Welcome, everyone, and thank you for joining us today. My name is Cathy Bowes, I'm the Marketing Manager here at Akoya Biosciences, and I'm delighted to be with you today for this webinar, which is one of the sessions in our Spatial Insights Driving Drug Discovery Series. This series explores the impact of spatial biology from early-stage discovery to the clinic, and we've been delighted to welcome some leading experts in this field to share how spatial biology is reshaping our understanding of complex biological systems, optimizing therapeutic development, and advancing precision medicine. You can see here all the speakers who are part of the series so far. Some of these sessions are available on demand, and some are still to be broadcast. Towards the end of the webinar today, I'll provide more information on how you can find out more about each of these webinars and either watch on demand or sign up, so please stay tuned for that. Today, we're delighted to welcome Professor Antoine Italiano from the Institut Bergonié in France, where he is Head of the Department of Medicine, and we're looking forward to his presentation as he provides some insights and perspectives on tertiary lymphoid structures in cancer immunotherapy. Professor Italiano, a very warm welcome to our session today. Thank you for joining us. You may now share your screen and begin your presentation. Thank you very much. I'm very pleased to share with you today some insights regarding the impact on tertiary lymphoid structures on cancer immunotherapy. You know very well that now immunotherapy represents a very important therapeutic tool to manage patients with cancer, but in the advanced setting at the end, only 30% of patients will really benefit from immune checkpoint inhibition. This means that we need better predictive biomarkers to select accurately the patients who are more likely to benefit from this kind of approach. In 2020, in the same issue of Nature, three seminal papers were published showing the major impact of the presence of tertiary lymphoid structures on response to immunotherapy to immune checkpoint inhibition in cancer patients, and particularly in three different indications: sarcoma, renal cancer, and melanoma. My team contributed, thanks to the work of Wolf H. Fridman, to the sarcoma paper. Basically, this paper showed that the presence of B cells, and particularly the presence of tertiary lymphoid structures, is the major predictor of response to IO. I will illustrate to you how these findings were able to modify the therapeutic potential in some cancers, and particularly in sarcoma. Sarcoma represents about 1% of cancer in adults, between 10% and 15% in children. It is a rare condition, but it is also a very heterogeneous disease with about 100 different histological subtypes identified in the World Health Organization classification. The cornerstone of treatment for patients with sarcoma is surgery. The problem is that despite good surgery, about between 30% and 40% of patients will develop metastatic disease, and patients with metastatic sarcoma will at the end die of this condition. In fact, the treatment possibilities for metastatic sarcoma are very limited. The standard of care is based on doxorubicin chemotherapy, and this standard of care has been established in 1973, so no improvement in the last 50 years. It is a pity because actually chemotherapy is very poorly efficient in this condition, with a response rate of less than 10%, and the median overall survival of this patient is less than 18 months. For what regard immunotherapy, you know that the cornerstone of immunotherapy today relies on antibodies targeting the PD-1, PD-L1 synapse, and that these antibodies are now approved in a lot of several different indications of cancer. Sarcoma represents an exception. Why do we have this exception in sarcoma with no approval immunotherapy so far? One of the reasons for this situation is summarized in this slide. Actually, this is the first study which was coordinated by my team investigating PD-1 inhibition in sarcoma. In that study, we designed a clinical trial enrolling a sarcoma patient with advanced disease based on an old commerce strategy, meaning that we did not implement a biomarker strategy, and the reason for that is that at that time, we did not know any biomarker that could help to select patients. In this clinical trial, as you can see in this waterfall plot, actually the efficacy results were very disappointing. Only one patient had an objective response as per RECIST criteria, which are the international criteria to consider a reduction in disease burden as significant from a clinical point of view. The response rate of patients in this clinical trial was only 2%. In front of this data, two situations were possible. The first decision could be to give up and to decide that immunotherapy is not relevant in sarcoma and to try to test some seagulls. The second approach is to try to understand why we obtained such a poor efficacy result in this clinical trial. This is the path we have chosen in collaboration with the team of Wolf H. Fridman in Paris, and we decided first to understand more the microenvironment of soft tissue sarcoma. To do that, we performed a transcriptomic analysis of several hundred soft tissue sarcoma cases, and by using bioinformatic tools and a deconvolution approach, we decided to investigate the immune landscape of this sarcoma. Basically, we estimated the level of immune cell infiltration based on the gene expression data. By using this approach, we were able to identify five different cases of soft tissue sarcoma: A, B, C, D, and E, from the less inflamed, so basically a desert in terms of immunological landscape, so the class A, versus the most inflamed with a lot of immune cells in there, and particularly B cells, which is the class E, which represented about 20% of soft tissue sarcoma. What we observed also in this inflamed case is that despite the high infiltration in T cells and B cells, there was also a strong expression of immune checkpoint. Based on that, which was first an important result, there is a small group of soft tissue sarcoma that is inflamed. The first question that raised was, does this classification have a prognostic impact? The answer was yes, but not all the immune cells had a prognostic impact. In fact, surprisingly, we found that the level of CD8 density was not associated with outcome, but this was not the case for B cells. Actually, we found that the level of B cells infiltration was strongly associated with outcome, with patients having a high level of B cell density in the tumor, having a lower risk of metastatic relapse, a lower risk of death from sarcoma. We found a group of inflamed sarcoma. We found that this group has a better outcome. The next question was, is this classification having a predictive value in terms of response to immunotherapy? You remember that in the clinical trial we did, there was only one responder, so it was not possible to make any correlation between the immune landscape and response to immunotherapy. This is the reason why we established a collaboration with American colleagues from the MD Anderson in Houston that actually conducted a similar study than the French one investigating PD-1 inhibition with pembrolizumab in sarcoma patients. Intriguingly, in very clinical trials, there were more responders than in all clinical trials. We established a collaboration, and the friends from the MD Anderson sent us the tumor samples from the patients included in the study. These tumor samples were analyzed by the team of Wolf H. Fridman, and what we observed is that the majority of the responders belong to the E group of the immune sarcoma classification we established before. The majority of the responders had an inflamed microenvironment, meaning that this classification has not only a prognostic value, but also a predictive value in terms of response to immunotherapy. We then looked at the microenvironment features of these tumors, and we were very happy to find that actually these samples belonging to the E group of the immune classification were in fact enriched in tertiary lymphoid structures, as you can see here, which represent again aggregates of B cells, T cells, and antigen-presenting cells. All these data suggested that TLS-positive sarcoma were more likely to benefit from immunotherapy. The problem was that all these data were retrospective, and we needed a prospective confirmation of these findings. This is the reason why we decided very rapidly to set up a new clinical trial in which the sarcoma patients were selected based on the presence of tertiary lymphoid structures. Only patients with TLS-positive sarcoma were allowed to be enrolled in the clinical trial. Here you have the statistical hypothesis of the study, and here the patient disposition. Basically, we had to screen 240 patients for TLS status. 48 were positive, so 20%, exactly the same proportion we observed in the immune classification, and 35 were included in the study. As you can see, these tertiary lymphoid structures were observed in several different histological subtypes of sarcoma. TLS were not restricted to one or two histological subtypes. They can be found in several sarcoma subtypes. Here you have the efficacy data. As you can see in the waterfall plot, a significant number of patients derived clinical benefit. The response rate was 30%. Remember, in the old commerce study in which there was no selection, the response rate was only 2%. Moreover, even more interesting, the responses were durable, and the median duration of response was about one year. Then, based on that, we decided to analyze retrospectively the TLS status of the first study. What we observed is that actually all the patients were TLS negative except the responders. We were very unlucky in this trial. We enrolled only TLS negative patients, and this is the reason why we observed so poor efficacy results. The pembrolizumab study was actually the first clinical trial using TLS as a biomarker to select cancer patients for immune checkpoint inhibition, and basically, it represented the first prospective confirmation that TLS is a reliable biomarker to select cancer patients for immunotherapy. The beauty of the story was, in fact, that the predictive value of TLS is not limited to sarcoma, but is universal, and we confirmed that in a subsequent study in which we decided to analyze the TLS status of all the patients treated with immunotherapy at our institution, and whatever the tumor type. Eleven different cancer indications were included in the study. What we observed is that in this study, which included more than 600 patients, the TLS status was the most powerful predictor of outcome in terms of response rate, in terms of PFS, in terms of OS, independently of the PD-L1 expression level, independently of the level of CD8. Here you can see that this study made the cover of the Nature Cancer issue, and you have a very nice representation of a tertiary lymphoid structure. This patient was a cancer patient. It was a pancreatic cancer patient, sorry. Pancreatic cancer is considered as a resistant disease to immunotherapy. This patient had an excellent response to pembrolizumab, and in fact, TLS can be present in about 7% of patients with pancreatic cancer. This means that this is again an illustration that investigating TLS can be a robust approach to select patients for cancer immunotherapy, even in indications that are considered generally as resistant to IU. We made a prospective confirmation of this universal value of TLS prediction in another clinical trial I presented at CDC in 2023 and last year at RCR. The paper is upcoming, in which we included a patient based on the TLS status, but whatever the cancer type. It was a histotype agnostic study, and this study again was positive, and we observed a meaningful response, again, even in indications considered as resistant to IU, such as MSS colorectal cancer, small bowel carcinoma, biliary tract cancer, sarcoma, etc. I hope now that I convinced you that TLS is a relevant biomarker, but obviously, not all the patients with TLS-positive tumors respond to IU. What we try to understand now is what are the determinants of resistance to IU in TLS-positive tumors. We try first to answer this question for sarcoma patients, and this data were in the Nature Medicine paper. First, we use a spatial transcriptomic approach to compare TLS-positive sarcoma responding to IU versus TLS-positive sarcoma resistant to IU. What we observed is that in resistant patients, there was a strong enrichment in Treg in the TLS structure, within the TLS structure, and outside the TLS structures. In the responding patient, there was a strong enrichment in plasma cells, actually plasma cell-producing immunoglobulin that stains at the surface of the membrane of tumor cells. We confirmed this finding by using a multiplex IF approach, and you can see here that patients with high levels of plasma cell infiltration, and we use here a MEM-1 marker with also immunoglobulin G staining, have a better outcome, PFS and better OS than patients with low infiltration in plasma cells. It is the reverse for Treg with CD4 FOXP3. We have a poor outcome of patients with high levels of Treg versus patients with low levels of Treg, both in terms of PFS and on OS. This impact of Treg is not limited to sarcoma, and we observed that also in other indications, including non-small cell lung cancer, in TLS-positive non-small cell lung cancer. You can see that there is a detrimental impact of high levels of Treg infiltration in TLS-positive tumors. What is interesting is that there are now new immunotherapy targets, new immunotherapy drugs that target specifically the Treg, and particularly CCR8 antagonists. Here you have data that we presented last week at ASCO, which demonstrated the detrimental value of a high abundance of Treg expressing CCR8. Again, we use a multiplex IF approach for that, both in terms of PFS and of OS. Obviously, there are other components of the microenvironment that are involved in resistance to IU in TLS-positive tumors. Here you have an example of a very recent paper we published also showing the impact of two specific populations of fibroblasts in resistance to IO in TLS-positive non-small cell lung cancer. To identify these two populations of fibroblasts, we used first, sorry, a spatial transcriptomic approach, which showed that basically the main difference between responders and non-responders relied on the stroma segment. Then by using several approaches, including multiplex IF, we were able to characterize this population of fibroblasts and to correlate their abundance with the immune contextual exclusion of CD8, and also with the Treg infiltration and the immune exhaustion profile. How can we improve the response rate to immunotherapy in TLS-positive tumors? I remind you that in TLS-positive sarcoma, there was a strong overexpression of immune checkpoints, not only PD-1, PD-L1, but also those other immune checkpoints: CTLA-4, LAG-3, TIM-2, etc. One way maybe to increase the response rate to IU in this group of tumors is to use a double immune checkpoint inhibition. This is something that we are considering in several clinical trials, and one of them is the CONGRATS study, which compared PD-1 targeting versus PD-1 plus LAG-3 targeting in TLS-positive suspicious sarcoma. We will be able to present the result of this randomized study at NexSmall in Berlin. There is also a strong expression of TIGIT and other immune checkpoints in TLS. TIGIT plays an important role in the biology of B cells, and probably this is an important target to consider in TLS-positive tumors. Obviously, if we have a biomarker, we have to know, we have to have also a standardized approach to evaluate this biomarker. This is the reason why my colleagues from the Institut Bergonié reported, published in 2023, guidelines for the pathologists to assess the TLS status in tumors. The first approach is to analyze the H&E slide. If you have a clear big germinal center, you can consider the tumor as TLS positive. If this is not clear, you will have to use CD20 and CD23 staining, CD23 for the follicular dendritic cells and to confirm the maturation of the TLS. If the tumor is a desert in terms of immune cell infiltration based on an H&E review, you do not need any staining, and you can consider the tumor as TLS negative. In conclusion, I hope that I convinced you that TLS is a robust biomarker to tell your immune checkpoint inhibition in patients with three tumors. The mechanism of action is probably in part related to the production of antibodies that are targeting tumor antigen at the surface of the tumor cells. In the next year, one way to improve immunotherapy in cancer patients is probably to try to induce TLS in the tumor microenvironment for patients with negative TLS tumors. Thank you very much for your attention. Thank you so much, Professor Italiano, for that insightful presentation. That was wonderful. We've got some interest from our audience, and we'll move on to the Q&A session. Just a reminder to our audience that if you do have a question for Professor Italiano, you can submit this now using the Q&A button on the Zoom panel. Professor Italiano, we have some questions already, so we'll make a start. The first question is, could you elaborate on which immune cell subsets within TLS seem most predictive of response to immune checkpoint inhibitors in sarcoma patients? Yeah, absolutely. TLS represents an aggregate of B cells, T cells, and antigen-presenting cells. What we observed is that probably what is the most impactful in terms of response to IU is the abundance of plasma cells. Basically, a maturation of B cells that allow them to produce a specific immunoglobulin that are targeting specific antigens at the surface of the tumor cells. We do not know what are these antigens, and probably this is an important research area in the next few years. I believe that this is an important mechanism of action of the TLS in the tumor microenvironment. The other one is the role of B cells as antigen-presenting cells, but the production of immunoglobulin by plasma cells is probably also crucial. Thank you for that. The next question is, what do we currently understand about the mechanisms driving TLS formation in soft tissue sarcomas, and are there tumor intrinsic factors at play? Yeah, it's an important question. Actually, we don't know yet what are the tumor determinants that can be associated with the formation of TLS. One first hypothesis, for instance, was that there could be a link with tumor mutational burden. Actually, there is no clear correlation between tumor mutational burden and the presence of TLS. In fact, TLS are also present in tumors with very low tumor mutational burden, such as sarcoma. This is something we are investigating in a large program, which is called ConDoR, in which we will analyze the genomic features, genomic transcriptomic features of more than 1,500 different sarcomas and correlate it with the TLS status. I hope I will have the answer to your question by next year. Fantastic. Thank you. The next question is, what are the key steps needed to validate TLS as a biomarker across different tumor types beyond sarcomas? TLS represents microenvironment features that can be easily assessed by a pathologist in the routine setting. In my institution, for instance, TLS are now part of the routine assessment of sarcoma and of non-small cell lung cancer. In the guidelines my pathologist colleague published, they recommend to use first the analysis of the H&E staining and then to do CD20 and CD23 staining. I am aware that there are now some vendors that are validating TLS assay based on CD23 staining to assess the TLS status in tumors. There are also other approaches using artificial intelligence and digital pathology workflow to assess the TLS status based only on the H&E slide without any staining. I do not think really that we need a companion diagnostic assay, but we have just to follow a rigorous approach with it based on H&E analysis if there is no clear germinal center, CD20 and CD23 staining. Okay, thank you. I think we have time for just one more question. We will finish with this one. Are there significant differences in TLS structure or function between soft tissue sarcomas and other solid tumors? It seems not. What we observed is that the predictive value of TLS was strong, whatever the tumor type. Possibly the respective role of cellular immunity versus humoral immunity could differ according to the tumor type, but at the end, the TLS represents, whatever the tumor type, the most robust machinery to induce anti-tumor immune response. That's great. Thank you so much. Thank you for the presentation that you did for us today, and thank you as well for taking the time to answer those questions. Thanks as well to our audience for submitting your questions. If we have not had time to address your question today, we will certainly respond to you after the webinar. Thank you very much. Thank you, Professor Italiano. At the beginning of today's session, I promised to provide some details on how you can access the other webinars in this series. Simply visit our website at akoyabio.com and navigate to the resources section. There you will find details of all our on-demand webinars so you can listen again, as well as details of our upcoming webinars and all the registration links that you need. Please do take a look. Finally, all that remains for me to say is thank you again to Professor Italiano for his presentation today, and also to you, our audience, for attending and for your engagement. If you've attended a recent webinar of ours, you'll know that as you leave the webinar, a short survey will appear. We appreciate it if you could take just a couple of moments to complete that survey, as your feedback will help optimize the webinar experiences we bring to you. That's all for today. I will now bring the webinar to a close. Thanks again, everyone. I wish you all the best and hope you can join us next time for our next webinar session. Thanks all. Goodbye for now.