TANIOS BEKAII-SAAB: Well, hello, and welcome to today's session. My name is Dr. Tanios Bekaii-Saab. I'm a medical GI oncologist with the Mayo Clinic Comprehensive Cancer Center in Phoenix and Scottsdale in Arizona.

Today, we're getting a glimpse into the future of medicine, and specifically in pancreas cancer, one of the most difficult, if not the most difficult, cancer to treat and achieve-- optimize the outcomes. We have some fantastic colleagues and experts in the area that will speak on optimizing outcomes for patients with non metastatic pancreatic cancer-- so we're going to focus on those patients with non metastatic disease.

From our Rochester, Minnesota campus, we have with us today our surgical oncologist who is internationally known for his work in pancreas cancer, Dr. Mark Truty; also, one of our international experts in radiation oncology focused on gastrointestinal and hepatobiliary cancers and pancreas cancers, Dr. Christopher Hallemeier; and then, again, a radiologist who specifically focuses on GI cancers, and more specifically HPB, hepatopancreatobiliary cancers, Dr. Ajit Goenka. So a fantastic panel that will guide us through in how to optimize care for our patients with non metastatic pancreatic cancer.

We'll have three topics, each one about 30 minutes, so we'd love to hear from you once we open the discussion to questions following each topic. So we'll have the opportunity to actually discuss some of your questions.

So we're going to start with the first topic, and that's essentially focused on staging and understanding some of the intricacies of how we stage our patients with pancreas cancer, going through a multitude of modalities from the anatomic to the biochemical to procedural.

So pancreatic cancer staging does affect the work of each of our panelists here today and likely all of you listening as well-- of course, our patients at large. From each of your practice and research backgrounds-- and I will go certainly through a number of these to understand the standard of care pancreatic cancer staging and what we may be doing differently in our practice at Mayo Clinic that ultimately would translate to changing the way we should be staging our patients across.

So I'm going to start with you, Mark, and I know you want to perhaps share a slide-- and then just at a high level, how do we think about anatomic versus metastatic staging of our patients with pancreas cancer?

MARK TRUTY: Well, thank you, Tony, for the introduction and for this webinar. So as anyone who treats these types of cancers and the patients and their families, they're well aware that pancreas cancer is the least survivable of all malignancies. Now, I mean, if you say that out loud, I mean, it's a pretty dramatic statement, but it's true. And the reason for that is because of its predilection for metastatic disease, whether that's obvious or whether that's occult. And I'm going to share a simple slide here that kind of exemplifies the point.

So although it is a highly lethal cancer, fortunately it's not terribly common. About 62,000 patients in the United States were diagnosed last year, half of whom have already obvious radiographic metastases. Those are the patients we are not necessarily discussing today. The other half of patients have no obvious metastases, and they've been traditionally classified as either resectable-- meaning the tumors are localized to the pancreas itself-- or classically unresectable-- we now know these as borderline or locally advanced, meaning the tumor has grown outside the pancreas to involve critical blood vessels.

Now, although these 50% of patients we don't see obvious metastases, if we follow these patients over time, the vast majority, 80% to 90%, do develop metastases-- either hematologists in nature, or through peritoneal dissemination, or through some combination thereof. Hence, it's critical that we stage these patients appropriately. Otherwise, we're going to have a poor outcome.

TANIOS BEKAII-SAAB: All right. Thanks, Mark. Yeah, absolutely. I mean, I think adequate staging makes the world of difference for our patients and understanding where they're at. You know, would one guide them through the right path to the type of therapy they are most eligible for, but also refine further who may be eligible for potential curative resection.

So Ajit, Dr. Goenka, thinking about the changes that we've seen over the last few years about increasing and enhancing the sensitivity and the accuracy of radiologic staging, can you go over what is considered standard? What kind of things we're doing a little bit differently at Mayo specifically with the PET/MRI? Pros and cons, what's on the horizon, please?

AJIT GOENKA: Yeah, absolutely. And thanks again, Tony, for the introduction and for the invitation. So I think when we talk about what's the standard of care, I think there is a clear consensus in the community that the standard of care is a multiphase pancreatic protocol CT for those patients who don't have obvious metastases at the time when the pancreatic cancer is diagnosed.

Now, what has happened over the years is that, of course, we have now the ability to do much higher resolution imaging compared to what we were able to do a few years back, and as a result of which, we've become much better at locoregional evaluation of the pancreas cancer. So what that specifically means is that, whether or not that tumor is encasing any of the major blood vessels, like the celiac axis, the SMA, or any of its branches, we can go down to the second and third order branches and give accurate information to our colleagues like Mark so that they can plan those complex surgeries.

Unfortunately, when it comes to metastasis, I think anatomic imaging has kind of reached its plateau. So despite the fact that we have this increasing higher resolution, the challenge is that on something like CT, we probably don't have the level of soft tissue resolution that we need to be able to pick those subtle metastases in the liver or in the peritoneum. And I think that kind of alludes to what Mark had mentioned earlier, is that many of these patients eventually develop metastases, and it's not that they did know they probably developed metastases, it's just that we are not able to detect them at a stage where we should be doing that.

So a few years back, when Mayo Clinic decided to invest into new technologies to improve the care of our patients, one of the things in which we invested heavily was a PET/MR, and Mayo Clinic in Rochester was the first center in North America that installed what we call a fully integrated PET/MR. And what we mean by fully integrated is that it is something which has the state-of-the-art solid state detectors, which is something called as photomultiplier tubes, and these tubes allow us to have simultaneous acquisition of the MRI data along with high signal to noise ratio PET data.

And because of the fact that we are able to acquire them simultaneously, what it allows us to have-- accurate coregistration of the anatomic data as well as molecular imaging data from the FDG PET component. And as some of the data that I will demonstrate to you in some of the slides that I will share, we have found that this PET/MR is much superior compared to CT scan for detection of those subtle peritoneal and hepatic metastases.

In addition to that, what it also allows us to do is that it allows us to accurately capture the molecular profile of pancreas cancer because the challenge is that, traditionally if you speak to people, they'll tell you that, well, pancreas cancer is not really a FDG cancer, so you should not be really doing PET in them. But I think one of the reasons why we are being able to utilize PET technology to the best advantage of our patients is because with the increasing detector technology, we are able to capture the FDG ability of the tumor much better than what we were able to capture with the first generation PET-CT scanners. And as a result of which, we are able to follow up this patients for things like response assessment.

So in short, what I would say is that high resolution CT of the pancreas protocol is certainly the standard of care for the locoregional evaluation. It's also very good for looking at metastasis in the liver, but still not cutting the mustard when it comes to detection of those subtle and early metastases. And for that, we have now been able to use FDG PET/MR here at Mayo Clinic in Rochester. We've published our data on that.

In addition to that, we also have the privilege of having two NIHR 1 studies now. One of them is prospectively validating what we have been doing in our practice, combining that with other biomarkers such as that from liquid biopsy, and the second one is that on a newer radio tracer called fibroblast activation protein inhibitor PET, which is also something that we will discover as we move along in the discussion.

TANIOS BEKAII-SAAB: Ajit, I mean, what are the goals ultimately that you would like to see being achieved with those new modalities that you're looking at towards the future?

AJIT GOENKA: Yeah, absolutely. So I think one of the main goals that we have is that for colleagues like Dr. Truty, we should be able to tell them as to which of the patients actually fulfills the criteria that they want them to fulfill before they decide to take them onto this extremely, highly morbid surgery, such as whipple procedure. And for that, what they want to know from us is that-- well, they don't want a patient where we miss out some metastases in the liver or in the peritoneum, and the patient goes through the whole drill of neoadjuvant therapy, and only goes to the OR to be able to find metastasis-- or they don't find metastasis, let's say, at the time of surgery, but shortly after two months later, the patient presents with hepatic metastasis.

So that's not the goal that any of our colleagues want to give to the patient. And that is even one of the main goals, in addition to, of course, characterizing whether or not the tumor is involving any of the smaller branches of the arteries, is also to be able to pick up those subtle and tiny metastases.

The third goal, which is more of a semi-futuristic, is to be able to predict the aggressiveness and the biology of the tumor. And again, when we discuss about response assessment, I'll show some of the biomarkers that we have identified that can help us make those decisions, that will help us to risk stratify this patient and prognosticate them in a much better way.

TANIOS BEKAII-SAAB: This is great. This certainly sounds very promising, and certainly we'll look forward for more results.

Mark, I want to get back to you. So we have our traditional radiologic staging-- how about biochemical molecular staging? What are we doing here that's a little different, but also thinking about our colleagues in the community who may not have the same access-- which we hope ultimately everyone will follow. But what kind of things are we doing and what kind of things you think should be done to adequately stage patients in addition to the standard radiologic staging?

MARK TRUTY: Yep. So there's something that can be done everywhere, and we're doing some things on top of that. The key thing is we want to risk stratify these patients. And a patient who otherwise has adequate imaging without any evidence of metastases, we still need a little bit more information, and so we use biochemical staging, specifically using tumor markers, such as CA 19-9, which everyone is familiar with.

Now, CA 19-9 is not just a passive molecule. It actually has a mechanism. It is the ligand for endothelial cells selected. Those patients with an elevated CA 19-9 have a higher predominance of hematogenous metastases.

Now, we've known it's not a very good screening tool because there are false negatives and false positives. Ten percent of the population lack the fucosyltransferese necessary for expression, so if you check, it reports as less than 1. Those are non secretors. One third of patients have normal levels, and then some patients, their CA 19-9 can be elevated for a variety of other benign reasons, such as biliary obstruction, et cetera. And everyone-- every single major institution has published their cutoff through an ROC curve, but no one seems to follow it.

And then I'm going to share a slide here, which is a very important paper that came out, and this was out of Germany. And in Germany, this is all data of every patient with pancreas cancer who is seemingly resectable. They take them directly to the operating room with an upfront operation, and they stratified their survival strictly by their baseline CA 19-9.

Now, look how beautifully this stratifies. It almost looks like fake data because it's so amazing. But as you can see, the higher the CA 19-9, the worse the subsequent survival. And those patients with the CA 19-9 above 1,000, at diagnosis, those patients had a median survival of less than one year after seemingly curative intent resection.

Now, most places consider over 1,000 as consensus criteria for very high risk. They should get neoadjuvant. But the question arises-- what do you do on someone who is less than 1,000 but higher than normal? Sort of that middle ground? What do you do with the patient with the big bulky tumor, you know? Does that explain the tumor marker elevation? And what about those patients who are jaundice, which a large proportion of patients are? How do we interpret their CA 19-9 in the setting of biliary obstruction?

So we specifically looked at that. This is a paper published by some of our colleagues in radiation oncology and Dr. Hallemeier. And what we found-- and these are all patients who underwent upfront resection with a tumor marker measured-- and we found that that CA 19-9 actually didn't correlate with bilirubin, OK? Nor did it correlate with tumor size. The only thing that CA 19-9 correlated with was their survival. And we found that those that were normal or non secretors had similar survival, but those patients that had any elevation above normal, regardless of extent, did significantly worse, and both were recurrence free and overall.

We then took a look at a larger data set. We looked at the NCDB over 100,000 patients of all stages, and we confirmed these results-- non secretors have a identical survival to those that have a normal CA 19-9, but any CA 19-9 above normal lends itself to a worse survival stage for stage, and the largest detriment was actually in those patients with localized tumors undergoing resection. What was that detriment? About a 50% reduction in their survival. That's the same reduction we see with the positive margin operation. So that's why we use CA 19-9 specifically in these patients. And anyone with an elevated level, we consider them high risk and indication for neoadjuvant.

Tony, you ask, are we doing anything on top of this? We are. We're in the era of genetic testing and technology where all of these-- deep sequencing is available at relatively low cost.

And a couple of years ago, I was in an elevator and one of the laboratory scientists was commenting on cell-free DNA. They had developed a cell-free DNA assay in blood for mutant KRAS. They were initially developed for colon cancer. And I said, really, we have this? They're like, yeah, it's clinically approved. And so I started using it.

And so we used a prospective trial in all patients with seemingly non metastatic cancer- it's a blood test. It measures a variety of mutant KRAS mutations-- and we found that one in five patients had a detectable mutant KRAS in their blood.

Now, you know, logically, you shouldn't have mutant cancer DNA in your systemic vasculature. That would suggest you're metastatic. And here are the results of that initial nearly 800 patient trial. Those patients that had detectable mutant KRAS in their blood had a significantly worse overall survival than those patients who was not detectable. And so this is now standard of care at Mayo Clinic in Rochester, and it's a simple relatively inexpensive test. Using it on all patients, and we're hoping to get this now pushed nationwide.

TANIOS BEKAII-SAAB: So Mark, this is very interesting data. Do we know what the sensitivity level of this plasma KRAS testing in early stage pancreas cancer?

MARK TRUTY: So in about 800 patients, about 20% of those patients were positive at diagnosis. When we tested patients who've already received some chemotherapy, that number drops significantly. So that does lower the sensitivity, any type of treatment, but that makes sense. If you're getting chemo, you're hopefully treating it.

Obviously, as the technology improves, we may find every patient has detectable mutant cancer DNA, based on obviously their known history of having occult meds. But this is something that's readily available and is very useful for prognosticating patients, whether we do upfront surgery or consider neoadjuvant.

TANIOS BEKAII-SAAB: So for those patients who may actually have that KRAS available, then it is part of your overall algorithm about how you decide on moving forward. And we'll talk a little bit more about how do you think about your treatment strategy and how you integrate those tests.

Chris, I want to just touch upon with you very briefly. SMAD4. There's been, in the past, a lot of talk about SMAD4 and how it relates to more locoregional disease versus metastatic disease, and it came from-- some of the data came from one of the RTOG study that included samples that measured SMAD4.

Any follow up on this? Is that an important biomarker that has a clinical utility at this point of time? Do you even care about it in your clinical algorithm?

CHRISTOPHER HALLEMEIER: Yeah, that's a great question, Tony. Most of the data looking at SMAD4 was from patients who had upfront surgery, and then predicting the risk of distant metastatic disease afterwards, or from autopsy studies-- patients who died from pancreas cancer and whether SMAD4 and the tumors associated with more locally predominant versus a metastatic phenotype. And it is associated with locally predominant versus metastatic.

I think it's of limited utility in clinical practice. It requires a fair amount of tissue in order to be able to do the test, which, if we're doing a diagnosis based on an EUS just to confirm that the patient has pancreas cancer with an intent to do initial upfront chemotherapy, I think it's of less relevance in the neoadjuvant era. And certainly there are much more exciting biomarkers-- blood-based biomarkers, as Mark Truty has discussed, in addition to better radiologic tests, which I think are going to be more clinically useful currently and moving forward.

TANIOS BEKAII-SAAB: Thank you. That's quite helpful.

Mark, in addition to radiologic staging, biochemical molecular staging, do we-- do you do peritoneal laparoscopies? So laparoscopies, do you integrate those? What's the role of endoscopic ultrasound? How much do you rely on it versus a good CT angiogram and a PET/MRI? Are these complementary?

How do you think about these procedures as they relate to the other forms of staging in pancreas cancer?

MARK TRUTY: So I'll start with your last question first-- EUS versus CT. If you have a good triple phase CT scan, pancreas protocol, that gives me everything I need to know basically about the tumor and vascular relationships.

Now, occasionally, we may find an EUS report that differs-- suggesting more vascular involvement than the CT would suggest. And I would defer to the EUS over the CT if it's worse because, again, a CT scan, if it has vascular involvement, it almost certainly does, but it's not a microscope, OK? That's what differs between a CT scan and-- and the EUS, it's a little bit more higher resolution.

Regarding do we do any other staging procedures in addition to good imaging blood work? We do. That's where we also differ from the rest of the country, in that we do staging laparoscopy on every single patient who has no obvious metastatic disease, prior to determining what we're going to do.

Interestingly, historically, about 30% of those staging laparoscopies decades ago were positive. They found hepatic or peritoneal metastases. But the performance of staging laparoscopy has really plummeted in the neoadjuvant era, presumably because we think we have better imaging.

The issue is the biology hasn't changed. People still have peritoneal dissemination. That hasn't at all dropped. And so we've recently looked at our experience in the modern imaging era of over 1,000 patients-- and I'll share a slide here just to show you that.

And so we looked at our last five years of our staging laparoscopy experience, and we found that one in five patients had a positive laparoscopy. A positive laparoscopy is either we see gross disease, liver or peritoneal, or those who don't have gross disease but we do washings and they have a positive cytology. That's a significant number, and this is in the modern era, and this is in patients who've had triple phase pancreas protocol, PET/MRs, et cetera.

When we look to see who are the patients at highest risk of having a laparoscopy, we did find several risk factors on multivariate analysis-- younger patients, less than 60 years of age; those patients with a larger tumor, greater than 2 centimeters; tumors in the body or tail; and then any patient with an elevated CA 19-9.

And you could see how our patient-- the 1,000 patients distributed here-- if you had no risk factors, your risk of a laparoscopy being positive, extremely low, but that was a small fraction of patients. Three quarters of our patients had two or more risk factors, and those patients had at least a 20%-- in some cases, over a 30% to 50% risk of having a positive laparoscopy. And that's something that would have otherwise been missed if we hadn't done that, and it helps to restage them better and then alters their treatment. And so this is something that we consider, and I think every institution should consider the laparoscopy.

We also do it before any neoadjuvant therapy. The other thing that I didn't show here is that those patients that had a laparoscopy after neoadjuvant, there was a lower yield. Why is that? Probably because we're masking some of those metastases with the neoadjuvant, and those are the patients that are more likely to subsequently get peritoneal recurrence. Thus, we want to do this early on as possible to stage our patients better.

TANIOS BEKAII-SAAB: Thanks, Mark. So we'll have a couple of minutes for questions. So if you have any questions for the audience, please make sure you include them in the Q&A and we'll try to answer them.

But I want to jump just quickly to Ajit. I mean, when we think about peritoneal disease, which is essentially the hidden disease in many ways-- so laparoscopy and potentially measuring KRAS and peritoneal fluid-- but radiologically, what are some of the more sensitive methods that can guide us into understanding whether the patient is at risk versus not?

AJIT GOENKA: Yeah. So I think peritoneal disease is a tough cookie to crack on any kind of imaging modality, but I think where imaging really shines is in its positive predictive value. So when you see enhancing nodularity in the peritoneum, or when you see lobulated cystitis in the pelvis, or you see enhancing nodules in the pelvis in the dependent areas, which you can especially see on the CD component of what we do, then I think the positive predictive value is quite high.

In terms of the newer imaging modalities that we have, the PET/MR, where we are not only combining the anatomic resolution, but also the molecular component, that is also improving our sensitivity to be able to pick up those disease.

And then finally, because of the imaging properties of the new radiotracer, which is FAPI, which is still under investigation-- and we do have an R1 for that-- that is where I think we are most optimistic about where, hopefully, it would decrease the number of laparoscopies or diagnostic laparoscopies that somebody like Mark has to do because there is certainly a cost-- there is involvement of the surgeon's time-- or time, so on and so forth. So in improving the outcomes, we also want to make it more cost effective.

And our hope is that once we finish those studies, once we have FDA approval, hopefully we'll reach a stage where imaging would be as good for detection of peritoneal disease than it is for liver disease right now.

TANIOS BEKAII-SAAB: Thank you, Ajit. Mark, quick question. This is mostly pertaining to adenocarcinoma. There's a question from the audience that says, do we think about pancreatic neuroendocrine tumors a little bit different? They understand that there are certainly a lot of nuances across the board, from the Ra story CA 19-9. But at least from the high level staging and how we think about neuroendocrine, is your thinking different in those patients?

MARK TRUTY: I think it is, primarily just because one, it's less-- certainly less common, but it's also a much better prognosis to have a neuroendocrine tumor.

Now, like any type of cancers, you could have a spectrum of behavior. There are some neuroendocrine carcinomas that are terrible. They're worse than adenocarcinomas. So those patients with higher grade neuroendocrine tumors, I would typically kind of approach them like I would a pancreatic adenocarcinoma, just because of their risk of metastatic dissemination and poor outcomes with upfront surgery. But for the garden variety neuroendocrine tumor, we definitely manage and address very differently.

TANIOS BEKAII-SAAB: So Ajit, from the radiological standpoint, can you tell the difference most of the time?

AJIT GOENKA: So I think we can tell the difference most of the time, and the reason is because, as you know, that majority of the neuroendocrine tumors, they tend to be hyperenhancing on the arterial phase, which is something that we don't see with the pancreas cancer, which tends to be hyperdense in the early arterial phase, as well as in the pancreas phase.

In addition to that, the nature of the vascular relationships with pancreas cancer, which is more of an infiltrative neoplasm, tends to be way different than neuroendocrine tumors, which typically tend to be large but usually more circumscribed. And likewise, the imaging appearance of the liver metastases is also very different because, again, in pancreas cancer, they tend to be more hyperdense, whereas with neuroendocrine tumors, they tend to be hyperenhancing.

So I think there is a world of difference-- and speaking of neuroendocrine tumors, I don't want to take much time there, but you do know that we also have gadolinium uptake now, which has completely revolutionized the way that we image and treat the patients with neuroendocrine tumors. So that is another big difference where the imaging tends to pivot off of kind of tumors.

TANIOS BEKAII-SAAB: That's great. Thank you. I mean, this was a great session. I think we come to the end of this first topic, but it's clear that we're getting much better with the way we're staging patients, integrating radiologic biochemical molecular staging, procedural staging-- you know, that this is not just one scan do it all, but it's a puzzle and we're including a lot of these pieces together to actually have a clearer picture about how to best optimize the care for our patients with non metastatic disease. So this is fantastic.

So moving on now to the second topic-- and primarily, this topic is focused on now that we had our patients staged and we certainly confirmed that they have a very low likelihood of micrometastatic disease or metastatic disease. And then we want to consider their surgical options.

One of the important elements that we have seen emerging over the last few years-- and certainly has become our standard for the overwhelming majority of these patients with a non metastatic disease-- is the value of actually neoadjuvant therapy for patients with resectable disease. And we'll go through some of the nitty gritty of what makes the most sense for these patients.

But as we started moving more and more to the neoadjuvant setting, based on data, that at least neoadjuvant therapy would certainly help us select better patients who biologically would benefit the most from surgery-- the other token would be who are those patients, even though they may have localized disease that a biology dictates otherwise. And so that becomes an important-- one aspect of neoadjuvant therapy.

So we know that this disease, as Dr. Truty mentioned initially, it's primarily systemic disease. Out of all the cancers we treat, this probably is one where even when all these elements we talked about point out to a locoregional disease, there's a good chance this is already metastasized-- a really good chance that this already metastasized, and we just are unable to detect that metastatic disease and neoadjuvant therapy will give us that time to test the biology.

MARK TRUTY: And Tony, you know the--

TANIOS BEKAII-SAAB: Yes?

MARK TRUTY: You know, the data-- the most data that we have, as you're aware, is in the adjuvant setting-- adjuvant chemotherapy. We've had, for a couple of decades now, data for that. It shows it improves survival.

The fact that adjuvant chemotherapy improves survival after resection proves your point. These patients are metastatic at the time of resection. So it's kind of silly when people say, oh, they're not metastatic. Well, they are, otherwise those adjuvant studies would have been not at all effective.

TANIOS BEKAII-SAAB: Oh, absolutely. And to your point, we know that about a quarter of the patients after surgery, if we haven't given them that piece before surgery, are unable to get adjuvant therapy postoperatively, and we know that that adds to their survival. And so I think as we move it more to upfront, not only that we're testing the biology and we're certainly picking the right patient for surgery, but we're also making sure that more patients are exposed to that other modality that enhances the likelihood of patients surviving that cancer.

You know, theoretically-- and Mark, you can talk a little bit more about this-- it does improve also surgical outcome-- in some ways, improving the R0 resection rate, lymph node negativity local recurrence. We downstage patients-- but whether that ultimately affects survival or not is unknown.

And to your point as well, some concerns always arise, and we know that patients come with these questions, or even physicians-- referring physicians-- is, what if we miss that window of opportunity for surgical resection? And we know that if a patient never makes it to surgery because of all the elements we're going to be discussing, they actually are the patients who should never get surgery anyways. Those are the patients who would not benefit, and arguably, those are the patients that may actually experience more detriment than a benefit. So there's quite a bit of good reasoning to move to the neoadjuvant and total neoadjuvant therapy.

And I think as we're getting more and more along the lines of genotyping, checking for inherited germline alterations, but also genotyping tumors, whether through liquid biopsies or tissue biopsies, we also have a good glimpse about what may be some of the drivers.

Now, chemotherapy remains the mainstay-- and we'll get to which patients may be eligible for radiation-- chemotherapy remains the mainstay for patients with non metastatic pancreas cancer that would proceed to surgery. We have two options-- two major options that are available to patients.

They seem to-- when they're put head to head, they seem to actually be relatively equivalent, but then when we start understanding a little bit the intricacies of where, say FOLFIRINOX versus gemcitabine nab paclitaxel performs better, we understand that, let's say in the presence of a BRCA one, two, or probably two alteration-- or HRT in general that perhaps FOLFIRINOX or gemcitabine cisplatin-- and arguably, until we see more data, gemcitabine nab paclitaxel and cisplatin could be favored versus those that are not, or these mutations are absent where gemcitabine nab paclitaxel may actually end up being a little preferred.

There are other-- there are other elements as well that are being looked at within the histology of the tumor that may help decide of one versus the other.

So more is coming to help us decide on what patients may be optimized for one versus the other. I would say that as we move more and more to the neoadjuvant world, we actually overall are improving the outcomes of our patients, and that's what we're going to continue seeing. There are studies that are looking at pre versus post-- although at Mayo, our standard is neoadjuvant for most patients.

Next generation sequencing testing is almost a must now, since it helps us not only with the pick of chemotherapy, but also may add biologic options, including on trying for our patients. But also inherited panel testing, that can help both, of course, with risk stratification for other family members, but also important for picking the treatment.

But mostly, what we have integrated into our neoadjuvant approaches is this whole concept of chemotherapy switched strategy. And Mark, you being one of the pioneers of this strategy, and it certainly yields significant benefit-- so I think this has actually revolutionized how we care for our patients and has led patients to improved outcomes. So I'm going to give this to you so you can go through what that means to our patients and how, not just at Mayo, but how do you think this will essentially translate to patients who may be cared for outside Mayo?

MARK TRUTY: So that's a critical point, Tony. We are-- I think across the country, there is a movement towards neoadjuvant, OK? And we're not really arguing that much about neoadjuvant upfront surgery as much anymore. So that's not the question to answer.

The question to answer is, if we're giving neoadjuvant, is it doing what we hope it's doing? Because we're giving that preoperative chemotherapy in the hopes that it would change an outcome compared to doing an upfront operation. Therefore, in order to get that outcome, we have to prove that it's actually effective.

So I often pose a question to my trainees. I said, if you had pancreas cancer, what neoadjuvant treatment would you want? Would you want six months of chemotherapy, or would you want six months of chicken nuggets daily? And they look at me, they think I'm joking, I said, I'm dead serious. And they all invariably say, oh, I'll take the chemo. I go, well, it actually depends because if you get six months of ineffective chemo, you might as well had six months of Chicken McNuggets and had the same outcome. And so the question is, how can we know that the chemotherapy we're giving is effective?

Now, any other tumor type, we asked Dr. Goenka-- is the tumor shrinking? Is it responding? And then hopefully Ajit can comment as well, but unfortunately, for pancreas cancer, radiologic response is a very poor surrogate of response. It has no correlation to pathologic response or survival, tumors that involve vascular structures never pull away-- they're always there-- and that data is so strong that neither the NCI nor the NCCN considers radiologic response a clinically relevant endpoint.

And so therefore, although we keep repeating these CT scans, we're really not expecting much to change. And so if it's not changing, is it working?

Our second way of assessing response is looking at those tumor markers we mentioned, that CA 19-9. And if it's elevated and it comes down, that is a good surrogate. However, as we mentioned, 10% are non secretors. Thirty percent of patients have a normalcy in 19-9. So right from the start, 40% of your patient population, you can't use it.

Also, once it goes to normal, it can't get any more normal than normal. So once you normalize that, and if you normalize it early, you can't use it any longer. And so thus, we're stuck. We can't use CTs very well, CA 19-9's helpful in some but not in all, and so thus we've had to come up with alternative ways, and that's where our institution has used metabolic imaging, specifically PET. Either PET-CT or PET/MR-- and then hopefully Ajit can kind of comment. But we routinely use PET at the time of diagnosis, and then we follow with PET after therapy, and we found that PET scan highly predicts response to chemotherapy, and that also has been shown to be pathologic-- shown to correlate with pathologic response.

Ajit, maybe you could comment because obviously you've looked at more PET scans than anyone for pancreas cancer.

AJIT GOENKA: Yeah, absolutely. And let me share some of our data, as well as our screen images. So let me start here.

Yeah. So I think Mark has already alluded to the fact that we have CA 19-9 that has very limited clinical utility for all the reasons that he has very eloquently described.

Now, what we traditionally use on any of the solid tumors or CDR/MR, is what we call the RECIST criteria, which basically measure the two-dimensional size of the tumor, and then we measure sure that after treatment. And as Mark has mentioned, unfortunately in the case of pancreas cancer, the tumor just gets replaced with fibrosis, and as a result of which, oftentimes there may not be a downstaging. And even if there is downstaging, as he has mentioned, that may not necessarily predict good outcomes, which is exactly why we want these patients to undergo neoadjuvant therapy.

So we clearly need reliable biomarkers, not just to guide therapy, but also to reduce toxicities and costs from an ineffective chemo, and that is the reason Mark asked his trainees is to rather have Chicken McNuggets instead of getting chemo.

In addition to that, we also need to think about the next generation of the chemotherapy and immunotherapy that is on the horizon. And for that, we need to be able to design clinical trials that have got optimal outcome points.

And finally, we also-- one of the reasons that, Tony, you mentioned, we are giving neoadjuvant is to select the right biology of a tumor that really needs to undergo surgery, and that's the reason we need to be able to have markers that not only predict response but also give us some information about the biology of the tumor.

So right now, what happens is that typically, they get the pre-op chemotherapy at most of the places, and those centers that don't have access to anything beyond CA 19-9 or CT, patients will inevitably go to surgery if they don't have any progression of the local disease or if they don't develop any metastases. And then it's only after the surgery that we come to know if all of this chemotherapy has actually worked or not. And that, in the era of precision medicine, is not really the ideal way to do things.

So what is it that we have done here at Mayo? So just like Mark mentioned about liquid biopsy with KRAS ctDNA where he had this elevated conversation with another basic scientist and they decided to do that, similarly we had an elevated conversation centered around PET/MR. And when we installed PET/MR here at Mayo Rochester in 2016, one of the indications that we started doing is pancreas cancer.

And it's not just something that is based upon anecdotal evidence, it is something that we have systematically evaluated. We published our results a couple of years earlier, and that paper got selected for the most impactful paper of the category in The Journal. Mark and I were also invited to present that as a webinar, and it also got selected as the Editors Choice Award.

So what exactly did we find in this particular category, and what is it that we are doing right now as part of our standard of care for response assessment in these patients?

So what we're doing is that what we found in this study, which is that the biomarkers that we were able to quantify from PET/MR, for these patients, they-- first and foremost, we found that those patients who eventually had major pathologic response, which is basically where either they have no tumor cells left on the specimen or they have got very limited tumor cells, they had significantly longer overall survival compared to those who did not have complete or major pathologic response, which is something that we already now know very well.

And the biomarkers that we specifically looked at was something called a CMR, or complete metabolic response; we looked at the relative change in the SUV max, which is a very commonly used metric; and then we also used what we call as a delta SUV gluc, which is basically an SUV max that is normalized to the blood glucose of these patients.

Now, one of the things that we also know, which we often don't talk about, is that majority of the patients with pancreas cancer, they tend to have either hyperglycemia or they tend to have diabetes. And as you know that for doing FDG PET, we need them to have good blood glucose control-- otherwise, the PET doesn't really work very well-- and that's the reason we normalize SUV max to the glucose of these patients, so that we would get an optimal idea of how that tumor is behaving.

And then what we found is that these biomarkers, they have an extremely high negative predictive value for major pathologic response. So what that effectively means is that if you have a patient who does not achieve complete metabolic response or who does not have a threshold decline in a SUV max or a SUV gluc, in response to neoadjuvant therapy, is unlikely to have major pathologic response, even though everything else may look OK.

The other important point for these kind of biomarkers is that they need to have excellent inter-reader agreement so that, let's say one of my colleagues reads the scan, they should not have a completely different result compared to, let's say, I reading the scan. And we found that these biomarkers had excellent anterior agreement, which is what makes it-- which just adds to our ability to consistently deliver the excellent results that we want.

And finally, in this study, we found that CA 19-9 was neither associated with response nor overall survival, and one of the reasons could be the different things that Mark mentioned, that compromises our ability to use CA 19-9 in this particular context.

Now, some of you may wonder why PET/MR and why not PET-CT? So one of the reasons that I earlier mentioned is that when you look at the silicon photomultiplier detectors that are part of this PET/MR, as well as the scanner geometry-- so what we mean by scanner geometry is that if you look at the axial field of view of this PET/MR, it's about 25% centimeters, compared to much smaller axial field of view for PET-CT.

So when you combine those technical advantages, it gives you a much higher ability to pick up these photons from the tumor, which makes it possible for us to interrogate pancreatic cancer, which yes, we know is not as FDG-avid as other soft tissue neoplasms, but we now realize that about 90% to 95% of these tend to be FDG-avid, which is how we define them as SUV max-- more than four.

In addition to that, because of the fact that we are acquiring the MR and the PET data simultaneously, we have excellent co-localization, and we are also able to use the full spectrum of MR soft tissue resolution to be able to pick up those subtle hepatic and peritoneal metastases.

So one of the unique parts of our PET/MR practice, which is what we use for pancreas cancer, is this imaging protocol that we have designed, that essentially what we do is that we first do a whole body interrogation for metabolic imaging with metastasis, and then we do what is called a focused PET/MR of the abdomen, where we acquire a long 15 minute PET of the upper abdomen. In addition to that, we do diffusion weighted images, we do EUS enhanced MR, and therefore we get the hepatobiliary 20 minute sequence, which, as you know, is considered to be the reference standard for looking at liver metastasis.

And it's because of these that our PET/MR practice has been steadily going in the context of PET/MR, and you see that majority of these patients that we have are due to the pancreas cancer, which is quite unheard of in other institutions that may have PET/MR.