ESMO 2026 | RNA sequencing to identify predictors of response to neoadjuvant IO in melanoma

So, at ESMO, I did this amazing discussion of this amazing abstract, but I also presented a poster, basic research, in fact, translational research. So, basically, I took a large cohort of patients, melanoma patients, treated with immunotherapy in the neoadjuvant setting, okay? So, stage clinical stage three patients were treated with a short course of immunotherapy in the neoadjuvant setting and what I did was to cluster those patients or big cohort in groups based on the interaction between the tumor and immune system using RNA sequencing data so based on on what? Based on the ability of expressing antigens, the ability of presenting antigens, the ability of activating T cells, the expression of immune checkpoints. So based on these features, then we cluster the patients. We identified five different clusters and each cluster, they have specific characteristics. There are the cluster that is the perfect one, that has everything good. There are the clusters that they miss, for example, the antigen presentation. So based on this, then we zoomed in to the non-responders. And we know that the non-responders, 40% of the non-responders, do not recur. Even though they are at high risk of recurrence, they do not recur so then with those clusters we were able to discriminate better so using those five clusters three of those clusters the patients did not recur so we are kind of the good ones within the best pathological group and two clusters they performed very poorly so then why is this important because then we can go back and say okay this patient did not respond and this patient is missing for example antigens right then we can suggest a vaccine because the vaccine gives the antigen but if for example the problem is antigen presentation well then the problem is bigger we can use for example cellular therapy or if the patient has everything there but has high expression of immune checkpoints, we can use immune checkpoint inhibitors like LAG3. So that’s kind of the concept. And this is the beginning. We just started by clustering. And now we are going to characterize each cluster with other layers, DNA sequencing data, and characterize the pre-treated immune profile.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.