 Thank you. It's a great pleasure to be here and have the opportunity to talk about adrenal cancer I've had an interest in adrenal cancer for at least 20 years and I've been banking every adrenal cancer That's come through the University of Michigan for for all those years So it's great to see all that effort of banking come to some some benefit So it's a really rare cancer only one or two cases per million per year So there's only about 500 cases in the United States a year the outcome is quite variable and of course like all cancers dependent on both the tumor grade and stage. What's interesting is it can be associated with various clinical endocrinopathy such as hypercortisol ism otherwise known as Cushing syndrome the therapeutic options are kind of limited and So having better ideas about how to approach this therapeutically is certainly one of our tasks in this project Fortunately, I'm at one of the places that have set up multidisciplinary clinics One of which is here at the NIH with Tito foho and others but Michigan and MD Anderson So we see a lot of patients come through Ann Arbor even though, you know, they're from New York or overseas The first thing to point out I'll spend a little time quickly on some background and These can be massive tumors. This one is a 20 centimeter tumor Usually there are about 8 centimeters and above and so they can get even larger than this So the first thing to notice is that there's an opportunity for heterogeneity So this is all necrosis and so all these clones are popping up and they can be quite different So the first thought is that the notion that you're gonna put a needle in this clone and and just understand everything about the tumor Is kind of naive when you're talking about a 20 centimeter tumor You know, this is a little controversial, but adrenal cancers just don't come from anywhere They come usually from in my opinion pre-existing benign tumors like this one, which is only less than two centimeters It is an adenoma So I believe in an adenoma to a carcinoma progression and the molecular data is starting to fill that in even though There's some people that are hesitant to believe this the histopathology in most cases is really not that difficult This is an adenoma the cells are very differentiated. They have tons of intercellular lipid and gives us clear clear appearance And cancers are much more cellular having necrosis nuclear pleomorphism and mitosis So the pathology is pretty well worked out but as you'll see there's there's some challenges in the pathology and Honestly like a lot of cancers resection is the only curative treatment so we we try to resect as many people as possible but In many cases it's it's not possible So stage pretty simple tumors less than five centimeters confined to the adrenal. These are quite rare Stage two more than five centimeters also confined to the adrenal and then locally invasive and distant Distant metastatic disease so a pretty simple staging and grading is actually not Architectural so a lot of times when you talk about tumor grade It's like does it form glands or are there squamous pearls in adrenal cancer? It doesn't really do any of that there are no glands it just grows in solid sheets, right? So solid sheets of tumor cells so we actually have developed a mitotic grading And we count mitosis less than 20 is low grade greater than 20 is high grade And so you can see that the stark difference between a low-grade ACC and a high-grade ACC and This'll this will come through in the genomic data and there's clearly a range of morphologies So an exceptionally low-grade tumor without mitotic activity or necrosis of more cellular tumor now with mitotic figures And then probably the most extreme manifestation an extremely high-grade tumor where you can see one two You know three probably four mitotic figures and several apoptotic bodies scattered about two so an extremely high-grade tumor So we have a full range of biology And then back to that intra-tumoral heterogeneity you can actually see it histologically So this is the same tumor a low-grade tumor with a low proliferation index Almost if they would all look like this I might be tempted to call it benign and adenoma But nonetheless it has areas that are more cellular with a higher proliferation rate So clearly we have a big sampling issue here for these tumors so what are the challenges occasionally they're still diagnostically difficult cases and Overall prognostic assessment low-grade high-grade is not the you know, it's not that granular could we do better? And then prediction and risk assessment both of who's going to occur locally who's going to spread Distantly who's going to respond to you know chemotherapy might attain some of the therapies that we have and then related to that What are some other therapeutic options that we could could bring to bear? So so those are the challenges that we sort of went into this TCGA project. Oops. That is not right. Oh Something out of order there. Okay, but and so what's nice is that you'll see oh, no, I'm sorry I am mistaken. So what's happening is there's this merging molecular classification some of the work from my lab and others from the NSAT group the European Network for the study of adrenal tumors In Europe obviously that we can separate adenomas from good prognosis and poor prognosis cancers And then some of the parameters that are driving this classification and you can start to see p53 and beta-continue mutations But what's nice about this field is that there's a really strong genotype phenotype correlation and so this Merging molecular classification is in pretty good agreement with this low-grade high-grade classification right so it's sort of it's even though it's a Big range of biology. It's kind of a more simpler cancer. So it's really well suited towards genomic studies Okay on to our project. So we had a global cohort. We needed a global cohort to get enough cases So we had North America South America Europe and Australia. We ended up with about 92 one had prior chemo I went out went out we did a super set for sequencing and then about up to about 80 cases past RNA we lost that one so we had a pretty good cohort for such a rare disease First first snippet of data Mutation density as Gatti will tell you I you know I pushed really hard on the thyroid project to to look at mutation density we discussed it many times But I think it's I think it's useful. So right smack in the middle of this of this plot and and it shows you can see a broad range of of mutation density and we tried actually to To make something of this so we looked at me 10 mutation density quite formally and you can see the high Intermediate and low and actually there's a fair number of things that correlate quite strongly with mutation density This is all stages and then we look just at at organ confined disease stage one and stage two because those are the ones where you'd Want to make a decision. Do I treat this patient? Do I watch this patient? And so even in even in organ confined disease you can see that there's a big range of mutation density From high to low and then again some of these Ex there's molecular correlates are shining through so there's just tells us there's an opportunity to use mutation density To do things clinically or just more broadly that you know the accumulation of mutations is clearly a bad thing in this cancer Onto significantly mutated genes usual characters beta-catenin p53 men1 in pretty much Expected frequencies I guess with men being more than we expected This is a shock This is the regulatory subunit of protein kinase a and this is this mutations of this gene are associated with Carni's complex and a rare Adrenal disease called PP NAD primary pigmented nodular adrenal cortical disease and it's really not cancer and Rarely do those turn into cancers But yet here we have Several mutations in this gene So that's it that's a not a shock But it's much higher frequency than expected and then RPL 22 which which is interferes Relates with p53 and that's novel for this for this disease So we spent a lot of time looking for gene fusions hoping that Gene fusions like in thyroid would explain some of the cases that don't have obvious mutations So we were very excited about this. I hired a summer student. We built tissue microase so we could do fish We used two methods with with rover-hake. We found lots of fusions a hundred and fifty six and 48 of the tumors And there's supporting data that they're real from copy number However, there were not not a single recurrent fusion So we decided you know doing fish to validate this was a big waste of time and we put the rays aside for another project So that was a little bit of a disappointment, but some of them are quite interesting In which they are known Cancer genes like mTOR and so it's possible that fusions play a minor role in this disease creating these private fusions We then turned to copy number and so this is data from Old cgh from circa 2000 and you can see there's lots of gains and losses in adrenal cancer So we were expecting to see something and sure enough we did we here's the focal changes Turk and CDK for turf to which we think may be unique within TCGA CC any one and on the deletion side we have RB and ZN ZN RF3, which is a regulator of The the wind pathway and thought to be an alternative way to to Inactivate the wind pathway So this was confirmatory because the NSAT group had already published that so so this was interesting with some some novel changes But so here's turf to and it just provides more evidence for the role of telomeres and in adrenal cancer Here's ZN ZN RF3 deletion. It's a negative regulator of wind signaling So the my hypothesis is that deletion represents an alternative way in addition to APC and beta-catenin mutations And so Gary Hammer our other co-chair in this project His lab is starting to work on this in a mechanistic fashion and it's present in about 20% of adrenal cancers So here's the the wind and you can start to see it starts to add up between APC and ZN RF3 and beta-catenin You can start to see it's almost about half of the cases that have some sort of wind pathway defect More copy number so copy number defines different classes of ACC We have a few cases that we called quiet that are Diploid and then you can see all hell goes you know everything Goes to hell and we have hypo diploid cases and hyper diploid cases with whole genome doubling noisy and a chromosomal group and it turns out that the noisy tumors in Both the width and without whole genome doubling are the most aggressive Okay, so that was kind of interesting We then put this in context of a pan cancer and we looked at our ploidy data Here relative to what was available and here's thyroid, which if you may remember has a very quiet genome and then Chromophobe kidney again with this hypo diploid so you can see the big range Employee present in this cancer and it's not really related to any type of purity because adrenal cancers as I showed Histologically are really quite pure. So so this is a really true biological observation We then turn to molecular classification And we have two large classes shown here and then form more granular Classes with profound differences both in survival shown here and in endocrinopathy And endocrine function so this is Recapitulate some of the work that my lab did through the director's challenge grant on adrenal cancers years ago Methylation we have three classes again with survival differences and other associations coming through and then we The NSAT data was available. We took the NSAT data showed that they're similar and Luda distilled a nice Luda and others distilled a nice 68 probe Methylation signature to show that we could get three classes and Validated it with the NSAT data. So we did this in response to reviewers asking. Well, you know, how are you going to? Really distill this into the clinic. So this is one possible method for clinical translation But we really like this result or I really like this result this cluster of cluster analysis with the different clusters here and the Associated data and this very nice result and remember I told you we do low-grade high-grade, you know Which is basically, you know, low-grade high-grade So if we could actually go from a two-class prognostication to a three-class would actually benefit You know what we do in the tumor and weekly tumor boards, right? So I'm excited about this and you can see some interesting associations with beta-catenin Most of the beta-cat mutations are in the more aggressive clusters So this is you know, I'll talk more about this is what I'm going to focus on as a pathologist Going forward and how do we translate this three-class into the clinic? So this is a really busy slide It's our our sort of supervised landscape view where we have the three cluster clusters all the associated mutations all the different clinical and other Genomic data and then some interesting things and I'll just point out a few things everyone can look at their favorite gene but we have a group here that are sort of You know free of mutations and so the question to me is Tom are you sure they're sure they're cancers and Because I told you sometimes we we had we struggle and I would argue that they are and the reason is is his here's IGF2 expression And it's well known that IGF2 expression is is only seen in cancers and goes up quite substantially And so all these tumors here that don't have drivers Clearly have high IGF2 expression and some of the ones that don't are actually in some of the more aggressive clusters So I think that's pretty strong evidence that in fact they are cancers They're just devoid of drivers and it also speaks to the fact that IGF2 is probably the main driver in these tumors Here's Ki67 and you can see that these are sort of the most low-grade tumors in terms of their proliferation Because Ki67 and mitotic counts correlate with each other So so we're sniffing out the earliest form of adrenal cancer, which I think is quite quite interesting And then you can see the accumulation of mutations as you go from cluster 1 to cluster 2 to cluster 3 And then a few other things in here notice though the fusions I was hoping the fusions might concentrate in there, but they don't and so you know you have great ideas But sometimes they work out so I think this is a great Figure I think it will be really well received by the adrenal cancer community We did a lot. I don't have much more time, but we did a lot. We looked at telomere length at the suggestion of Matthew Meyerson I always like to put things in the context of our Clinicians so we developed this adrenal differentiation score that sort of puts the genomic data in the context of adrenal function and you know And endocrinopathies we did hotnet. Sorry hotnet to file Oncocyan and then pan cancer genomic processes. I don't have time that I apologize I don't have time to show you all of these, but I'll show you that the telomere length analysis I think it worked out pretty well. This was done by C1 Zang and Rover Hake in which they they did sort of off-target Sequencing they took that data and don't ask me the details But they they took the whole like whole exome data and met and got telomere length data shown here And then they validated it with some glioma data where they had whole genomes And so there's some supporting evidence that this is believable And then on top of that you can see where the telomere length is increasing We have enrichment of ATRX and Dax mutations and also MLL And suggests that alternative lengthening of telomeres is in play for adrenal cancer in a subset of tumors So Matthew, thank you for the suggestion I think it worked out pretty well and we bumped this up to one of the main figures in the paper And then finally we because we knew that there were other Papers we really wanted to take advantage of the pan cancer data So we actually closed the paper with a pan cancer analysis in which we looked at all the different classes of Mutations that define oncogenic processes and we're able to put adrenal cancer in largely three of them Shown here and and this is more interesting where you can now plot it this this big C of this plot Where you have the mismatch repair signature a smoking signature from lung cancer And you can now see that adrenal cancer looks more like a scatter plot But nonetheless there are some cases that are in the mismatch repair signature We had published within a year or two ago that adrenal cancer is part of Lynch syndrome and that they rarely do have German line mutations of mismatch repair genes. So this confirms what we published and And more importantly though there there was some Suggestion in the literature that it that smoking may play a role And so here we now have some data that at least in some cases that smoking may play a Causative role so that's we put that in there. It's a little speculative We tried to get TCGA data the smoking history to sort of corroborate that and sort of quickly gave up because it just became too difficult So what have we what have we done in the last minute? We created I think an outstanding genomic resource for adrenal cancer research We do have new somatic alterations We expanded the role of of went the hotnet analysis really put pulled that together nicely We have this copy number and whole genome doubling story that is quite elegant and well done this copy No, this cluster of cluster analysis with three classes of tumors, which I really think will be advanced for the field We did the pan cancer analysis the papers Honestly was rejected from nature genetics, but we've repealed and they've they said you know actually made some good points So we're hopeful that we'll have a chance with nature genetics and we're getting ready to send it back probably next week So looking forward as I said I will be looking for ways to deliver this three-class solution to routine cases Because how it works now is very few people get operated on at Michigan But they all show up at our door with the paraffin blocks and I need to you know If I'm gonna do something if we're gonna be a center of excellence I need to do something and so how can we deliver this to those patients and then therapy I think it's clear though that IGF 2 is still there. We've run IGF 2 trials We only have a few responders in the eyes of big pharma They failed but nonetheless there are patients in there and what I think what I think is happening This is a little speculation allow me to do this those quiet cases I think are the ones that are responding to the IGF 2 inhibition and once you accumulate all those other mutations That signal sort of gets drowned out But I think the combination of IGF 2 and when I think holds the secret for treating high-grade adrenal cancer I understand how hard it is to develop good inhibitors of beta-catenin So I'm not hopeful this is gonna happen anytime soon But I think the paper is telling us what we have to do and then there's other pathways in smaller subsets This was a huge amount of work led my co-chair by rollover hake and his postdocs He won zang they did a phenomenal job and then of course that many other people that contribute to this We really did look at that right on time we really did gather the world's group of endocrinologists to work on this and My co-chair gary hammer and I were chatting when we started I said if we're all talking to each other at the end of this It'll be a miracle But we've sort of done it and there's really nobody annoyed with each other Even though you know some of the issues of authorship have surface But I was able to sort of tamp them down You know without causing too much troubles and then just many other people from the TS sets the program office And then I just have to give a shout out to Kenna because when you know the TCJ announced rare projects I kept jumping up and down and let's do adrenal cancer and you know at first it was like okay, Tom That's great, but she did listen and so we got a ton. Thank you very much Sort of a whirlwind tour Yeah, so the clinically they've gotten much better there's a radiologist to Dedicated adrenal scans they do this wash out they look for heterogeneity so Having said that you know, it's still a mystery at surgery sometimes right so I think pathologically we've gotten much better Right, so there's there's this mystique which I think you might be referring to in the path literature You go back 30 40 years people said you couldn't tell adrenal Adnomas and carcinomas, but I think you know the work of Larry Weiss and other groups have really sort of distilled the Histologic features down so truthfully, it's not that hard most of the cases across my desk Are pretty straightforward now But again, I'm in a unique situation that I get to see two of these a week Whereas most private practice pathologists will see one every couple of years, so it's a I benefit from that Does it affect the localization of beta-catalan? Oh, yes so so what we're doing now is we're actually looking carefully at the ZNR of three mutations to see what the localization in those cases looks like because They it's clear that they don't look you so with beta-cat you get a strong nuclear signal You don't necessarily get that so it may mean that the the winch signaling is you know Not as potent in the ZNR of three mutations So we're we're exploring that with our tissue arrays that we didn't use for the fish studies Yes, do you have any information on intratumor heterogeneity and does that affect prognosis? No So that's my next project right is to wait for a big 20 centimeter tumor and sample it in 15 spots and then you know Get roll to construct the you know the evolution of the mutations, so that's there Unfortunately, I have many samples in the freezer, but nothing that's really tailored to try to reconstruct the evolution of a tumor what I am collecting though are tumors that have both adenomas what it would appear to be adenomas and cancers and Most of those are in paraffin But I am collecting a series of those to try to build a molecular, you know argument that this progression really does happen Okay, thank you very much Thanks, Tom and our second speaker Tonya Davidson from NCI she's going to talk about NCI cloud pilot report You