Dr Matthew Wilson - Update on and management of diffuse large B-cell lymphoma

Foreign. [Music] Academy for asking me to do this. This webinar today. Uh, my name is Matt Wilson. I'm a consultant hematologist in Glasgow in Scotland. And so I'm going to be covering, uh, the management of diffuse large B cell lymphoma in in 2022. Uh, these are my disclosures. So this is what I'm going to try and cover today, probably, um, over the next 40 minutes or so, 40, 45 minutes. Um, it's obviously a very big topic. Um, and what I'm, what I've tried to do is really focus on, uh, the most relevant areas for decision making, uh, when we see patients with with diffuse large B cell lymphoma, and focusing on perhaps some of the more recent updates over the last year or two. But we're going to go through it in a kind of stepwise manner. I'm going to start with some updates on the classification, which, um, is often quite confusing, and there have been some updates and, uh, in that, as we'll, we'll see shortly. Um, and then we'll spend a wee bit of time just looking at how you work patients up with this lymphoma subtype, how they're staged, and on prognostic models, because that is very important, um, in terms of how we treat people. And then we will move on to treatment, looking and focusing probably mainly in front line management, and, and within that, will include specific scenarios such as early stage disease, CNS prophylaxis, which is always a contentious topic in this, uh, in this subject, and management of older patients. And then at the end, we'll finish with some, uh, discussion on relapse refractory DLBCL, because that's a very fast-moving area with some potential change in the next few years. So, attention to introduce, uh, DLBCL. We know that this is the most common subtype of non-Hodgkin lymphoma. It comprises, and we're around about a quarter of all NHL cases. Um, and as with many of the former subtypes that we treat, there is a steadily increasing incidence with age, as you can see in this figure at the bottom from the HMR, HMRN and statistics, showing a very steady, uh, increase in incidence. Um, and within median age of diagnosis being about 60 years old. And obviously, that's very relevant because, and that has to be factored into our decision making for for treatment. Or it can result as a transformation from an indolent lymphoma, such as follicular lymphoma, CLL, where it would be called a Richter transformation, or marginal zone lymphoma. And as we'll see, in the DLBCL, typically presents a very aggressive manner, and, and there is an urgent need for therapy, and there really isn't a rule for for watchful waiting, as we sometimes see in other, uh, more indolent lymphomas. So, as I said, I wanted to start with updates on the classification of DLBCL, because I think this does cause confusion, and, um, I would argue maybe that as a field, it has been perhaps made a little bit more confusing than it needs to be with some recent updates. So, for the last few years, we've, we've been working from The WHO 2016 classification of lymphoid neoplasms, and this table is the, the, uh, category basically of high grade B cells from WHO 2016. Um, and what you'll see is that DLBCL, um, at the top here, there's it is a specific category in its own right, DLBCL NOS, um, but within that, in WHO 2016, there was a recommendation that it should be subclassified further according to the cell of origin, that's germinal center, ABC subtype, and I'm going to talk a wee bit about that in a minute, and, and they suggested that that should be done using an immunohistochemical algorithm, and, and also that the core expression of Mick and BCL2 should be considered, um, as a new prognostic marker, this so-called double expressor lymphoma. And because that's recognized as a problem of CIS histology, but you can see as well that obviously there are many other subtypes of of large B cell lymphomas, many of which, to be honest, are treated in a very similar way to DLBCL NOS. Notable exceptions, I suppose, being primary mediastinal B cell lymphoma, primary CNS lymphoma, plasmablastic lymphoma, and, but a lot of these subtypes actually, you would really encompass within the kind of umbrella of the DLBCL. And obviously, that wouldn't include Burkitt lymphoma. Um, and we're going to talk a little bit about double hit in a moment, because it's worth its own, uh, discussion. But this year, we've seen, um, an update to the WHO classification, the 2020, uh, two classification, which, um, is great, but unfortunately, it's been the waters have been muddied slightly, because at the same time, we have seen another classification, uh, published, um, from this International Consensus Classification group. Um, and this is not just for lymphoid neoplasms, but for all hematological malignancies. And this, as I say, has really created a bit of a fuss, and because the question is, which one do we use? Um, I would summarize by saying, we don't know yet. There is, this is an ongoing discussion. It is potentially problematic for both in the clinical setting, but also in terms of how we classify diseases, and, and for things like trial eligibility. But the good news is, for DLBCL, they're almost identical. I've had a good look across both the, the, the, the classifications, and really the changes are very similar. So, what here, I'm just going to summarize the changes in WHO5 or WHO 22. So, the first thing is, uh, on the right here, sorry, I should say, is this 2016, and on the left is the new classification. So, double hit lymphoma, previously, and was known as DLBCL with Mick, with Mick and BCL2, and or BCL6 rearrangements. The big change now is that the BCL6 rearrangement is no longer thought to be as significant or prognostic, and now it is really only if you have a DLBCL with a Mick and a BCL2 rearrangement. But then, this is a specific category in its own right, and that is it. That is probably the most significant change. Burkitt lymphoma with 11q aberration is now called High Grade B cell lymphoma with 11q aberrations, reflecting that it is probably a slightly different disease to Burkitt lymphoma, although shares many of the same characteristics histologically. Um, fibrin associated large B cell lymphoma is a new category, which is previously called lymphoma associated with chronic inflammation. And there's this new umbrella term of primary large B cell lymphoma of immune privileged sites. And this is an umbrella term again, which encompasses, uh, primary CNS lymphoma, primary testicular lymphoma, and mediastinal retinal lymphoma. Now, you could argue it's not overly helpful, because of course, these subtypes are, are certainly treated differently. But I think the, the classification is really to reflect that they have a very common, they certainly have a common molecular and biological basis. Um, and mediastinal gray zone lymphoma, um, is now what used to be DLBCL, and classical with features overlapping between the two. So, I think that's probably what most people have been calling this subset, this subtype, uh, anyway. So, I think it's just worth quickly, um, addressing this DLBCL, high grade BCL, Burkitt, double hit, what's the difference? It can be very confusing, and this is quite a helpful diagram, I think, from the, from the WHO classification, which explains it a little bit further. The R here stands for rearrangement, and the, and the G stands for germline, so essentially no evidence of rearrangement. So, the first classification is actually made on a morphological basis, and, and, uh, B cell lymphoma has its own typical morphology, which is different to Burkitt lymphoma, and high grade B cell lymphoma, which have a very similar, but, of a very similar histology. But in terms of DLBCL, if you have a Mick rearrangement, as I said, and now the only time, the only way you can be categorized is having a double hit lymphoma is if you have a Mick and a BCL2 rearrangement, and that can have been with or without a BCL6 rearrangement. But if you only have Mick and BCL6, that is not double hit lymphoma, that is just simply called, if it starts B cell lymphoma NOS. In terms of Burkitt lymphoma, and again, there's sometimes there's confusion about this, because of course, Burkitt lymphoma is categorized by the presence of a Mick rearrangement, but in order to be called Burkitt lymphoma, you have to have histology that is in keeping with Burkitt lymphoma, you have to have a Mick rearrangement, but importantly, without a BCL2 or BCL6 rearrangement. And, and if that, if all those criteria are met, then it is Burkitt lymphoma. And some really high grade B cell lymphoma, and which may have similar histology to Burkitt lymphoma, if you have a Mick rearrangement, then, um, it can be high grade B cell lymphoma with Mick and BCL2, if that's present as well, or it can just be high grade B cell lymphoma NOS. And as we said, um, high grade B cell lymphoma with 11q aberration is an entity within its own right. So, we mentioned cell of origin, which previously was thought to be something pretty significant in DLBCL in terms of prognostication and potentially guiding management. And the recommendation from 2016 was that we should classify it according to the an immunohistochemical algorithm. And the Hans algorithm here is showing is the most frequently used, using CD10, BCL6, and 112 to identify cases as either germinal center or non-germinal center subtype. But we know that this can be done in a much more accurate way now using gene expression profiling. And this is a, um, uh, this diagram summarizes this one and shows that, um, cases can be categorized as either ABC or GCB subtype based on their very, uh, individual gene expression profile. And the other thing we know is that when, when done using gene expression profiling, it is very clearly prognostic. As a starter from the GOYA prospective trial, showing a significantly worse, uh, outcome in patients with ABC subtype. However, although we now know cell of origin is prognostic, and unfortunately, we haven't really seen a therapy that has been able to overcome that adverse prognosis, despite many attempts, and we're going to discuss those later in the talk. The other issue is that if we just use the Hans algorithm or the immunohistochemical algorithm, this is really, it is not particularly precise. It's not that well reproducible between pathologists, and it also results in quite a few cases which can't really be neatly put into one category or the other. The other, so gene expression profiling is better, but it's not routinely available in clinical practice. So, whether or not this actually becomes an important part of our subclassification going forward, I think remains to be seen. And finally, just on the classification of DLBCL, the, the last few years have seen some very exciting publications that have really delved even further into the molecular taxonomy of DLBCL. And these are the kind of seminal publications, and that were very, uh, that were published very close together in Nature and New England Journal of Medicine, um, and these two labs, and actually there's been in another group since, created this new classification of DLBCL based on, uh, sequencing of many DLBCL tumor samples, and they were able to categorize, uh, different genetic subtypes based on the, on recurrent mutations, uh, seen between samples. Now, the different labs produced slightly different, uh, categories of subtypes, but there was a lot of similarity between them. And I think again, going forward, as a field, we probably need to come up with a combined and agreed way to, to subclassify these different genetic subtypes. But on the right here, you can see one example, whereas before they may have been, uh, categorized only according to their cell of origin, now we have these various different genetic subtypes, which have been shown to be prognostic. But the other important feature is that they contain specific mutations, which are potentially targetable by by novel agents, and that's got to be an area that we focus. And indeed, this is going to be incorporated into prospective trials going forward, seeing whether we can, can adopt a more personalized approach to DLBCL, and, and not just treat it as one entity. So, I think that's got to be an area of excitement going forward. So, what do you do when you see a patient with DLBCL? I don't think it's worth covering this quickly. Um, we probably all know this now. I've got a core biopsy, really is essential. And FNA, and will very frequently not get you the diagnosis, and your pathologist will not be happy, and if that is sent. So, a minimum of core biopsy, and an excision is is always preferable. Um, and thereafter, an immunohistochemical panel will be run. And this is an example of one, which allows, uh, a classification of, uh, of cell of origin, and also allows them classification of whether it can BCL2 expression is there. FISH is then performed on, as we've discussed, and that has been to identify whether or not this is a, a double or previously triple hit lymphoma, although I would argue that perhaps BCL6 rearrangement is is no longer going to be, uh, really important, and really it's BCL2 rearrangements that are, are important. Tonight, um, security team floods, obviously should include an LDH and a urate, thinking about tumor lysis risk, and also bloodborne virus screen, because DLBCL can occur, as a, as a manifestation of bloodborne viruses, such as HIV, but also, and this is important for, uh, looking for potential reactivation with, with treatment. I think we've all, we've all agreed the PET CT is a gold standard for baseline imaging, and DLBCL, and certainly upstages quite a few patients who, uh, may have had CT only. And whereas previously bone marrow biopsy was recommended for full staging, it's not fairly widely accepted that that's not required if a PET CT is performed, because that typically picks up marrow involvement. And echocardiograms are needed if planning out to cyclophosphamide therapy. And CNS evaluation, we're going to come on to CNS prophylaxis later in the talk, but I think this is something we probably should be focusing a little bit more on than previously. Certainly, if patients have signs or symptoms of central nervous system disease, but even if patients have high risk features, we're having CNS relapse, and potentially we should be evaluating them. That'd be a slight, and we'll discuss that later. I think it is really important just to, to reiterate the staging of DLBCL, because as we'll discuss, there are differences in the way you approach early and late stage. And just as a quick reminder, using the Ann Arbor staging, stage one is one single lymphoma region only, um, and that can include the spleen. Stage two is two or more, but on the same side of the diaphragm. Stage three is involved on both sides of the diaphragm, and stage four isn't as, um, at both sides of diaphragm with involvement of an extranodal site. Now, sometimes it gets a bit confusing, because if you are an earlier stage and you have one single extranodal site contiguous or beside an extranodal site, then that would be delineated stage one, two or three E. But if the extranodal site is not contiguous with an extranodal site, or there are multiple extranodal sites, then it's going to be classified as stage four. And I think that it is just important to keep that in mind and make sure that's done correctly at the outset. And how do we prognosticate in DLBCL? So, this is quite a busy slide, but it's really just to present a summary of the different prognostic systems we have in this disease now. And the IPI was the first that was developed, and just in the top of this table, so this is actually in the, in the previous era, but actually, has been shown to really still actually be prognostic even in this reduction era. And the revised IPI, um, has demonstrated that. So, these, the IPI and the revised IPI, um, uh, include age, LDH, stage, ECOG, and number of extranodal sites. And you can see very clearly a worsening, uh, prognosis for patients with more risk factors. The age-adjusted IPI is was brought in for patients, younger patients with DLBCL, really to, to remove age as a factor, and to look at other specific high-risk features. And, and this is incorporated into some prospective trials, which target this, this subgroup specifically. And the NCCN IPI at the bottom is probably the most recent, which takes into account some of the kind of continuous nature of some of the variables, such as LDH and age, and also references, um, to specific adverse prognosis of some particular extranodal involvement, as you can see here. So, IPI, to be honest, is still the most widely used, and arguably is still valid, but, but the others can, can also be used, and, and are often reported in trials. So, now we're going to move on to DLBCL management, um, and you might argue that this could be done in one slide and say that this is R-CHOP, and that's the end of the story. And, and hopefully, what I'll show you today is that that might be true in part, but I think that, um, there's, there's a lot more to it, and things are potentially changing in the next few years. I think it is really important to think about where R-CHOP came from, because we are still using it for the vast majority of our patients, and probably will still continue to use it for the coming years. Um, and just to, to remind everyone that this was, um, where the, the CHOP regimen was demonstrated to be superior, um, for, for DLBCL. This is prior to it took some upcoming loans. This is in 1993, and CHOP was compared to other more intensive, um, chemotherapy regimens here, and you can see that absolutely no, no benefit with the more intensive regimens was seen in terms of PFS or OS, and importantly, as well, there were increased treatment related deaths with some more intensive regimens. So, that established CHOP as the best chemotherapy backbone for this disease in terms of balancing efficacy, toxicity. And then we had this seminal paper in 2002 from Benz-Rohn, Coffee, and which was a randomized controlled trial of CHOP versus R-CHOP. This is where rituximab now came along, essentially. And it's really, it's really staggering, even still, I think, to look at the magnitude of benefit that we see with R-CHOP versus CHOP in terms of progression-free survival, and overall survival, complete response rates of 76% versus 63%, and overall survival of 70% versus 57%, two years. And that has established R-CHOP as the, as the king of chemotherapy, immunotherapy, and this disease. And I would argue, we will probably never see an incremental benefit with anything that we do going forward, um, compared to this. But of course, as a field, we have continued to try to improve on R-CHOP, um, in many ways, and we're going to discuss some of that. And one of the, the first ways to try and do this was to see whether intensification of R-CHOP dosing could potentially, uh, improve survival for patients. And here's a, this is a summary of the, some trials that have looked into this. So, and without going through all the details, um, within the, you can see I've put a summary column here for each study, and I'll just go through that. The top, on the LNH032B study, is actually the only trial until very recently, which did actually show a survival benefit with the regimen R-ACVBP over R-CHOP. However, this is a really pretty toxic regimen, very dose intense, and significant increase in toxicity. The R-CHOP arm in this study performed quite poorly, and this has not been widely adopted, and because mainly because of the, of the additional toxicity seen. The NCIC NCI R-CHOP 14 study, then looked to see whether or not, get intensifying R-CHOP by giving it more frequently over 14 day instead of 21 day cycles would improve, would improve outcomes. But this showed that giving it over 21 days was not inferior to giving it over 14. And giving R-CHOP over 14 day intervals is actually pretty challenging for patients, and, and can be quite toxic. So, um, that, as well as the next study, which was, uh, the French GELA study, also demonstrated that R-21 was not superior to R-CHOP 14, and that is why we rarely would use R-CHOP 14 nowadays. These studies, you often used eight cycles, and that was, as you might have noticed in the original CHOP or R-CHOP papers, that's what was, uh, traditionally given. But the RICOVER-60 trial, which was the end of patients over the age of 60, subsequently demonstrated that giving only six cycles was not inferior to giving eight cycles. And finally, giving an autologous transplant in first CR was postulated as something that might improve disease control. But as you might expect, it, it did show some improvement in terms of reducing relapse, but again, a significant additive toxicity and no overall survival benefit. So, in summary, all of these trials have really established that giving six cycles of R-CHOP 21 is standard of care for advanced stage, uh, DLBCL. The other way that we've tried to improve on R-CHOP has been with the regimen DA-EPOCH-R. And there's always a lot of discussion about this, and I think is where they just have a slide by itself. So, just to, to summarize what the difference is between these regimens, and you can see the, the, this is summarized in the table. The rationale behind DA-EPOCH-R is that, um, some of the drugs, like vincristine, are given as continuous intravenous infusions instead of boluses. And the rationale being that that might catch more cells that are in cell cycle at any one time. That also adds a to EPOCH, which is not in R-CHOP. And the, the other key feature of this regimen is that the dose is escalated using the neutropenic index as cycle one and onwards, as a surrogate marker of, uh, how the patient is, uh, processing and eliminating the chemotherapy. So, in theory, it provides a more, um, personalized approach to, to ensuring sufficient chemotherapy potency. But a lot of this was demonstrated to potentially be pretty effective in some single arm studies, and particularly for higher risk patients. It out there has not been this randomized trial of six cycles of DA-EPOCH-R versus six R-CHOP published, um, uh, uh, three years ago now, and this showed no improvement in progression-free survival and no improvement in overall survival. And you can see these lines, you could almost draw one on top of the, on top of each other. So, convincingly no overall benefit. Now, in the bottom left, you can see they did do subgroup analyses, and, and there was a suggestion that those with, that with high risk features or a high IPI, and potentially benefited from DA-EPOCH-R, but of course, we have to be very careful with these subgroup analyses in terms of the very small numbers involved, and it wasn't powered specifically to look at this group. So, take home message, no benefit in terms of PFS or OS overall. May be a benefit for high risk patients, but the one thing we do know from this study is that it's more toxic. The rates of infection, mucositis, and peripheral neuropathy were significantly higher with DA-EPOCH-R. So, personally, my take on this is that it could perhaps be argued for selected patients who have high risk features, but we certainly know that the toxicity is increased, and that's something that has to be carefully considered on an individual patient basis. So, more recently, instead of, you know, we've tried to improve on the intensification of the R-CHOP dosing itself, and that hasn't really worked. So, the focus over the last five to ten years has been incorporation of novel agents to the, to an R-CHOP backbone. And again, I've got a summary of the main studies here. So, the CORAL study, them at the top, this, this used a more potent anti-CD20 antibody, obinutuzumab, with R-CHOP and compared it to rituximab with R-CHOP. Um, and this was a negative study and showed no improvement in progression-free overall survival. So, obinutuzumab is not used in this disease. The REMODEL B trial used gene expression profiling upfront to, to classify patients as ABC or germinal center, and then randomized patients to receive obinutuzumab with their R-CHOP if they were ABC subtype. Now, at the time that I wrote this slide, this was as reported a few years ago, a negative study with no PFS or OS benefit. But very interestingly, we're going to see Dr. Ash from extended follow-up, which has potentially suggested a survival benefit with the obinutuzumab. And I think that will be very interesting to see the detail of the PHOENIX study and incorporated a bridge in it with R-CHOP. And again, was a negative trial. There was a suggestion towards benefit in younger patients who could potentially tolerate the bridge in a better, but again, this has not been widely adopted because of the lack of clear benefit. And the ROBUST study was the probably the largest study, I'm incorporating lenalidomide with R-CHOP, again for patients of the ABC subtype. Um, and, um, unfortunately, and disappointingly, again, no benefit in terms of survival seen. And finally, what we have seen in the last year is the POLARIS study, which is R-CHOP with polatuzumab vedotin, with the vincristine and as we're going to see in the next slide, this has been really the first study to date that has shown any form of benefit over R-CHOP alone. So, is this really an improvement over R-CHOP? This is the design of the study, nearly 900 patients, and notably aged between 18 and 80. And they were relatively high risk in terms of nearly two-thirds having an IPI of three to five. They were randomized either R-CHOP plus placebo, or so-called POLAR-CHOP, so R-CHOP with polatuzumab vedotin, and the vincristine omitted. Firstly, it was deemed to be safe with really equivalent rates of, um, of toxicities between the two arms. And peripheral neuropathy, in particular, would have been an interest with polatuzumab, but that really didn't seem to be increased. And as I said, in the last slide, there was actually a significant PFS benefit, 76.7% versus 70%, and that was, uh, statistically significant. However, no overall survival benefit. So, there has been a lot of discussion about this over the last year. My take-home messages would be, yes, there is an improvement in PFS with POLAR-CHOP, but there is no OS benefit. No, why that is? You could argue that probably having improved therapies in the relapse refractory setting may have influenced this lack of OS benefit, and maybe actually with more long-term follow-up, an overall survival benefit may emerge. But this is a quite a significant increase in cost with the additional polatuzumab vedotin, and I think that, as well as the lack of OS benefit, and may limit its kind of widespread adoption. But I'm sure it will be reimbursed in, in many countries going forward. Uh, I think it was important to think about double hit lymphoma separately as well, because this is another topic that is frequently debated about whether we should be treating these cases differently. So, again, just to highlight double hit lymphoma, and this left figure is perhaps slightly out of date now with the new classification, because as I said, BCL6 rearrangements are no longer really deemed to be, uh, of of particular importance. It's really about BCL2 rearrangements with Mick. Um, and this comprises of around about five to eight percent of DLBCL. And one thing we know for sure is that overall, these patients unfortunately have poor outcomes with R-CHOP. And this data from a retrospective analysis of patients receiving a number of different, uh, chemotherapy regimens clearly demonstrates that none of these regimens do particularly well. With R-CHOP, in particular, does not appear to be, uh, very effective. There's been a really interesting study from the Lunenburg Consortium that have looked at the drilled down into the kind of which patients within this double hit umbrella are actually the ones who have the worst prognosis. The first thing they demonstrated with this, in this bottom left curve here, is that if you only have a Mick rearrangement and not with a concurrent BCL2 rearrangement, your prognosis is slightly worse, but certainly not as bad as if you're a true double hit. The other thing that is interesting to note is that within the, this group of double hit lymphoma patients, some actually do well. And those patients who present with low risk features according to IPI actually have outcomes, um, similar to non-double hit DLBCL. So, it is a heterogeneous group, and many of them present with these high risk features and high IPI features, and we know their outcomes are poor. But arguably, if a patient presents without these high risk features, just because they're double hit, does not necessarily mean that their prognosis is going to be as bad as might be feared. And the other thing we now know from the Lunenburg Consortium is that the Mick partner gene matters. You can see here, this bottom line, that is clearly far inferior in terms of survival to the others. This is patients who have a Mick translocation with an immunoglobulin gene partner, and those cases which are double hit with Mick translocated against a different partner don't appear to have an adverse prognosis. So, I think this is something that can be factored into our decision making when we consider whether we should be giving escalated therapy to our patients. But arguably, it isn't really done a, a routine clinical practice. Okay, and what can we do if we do want to intensify treatment? Well, DA-EPOCH-R is one option, and as we've discussed this before, in terms of the PS3 study, which had very small numbers of double hit patients, but this was a phase two study that we're seeing, um, in patients with a Mick rearrangement, and within that, there were nearly 60 that had double hit DLBCL, and this study did show a fairly promising event-free overall survival of over 70% for what was deemed to be a pretty high risk patient group. And, and as a result of that, many centers would actually, and particularly in the US, many centers would give DA-EPOCH-R for patients with double hit DLBCL. In the UK, we probably use a little bit more of the R-CODOX-M/IVAC regimen in this setting, largely based on the UK NCRI phase two trial, not just of double hit lymphoma, but of high risk DLBCLs when IPI score four or more. And within that, there was a relatively small proportion of patients with double hit lymphoma, but again, for this very high risk patient group, the survival figures seen were very promising, although it is pretty toxic. Treatment related mortality of 4%, and that, and all patients who, who had a treatment related death were over the age of 50 with the performance status. And arguably, this is not a therapy that should be given to patients who fall into that category due to toxicity. And certainly, in our center, for patients with very high risk features, high IPI, with or without double hit, uh, cytogenetics, we would consider this regimen in somebody who's young and fit enough. So, I'm not going to be able to talk about CNS prophylaxis, which, as I said, at the beginning, is, is really just an area that probably gets debated most in DLBCL management. Um, and we've seen an increasing data over the last year or so, but unfortunately, as we'll see, that I don't think the question is, is definitively answered. The thing that we must remember is that CNS relapse has a very poor prognosis when it occurs in this disease. You can see here from this retrospective analysis of secondary CNS lymphoma, the median overall survival, and was less than four months. And we know I have doctor from the MARIETTA, which is a prospective trial of secondary CNS lymphoma, which showed significant improvement in outcomes compared to historical controls. But when you look at the subset of MARIETTA, this red line is patients who were enrolled where they had previously had R-CHOP, and then developed CNS relapse. Now, that is the group that we are targeting when we think about CNS prophylaxis, and we can see that that group really does particularly poorly with a 28% PFS. So, this is why there's always been this desire, and it should always be this desire, to try and prevent this pretty devastating complication. The CNS IPI has been a step forward in terms of how we stratify patients for this complication. Um, and there's a summary which shows that when we calculate a score out of six for patients, which takes into account the five traditional IPI risk factors, as well as kidney or gonadal involvement, there's a very clear increase in risk of a CNS relapse with each additional point. So, as a result, this high risk group of those with a score of four to six, and has often been targeted as a group requiring prophylactic intervention. But what I would point out here, on the right, is that you can see that if you only limit your prophylaxis to those of the high CNS IPI, you're actually missing quite a significant proportion of relapses occurring out with this group. So, this score does not have the positive predictive value that we really need to, to adequately target, uh, a true high risk group. And the other point is that if you have a high CNS IPI, as a, as a result, you automatically have a high IPI and a high risk of systemic treatment failure. And that risk is actually greater than the risk of CNS relapse. So, this is another important theme, which is not compromising your systemic treatment. So, we, we produced this UK good practice paper a couple of years ago now, and really trying to give some pragmatic guidance on what is an area with very poor quality of evidence. And in this, in this guideline, we recommended delivering CNS prophylaxis to high CNS IPI patients, as well as those with multiple extranodal sites, and particular high risk extranodal sites, and high dose methotrexate. It was recommended, although it wasn't really clear exactly how much you needed to give or when you needed to give it. But since this was published, there's now been increasing evidence to question the use of high dose methotrexate. We're going to see, and this guidance will, I think, have to be revised in the next year. So, so we carried out this study, which was looking at the timing of high dose methotrexate, whether you delivered it in between cycles of R-CHOP, so-called intercalated, or delivering it at the end of treatment. And after a UK study where we showed increase in R-CHOP interruption and delay, a with a intercalated approach, we then carried out this larger international study, really trying to see was there any difference in CNS relapse between the two groups. And then briefly find no difference. We saw that there is no reduction in CNS relapse delivering intercalated high dose methotrexate compared to compared to end of treatment. And we had data over 1300 patients treated with high dose methotrexate. But the other important finding was that our overall rate of CNS relapse, and our high CNS IPI group, which is over 600 patients, it was 9%. And that was despite the use of high dose methotrexate. So, we've demonstrated the time it doesn't matter, but actually, is the bigger question, does it actually do anything at all? And there has been an additional, uh, study addressing this particular question, presented by Kat Lewis and colleagues about last year. And this was a large, again, retrospective analysis of high risk patients, as you can see according to these criteria on the slide, who either did or did not receive high dose methotrexate. And they had a pre-planned power calculation to, to give them sufficient power to detect a significant reduction in CNS relapse with high dose methotrexate. Although they had, they had started over 2000 patients, it's notable that they only had about 400, uh, 480 patients who received high dose methotrexate. However, accepting that caveat, they find no statistically significant benefit in terms of reduction CNS relapse with high dose methotrexate, um, compared to those who did not receive it. So, there's again, there's been a lot of discussion about this data, and the full paper is going to be published soon. But my take on how we need to beat this, uh, problem into 2022, I think that, um, you could argue that maybe there isn't a rule for high dose methotrexate at all. My personal feeling is that we don't yet know with enough certainty that there are not some patients who would benefit from it, those particular high risk groups, so those are the score of five or six, and those with multiple extranodal sites. And I think even the two studies that we saw, and I don't think are sufficient power to say there's no benefit there. And maybe we just need to be more restrictive in our use of methotrexate. I think we need to focus more on trying to identify patients upfront who have occult CNS disease, because those patients, I think giving them CNS prophylaxis really isn't, um, enough. I think they probably need intensified treatment. And we've seen a lot of interesting data looking at CSF circulating tumor DNA, which may improve our ability to protect these patients with occult disease. Our study showed that certainly, if you're going to give it, I think it should only be given at the end of treatment, because you do not want to compromise your systemic therapy. I think there's only really a rule for it's I've not really mentioned in the interest of time, but my personal feeling is that the, the evidence would suggest it's really not effective, and probably only has a rule in to stick your DLBCL, because there's prospective trial adapter in that subtype, uh, incorporating IT therapy. And I think we need to think, we need to think beyond high dose methotrexate now, and think about the use of novel agents. And there are some studies ongoing, such as REMODEL A, which may be interesting in that regard. So, next, we're going to move on to talk about early stage DLBCL, because as we'll see, this should be considered, uh, separately to advanced stage in some cases. About 25 to 30% of DLBCL cases are actually localized, and, and the important thing is that these patients do very well. They have a very good long-term survival, and as we see in other lymphomas, such as Hodgkin lymphoma, actually a lot of the difficulty comes in making sure we're not over treating these patients. The definition of early stage disease can vary, but typically at stage one to two disease with a a mass, as long as they don't have a mass that's more than 10 centimeters in diameter. And until recently, the two options were essentially to give combined modality therapy, giving abbreviated chemo with radiotherapy, or just to treat them as per advanced stage. And the rationale for a given combined modality comes from this, uh, study from the S8736 study, which is first published by, um, over 20 years ago, but more extended published more recently. But this demonstrated, as you can see on the left, initially, uh, an improvement in survival patients who had CHOP with radiotherapy versus CHOP alone. A little interestingly, with more extended follow-up, this progression-free survival benefit, um, appears to have been lost. But this is still a valid approach for patients with localized disease where radiotherapy is possible. But we've seen another couple of ways to approach this recently, and one is to, and one has been summarized in the OR ANALYZED in the FLYER study. So, this is published a couple of years ago now, and this was a randomized trial between, as you can see in the schema here, six cycles of R-CHOP as standard versus only four cycles of R-CHOP with two additional cycles that we're talking about. And importantly, this was for patients who fitted into very strict criteria, which was young, under the age of 60, early stage, stage one to two, with no age-adjusted IPI risk factors, or no bulk, so a very good prognosis group. And you can see here that the PFS and the OS were were not different between the two approaches. So, the take home messages from this are that giving this for, for R-CHOP was two, two cycles of rituximab alone appears to be non-inferior to giving six cycles of R-CHOP, and young patients with no additional risk factors. And of course, this spares the potential additional toxicity of more R-CHOP, and also of radiotherapy. And I think you could question what these exercises of rituximab are actually adding. And the other issue is that what we see now is probably moving towards more studies using interim PET in early stage disease, as we've seen in Hodgkin's lymphoma, and that wasn't incorporated into the FLYER, but I think this is a valid approach for this very particular subgroup of patients. Looking at this PET predictive approach, this is the, the largest study addressing this, uh, so far in DLBCL. The S1001 study, it was a phase two single arm trial of early stage DLBCL, and patients had three cycles of R-CHOP, followed by an interim PET. And if that was negative, they'd have one further R-CHOP and nothing else. If it was positive, they got radiotherapy as well as, uh, uh, radioimmunotherapy treatment. And what you can see here is that there was no difference in event-free survival or overall survival according to whether you were in the interim PET positive or negative group. And therefore, that would suggest giving four R-CHOP alone in patients who managed to achieve a negative PET after three cycles, again, it appears to be a valid approach, and, and similar in a way to the FLYER, to the FLYER findings. I don't think we really know what the role of the radioimmunotherapy is here for the positive group, and whether it was actually radiotherapy that has been, uh, that has been the important addition to this group. And there were slightly worse outcomes seen in high risk patients in this study. So, I take the jury is out as to whether if you've got a patient with some high risk features, even if they have a negative interim PET, is it really safe to omit radiotherapy, um, if it's a site that's very amenable, um, with potentially low toxicity, and I think that would still be debated. But this finally on early stage DLBCL, this is a very nice, um, algorithm, um, from Eliza Hawks and colleagues, written a really, uh, comprehensive review of this in Blood this year. And it really talks about either using a non-PET adapted approach or a PET adapted approach. And if you're not using a PET adapted approach, then on the left here, this is the FLYER approach, which can be considered for this low risk group. But of course, giving R-CHOP and radiotherapy is, as we have traditionally done, is still a very viable option, as is simply treating us advanced stage, if we are worried about certain high risk features, with or without any treatment radiotherapy. And then, as we've just seen in the previous trial, using a PET adapted approach, and giving only four cycles of R-CHOP with a negative interim PET is also valid. So, a lot of what I've discussed so far has been about, um, younger patients with DLBCL, and it is actually really important to think about older patients, because as we've seen, the median age, and this disease is over, well over the age of 60. And R-CHOP can be well tolerated for patients even up into their 70s. But, um, for patients over the age of 80, the, the evidence really becomes quite, uh, scant for how, how we should optimally manage them in terms of delivering chemotherapy. And this, so-called R-miniCHOP, does reduced R-CHOP, is frequently used for older patients. But this is not based on any randomized trials, based on one phase two prospective study, which did show promising results. But the difficulty with these patients is always that balance between toxicity and efficacy. Um, and really, you want to try and give them the best chance of cure, uh, upfront, but they're potentially less likely to tolerate therapy well, and it becomes pretty challenging. And I wanted to have one study, which I think has been informative in this area, which is this systematic review of R-CHOP dose intensity in older patients, published a couple of years ago. So, they, they did a systematic review of 13 studies and over 5000 patients. Um, and they, this did include, um, retrospective analyses and observational studies, as well as, as well as prospective trials. It must be, it must be said, what they found was that lower dose intensity of R-CHOP was associated with inferior survival overall. However, when you look to patients who were over the age of 80, these patients did not have worse survival by having reduced dose intensity. So, it's a complex trial, but my take-home message from this would be that generally speaking, higher dose intensity improves survival in those who are under the age of 80, and therefore, if possible, it is probably best to try and deliver full dose R-CHOP to these patients. Of course, if they're generally fit, and comorbidities don't dictate otherwise. But in, in if patients are over the age of 80, even if they're, uh, they're fit, the, the literature would suggest that using miniCHOP, as we have been doing, is probably the right, the right approach for this patient. So, as a result, our miniCHOP is good, is is used as the control arm for, for prospective trials going forward, looking at management of elderly DLBCL, which I think is really, really important, because this is, uh, uh, this is an expanding patient population where we, we often have these kind of dilemmas. And similar in a similar way, and there may be situations where, um, anthracycline chemotherapy can't be given, and that might be in older patients, but it might be in younger patients where, um, there's specific cardiac comorbidity which prohibits it. And that

There are two general approaches using this scenario. One is to use the regiment RCL. I wanted to highlight this study from a couple of years ago, which is interesting. This is a retrospective analysis from British Columbia of 70 patients receiving RCL, excluding contraindications, and they actually compared these to a case-matched control group who received R-CHOP. Now, you could very much criticize this as a statistical approach, and there are many caveats, but it did suggest that the outcomes of RCL were approaching equivalence with those who received R-CHOP. And I think what's important to see is that down in the bottom here, with RCL, with very long follow-up of 10 years, 58% disease-free survival, and I think that's really actually quite promising and suggests that this can be a curative approach for this group if you can't deliver an anthracycline.

And then in the UK, we've maybe used a little bit more of R-GCVP in a scenario based on this prospective, single-arm phase two trial from the NCRI group, again targeting patients who couldn't receive anthracycline and showing pretty promising survival outcomes with a clear plateau of the curve here, suggesting that this can be a curative approach. And so, I think either of these approaches are valid for patients who can't receive anthracycline.

So, I've covered frontline therapy and DLBCL and quite a few topics within that, but I think now we should move on to look at relapsed refractory DLBCL. And of course, this could be a topic within its own right, but I'm going to try and highlight some key themes from it and increasing data that we've seen over the last couple of years. So, this algorithm quite nicely summarizes our approach until recently for relapsed refractory disease. The first decision being whether or not patients are eligible for an autologous transplant. And if they are, then the patients proceed to this if they demonstrate chemosensitive disease or six-nine therapy. But if they do not, then what we can now do is put these patients forward for CAR T-cell therapy. Or the same could be said if they relapse after having an autologous transplant. And these patients, traditionally, outcomes were extremely poor, but CAR T has obviously been a big move forward. And without going into the detail of the seminal papers of axicabtagene ciloleucel and tisagenlecleucel, which are the approved products we use in the UK, and these results in a PFS of around 30 to 40%, which is much better than historical outcomes. So, CAR T-cell therapy is there for well-established as a third-line therapy, certainly in the UK.

I think we shouldn't forget about how important autologous transplant is and why we traditionally have tried to deliver this in the second line. And that is because it does improve survival and can actually cure quite a significant proportion of patients. And this paper from over 20 years ago now shows the very significant survival benefit with giving an autologous transplant after high-dose chemotherapy versus high-dose chemotherapy alone. And I think we really need to be mindful that this can cure quite a few patients, and I'm not sure that we're in a position to say it doesn't have a role.

Another question that frequently comes up is what second-line therapy do you give prior to an autologous transplant? And there have been a couple of prospective trials over the last few years looking at this. And the first was the CORAL study, which showed no difference in terms of survival between R-DHAP and R-ICE. And then the Ly12 study, which compared R-DHAP to R-GDP. This is of particular note because R-GDP is really easier to deliver and often delivered as an outpatient, and actually appeared to be equivalent to R-DHAP with less toxicity. So, as a result of this, certainly in the UK, R-GDP is often preferred as our second-line regiment. But in summary, really, any of these regimens can be considered, and there's often institutional preference, and there are additional ones not even shown here, such as R-ESHOP, which are frequently used.

But the question over the last year has been, what about replacing high-dose therapy and transplant with CAR T at the second line and moving CAR T forward in the algorithm? And there's been a lot of debate about these three studies, which were presented last year and have subsequently been published. Now, the schema here shows that the general structure of the trials is fairly similar. So, it's Zuma 7, which was investigating axicabtagene ciloleucel, Belinda with tisagenlecleucel, and TRANSFORM with lisocabtagene maraleucel. And basically, they all enrolled patients who had DLBCL who were either primary refractory or relapsed within one year of their first-line treatment, and that's an important point. And patients then got randomized either to have CAR T-cell therapy or to have standard of care, which is high-dose chemo and autologous transplant.

And I've summarized the outcomes in this table from these three studies. And maybe not going through them in a huge amount of detail, but the important things I wanted to highlight was that the Zuma 7 study did not allow bridging other than steroids only, whereas the other two did allow bridging, and that's been a criticism of Zuma 7 and potentially excluded some higher-risk patients as a result. They all used event-free survival as a primary endpoint, but there were differences in terms of how they defined event-free survival, and in particular, with the Belinda trial, that might have been part of the explanation as to why the results were not as good as in the other two studies. You can see that Zuma 7 has the longest median follow-up so far, and both Zuma 7 and TRANSFORM showed a significant improvement in both EFS and CR8 compared to standard of care. However, Belinda with tisagenlecleucel did not. The toxicity rates were probably higher with Zuma 7, as we've come to see with axicabtagene ciloleucel. And actually, I think that lisocabtagene maraleucel, although we can't access it in the UK, is probably the more exciting product out of the three because it appeared to have this same improvement compared to the control arm that has less toxicity than axicabtagene ciloleucel.

So, what are my take-home messages from this? Well, as I said, axicabtagene ciloleucel and lisocabtagene maraleucel have positive studies in terms of their primary endpoint. Tisagenlecleucel did not. We need to remember that this is only applicable to those patients enrolled who were primary refractory or relapsed within one year, and I don't think this data can be extrapolated to patients who relapse further down the line because I think there really is still a role for high-dose chemo and transplant for those patients. Importantly, we've not seen a clear overall survival benefit yet from these studies. Now, you could argue that that is perhaps because of crossover to receiving CAR T-cell therapy, either within or with the trial, and perhaps more follow-up there will be an overall survival benefit seen. I think we have to remember the cost implications are huge if we do move this forward in the therapeutic algorithm to the second line, and that I know is something that's being assessed in the UK with NICE appraisals at the minute. And I think we must remember that high-dose chemotherapy and autologous transplant is still a very effective and curative treatment for many patients if they demonstrate chemosensitive disease, and I don't think that we're going to be in a position of entirely replacing this therapy with CAR T, certainly in the near future.

So, what about other options in relapsed refractory DLBCL? And this is a rapidly expanding field, and I'm just going to summarize very quickly a few options. So, it talks about polatuzumab vedotin. This is actually the only licensed and approved additional therapy for relapsed refractory DLBCL in the UK as things stand. This is the progression-free survival curve from the longer follow-up from the study which compared R-CHP to polatuzumab vedotin alone. So, it can give quite durable response for patients. It is not really considered to be curative. You can see a pretty steady pattern of relapse in these patients, and this treatment is often used more as a bridge to CAR T in the UK.

There's some interest in this combination, tafasitamab plus lenalidomide, an anti-CD19 enhanced antibody, and this is data from the single-arm L-MIND study, again showing pretty promising progression-free survival at 12 months, and this is not yet approved in the UK. Loncastuximab tesirine is a CD19-directed antibody-drug conjugate with again some promising data from the LOTUS 2 study, which actually recruited a probably higher-risk group to what was seen in the L-MIND study, so quite promising results there. And but again, that is not currently approved in the UK, although can sometimes be accessed on an inpatient basis. And I think probably the most exciting area is bispecific antibodies, which again could be a talk in its own right, but as we probably know, these antibodies such as mosunetuzumab and epcoritamab in particular, having this two-to-one configuration which allows bivalent binding to the CD20 on the tumor cell and then engages T cells via a CD3 arm. And mosunetuzumab in particular has given us a fixed duration, and we saw some really interesting data at conferences earlier this year, in particular with this phase 1-2 study in relapsed refractory DLBCL over 100 patients, many of whom had actually had prior CAR T, and there were CR rates seen of over 30%, and that was even in patients who had relapsed after CAR T-cell therapy, so a really promising treatment for patients. And whether or not, in fact, going forward, they will replace CAR T-cell therapy or I think remains to be seen.

Okay, so that's really all I was going to cover today, and I'm aware that's been a bit of a whistle-stop tour, but these, I suppose, are my main conclusions from what I've told you. I think what we've seen is that for many patients, still six cycles of R-CHOP does remain the standard of care for DLBCL, but we have seen that we've got work to do, and particularly for those patients with high-risk disease who have poor outcomes. And as I've shown you, there may be an argument for using intensive regimens in this group, but certainly, they need to be looked at in respective trials going forward. We have seen a progression-free survival benefit with R-CHP, but as we discussed, whether or not it is widely approved, I think remains to be seen. And I think ongoing studies really have to be focused on not just adding novel agents to R-CHOP, but probably using our new knowledge of the molecular biology of all the different subtypes of DLBCL, incorporating specific novel agents, and trying to have a more precision medicine approach to DLBCL. I think that's got to be where the field looks in the next few years. I hopefully I've demonstrated that CNS prophylaxis does remain contentious, unfortunately, but probably we're going to see less use of rituximab for CNS prophylaxis, and certainly delivering it at the end of therapy only. And CAR T-cell treatment has clearly transformed the relapsed refractory landscape in DLBCL, and as we've seen, we may start to see that move forward in the therapeutic algorithm, although that's got still got a lot of discussion and particular cost involved going forward. And we have, you know, there are a lot of other promising agents out there, in particular bispecific antibodies in the relapsed refractory settings. So, much, much still to be achieved in this field. Thanks very much for listening, and I'd be more than happy to take any questions, and if anybody wants to get in contact with me, and thanks again to Blood Academy for the invitation.

Thank you.