Great, we're on the home stretch. As you heard, thank you very much, of course, for giving me this invitation. It was really quite tricky because everybody said, well, Sandy's going to tell you more about this and Sandy's going to tell you more about that. So guess what? I'm not going to tell you more about this because I didn't want to give a talk that everybody else had given. So let's see what we can learn about thinking about precision immunosuppression. And again, I'm trying to talk about things that I think I can know something about, but hopefully it's not super repetitive. Here are my disclosures. So I'm going to start off with tolerance because there has been talk about tolerance, but I'm going to also hopefully show you one or two new tidbits of data and another way to think about it. I think tolerance is really the holy grail and patients want it. Patients don't understand why we are still giving them drugs to keep their grafts going. I don't understand why the body can't accept a new organ and I worry about how long my grafts will last. And I think tolerance is the holy grail because it's simultaneously freedom from immunosuppression and true tolerance is a guarantee that your graft is going to survive indefinitely, certainly from the immunologic perspective. However, the problem is that the absence of tolerance, I think you've begun to hear today, may be pitting the patient and the graft. To be absurd, from the graft's perspective, it may want just the most immunosuppression that you can give. Why not? It's going to prevent the DSA, it's going to prevent the TCMR, it's going to prevent the chronic AMR, but maybe that's not the best for the patient. And so what we're really talking about for each patient is what is the balance between which error is better or least bad to make, to err on the low side of immunosuppression or on the upside of immunosuppression. And I think that's very different for each patient and it's also very different for each clinician. So what is the evidence that there's too much immunosuppression? And I'm going to say operational tolerance, because what we've done to take children or adults off of immunosuppression, we don't know that they're truly tolerant. They're almost certainly not truly tolerant, but they're operationally tolerant. So of course, there's the WSPR trial, which is super old, 12 out of 20 patients were actually able to come off of graft and pass their liver tests and pass their biopsy tests. And I just want to remind everyone that these children did have five biopsies over an eight-year period with the last biopsy eight years after entering the study. And just very briefly, there was no progressive change or systematic change in inflammation or fibrosis. So each column, each sort of block is a patient. The upper panel is inflammation. Each row is a year and each tiny little column is a compartment of the liver. And the lower graph shows fibrosis by compartment. Now mind you, some of these children have DSA. Some of these children have high MFIDSA, but again, there's no systematic progressive change in inflammation or fibrosis. So then you've heard about IWF. The punchline is in the end, we have 38 subjects, I mean, 38% of subjects who are tolerant. We have 39 rejecters, the right group, the two groups of red people. And we had what we call 16 that were non-tolerant based on biopsy. And so when I say, gee, we've been able to take these people off of drug, people then ask, who cares? Did you help them? Does immunosuppression withdrawal yield benefit? And I don't know how to answer that question. I think most of us in the room would say intuitively being on no immunosuppression for one year, five years, 10 years, and those kids in Whisper have been off of immunosuppression now going to college, drinking, and getting ASTLT abnormalities, and they're still off of immunosuppression, and most of them are off for somewhere between, all of them are off for between 10 to 12 years. So you would think that that is generally probably better. But do I have proof that immunosuppression withdrawal yields benefit? Not really. What is there, however, to think about? In the I-WITH trial, and this was stated for the adult trial as well, 70% of the children rejected at less than or equal to a quarter of their baseline dose. So while these subjects rejected and they were placed back on immunosuppression, it's possible that they might not have rejected if we had reduced their immunosuppression by one-third, by 50%. So perhaps there are people even beyond the tolerant subjects who are on just a little too much immunosuppression. Our poster here actually shows a quality of life benefit. So the left panels are the 55 non-tolerant children, the right panels are the 33 tolerant children, and what you can see is an upward slope of the quality of life measured for both the child in various instruments of quality of life, as well as for the parent. The other piece of data that I would share with you that responds to the question, is reducing or stopping immunosuppression beneficial, is that it's very hard to measure the toxicities of immunosuppression, and in children in particular, they emerge very late after transplant. Cancers generally emerge two-plus decades after transplant, and this is when the pediatric hepatologists are no longer taking care of these patients, they disappear into the adult world, and at UCSF they become one or ten of 2,000, 3,000 people, and it's just hard to kind of know what's going on, but the bad things are not happening for about 20 years post-transplant. There's also recent data about the impact of long-term calcineurin inhibitor therapy on brain function, and finally, Henkken-Virchow just presented in a poster here that there's a steep increase in renal insufficiency, chronic kidney disease, 20 to 25 years post-transplant. So I think it is very difficult to demonstrate that there's a benefit of immunosuppression minimization because the medical complications take so long to develop. I think much of this session has shown that kids and or adults may be on too little immunosuppression from the perspective of chronic graft injury. I'm not going to talk about that, but I wanted to show you the prevalence of funny-looking, not great-looking biopsies in three withdrawal trials. The left panel is the pediatric eye-width trial. The upper-right panel is the LIFT trial, which is based in Europe, led by Alberto Sanchez-Foyo, and the lower-right panel is the optimal trial that we're conducting in the United States. So this chronic graft injury is happening in children. It's happening in adults, and it's happening at pretty much the same frequency. You've seen this. We're going to skip. You've seen this. We're going to skip. So now we're going to now take a look at too much and too little immunosuppression and talk about how are we going to get to precision immunosuppression. So this is stuff you live every day, so we don't have to explain it. Our current approach to liver immunosuppression, whether it be adults or children, is center-based protocols. And we think about what happened before transplant, how old were they at the time of transplant, what disease did we do this for, and what happened since transplant. Time, rejection, infection, malignancy, medical comorbidities. We look at some data, predominantly liver tests. Some people might be looking at DSA, so we should discuss whether that's appropriate or not, and that's already come up. And then occasionally we might get peppered with some allograft histology. But the realities of decision-making is in the absence of any pressing clinical need on any given day, it's easiest to do nothing. How many times have we said to a patient, and I'm guilty of this myself, doesn't seem like anything much is going on. Let's not rock the boat. We don't want the boat to rock. So there's a reluctance to reduce immunosuppression, because we don't want to tip anything over. We don't want to bumble into anything. It definitely engenders a temporary burden of probably some increased monitoring, which is a headache all the way around. I have to do more work. They have to go to the lab and stick out their arms, no fun for everyone. And again, the benefits are theoretical, and you often can't appreciate any benefits. So why take the risk when you can't put your finger on why exactly you're doing this? If there's a pressing clinical need, we all do it. But in the absence, we just stay the course. We're reluctant to escalate immunosuppression. When you see that ALT of 48 or 62, do you want to increase immunosuppression? Oh, let's just check it next week or two weeks. Oh, you have the sniffles. It's probably the sniffles. We look for any reason to explain it away, because it's just unpopular. Who likes to increase immunosuppression? But it's very interesting. This is a call to action editorial written by Chris Wiebe and Peter Nickerson. And this is my segue into no longer talking about the liver, but talking about the kidney. I had to find something to talk about for 30 minutes after all these great speakers. So I'm now going to shift to the kidney. But this is a general statement. In 2016, the FDA held a public meeting on patient-focused drug development. And during that meeting, transplant recipients highlighted the need to individualize and to simplify. We heard from Jackie about they want simple, right? Maybe one pill. And so this is really something that our patients want, even though they themselves are also reluctant to rock the boat. They think that they want this. So what I want to show you is what the kidney transplant community is doing to move towards personalized immunosuppression, because let's face it, they're way ahead of us. The second thing is they have way more to lose. They know that the kidney is just day by day fizzling away. Their graft half-lives are way shorter, and they're just in more trouble. So what are they doing? Well, I'm really impressed from the kidney transplant perspective about epilet mismatching. So I'm hoping that people will learn something from staying for the rest of this 15-minute talk to learn about what epilet mismatching is and what the kidney transplant are doing. The kidney transplant community had sort of a sentinel event that the current way that we approach immunologic risk assessment was inadequate. The CTOT 09 study, part of the study was to randomize low-risk, like low PRA, first transplant, non-African American, you know, the kinds of things that we think about, you know, pre-transplant course, age, you know, race, etc. Low-risk, living donor kidney recipients, and they were randomized at six months, if they never had rejection in that six-month period, they have no DSA, they have no polyomavirus, they're on a decent dose of MMF, and they have a quiescent kidney biopsy. So they did everything. And the randomization was to stop CNI or not stop CNI, okay? This was, again, that's why they wanted people to be tolerating a reasonable dose of MMF. And this is what happened. The study got stopped by the Data Safety Monitoring Board. The upper panel is the control patients. The lower panel are the CNI withdrawal patients. Of course, this is super early post-transplant, but what this found is that if you didn't mess with those patients, they stayed on CNIs, very little happened. One person developed DSA, that one person with the blue triangle, and a second person was found to have DSA, but retrospectively, that DSA actually happened, was actually present before randomization. So that person has two blue triangles. The X is the time of randomization. You can see that the patients who underwent withdrawal had bad things happening. There are lots of blue triangles, so DSA's showing up. There's lots of red triangles. People were having rejection. So this was a message to kidney transplant that standard immunologic assessment by which we're dosing immunosuppression is inadequate. So then they went back and looked at various factors, more novel ways of assessing immunologic risk, and then they looked at what's called epitope mismatch. So that's the HLA molecule. If you look at HLA matching, you're looking at the HLA, the entire molecule. You can look at the polymorphic amino acid, which is the little yellow patch, and epilate, which is what is the current way to go. But this time, back in 2015, they were looking at the larger patch, the green patch. And what they found here is that basically, if you were low risk by the epitope matching, nothing bad happened. So if you were less than 16 mismatches, nobody developed DSA. But if you were high risk and you had more than 16 mismatches, half of them, seven people developed DSA, and six of them did not. Now we have, for kidney, this is a disaster. Seven people developing DSA, not good. So we have to ask ourselves, if we were to use epilate and we knew that seven people would develop DSA in the high risk, is that something we think would be an F? Because we're liver people. Is the DSA, is that a biomarker that's sufficiently good for us? So I think this is very, very helpful for kidney. And the punchline in terms of actual function was that there was actually no difference at the two year time point. They were randomized at six months. There was no difference in the GFR at two years, and there was no difference in the change in GFR between six months and 24 months. So these people got rescued, but seven of them are walking around with DSAs and were concerned about that. This shows that compatibility is determined by molecular mismatching, and molecular mismatching is not reflected by the types of mismatching that you and I are used to thinking about. So what you can see is the left panel is DR, the right panel is DQ. One and two is the usual mismatch that we talk about, HLA-DR, one mismatch, two mismatches. And what you can see is that the molecular mismatch distribution, while the two is higher than the one, you can find that people with only one conventional mismatch have very, very high molecular mismatch scores. So that the molecular mismatch is not the mismatch we're talking about. For kidney transplant recipients, the development of de novo DSA is associated with trough TAC levels of less than five. So here what you have is the amount of time that any person is spending at each TAC threshold. So the first set of bars, blue and red, is less than three. They're spending not very much time at less than three, and it goes all the way up. And what this graph shows you is that the people who develop DSA, the red bars, have spent more time at low TAC levels, and that it's significantly different. But if you get to high TAC levels, the two pairs of bars on the right, there's actually no difference in the prevalence of DSA. So that if you have high exposure, if you're frequently at low TAC levels, you're more likely to develop DSA. But if you are mostly at high TAC levels, then there's no difference in whether you develop DSA or not. And this basically, it's very confusing, these graphs. So if you're interested in this, spend some time with the papers. This basically shows that the risk of de novo DSA is modulated by molecular mismatch. This has no mismatching, and this shows the modulation by the molecular mismatch. So for any trough level, among high risk patients, which is the red and the blue, those without DSA, the blue bars, spend less time at low trough levels. So you can see the first set of bars, the blue bar is lower than the red bar. The blue bar is the amount of time that they spent at less than three nanograms. So the people who spend 5% of their time, less than three, have DSA. But the people who spend only 2% of their time at less than three don't have DSA. And so again, among high-risk patients, those without DSA spend less time at low trough levels. Low-risk patients can remain DSA-free even when they spend the same amount of time at low trough levels, okay? So again, mismatching relates to immunosuppression dosing relates to appearance of DSA. Now what about the liver? The similar message is emerging. Epileptic mismatching associates with DSA development, and we have a couple of papers shown here that show this. But again, I ask you the question, does de novo DSA have the same impact in liver as in kidney transplantation? If we develop data that shows that the impact of DSA is as bad for liver as for kidney, then we should use epileptic matching, we should use these trough TAC levels to guide our immunosuppression. Because the body is chucking out DSA. The body does not know whether there's a liver or whether there's a kidney. Everything about DSA, the fact that it's class 2, the fact that it's DQ, the fact that it emerges as immunosuppression is reduced or withdrawn, everything about DSA is the same. The reason that liver patients have more class 2, I believe, is more liver patients are lurking at low TAC levels. And so if we think DSA is a good surrogate endpoint, then we can just use all of the information from the kidney people, because there's thousands more of them, and dose our immunosuppression in that manner. So what we have to figure out is what is the meaning of DSA to us. The last thing I want to talk about is this IBOX thing, which is super, super powerful. So IBOX is a score that's been developed by the French kidney group, led by Carmen LaFosher and Alex Lupi, to predict graft failure. This is based on over 7,000 kidney transplant recipients, of which over a thousand of the grafts fail. So in order to predict failure, you have to have failing grafts. So that's one of our problems, especially in pediatric liver transplant. You know, we just don't have that many long-term, and we don't have that many grafts failing in the long term. So it's going to be very hard for us to do anything on our own. They have a derivation cohort of 4,000 at four French centers, and then the rest were in two separate validation cohorts. They identified eight independently associated prognostic factors. They have a C-index of 0.8, confirmed in two validation cohorts, one in Europe, three centers, one in the US, three centers. Look at those C-indices, look at those confidence intervals. It was then confirmed in three randomized trials, and it's accurate, no matter when you look at it. And this was the time frames at which they were looked at. A lot of this was based on scores that were collected based on data within the first one to two years, that's where the hump is. But you can see there's a pretty nice tail of long-term assessments, and the score being rechecked at those times, and the score is accurate, not only at six months, but at five years. Now, what are the criteria? I have a little boo-boo here. You can see it's a little bit difficult, but you'll have to imagine. It's based on glomerular filtration rate, based on the MDRD, proteinuria, circulating DSA, and kidney biopsy findings. And what I wanted to show you is what the kidney community expects for every biopsy that the pathologist produces. These are the number of scores that are given to every single kidney biopsy, and I would argue that we in the liver community would be facilitated by having more scores, particularly for late biopsies, where findings are much more subtle, and that that jig effort that Jake and Deirdre referred to, where we were looking at all those parameters, that might really be the start of moving us into a more quantitative assessment of the liver biopsy. Another feature of the IBOX score that is great is that it is able to show a response to treatment. 844 recipients from the derivation cohort received standard of care treatment for antibody-mediated rejection, TCMR, calcineurin-inhibited weaning for toxicity, and the IBOX prediction capability after treatment was accurate for actual graft loss after treatment. So we need precision immunosuppression. Not only do we need a marker that tells us what to do, but it would be great to have a marker that can then verify that what you did was good or bad. And look at that C-statistic again. It's just amazing. Now this score, again, is GFR that's calculated from a creatinine, proteinuria, DSA, and pathology. Okay? These are all readily available, and that's another beauty. It's not some sophisticated complex nobody can understand. But let me just show you one sobering thought, which is my second-to-last slide. In 2017, Harihara at UPMC published this survey of surveillance kidney biopsy practices. And he surveyed 238 centers, of which 45% or 106 responded. 40 centers performed surveillance biopsies. So you can see you need the biopsy to calculate the IBOX score, but only 40 of 106 centers performed surveillance biopsies. Okay? You can see that the 40% is made up of the blue pie and the red pie. 17% do surveillance biopsies in all transplant recipients, whereas 20% only do them in some selected subset. So maybe the higher risk. And is there assessment of that higher risk, really higher risk? Probably not, but those are the realities. You can see the percentage that are being performed at the various time points. 60% of those that were performing biopsies performed one or two, whereas 28% performed three, and a small number of centers performed more than three biopsy. And what were the common reasons that people did not do biopsies? Number one, low yield, meaning didn't change anything, didn't find anything out, and or number two, no impact on outcome. So I think even when the kidney world is pretty far ahead of us, the surveillance biopsy practices are really lagging behind. So I think I just want to close by saying what does the liver community need? What does the liver physician need? And what does really the liver patient need? We first have to decide, is there a problem with the current approach? And is the problem a problem with the graft or is it with the patient? Do we have sufficient data? I would argue to say yes, but maybe some people are still asking me, show me why immunosuppression is really bad, prove that to me. Too much immunosuppression is deleterious. Do we have data to show that too little immunosuppression is deleterious? And I think the answer to that is yes, but that's my opinion. I think we have to ask ourselves, which direction of error is worse? And is that the same for adults and for kids? And is that the same for Jane or for Joe? You know, patient by patient. Do we need other parameters of liver health beyond liver tests and liver biopsy to guide long-term immunosuppression management? And I think if we feel that what we have is not enough, then we really need to work together. Because unless we work together, like the kidney transplant community is, we're not going to get to precision immunosuppression. That's it. Thank you. We can have Deidre here too and we'll open this part of the program to discussion. So, Lady Deidre, I just came up here to call you that, by the way. You didn't show any data on patient or graft survival with fibrosis. Does fibrosis have any impact on patient graft survival? That's a really good question and what we found, but I didn't present the data because it's just still being evaluated, is it looks as though after the 15-year, there's a stability. But in 10 of those patients, they lost their grafts. But we hadn't completed the connections. That's why I didn't show it. So I think the answer is yes. I think in the long run, it may well influence outcome. Jack, they looked at the SRT, either the UNOS OPTN or the, I think it was the UNOS OPTN, and it seems as if beginning around 15 to 20 years, there's a distinct change in the slope of graft survival. So I think this is yet another problem where we might be nearing, you know, beginning to figure this out because we have enough patients that are getting to 20, 25 years. But I think it's hard to understand if you have progressive fibrosis, that it won't be damaging. So the question is the stability. And Sandy, I had a question for you and maybe Jackie cover this in the talk. I unfortunately had to step out. But I always see elephant in the room when you talk about DSA as noncompliance. And it could fit with even some of your data on the very low TAC levels and your data as well. Do you know how much of the de novo DSA that occurs is related to noncompliance? No. No is the answer. Because we haven't been doing DSA routinely in the UK because of the cost. They won't agree to do it as a clinical test. So we're doing it as a research study. But it will be associated with the protocol biopsies. And because there's just a hundred transplants done in the UK a year, it will take a while before we get meaningful data. I would actually object to the word noncompliance because about 50% of the patients who are coming into IWITH have DSA. And I think that they have it because they are being run at very low trough TAC levels. They are the most compliant patients on earth because their moms and dads and guardians are bringing them to the clinic or you know to the lab every two weeks for liver tests. So I think that it reflects low CNI exposure. And I think withdrawal then precipitates DSA in another subset of patients. So I think there is zero question that if you lower CNIs beyond a certain degree and I would say we can look at the kidney transplant world and use that threshold because I think there's nothing different about the immune system between the liver and a kidney recipient. It's just what's different is how the graft is responding to it. I think it's not non-compliance but it's as much running at low CNI levels. I think the results of your recent analysis will be extremely valuable because the SRTR data that you referred to, we don't have the knowledge of subclinical inflammation which again goes back to whether they're kept on minimal immunosuppression, suboptimal immunosuppression. But you have the data and you have the luxury of the data that you had previously with people who were off steroids. But now the second era is with CNIs and steroids and you show clearly that the subclinical inflammation is gone. It will be extremely valuable to see whether the fibrosis in those patients will progress after the first 10-15 years or is it just the lack of the inflammation will just prevent that? Well in the cohort that I showed you, they're the same patients who had it at 5 and 10 years and that didn't seem to prevent the graft fibrosis. Continuing the steroids got rid of the inflammation but it didn't significantly reduce the incidence of graft fibrosis at 10 years. We've got another cohort which we haven't studied yet who've never had steroids, who've had a different immunosuppression regime with MMF or azathioprine instead of steroids and we'll soon have enough of them at five years to see whether it made any difference whether it's just immunosuppression or not. We look forward to that. That'll be very helpful. Thanks for those great talks. I think one of the first things I ever learned in transplant was there's only two levels of immunosuppression, too high and too low. Each of these studies have been looking at sort of single outcomes. Do you have any work or suggestions on how we can develop a composite outcome that embraces all those bad things both from too much and too little? Do you mean all the the side effects as well as all the immunological parameters? Yeah and the growth effects of steroids which were demonstrated early and the neurocognitive issues and really looking for a composite outcome that can try and balance what is almost an unbalanceable conundrum. I mean I think if we could work with machine learning company to put all these items in because we've got a lot of the retrospective data we've over 500 patients. I mean we had 365 that were had the 10-year biopsies but we've got 500 patients across Europe and North America. If we were able to interest someone to put all those data in all the side effects as well as some of the immunological ones we might come up with some some factors. It's something that I've got that I'd like to do but you need to have the right people with you and the right questions and the money. Yeah I am Estella may want to speak to this you know we had a much much bigger or Estella had a much bigger package of neurocognitive tests all kinds of stuff with the I with trial. But number one you know the trial itself is expensive. Number two neither not I mean NIAID just didn't see that as part of their mission right and I'm think DK also doesn't really see that as part of their mission. So it's very hard to have a trial which has all those dimensions because the funders just lop things off which don't address their primary mission. And we had the same experience I'm sorry Estella I know you were in in the UK because we were putting together a grant which looked at neurocognitive outcomes physical outcomes graft injury and look at some of the mechanisms and they said thank you very much we're not interested in funding epidemiological studies we want something that's going to change how we manage our health service which was kind of disappointing. But well I would follow up on that idea though I think what Simon was suggesting I'm not sure we would ever be able to get to that point because although you could load in probability of diabetes probability of hypertension probability of coronary artery disease when I'm 50 years old the individual's preference for those different diseases is going to be very different so that one person may feel that diabetes is the scariest thing and another person may feel that something else is the scariest thing. So I think in order to really achieve precision medicine we have to pull in what the patients and the families want as well. It was interesting the original comment I was going to make was about you know the again the Goldilocks hypothesis what is just right immunosuppression and what was very impressive during this trial during the IWITH trial was that where the patients tended to fail was not when they were off completely and not when you made the first couple of jumps down but when you made jumps that got them to about between 50 and 25 percent of their original immunosuppression and just out of convention trial and error we've for more than 15 years adopted this approach where if at four or five years after transplant you've got a pristine looking biopsy and your liver enzymes have always been normal maybe spare a little rejection in the first six months we'll try you on once a day one dose of TAC at nighttime with a trough level that's detectable the following morning and then follow you with enzymes and a follow-up liver biopsy and many of those patients do do well there's no question that they're probably at higher risk to develop DSA and they're definitely at higher risk to get in trouble if they become marginally non-adherent so that when they start this process of once a day that's once a day that's not five days a week okay and so I think that if you're taking twice a day and you miss two doses out of those 14 doses it may really be very different and I think what we need to do in thinking about precision medicine for these patients is think interval to interval what's right for them now they may be doing well on minimized immunosuppression when they're seven or eight or nine where there's very careful supervision that it's seven doses per week given at precisely the same times and they may need to go back on a different drug like an extended release to chrolimus or tacrolimus with an anti metabolite if you're worried about their kidneys when they run into a period during adolescence and young adulthood where life is busy and it's easy to forget well on behalf of both six Vena and I thank you all for being here for excellent presentations from presenters

precision immunosuppression

organ transplant recipients

immunosuppression balance

individualized approaches

immunological risk assessment

drug dosages

graft survival

machine learning