ESMO 2026 | Examining difficulties in patient selection for investigational immunotherapies
Sebastian Kobold, MD, PhD, Ludwig Maximilian University of Munich, Munich, Germany, comments on the challenges of pre-selecting patients who are most likely to benefit from novel immunotherapies, highlighting the push towards generating biomarkers such as mutation load and PD-L1 expression, but notes that these approaches have limitations and inconsistencies in predicting patient response. Despite these efforts, it remains difficult to accurately pre-select patients, resulting in responders and non-responders across different groups. This interview took place at the European Society for Medical Oncology (ESMO) 2026 Congress in Madrid, Spain.
These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.
Transcript
This is clearly the elephant in the room, right? So how can we better pre-select the patients that are the highest chance of benefiting from these treatments? And I think this is clearly recognizable that there is a push towards generating these biomarkers, but it’s easier said than done. So what we’ve seen is, of course, that people consistently revolve around known biomarkers like mutation load, tumor mutation burden, PD-1, PD-L1 expression, or now in the frame of this targeted bispecific or T-cell receptor engineered approaches where they’re really looking for the expression, for the level of expression of the target antigen...
This is clearly the elephant in the room, right? So how can we better pre-select the patients that are the highest chance of benefiting from these treatments? And I think this is clearly recognizable that there is a push towards generating these biomarkers, but it’s easier said than done. So what we’ve seen is, of course, that people consistently revolve around known biomarkers like mutation load, tumor mutation burden, PD-1, PD-L1 expression, or now in the frame of this targeted bispecific or T-cell receptor engineered approaches where they’re really looking for the expression, for the level of expression of the target antigen. But as you can imagine, so this is relatively low-key, but also very difficult to implement at the scale and only inconsistently manages to really pre-select patients based on that. So meaning that you still end up with responders in the lower groups and non-responders in the higher groups, which means that we’re still not there yet.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
Read more...
