
30:50Mismatch repair.
30:51Deficient cancers have a lot of antigens.
30:54They have a lot of lymphocytes and
30:56they have a lot of PDL one expression.
30:59So I'm going to show you this.
31:01Particular publication here looking
31:04at PD1 blockade in the setting
31:07of mismatch repair deficiency.
31:09There's 10 pathologists on this manual.
31:15And somebody please mute who's
31:17ever talking neuter phone.
31:20Somebody is somebody talking. So
31:23this study was presented in 2015
31:27and one what they collected
31:29were mismatch repair. Deficient cancers.
31:35About 25 of each. They treated them
31:38with PD one and it's no surprise that.
31:45They responded well.
31:46They were responded super well and
31:49the thing that really catapulted
31:52this particular finding was that.
31:57The response rate here 62
31:59versus 60 or 92 versus 70.
32:02It didn't matter if they were colon cancers.
32:05What mattered?
32:08Was that they were mismatched
32:09for pair deficient,
32:11so suddenly now we move forward,
32:14it's not mismatch repair,
32:15deficient colon cancer,
32:17it's mismatch repair, deficient,
32:18any cancer, and that quickly catapulted.
32:22Monk ologists at at at at Johns Hopkins.
32:26Doctor Lee and Doctor Diaz to put
32:30together a study in which the
32:33enrollment criteria was mismatch
32:36repair deficiency or MSI instability.
32:39However you wish to call
32:41it in any tumor type.
32:43This is crazy.
32:44We never thought that this would ever, ever.
32:49B. Away we enrolled criteria patients right?
32:54If you have a colon cancer,
32:56you get one treatment.
32:57If you have a pancreatic cancer,
32:58you get a different treatment.
33:01This was based on molecular diagnosis
33:05and you know what it goes even crazier.
33:08Because the FDA approved it.
33:13Now I have sat with the FDA
33:16for prolonged periods of time.
33:19They are not very exciting people.
33:24And the fact that they approved a
33:27therapy independent of Histology
33:29First off has never been done before,
33:32so that's worth mentioning.
33:34But the fact that they approved a
33:37treatment independent of Histology,
33:40and based on a non FDA approved test.
33:45Are you kidding me?
33:47But yet it's the standard of care today.
33:51And it all came back.
33:54To the fact that we were. Looking for.
34:00Patients who did respond.
34:03In a setting of patients who.
34:06Largely didn't respond.
34:11OK, so the thing that happens to me
34:14most and it is probably happening
34:17to doctor Shelper and Doctor Rim,
34:20is let's find more biomarkers.
34:24You guys are the pathologist tissue is
34:27important. Let's find more biomarkers.
34:30So how do we find more?
34:33File markers. So.
34:37Anyone who walks in my office.
34:40The first thing that they hear is
34:42there's no substitute for quality,
34:44and that's true in terms of cancer biology,
34:48and it's really a science of corollary.
34:53And you know, for me,
34:55how do you cure cancer?
34:56You make high accurate or
34:59highly accurate measurements.
35:04And the emphasis here on
35:07accurate and measurements.
35:09Well then it brings up the
35:11question how many measurements?
35:13Do we need to make hundreds of measurements?
35:16Four measurements,
35:17you know we have that issue in front of us.
35:20What are we going to do?
35:21And that was the subject of a
35:24beautiful paper done by one of our
35:26graduate students here Steve Lu.
35:28And what I'm showing you is
35:31the receiver operator curve for
35:33a number of different tests.
35:36And for those of you who might not
35:38be familiar with a narrow C curve,
35:41the closer you get to this one.
35:44The better so the green line
35:47is better than the dashed line.
35:50To put it simply. OK,
35:53So what do these different lines represent?
35:56And so this is the, UM.
36:00This is a meta analysis in which
36:03Steve Lew collected 55 studies
36:07from 10 different tumor types,
36:09encompassing over 8000 patients.
36:11So this is about as comprehensive
36:15as we could get in in 2019,
36:17and what was determined.
36:22Four response was either PDL.
36:26One expression by immunohistochemistry,
36:29that's the blue line.
36:32Total mutational burden.
36:35So that's you know the number
36:37of mutations in a genome.
36:39That's the red line. And then.
36:43The expression of RNA markers.
36:46Generally these are inflammatory markers.
36:48That's the yellow wine.
36:50And then in the green line.
36:53Which is showing the best performance here?
36:56All we needed. Was a minimum of two.
37:02Emphasis on 2 markers.
37:06And you can see the significant shift in
37:10the arosi curve here with as few as two.
37:16Immunohistochemistry markers.
37:21So based on that,
37:23on those types of findings and in
37:27a lot of the work that we've done,
37:29Doctor Janice Taube and I were
37:32were charged with setting up the
37:35tumor and microenvironment core.
37:37We have a very experienced lab manager here.
37:44It was set up through the Bloomberg Kimmel
37:47Institute for Immune based Cancer therapy.
37:52Essentially it's a Cancer Center resource.
37:54Here the idea is that we would help
37:57you measure tumor markers in the
38:00tumor microenvironment and our current
38:02holdings are are explained here.
38:05We we offer hundreds of stains,
38:08some of which we can do in two color and
38:11some of which we can do in in multiple color.
38:14We have a lot of machines in this
38:19particular laboratory that we run.
38:22We have 4. Our scanning machines.
38:26We have two auto stainers.
38:28We have an optical scanner and
38:31a laser microdissection scope,
38:33so that's to say you know we're
38:36we're we're equipped as anyone
38:38would expect us to be in order to,
38:41you know, perform investigations
38:43into the tumor microenvironment,
38:46and you can see this is what
38:48we're able to produce.
38:49And again,
38:50Doctor Shelper and and RIM have really
38:53been leading the way in terms of.
38:55Multi color immunofluorescence,
38:57which is what we see here now.
39:01The interesting thing is.
39:03That when we broke down the the steps
39:09of evaluating us a slide that's been
39:14staying with immunofluorescence.
39:17Janice and I found out there's
39:19a lot of errors.
39:21And those errors make evaluation
39:24very very problematic.
39:27So I'm summarizing the steps of of
39:31staining something with multiple markers.
39:34And and summarizing the errors here.
39:39Such that the image acquisition off
39:43of the microscope it was it turned
39:46out that the spherical you know sort
39:49of band or aperture of the objectives
39:53was contributing almost to a two fold
39:56error in in in image acquisition.
39:58Well,
39:58it's really hard to be sure what you're
40:02measuring when there's a two fold error.
40:05And again we systematically.
40:08Went through and and looked at these
40:11different errors the the method of
40:15immunofluorescence is is indicated here.
40:18It's a. It's a TSA method, it's actually a.
40:24It's a covalent bond.
40:27Such that.
40:28When the two antibodies are present,
40:31so this could be a CD3 antibody
40:34and this would be an antibody
40:37to detect this antibody.
40:39In the presence of horseradish peroxidase.
40:44This tyramide becomes inactive and
40:48becomes active and covalently links
40:52with the proteins that are nearby.
40:55This is what allows us to do reciprocal.
41:03Antibody staining because these
41:05antibodies are then rinsed off while
41:09their covalent binding is in place.
41:14Uh, it's not easy to to acquire these images.
41:19What I'm showing you here is
41:22a typical lung cancer case.
41:24Each square is a high power field.
41:28Each each cancer then would be
41:31represented by over a 1000 fields
41:35which which is 75 gigabytes of data.
41:40That's a lot of data.
41:42For one patient.
41:45So Janice and I were,
41:46as pathologists were quickly overwhelmed
41:50and we were so fortunate by Doctor
41:54Jaffe to be pointed in the direction.
41:59Of Doctor Alex Szalay who helped us
42:03develop essentially the Astro pathology
42:06platform and the idea is that Doctor Salay.
42:09Here he is an expert in establishing
42:14the database that's responsible
42:17for imaging the entire sky.
42:20Not your pretty daunting idea here and
42:23then what Doctor Salay quickly realized
42:25is that what Janice and I were trying to do.
42:29On the microscopic level,
42:31was really exactly what he had done at
42:34the macroscopic level in establishing
42:37the the Sloan Digital Sky server,
42:40and if you've never done this before,
42:42it's super cool.
42:43You don't have to be in
42:45astronomer or anything,
42:47you can just log in as a regular
42:49you know interested scientist,
42:52person and and look at the sky.
42:56And in fact,
42:57there have been very high level publications.
43:00Nature level publications that have
43:03come out of. You know, sort of.
43:07Public scientists going into this database
43:11that doctor Slate helped establish
43:15and and and finding things in a straw.
43:20In the night sky that we
43:22didn't even know existed.
43:23So probably after the telescope.
43:28This is the most important piece of.
43:33Technology that we've had to look into
43:36the skies and we again had just been so
43:40so fortunate that this individual doctor
43:43Salay was interested in our problems.
43:47And again, I don't think it's
43:49too complicated to think of.
43:52Microscopes and telescopes are just
43:54sort of the same problem, inverted.
44:00So that led to this publication.
44:04Where we described the method.
44:07Now let let let's just step
44:08back for a second here.
44:10This is essentially a method describing.
44:16How to interrogate in multiple colors?
44:19Oh it was only seven or eight
44:21colors in into human tissue and
44:23it ended up with a high impact.
44:26You know publication, I think.
44:30That's really a testament to the
44:33fact of how difficult this is.
44:35What did we do in this paper?
44:38We looked at metastatic melanomas
44:41and found predictive biomarkers.
44:43The discovery cohort came out of Johns
44:46Hopkins in the validation cohort was
44:48so graciously provided by Doctor Rim.
44:52What? What was the what was the magic?
44:57It was, it was cells that you and
45:00I as tumor immunologists or even
45:03as pathologists would predict.
45:05It's you know PDL, one expression
45:08Fox P3 expression CD 8 expression.
45:12You know things that we know are important.
45:14These are the the the predictive biomarkers.
45:19Alright, so how can we improve upon multi
45:23marker detection and cancer tissues?
45:27So we're exploring adding more markers in,
45:30but I don't actually think that is the key.
45:33Using a codex of.
45:37Method there's also RNA based methods.
45:43Let's just get back down to you know,
45:45I've been doing this for over a decade now.
45:48And and here's my thoughts on on really
45:50how to find predictive biomarkers.
45:52First off, there's no substitute for quality.
45:56If you don't know what you're
45:57measuring or there's too much,
45:58you know error and what you're measuring,
46:00you're just looking at noise.
46:02It's easy to generate noisy signals,
46:05but it's difficult to, you know,
46:07sort through that noise.
46:10Don't overthink the analysis.
46:13We need T cells and we need B cells
46:15to make immune based, you know.
46:19Cancer therapies that's sort of
46:21what they're what they're aimed at.
46:25I think we should focus on on spatial
46:29relationships and that's described or or
46:31or demonstrated here in the bottom figure.
46:35So what I'm showing you here is
46:37a tumor cell that's in white and
46:39the T or B cells. I don't know.
46:41Pick your favorite are are the yellow
46:44dots here and something tells me.
46:47That the image on the right.
46:52With T or B cells surrounding the tumor cell.
46:57He's probably more indicative of a tumor
47:01immune response than the one over here yet.
47:05By simple density,
47:06we would report out the same numbers here.
47:09So I think the spatial
47:11relationships which interestingly
47:15we're not part of this.
47:17Publication this was simply a density.
47:21Publication so I think spatial
47:23relationships are very important.
47:26The other thing is have a validation plan it.
47:28There's no sense in in in measuring
47:31a bunch of markers that you can't
47:35secondarily validate like you know
47:38that that's called doing science
47:40without having a hypothesis.
47:45And I think I think finding low frequency
47:48markers at this stage is not the key.
47:50Uhm, there's not some super
47:53secret cell that's hiding deep
47:55deep in these tumor tissues.
47:58Like it's CD three,
47:59it's CD 8 and it's the relationship of
48:02those CD3 and CD8 cells to the tumor cells.
48:05So I'll I'll wrap up here
48:09with some conclusions.
48:10You find cancer biomarkers by looking
48:12in cancer tissue that should be great.
48:15That should be great news to all of us.
48:17As pathologists,
48:18we can improve upon the evaluation
48:21of PDL one as a biomarker,
48:24and there are a lot of
48:26efforts out there to do that.
48:29Look for rare responders.
48:31They those are individuals that
48:34are really trying to tell you
48:37something important and you know,
48:39in our case here they were MSI.
48:43Trust me, nobody wanted to fund this.
48:46Nobody was interested in this.
48:49Nobody, despite the fact that
48:51we were standing there saying
48:53these are MSI associated cancers.
48:56So you really need to
48:57make your argument here.
48:59Focus on the people who do respond.
49:02And then I think high quality analysis
49:05are important and you're all very
49:07lucky to have individuals again,
49:09like Doctor Rim and Shelper who are
49:12interested in making high quality.
49:14You know tissue biomarkers.
49:16I'll end here.
49:18It's impossible to do this all in isolation.
49:22I tried to point out the
49:26key people along the way.
49:28And and those in and and many
49:30more of the individuals who work
49:32in my laboratory or who have
49:34collaborated with me over the
49:36years are are listed here.
49:38And I'm gonna stop right there.
49:39And thank goodness because I'm exhausted.
49:46Thank you Bob for that really,
49:50really elegant and and sort of.
49:53Focusing on very high points and meticulous
49:57points of quality that we might not
49:59always here in talks about Peter one
50:02another biomarkers and I'm I'm, you know,
50:04there's a silent applause going on,
50:06which is so missed in zoom,
50:08but it's definitely happening.
50:09We have a few minutes for questions
50:12and please, since I can't see,
50:14I don't know if you wanna stop your share.
50:16Or maybe you want to keep it on 'cause
50:18people might want to go back to slides so,
50:20but please just speak out
50:22with your questions.
50:24And I'll start with one just
50:25to get the ball rolling.
50:26I actually have two things
50:28I wanted to discuss,
50:29but one is I'm just gonna start with.
50:31I was really intrigued by and
50:34thank you for talking about.
50:37You had a slide on errors in Multiplex
50:41immunofluorescence and you showed
50:43multiple steps along the way where
50:45you found errors and I'm curious.
50:47I wonder if you can say just
50:48a minute about that.
50:49How did you detect the errors and
50:52did are you able to fix the errors?
50:54And is this something that
50:56everyone who's publishing in this
50:57or are they as meticulous?
50:59You know they doing this careful
51:01detection that you're doing?
51:02Can you just talk a little
51:03bit more about this?
51:04Yeah, sure great great question so.
51:07Some you know, Janice and I are
51:09ultimately were pathologists, and,
51:11you know, for us what we would do is
51:14everything had to be referenced back
51:16to a a brown state or an IHC state.
51:20So anything that we were doing,
51:23whether it's digital image
51:25analysis or image acquisition,
51:27they were always put against
51:30the gold standard and the gold
51:33standard was immunohistochemistry.
51:35So that's kind of how we were
51:37able to go back and say, oh look.
51:39The the the images are not being
51:43acquired in a in a flat manner, or.
51:46There is over acquisition of cells in it,
51:51you know, and any reasonable observer
51:53learned weren't deemed to be positive,
51:56you know? In the end, you know.
52:01That's a science paper,
52:03and it's a methods paper which
52:05is not super common. Right?
52:07Like when we when we were putting this
52:10together, a lot of people said, oh,
52:12you know that that's just a methods paper.
52:15But but I think.
52:17I think it's recognized how difficult it is.
52:20It took us five years to do it.
52:23To to kind of sort out the
52:25errors at each step.
52:29And and then it is. This are those
52:32errors likely to be possible in
52:36in anyone doing Multiplex IF.
52:39If you use our,
52:40if you use the same technologies
52:43which is akoya Biosystems, yeah.
52:46That you know they're the good.
52:48The good thing is,
52:49once we figure out the errors,
52:51we can send you the code that will fix them.
52:55You know it's you know if a
52:58if a if an objective is not.
53:01Perfectly round or perfectly
53:04flat in acquiring there,
53:06there are easy easy.
53:10As a non computer person there are
53:12easier ways to to fix that, so yeah,
53:15well as you say we have experts here
53:17and some of them may be on the the
53:20grand rounds so I'm sure they can.
53:22They can speak to this as well,
53:23'cause they're excellent pay.
53:25I see you have your hand up.
53:27Go ahead and then Dave who's done
53:31this? Can you hear me?
53:33Yes, OK thank you.
53:34Thanks for the excellent talk.
53:36I enjoy a lot so I have
53:37a question about the MSI.
53:40NMR question about the
53:42testing at your site. Do
53:45you run both
53:46tests in molecular or
53:48immunohistochemistry? Or you
53:49select just one?
53:51Yeah, great question.
53:53It's probably after is PDL 1A.
53:56Good biomarker,
53:57it's which MSI should I or Mr?
54:00Should I be testing so you know,
54:01Doctor Vogelstein,
54:02you know one of our most
54:05eminent cancer biologists here.
54:07You know he was really instrumental and.
54:10You know, in in in figuring out
54:12the MSI along with Stan Hamilton,
54:14another GI pathologist,
54:16I see Dave Pat Jane is on and
54:19a few other GI pathologists.
54:24We're very much a PCR based system.
54:30I do believe that if a
54:33patient is Ms is PCR negative,
54:36it's probably reasonable to go back and
54:40and do a stain for MMR just because.
54:43You're almost talking about a
54:46cure in 50 to 70% of patients.
54:49And and I don't think we can.
54:53I don't think we can miss.
54:55Potential cure.
54:56I do believe that the performance of of
55:00PCR and MMR testing are about the same,
55:05but they both have their their weaknesses.
55:08Thanks for that.
55:10Our notion so actually
55:13hear different services
55:15actually approach differently.
55:18GI service using them are
55:20sold at giant service.
55:22Yeah, exactly what you just said.
55:24I think our geologist they
55:27don't want to miss a single patient
55:28so we know very well you know
55:31the some of the MSI high tumors
55:33may turn out to be a memory.
55:35You know, you know if you you
55:36know can be normal. Yeah,
55:38and it seems like it happens the same way.
55:40That's why I think now in
55:42the answer is here we do.
55:44Simultaneous testing for both methods.
55:47So with the sense to
55:48capture. All possible patients?
55:50Yeah, especially, especially since you know
55:52it has to do with the turnover of the cells.
55:56You know baseline and you know
55:59you wanna get it. Oncologist mad.
56:01Tell him that you know MSI PCR or MMR.
56:06You know HC hasn't really been.
56:09I mean none of its FDA approved number one,
56:11but it's never really been vetted
56:13on all of these different tissues
56:16yet we're doing it all the time.
56:18You know that'll make their brain melt, so
56:20I just wanna go to David you,
56:23you did a show us yourself did you one?
56:25Then I see curtain Uma Dave please
56:27go ahead. If you had a comment
56:29I just wanted to follow
56:31up with your concerns about.
56:33Standardization of quantitative
56:35fluorescence and bobs showing of
56:38the problems with each one and
56:39and comment about the MITRE study
56:41which is actually being led by
56:42Janice Bob's colleague at Hopkins,
56:44which is a standardization study
56:46of quantitative fluorescence,
56:47and I think that will address that
56:49question that you asked Marie exactly,
56:51but as the minor one study was where
56:55the data analysis was all done
56:56at one site and the
56:57scanning was done at 6 sites might
56:59or two study will have the scanning
57:01and analysis done at 6 sites.
57:03And that will answer your question,
57:05and Bob will probably participate in
57:06that since it's being led by Janice.
57:09Well, I have complete faith in
57:11our unit here, that's for sure.
57:13Thank you. I see Kurt and and then Uma.
57:18OK, then can you hear
57:20me? Yes, Yep we can hear it.
57:22So when one thing we
57:23touch on before Bob and I think
57:26it's becoming more and more relevant
57:27is the rapid shrinkage in and
57:30sometimes the disappearance of
57:32tissue samples in the eminent role
57:35of circulating toward DNA to call.
57:37You know, Microsoft instability status,
57:40TMB and other things?
57:41How do you foresee some of these
57:45immune biology or immune biomarkers?
57:46Adapting to circulating tumor DNA?
57:48Do you think that something?
57:50That we'll pair will and and how to,
57:52you know, marry with our usual, you know,
57:55diagnostic operation that is tissue based.
57:58Yeah, that that's a great question
58:00because you know not not every
58:02patient is biopsy able right?
58:04I mean, we we we kind of wish they were,
58:07but it's it's not. And you know,
58:09I you know I would like to see.
58:11I would like to see a study of just
58:13immune cell activation, right?
58:15Like like someone who gets CMV if you can't
58:19measure CMV like can you just tell that
58:22their immune system is activating right?
58:25And and I kind of think that the
58:27same might be true. Then you know,
58:30for for tumor immunologists it's just.
58:33Hey we gave PDL one is the
58:36patient's immune system activating?
58:37You know, I don't.
58:38I don't know what that entails,
58:40but you know is there a way to
58:43to not relatively non invasively
58:45you know with a blood sample?
58:51Measure that because again,
58:53not every patient.
58:54You know we're pathologists,
58:56we love tissue.
58:57I just told you how awesome tissue is,
58:59but it's not always available,
59:02so I I'm not sure if I
59:03answered your question.
59:04Are you currently testing MSI
59:06status by city DNA at Hopkins now?
59:10Even I haven't seen that outside
59:13of research. OK, thank you.
59:15Uma yeah thank you Doctor Andrews.
59:18That is a wonderful talk.
59:19My question was on heterogeneity
59:21in a study I did with 100 plus you
59:26know cervical and anal cancers.
59:2895% were heterogeneous in their
59:30staining with Foo side dead
59:32negative tofu side with 100%.
59:34So for an individual patient,
59:36is it going to be a luck of the
59:38draw when you have cut off someone
59:40and 10 how much should they be?
59:42You know what will be the guideline?
59:44What are your thoughts on
59:45that? Yeah, yeah, that's that's really.
59:48That's really important, right?
59:50I kind of think it's like one of these
59:53things where if it's positive we learn
59:56something, but if it's negative.
59:58We didn't really learn anything and and
01:00:01and I think that you know that's where we
01:00:03need to help our immunologists or our.
01:00:06I'm sorry our oncologists to say.
01:00:09You know, a negative result
01:00:10shouldn't really dissuade you here,
01:00:12but we're a long way from that. And yeah,
01:00:15any any look at heterogeneity and I,
01:00:18I know Kurt and I talked about.
01:00:20You know, some actually mathematical
01:00:22models of of heterogeneity.
01:00:24I think it's all important.
01:00:27Thank you.
01:00:32Other questions.
01:00:34And I'll, I'll I'll ask my last one,
01:00:35which just pivots off of Kurt's
01:00:39comment about a circulating tumor.
01:00:41You looking for markers and
01:00:43circulating tumor cells.
01:00:44And so I was just going to ask the
01:00:47classic question, what's next?
01:00:49What is the future of this?
01:00:53Especially since relying
01:00:54on MMR is is is easy.
01:00:58Staying pretty easy,
01:00:58stain to read but PDL.
01:01:00One has challenges and and then there
01:01:02are just things deeper than this.
01:01:04That Kurt and Dave and you are working on
01:01:07that are deeper than than just PDL one.
01:01:10So how far are we from getting to a
01:01:13blood test talking about not being
01:01:16people? My talk about blood tests
01:01:19'cause I you know, I don't know that
01:01:21you know I was talking with Kurt a
01:01:23little bit earlier as like you know I.
01:01:25I would like to see some CD 8 cells and
01:01:28some CD 20 cells in a tumor. Right?
01:01:31Like are they close to the tumor?
01:01:33Are they not close to the tumor,
01:01:35or are they present?
01:01:36Are they not present like?
01:01:39I I think I think we have so many basic
01:01:41things that we could be measuring
01:01:43here and and we're all and and I'm
01:01:46not blaming anyone on the call here,
01:01:48but we're all like can you
01:01:50do 4000 markers right?
01:01:53Like and you know,
01:01:55I bet you we could spitball
01:01:57four or five important markers.
01:02:01So yeah, dhanpat, sorry.
01:02:04I think we are all your time.
01:02:06I was in the liver tumor board
01:02:07so I had to attend something.
01:02:09Great. Talk Bob. Nice seeing you.
01:02:12I have a question which is sort of
01:02:14follow-up question to what Doctor
01:02:15Who was asking and I want to clarify.
01:02:17So I understand that you guys do
01:02:19only the MSI as the starting test,
01:02:21but you don't do them are to begin with.
01:02:24Is that correct or you do both?
01:02:26We
01:02:27we tend to heavily rely on MSI.
01:02:31I see the issue I think is I think
01:02:35what Doctor Who was asking was,
01:02:37you know whether do both
01:02:38or do you only one person?
01:02:40As I understand, most of those
01:02:42centers do either one of the tests
01:02:44and what you said is correct that
01:02:46none of the tests will ever be 100%.
01:02:48Like many other things that we do
01:02:50and you always accept the yeah Greer
01:02:53false positive or false negative you
01:02:55get with each test is like how many
01:02:57blocks will you stay in for PDL one
01:03:00or a given case you end somewhere.
01:03:02So, and I think a lot of people
01:03:05tend to go with DNA mismatch repair,
01:03:08yet see for two reasons that the
01:03:11sensitivity is basically almost
01:03:13as close to MSI.
01:03:14But in addition also identifies the germ
01:03:18line defect that one should be looking for.
01:03:20If you're screening for is syndrome,
01:03:22so there is added advantage.
01:03:24But I I agree.
01:03:25I mean there are pitfalls of each methods
01:03:27and you will miss a few no matter what.
01:03:30Yeah, yes, we're we're on the same page.
01:03:34You know, I you know,
01:03:35sometimes with a limited biopsy,
01:03:38you don't really know if
01:03:39you're getting that you know.
01:03:40Are you really getting the cancer DNA,
01:03:43but with an IHC you can, you know?
01:03:45Clearly see whether it's
01:03:47the cancer or not. So yeah.
01:03:51Oh, I'm sorry. Joanna, and then
01:03:54I think Jeff might. I don't know,
01:03:56or I I actually.
01:03:57I like this discussion that
01:03:59people have been having about MRI,
01:04:01HC versus MSI PCR.
01:04:03I think we don't really need to
01:04:05worry about which method is better,
01:04:08but I do think that we need to understand
01:04:10what information each method provides
01:04:12and why each method was developed.
01:04:14And also I really liked your
01:04:16comment towards the end of your
01:04:17talk where you said we're doing
01:04:19MSI testing on all of these tumors.
01:04:21MRI HD on all of these tumors
01:04:22now because of treatment.
01:04:23Options and we really don't know.
01:04:25What you know? How?
01:04:28How this how these MRI,
01:04:30HC behave and tumors outside
01:04:32of endometrial and colorectal
01:04:33cancer to the same extent that we
01:04:36understand in those original pores.
01:04:37And I think you know,
01:04:39I think doing doing MSI PCR in
01:04:43addition like is not really a
01:04:44problem for those cases where it
01:04:46is really important for treatment.
01:04:48So I think we just need to really
01:04:49keep in mind why are we doing this
01:04:51task right now at this very moment?
01:04:53What is the purpose of it?
01:04:55Is it for lynching of screening?
01:04:56Is it? For. Ms for treatment?
01:04:59Or is it for both and then we really
01:05:00need to sort of be cognizant that
01:05:02what information we get out of
01:05:04it and how it's gonna be helpful.
01:05:05So I I,
01:05:06I really liked your summary and your
01:05:08and and the fact that you brought
01:05:09up my side. Yeah, I think we're,
01:05:11you know we're speaking the same,
01:05:13the same language and you know
01:05:15I'm I'm I'm on a few of these,
01:05:17you know like advisory boards and
01:05:19you start telling oncologists that
01:05:21none of these tests have been valid.
01:05:23I mean number one not nothing is FDA
01:05:25approved but they haven't been validated
01:05:27in other cancers and they just.
01:05:30Their their mind is. Dissolves.
01:05:34Just one, there's a limitation.
01:05:37I have a question.
01:05:38I want to go back to leave in the middle
01:05:40so I might have missed some things,
01:05:41but I'm gonna go back to a point you
01:05:44made early on about the negative
01:05:46predictive value of absence of
01:05:48ligands or PDL one PD one in tumors.
01:05:53What is the thinking about
01:05:55those patients who respond?
01:05:58When they have no detectable antigen,
01:06:01no detectable PDL want.
01:06:02Is it that the test is
01:06:04insufficiently sensitive?
01:06:05That doesn't explain at all,
01:06:07I don't think.
01:06:08And is it a matter of these
01:06:10other factors that they had high
01:06:12mutation rates but nevertheless
01:06:13had low PDL one so that compensates
01:06:16for the absence of PDL one?
01:06:19What's understood about the
01:06:20biology of these responders?
01:06:21Who were PDL?
01:06:22One negative?
01:06:23Yeah, I mean I think number one,
01:06:25it's just you know it.
01:06:26It is a concern that it's
01:06:27a sampling issue, right?
01:06:29We have a tiny piece of a large
01:06:31tumor and the second thing would
01:06:33be you know if the TMB is high.
01:06:35There's not an oncologist who won't.
01:06:38Go for, you know,
01:06:39won't treat those types of patients.
01:06:44So I you know, I don't.
01:06:45I don't think we have detailed
01:06:48answers about I, I'm actually.
01:06:50I'm actually more interested in
01:06:52the people who are PD L1 positive
01:06:55who don't respond or or design
01:06:58positive who don't respond right.
01:07:00Everybody is interested in those,
01:07:01but I think that there might be value in
01:07:04the ones who are the atypical responders.
01:07:07Yeah, yeah, the drug has the drugs
01:07:10I assume are incredibly specific.
01:07:13It's not that they're often responses.
01:07:17Sorry, can I come in on that quickly?
01:07:18No, go ahead, Kurt, yes.
01:07:20So so there is a story published a
01:07:22couple of papers showing that PDL one
01:07:25can get glycosylated and when it gets
01:07:28glycosylated antibodies may not recognize it.
01:07:30So that's at least one direct
01:07:32biological ground by which patients
01:07:34may be false negative for PD L1.
01:07:37I could explain why some
01:07:39negative cases respond.
01:07:40That's one thing that's been sort
01:07:43of well established and and this is
01:07:45a work from Indy Anderson Group.
01:07:47And then the second layer is that
01:07:49tumor mutational burden and PDL one
01:07:52are not correlated in most tumors.
01:07:54So to your point it is likely that
01:07:55some of those period one negatives may
01:07:57actually be TNB high and that could
01:07:59be the driver of a clinical benefit.
01:08:01So it could be you know around those
01:08:04topics and in some unknown things for sure.
01:08:08But presumably response has to
01:08:10be through the PDL 1.
01:08:13PD One you know circuit because.
01:08:17That's what's with the drug works, right?
01:08:19That's the part I'm doing.
01:08:21There may be other things,
01:08:22yeah?
01:08:25OK thanks.
01:08:29Great, great discussion. Well,
01:08:31if there are no further questions,
01:08:34we're 10 minutes over which I always
01:08:36love because it means that this
01:08:38is really an interesting topic.
01:08:40I want to thank you Bob for spending
01:08:43time with us today and for meeting
01:08:45with several of us this morning.
01:08:47It's been a real pleasure
01:08:50and an education as well.
01:08:52Congratulations on your wonderful work.
01:08:55Thanks for the invite and all the
01:08:57time this. Yeah. With everyone,
01:09:00thank you. Bye bye OK bye.